|Home | About | Journals | Submit | Contact Us | Français|
Under exposure to UV radiation, some cyanobacteria synthesize sunscreen compounds. Scytonemin is a heterocyclic indole-alkaloid sunscreen, the synthesis of which is induced upon exposure to UVA (long-wavelength UV) radiation. We previously identified and characterized an 18-gene cluster associated with scytonemin biosynthesis in the cyanobacterium Nostoc punctiforme ATCC 29133; we now report on the expression response of these genes to a step-up shift in UVA exposure. Using quantitative PCR on cDNAs from the N. punctiforme transcriptome and primers targeting each of the 18 genes in the cluster, we followed their differential expression in parallel subcultures incubated with and without UVA. All 18 genes are induced by UVA irradiation, with relative transcription levels that generally peak after 48 h of continuous UVA exposure. A five-gene cluster implicated in the process of scytonemin biosynthesis solely on the basis of comparative genomics was also upregulated. Furthermore, we demonstrate that all of the genes in the18-gene region are cotranscribed as part of a single transcriptional unit.
As phototrophs, cyanobacteria must live in environments where they can access radiant solar energy. In addition to the visible wavelengths used for photosynthesis, cyanobacteria are also typically exposed to deleterious UV irradiation, which causes DNA and protein damage, as well as photooxidative stress (23). Since many cyanobacteria live in environments such as soil and rock surfaces, where UV irradiation is a major environmental stress (6, 13), they have developed numerous UV defense mechanisms. Many of these are active mechanisms, such as the expression of UV shock proteins, like superoxide dismutase (33, 34); the excision of mismatched DNA sequences with subsequent photoreactivation to repair the excised sites (25, 39); and the use of UVB radiation-specific photoreceptors (28) to orchestrate some of these responses. Other mechanisms are behavioral, such as the photophobic response to UV displayed by some motile cyanobacteria (4). Alternatively, some cyanobacteria produce and accumulate UV sunscreen molecules (11, 16), which passively prevent high-energy photons from reaching the cell. These sunscreens absorb incident UV radiation and release it through thermal de-excitation. The use of sunscreens is considered a passive mechanism of defense because they minimize UV damage without the need for active metabolism or the additional input of energy (6, 14). Scytonemin, a yellow-brown, lipid-soluble, heterocyclic indole-alkaloid (29) that absorbs best in the UVA range (315 to 400 nm) (16), is widespread among cyanobacteria. In fact, with the exception of planktonic species, it has been found in members of all groups of cyanobacteria, including the unicellular and filamentous morphotypes (14). It is produced in response to UVA exposure by some strains, in which it accumulates in the extracellular sheath (14). Previous research suggested that tryptophan and tyrosine are precursors in the biosynthesis of scytonemin, as is also suggested by its chemical structure (26, 29).
We previously identified a genomic region associated with scytonemin biosynthesis in the filamentous cyanobacterium Nostoc punctiforme ATCC 29133 (PCC 73102) (35). This genomic region contains 18 contiguous genes (Fig. (Fig.1A)1A) transcribed in a single direction (NpR1276 to NpR1259 as annotated in the N. punctiforme genomic database [available at http://www.jgi.doe.gov]). Most of the genes in the upstream region of the cluster, with the exception of scyA (NpR1276) and scyB (NpR1275), code for novel proteins without any easily predictable function. It is very likely that these proteins are directly involved in scytonemin biosynthesis, considering that the enzyme activities of scyA and scyB confirm their involvement in the early stages of scytonemin biosynthesis (3) and that a transposon inserted in scyD (NpR1273) effectively eliminates scytonemin biosynthesis under otherwise inducing conditions (35). Most of the genes in the downstream region appear to function as a redundant set of the aromatic amino acid biosynthesis genes, including the first two enzymes of the shikimic acid pathway (AroG and AroB), the first enzyme of tyrosine biosynthesis (TyrA), and all five enzymes of the tryptophan biosynthetic pathway (TrpECABD). Several other genes were recently identified in the genome of N. punctiforme by their conserved localization in the scytonemin clusters of several other cyanobacterial strains (Fig. (Fig.1B,1B, genes in the scytonemin five-gene satellite cluster). These genes are present as a five-gene satellite cluster in N. punctiforme (NpF5232 to NpF5236), and although their annotations are unclear, they may also be involved in scytonemin biosynthesis (T. Soule, F. Garcia-Pichel, and V. Stout, unpublished data). Additionally, two genes immediately upstream from the scytonemin-associated cluster (NpF1278 and NpF1277, encoding a two-component sensor kinase and a response regulator, respectively) are highly conserved in sequence and location among cyanobacterial strains that possess the scytonemin cluster, leading us to hypothesize that they may be involved in regulating the expression of the scytonemin-related genes. Little is known about the regulation of the genes associated with scytonemin biosynthesis. These genes were not differentially expressed under conditions of nitrogen starvation or under the induction of cellular differentiation (akinete formation) compared to control cultures, as demonstrated in whole-genome microarray-based analyses of global gene expression patterns in N. punctiforme (5). Furthermore, the predicted protein products of scytonemin cluster genes were not detected in a whole-proteomic analysis of N. punctiforme cells harvested at different developmental stages in the absence of UVA exposure (2).
In this study, we report on the temporal expression of each of the genes putatively involved in scytonemin biosynthesis from N. punctiforme cultures exposed to UVA irradiation using quantitative PCR (qPCR) (31). We also discovered evidence of cotranscription and the presence of transcriptional units among these genes.
N. punctiforme ATCC 29133 cultures were grown and harvested during a transition from exposure to white (visible) fluorescent light alone (10 W m−2) to illumination by (unchanged) visible plus additional UVA radiation (5 W m−2). Filter cultures were prepared to minimize self-shading and to promote homogeneous exposure to the majority of the cells. To prepare the cells, a liquid batch culture of N. punctiforme was grown in Allen-Arnon (AA) medium without combined nitrogen (1), since a previous study found that the rate of scytonemin production in N. punctiforme increased under diazotrophy (12). Equal amounts (ca. 5 ml) of this homogenized culture were filtered onto 90-mm-diameter polycarbonate membrane filters and placed, floating on liquid AA medium, in glass petri dishes as described previously (15). The cells grew on these membrane supports for 3 days, at which time a subset of filter cultures were placed under white light supplemented with UVA (treatment series) while the rest were kept under white light without UVA (control series). Three culture membranes were retrieved for each time point thereafter (beginning with time zero) daily for 4 days, with a final sampling after 7 days. The cultures were processed immediately after retrieval as follows. Each filter was transferred with blunt-ended forceps to a sterile 50-ml centrifuge tube containing 25 ml AA medium and vortexed to wash the cells off of the membrane. The membrane was then removed with sterile forceps, and a 2-ml aliquot of the homogeneous suspension was removed from each sample to quantify the scytonemin. The remaining cells were washed with fresh AA medium and centrifuged to concentrate them into a total volume of 0.5 ml, which was flash-frozen in liquid nitrogen and stored at −80°C until it was processed further. The scytonemin contents were quantified spectrophotometrically in acetone extracts as previously described (14).
The total RNA from each sample was extracted using a LiCl precipitation protocol (32, 37), treated with DNase, and purified using a commercial kit (Qiagen Sample and Assay Technologies). The RNA integrity was verified using an Agilent 2100 Bioanalyzer, and the absence of genomic DNA contamination was confirmed by PCR using several of the primer sets described below on the RNA extracts, with 20 ng of N. punctiforme genomic DNA as a template for a positive control. This high-quality RNA was used to synthesize cDNA, which was primed using random hexamers (Invitrogen Co.) and tandem primers (N18) and extended by SuperScript III polymerase (Invitrogen Co.) (5). All nucleic acid extracts were quantified on a Nanodrop spectrophotometer (Thermo Fisher Scientific).
Primers specific to each gene in the 18-gene cluster and each of the 5 satellite genes, as well as the housekeeping copy of 2 of the aromatic amino acid biosynthesis genes, were designed (Table (Table1)1) and used to amplify ca. 0.2-kbp fragments from the cDNA in each extract using a real-time qPCR procedure (31). Before the primers were used in qPCR, they were tested through standard PCR on an N. punctiforme genomic DNA template to verify the thermocycling conditions and specificity. This was done by obtaining a single product of the expected length, confirmed on 1% agarose gels using a molecular mass standard (Bio-Rad Laboratories). For standard PCR, 20 ng of genomic DNA was used in 50-μl reaction mixtures consisting of 1 μM of each specific primer, 5 μl 10× Ex Taq DNA polymerase buffer, 4 μl deoxynucleotide triphosphate mixture (2.5 mM each), and 1.25 units Ex Taq DNA polymerase (all reagents for standard PCR in this study were from Takara Bio Inc.). PCR was performed in a Bio-Rad iCycler thermal cycler with the following parameters: 95°C for 5 min, and then 30 cycles of 95°C for 1 min, 53 or 55°C for 1 min (Table (Table1),1), and 72°C for 2 min, followed by an extension at 72°C for 10 min. All primers in this study were from Operon Biotechnologies Inc.
A single batch of 36 qPCR reactions was carried out for each gene, which included the three experimental replicates and the three control replicates for each of the six time points, in addition to three negative controls devoid of cDNA template and three positive control reactions that used N. punctiforme genomic DNA as a template. Furthermore, primers specific for the reference gene, rnpB (the RNA component of RNase P), were used on cDNA extracts from each time point. rnpB was chosen as a reference gene to normalize the raw expression data because its expression does not change significantly with UV exposure (7). In each qPCR mixture, 1 ng of cDNA was combined with 10 μl 2× SYBR green Supermix with Rox (Bio-Rad Laboratories, Inc.), 250 nM of each specific primer, and sterile water to 20 μl. qPCR was performed on Ambion's ABI7900HT thermocycler using a 384-well format with an initial denaturation at 95°C for 150 s and then 35 cycles of 95°C for 15 s, 55°C or 53°C (Table (Table1)1) for 20 s, and 72°C for 15 s, followed by a melting curve analysis. The levels of differential gene expression in extracts from the UVA-induced cultures versus those from the white-light control cultures were calculated as ratios from the rnpB-normalized real-time amplification data.
To evaluate cotranscription between adjacent genes in the 18-gene scytonemin cluster, we did reverse transcriptase (RT) PCR on the RNA extracts obtained as described above. For this, we designed a primer pair for each gene pair that spanned the downstream part of the first gene, the intergenic regions, and the upstream portion of the second gene (Table (Table2).2). Each cDNA synthesis reaction mixture contained 250 nM of a specific reverse primer, 1 μg total RNA extracted from the UVA-induced cultures, and reagents supplied in the iScript Select cDNA Synthesis Kit (Bio-Rad Laboratories, Inc.). Samples were processed according to the manufacturer's instructions by incubating them at 42°C for 1 h. For PCR, each 50-μl reaction mixture contained about 20 ng specific cDNA, 1 μM of each appropriate primer, 5 μl 10× Ex Taq DNA polymerase buffer, 4 μl deoxynucleotide triphosphate mixture (2.5 mM each), and 1.25 units Ex Taq DNA polymerase (Takara Bio Inc.). N. punctiforme genomic DNA was the positive control, while the negative controls had no template DNA. To assess false-positive reactions that may have been a result of contaminating genomic DNA, the corresponding RNA extracts were used as negative control templates. PCR was performed in a Bio-Rad iCycler Thermal Cycler with the following parameters: 95°C for 5 min and then 35 cycles of 95°C for 1 min, 55°C for 1 min, and 72°C for 1 min (3 min for large products), followed by an extension at 72°C for 7 min. The PCR products were examined on 1% agarose gels run at 110 V for 2.5 h and subsequently stained with SYBR gold (Invitrogen Co.). The cDNA band of the appropriate size was determined for each sample by comparing the bands to an N. punctiforme genomic DNA fragment amplified with the same primers (Fig. (Fig.2).2). The appropriate bands were excised using a sterile scalpel, gel purified using the QIAquick Gel Extraction Kit (Qiagen Sample and Assay Technologies), and sequenced commercially (Applied Biosystems) to verify that the correct products had been obtained.
As expected, N. punctiforme cells exposed to UVA and white light underwent induction of scytonemin synthesis, whereas the controls grown in white light alone did not (Fig. (Fig.3).3). The level of scytonemin produced by N. punctiforme in this study reached approximately 1.3% of the total dry weight of the cells. Although a comparison of scytonemin levels across strains is difficult to assess due to varying culture conditions, the level in this study reached those typically obtained in the laboratory (14).
In order to determine the effect of UVA irradiation on the expression of the genes associated with scytonemin biosynthesis, we used qPCR to assess the relative mRNA transcript abundance levels of the genes in the 18-gene cluster (NpF1276 to NpR1259), the 2 putative regulatory genes (NpF1278 and NpF1277), and the 5 genes in the satellite cluster (NpF5232 to NpF5236). We also evaluated the expression of the shikimic acid pathway genes that are not found in the scytonemin cluster (aroQ [NpR4569], aroE [NpR0575], aroK [NpR6293], aroA [NpF1315], and aroC [NpF3573]), as well as two standard housekeeping aromatic amino acid biosynthesis genes (tyrA [NpR0978] and trpA [NpF6623]) with homologs in the 18-gene cluster. Our results demonstrated a general trend of expression for all 18 of the genes in the scytonemin cluster, which reaches a maximum after 48 h of continuous UVA exposure (Fig. 4A to C) with relative transcript abundance levels elevated 2.8- to 5.2-fold over those found in the white-light control series. Two of the genes, however, did not seem to conform to this temporal pattern. NpR1261 (trpD) increased to a 3.9-fold change after 1 day, and NpR1269 (tyrA) rose to an 8.4-fold change after 4 days, the largest increase seen for any gene in the experiment. The analysis for tyrA was repeated for verification. The genes in the five-gene satellite cluster, except for NpF5233, showed maximal relative expression after 4 days of UVA exposure, a delayed response with respect to those in the main scytonemin cluster. Additionally, their levels of induction were generally more subdued, with increases ranging from only 1.9- to 3-fold (Fig. (Fig.4D).4D). The relative expression of NpF5233 peaked on day 2 with a threefold increase, in unison with the scytonemin genes in the main cluster. No difference in the levels of expression under UVA versus the white-light controls could be detected for the putative regulators, the shikimic acid genes absent from the main gene cluster (aroE [NpR0575] is not shown, since its relative expression levels were below the detection limit), or the standard housekeeping genes (Fig. 4E and F). Genomic DNA was not detected in any of the RNA extracts, confirming that our qPCR results could not be attributed to false positives.
To demonstrate cotranscription of adjacent genes in the scytonemin-associated gene cluster, RT-PCR products were obtained from RNA extracts of UVA-exposed cultures using primers targeting adjacent genes. After confirmation by sequencing, our analyses indicated that there is a single large cotranscriptional unit in the 18-gene cluster, since we could concatenate all of the products (Fig. (Fig.5).5). Although the aroB and trpE genes did not show cotranscription with the initial set of primers, RT-PCR was attempted with two other sets of primers, and cotranscription was detected with the last set (Fig. (Fig.6).6). The most likely interpretation of these results (other than technical difficulty) is that there is little read-through in the operon between these two genes. As a control, a parallel analysis of NpR1259, the last gene in the main cluster, and NpF1258 (copZ), the most immediate downstream gene outside of the cluster (and with an opposite transcription direction), yielded no products, as expected. Cotranscription analysis between the most immediate gene upstream of the cluster, NpF1277, and NpR1276 was not done due to the large intergenic region. It is important to note that several attempts to use Northern blotting in this analysis were unsuccessful. This was probably because the mRNA levels of our target genes were too low for detection (Fig. (Fig.44).
Given that scytonemin biosynthesis is induced by UVA radiation, we hypothesized that all of the genes in the 18-gene cluster would increase in transcript abundance in response to UVA irradiation. This was correct, and the direct response of these genes to UVA provides further experimental evidence to support the claim that all of them are involved in the biosynthetic pathway. Indeed, a recent publication has determined the activities of the first two enzymes encoded in the pathway (ScyA and ScyB) and described their direct roles in scytonemin biosynthesis (3). Although it could be considered that the overall increases in the transcript abundances of these genes, with respect to the controls, are low (2.8- to 5.2-fold), these levels were consistent throughout our analyses. Furthermore, subtle regulation of gene expression is not abnormal for genes involved in secondary metabolism (19), which are generally expressed at lower levels than those genes involved in primary metabolic processes. We were also interested in the expression of the five satellite genes, since their only association with scytonemin biosynthesis was based on genomic comparisons. The fact that they were upregulated provides the first direct evidence of their involvement in scytonemin biosynthesis.
The only genes in these regions that did not appear to be affected by UVA treatment were the putative regulators. Thus, we could not show evidence of their involvement in the regulation of the cluster downstream. However, one must consider that they could still act as a regulatory system for scytonemin biosynthesis if (i) they are expressed constitutively and activated through phosphorylation (36) or (ii) they are induced and returned to basal levels sometime before 24 h. The former is common among two-component regulatory systems, as many histidine kinase proteins are available and ready to receive the environmental stimulus (36). The latter is also a plausible interpretation; since scytonemin is present by 24 h, any regulators would have to be transcribed before that time in order to act on the scytonemin biosynthetic genes.
Since scytonemin requires the aromatic amino acids as building blocks, we expected that the aromatic amino acid biosynthetic genes in the cluster would increase in expression in response to UVA. Indeed, while this was the case for the scytonemin-associated copies of these genes, it was not for their housekeeping homologs. It is unlikely, however, that the scytonemin precursors are drawn solely from the standard cellular aromatic amino acid pools provided by the housekeeping homologs. For example, the cellular levels of free tryptophan in Escherichia coli are only about 0.0136% of the total dry weight (22) while the scytonemin content in this study reached 1.3% of the total dry weight (Fig. (Fig.3).3). Even at half that amount (0.65%), given that the tryptophan-derived moieties constitute about half of the molecular weight of scytonemin (29), the cellular levels of free tryptophan available for scytonemin biosynthesis would need to be turned over 50-fold. The additional demand for free tryptophan in the cell to synthesize scytonemin precursors suggests that the set of proteins encoded by the redundant copies of aromatic amino acid biosynthetic genes in the scytonemin gene cluster are dedicated to this purpose.
Since only the first two genes (aroG and aroB) of the common aromatic amino acid biosynthetic pathway (shikimic acid pathway) have dedicated copies within the scytonemin gene cluster, we became interested in the expression of the rest of the shikimic acid pathway genes that are present elsewhere in the N. punctiforme genome. The fact that the genes in the rest of the shikimic acid pathway are not upregulated in response to UVA suggests that they are not essential to boost the production of aromatic amino acids. Indeed, aroG is a major point of regulation in the initiation of the shikimic acid pathway. It is subject to feedback inhibition by tyrosine, phenylalanine, and tryptophan (20, 21), and feeding experiments with labeled tryptophan suggest that this inhibition also applies to the synthesis of scytonemin (26). Furthermore, the enzyme encoded by aroB is one of the rate-limiting enzymes in the shikimic acid pathway (10). Thus, redundant copies of aroG and aroB in the scytonemin gene cluster are likely sufficient to boost and control the overall rates of the shikimic acid pathway to accommodate the increased demand for scytonemin precursors under UVA irradiation.
The relative timing of gene expression was also interesting. Since our cotranscription data suggest that all 18 genes in the main scytonemin gene cluster are part of a single operon, all of the genes should have shown similar temporal expression patterns. Indeed, the expression of all of these genes (except tyrA) peaked in unison soon after the UVA cue. The unusual delayed response of tyrA was surprising and is difficult to explain, as it is cotranscribed with the other genes and encodes a protein expected to be involved in the early stages of biosynthesis. As for the genes in the five-gene satellite cluster, their delayed response (except for NpF5233) may be due to their involvement in post-scytonemin biosynthesis processes, such as secretion, posttranslational modification, or protein degradation. It is difficult to assign roles to these proteins, as the annotations of the genes that encode them are generally unclear, and further research is required to elucidate their protein functions.
The main scytonemin gene cluster is a single transcriptional unit (operon) that is approximately 24.4 kb in length, spanning over 18 open reading frames (ORFs). While operons of this size may not be common, they are certainly not unprecedented. For example, the motility operon of Borrelia burgdorferi is composed of 26 ORFs that total 21 kb in length (17, 18). Furthermore, enzymes for the biosynthesis of the antifungal lipoprotein iturin A, a secondary metabolite, are encoded by the 38-kb itu operon in Bacillus subtilis RB14 (38), and an 11-kb operon encodes proteins involved in carotenoid and bacteriochlorophyll biosynthesis in Rhodobacter capsulatus (40).
It is important to note that there are two putative transcriptional terminators in the main scytonemin gene cluster, one after trpC and another after aroG (Fig. (Fig.1),1), detected through computational analyses as putative intrinsic transcriptional terminator sequences because they contain characteristic GC-rich stem-loop structures (9, 27). Our data do not support this prediction. Experimental evidence has contradicted computational predictions of rho-independent terminators in other instances, as a computational search for transcriptional terminators in B. subtilis returned a false-positive rate of 6% (24). It may be that the transcriptional terminators after trpC and aroG were inaccurately defined through bioinformatics. Alternatively, it may simply be that they act with low termination efficiencies. For instance, Reynolds et al. found terminator efficiencies that varied from 25 to ≥90% in vivo in an analysis of 13 transcriptional terminators from a variety of bacterial and phage DNAs (30).
At this point, we cannot determine the termination efficiencies of the putative terminators in this gene cluster; either they are not efficient and allowed the transcriptional read-through detected in our analyses, or they are not true terminators.
To complement our results, it would be beneficial to identify a promoter upstream of scyA, since one cannot be identified based on the genomic sequence. Since cyanobacterial promoter sequences are inconsistent and lack reliable consensus motifs (8), any promoter associated with this region will have to be determined experimentally in the future.
Our data suggest that each of the 18 genes in the main scytonemin cluster, as well as the 5 satellite genes, is involved in scytonemin biosynthesis, given their induction by UVA and previous analyses linking them to scytonemin biosynthesis. Further research should examine the transcriptional regulation of these regions. Perhaps NpF1277 and NpF1278 (sensor kinase and response regulator, respectively) are directly regulating their expression, or perhaps there are additional regulatory elements within the gene clusters (i.e., aroG). A thorough study of the promoter elements with further insight into the transcriptional terminators discussed above would also contribute to the general understanding of transcription initiation and termination in cyanobacteria. Furthermore, a follow-up study of the global UVA stress response using N. punctiforme genomic DNA microarrays is currently being undertaken in our laboratory. This additional study will further validate the results of this study and reveal other transcriptional responses to UVA radiation.
We thank Jack Meeks and Elsie Campbell for providing us with RNA extraction and cDNA synthesis protocols. We also thank Scott Bingham, Jeffrey Hock, and Cosmin Sicora for their qPCR support on this project and Kendra Palmer for assistance with various aspects.
We thank the Arizona State University Graduate and Professional Student Association for funding.
Published ahead of print on 8 May 2009.