Characterization of nifH gene expression, modification and rearrangement in Nodularia spumigena strain AV1

Authors


  • Editor: Riks Laanbroek

Correspondence: Rehab El-Shehawy, IMDEA-Water, C/Punto net 4, 28805 Alcala de Henares (Madrid), Spain. Tel.: +34 91 830 59 62; fax: +34 91 830 59 61; e-mail: rehab@imdea.org

Abstract

The annually reoccurring blooms that characterize the surface waters of the Baltic Sea are dominated by filamentous, heterocystous cyanobacteria such as Nodularia spumigena. In a previous study, we have demonstrated that N. spumigena strain AV1 differentiates heterocysts in the absence of detectable nitrogen fixation activity, an unusual physiological trait that is clearly distinct from other well-studied cyanobacteria. To further analyze the uncoupling between these two processes, we analyzed the gene expression and modification of the nitrogenase enzyme (the enzyme responsible for nitrogen fixation) in N. spumigena AV1, as well as in several other N. spumigena strains. Here, we demonstrate the occurrence of two nifH gene copies in N. spumigena strain AV1, only one of which is located in a complete nifHDK cluster and several NifH protein forms. Furthermore, we demonstrate the occurrence of a DNA rearrangement mechanism acting within the nifH gene copy located in the nifHDK cluster and present only in the strains exhibiting the previously reported uncoupling between heterocyst differentiation and nitrogen fixation processes. These data stress the existence of a distinct and complex regulatory circuit related to nitrogen fixation in this ecologically significant bloom-forming cyanobacterium.

Introduction

Noxious summer blooms are a common phenomenon in the Baltic Sea. Nodularia spumigena is one of the filamentous, heterocystous cyanobacteria that dominate these annual blooms. Together with Aphanizomenon, these photoautotrophic, nitrogen-fixing cyanobacteria contribute significantly to the primary productivity of the Baltic Sea by introducing two key nutrients (carbon and nitrogen) into the food chains of this brackish water ecosystem. The increasing summer blooms aggravate the effects of the ongoing eutrophication by increasing the nutrient load (Kahru et al., 1994; Finni et al., 2001).

The reduction of atmospheric nitrogen to ammonia is catalyzed by nitrogenase, an enzyme complex irreversibly inactivated by oxygen. The nitrogenase complex is a multimeric protein complex found in many bacterial groups. It is composed of two component metalloproteins, dinitrogenase encoded by nifDK and dinitrogenase reductase encoded by nifH, with the nif operon highly conserved between species (Buikema & Haselkorn, 1993; Flores & Herrero, 1994).

Heterocysts are specialized, terminally differentiated, unique cells in certain cyanobacteria that provide an anaerobic atmosphere for the nitrogenase enzyme complex (Fay, 1992; Adams & Duggan, 1999; Haselkorn, 2007). The development of heterocysts is a tightly regulated process that involves numerous genes and transcription factors as well as highly specific DNA rearrangements within the nifDK operon before transcription (Wolk et al., 1994; Carrasco & Golden, 1995; Wolk, 1996; Zhang et al., 2006). In addition to transcriptional control, the nitrogenase proteins are also modified post-translationally (Pope et al., 1985; Reich & Böger, 1989; Stal & Bergman, 1990; Ohki et al., 1992; Chow & Tabita, 1994). In Rhodospirillum rubrum (a purple bacterium) and some other species of nitrogen-fixing bacteria, nitrogenase activity is regulated by reversible ADP-ribosylation on one of the subunits of the Fe-protein in response to darkness or the presence of ammonium ions, the ‘switch-off’ effect (Nordlund & Ludden, 2004). A post-translational modification also exists in nitrogen-fixing cyanobacteria, such as Anabaena sp. PCC 7120, Nostoc PCC 73102 and Gloeothece (Durner et al., 1994; Gallon et al., 2000; Ekman et al., 2008). However, ADP ribosylation has not been proven as the modifying mechanism of nifH gene products in the cyanobacterium Anabaena sp. PCC 7120, although no other modification mechanism was proposed (Ernst et al., 1990). Later, palmitoylation was proposed as a modification mechanism in the unicellular cyanobacterium Gloeothece sp. (Gallon et al., 2000).

A recent study by Boström et al. (2007) showed that the majority of nifH transcripts detected in the Baltic Sea proper belonged to Aphanizomenon and Nodularia, indicating that these are the dominating nitrogen fixers. It is therefore of great ecological interest to study the regulation of nitrogen fixation in these dominating species. Recent findings from our laboratory demonstrated that N. spumigena strain AV1 continues to differentiate heterocysts in the absence of detectable nitrogen fixation. This unusual behavior of N. spumigena strain AV1, as compared with other well-studied heterocystous cyanobacteria (such as Nostoc PCC 7120, Anabaena variabilis and Nostoc punctiforme), indicates an uncoupling between nitrogen fixation and heterocyst differentiation (Vintila & El-Shehawy, 2007) and may be of significance in our understanding and the future management of the Baltic Sea. To further analyze this unusual behavior, we have analyzed the regulation of nitrogenase in N. spumigena strain Av1 at various organization levels. Here, we demonstrate the occurrence of two nifH genes, and three forms of NifH protein as well as DNA rearrangement within one of the nifH genes.

Materials and methods

Growth conditions and nitrogen supplementation

Axenic batch cultures of N. spumigena strain AV1, isolated from the Baltic Sea, were grown in 500 mL Z8XN0 (nitrogen-free) medium (Sivonen et al., 1989) on an orbital shaker. Cultures were grown at 20 °C and at an irradiance of 45 μmol photons m−2 s−2 (white fluorescent light) on a 16 : 8 light : dark cycle. Other strains of N. spumigena (Table 1) were grown in 50 mL Z8X with or without nitrogen supply under the abovementioned growth conditions.

Table 1. Nodularia spumigena strains used for DNA extraction and PCR of nifH1
Nodularia spumigena strainToxicitySite of isolationIsolated by
KAC7ToxicKalmarsund, Baltic SeaEsplund C. (2000)
KAC66ToxicAskö, Baltic SeaGisselson L.-Å. (1996)
KAC71ToxicMörtviken, Baltic SeaEsplund C. (2002)
NSGG-1ToxicGulf of GdanskMisiał A. (1997)
NSZG 0205ToxicGulf of GdanskKobos J. (2005)
NSOR10ToxicOrielton lagoon, Tas., AustraliaBlackburn S. (1993)
NSBL05NontoxicLake Bullenmeri, Vic., AustraliaBlackburn S. (1993)

Chemostat cultures were grown in 1-L PC flasks (Nalgene) in a continuously supplied Z8XN0 medium, at a flow rate of 0.3 mL min−1, and bubbled with filtered air provided by an aquarium pump. Cell density was determined by fluorometric measurements of chlorophyll a (chl a) at 665 nm as described previously (Meeks & Castenholz, 1971). When the cultures reached a steady state of cell growth, as indicated by chl a measurements, the cultures were supplied with Z8X containing 1 mM NH4Cl. When cells stopped fixing nitrogen, the cultures were filtered and resuspended into Z8XN0. Additional strains used in this study are listed in Table 1.

Nitrogenase activity measurements

Nitrogenase activity was measured using the acetylene reduction assay (Capone, 1993), with modifications as described previously (El-Shehawy et al., 2003).

DNA and RNA isolation

Isolation of genomic DNA from N. spumigena filaments was performed using phenol as described previously (Wilson, 1998). The preparation of genomic DNA from heterocysts of N. spumigena was performed as described previously (Carrasco & Golden, 1995).

Isolation of RNA was carried out using the RNA mini prep kit (Qiagen) as described previously (Vintila & El-Shehawy, 2007). Total DNA/RNA were quantified using the NanoDrop® ND-1000 Spectrophotometer (NanoDrop Technologies).

For cDNA synthesis, 500 ng RNA from each sample was used using the iScript cDNA Synthesis Kit (Bio-Rad) and following the manufacturer's instruction.

PCR

The primers used in this study (Table 2) were designed using the primer 3 software (http://workbench.sdsc.edu).

Table 2.   Primer sequences used in this study
Primer namePrimer sequence (5′→3′)
nifH1FwCATCCTGATCGTTGG
nifH1ReGTTAGCTTGGTTGCT
nifH2FwGCGATGTCCGTTGTG
nifH2ReCGCTTTACCTTCTCT

To verify primer specificity, PCR products were sequenced (DNA Technology, Denmark). The resulting sequences were analyzed using blastn (http://www.ncbi.nlm.nih.gov/BLAST) (Altschul et al., 1997) and deposited in GenBank under accession numbers GQ456132 and GU062791.

Semi-quantitative PCR was performed on total genomic DNA and DNA from heterocysts using equal amounts of DNA for each reaction, only running the PCR reaction for 24 cycles and ensuring that amplification of the 16S rRNA gene resulted in equal amounts of product. The PCR program used was as follows: 95 °C for 15 min, 24 cycles of 95 °C for 30 s (using 16S primers)–45 s (using nifH1 and nifH2 primers; see Table 2), 56 °C for 30–45 s and 72 °C for 30–45 s, followed by a final elongation step of 10 min at 72 °C.

Real-time PCR was performed in duplicate in an iCycler Real-time PCR machine (Bio-Rad) using the iQ™ SYBR® Green Supermix and the QuantiTect SYBR Green PCR Kit (Bio-Rad and Qiagen, respectively). The primers used to amplify 16S and nifH1 were as described previously (Vintila & El-Shehawy, 2007).

Standard curves were generated using genomic DNA from N. spumigena strain AV1 in a series of 10-fold dilution. The cDNA quantities of nifH and the 16S rRNA gene were determined using the standard curves. The nifH quantities were normalized by the relative cDNA quantities of the 16S rRNA gene.

Protein extraction procedures

Samples for sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and subsequent Western blot were collected by filtration (Whatman, PC filters 5.0 μm) and resuspended in 500 μL of Laemmli buffer (Laemmli, 1970) containing 1 tablet of protease inhibitor cocktail (Roche Diagnostics). Samples were stored at −80 °C until further analysis.

Proteins were extracted by grinding the samples with a plastic pestle in tubes containing acid-washed glass beads (Sigma, 212–300 μm), followed by heating the sample to 99 °C for 5 min. The samples were then centrifuged for 15 min at 15 000 g.

Samples for two-dimensional (2D) gel electrophoresis were collected by filtration (Whatman, PC filters 5.0 μm), resuspended in a rehydration buffer containing 7 M urea, 2 M thiourea, 4% CHAPS, 1% ABS-14, 1% IPG buffer 4–7 (GE Healthcare), 50 mM dithiothreitol and protease inhibitor cocktail tablets (1 tablet/10 mL buffer) (Roche Diagnostics) and frozen in liquid nitrogen. Samples were stored at −80 °C until further analysis.

Cell lysis was performed by grinding the samples in liquid nitrogen four to five times. Lysis was verified by light microscopy. The lysate was centrifuged at 15 000 g for 15 min to remove cell debris.

Samples for anaerobic extraction of proteins (to protect the nitrogenase enzyme from degradation under aerobic conditions) were collected by filtration (Whatman, PC filters 5.0 μm), frozen in liquid nitrogen and stored at −80 °C until further analysis. Cells were resuspended in 10 mL of degassed Tris buffer (100 mM) containing 1 mM of phenylmethylsulfonyl fluoride (Boehringer-Mannheim) and 1 tablet of protease inhibitor cocktail (Roche Diagnostics). Lysis was performed by passing the cell suspension four times through a French pressure cell at 18 000 psi to ensure the total breakdown of heterocysts. The lysate was centrifuged at 15 000 g for 15 min to remove cell debris. Proteins were precipitated using four volumes of acetone overnight at −20 °C and then centrifuged for 20 min at 15 000 g. The protein pellet was washed twice with 90% acetone, air-dried and resuspended in an appropriate amount of rehydration buffer or Laemmli buffer.

Protein fractionation was performed as described previously (Ran et al., 2007).

All protein concentrations were determined using the RC DC Protein Assay kit (Bio-Rad).

2D gel electrophoresis

Twenty micrograms of protein was loaded by cup loading onto the acidic ends of rehydrated 7-cm Immobiline gel strips (pH 4–7, Amersham/GE Healthcare) and the isoelectric focusing of the proteins was performed on an IPGphor system (GE Healthcare). The focusing time was adjusted to a total of 8000 Vh. For gel strips of 18 cm, 360 μg of protein was loaded by cup loading and the focusing time was adjusted to a total of 75 000 Vh.

Equilibration steps and second-dimension electrophoresis were carried out as described previously (Ran et al., 2007).

The gels were either stained with Blue Silver as described previously (Candiano et al., 2004) or transferred to membranes for subsequent Western blot analysis.

Western blot

The Western blot procedure was carried out as described previously (Braun-Howland et al., 1988). The membranes were incubated for 1 h with polyclonal anti-dinitrogenase reductase from R. rubrum raised in a rabbit at a 1 : 5000 dilution in phosphate-buffered saline (PBS)–Tween. Membranes were then incubated for 1 h with the secondary antibody (affinity-purified polyclonal pork anti-rabbit/HRP antibody; Dako) at a 1 : 5000 dilution in PBS–Tween. Detection was performed using the ECL Plus system (Amersham/GE Healthcare) according to the manufacturer's instructions. Visualization was performed on the Chemidoc system (Bio-Rad). The same procedure was applied to the Western blot against D2 protein (originally from spinach), raised in a rabbit at a dilution of 1 : 2500.

In-gel digestion and MS

Protein identification was performed by Alphalyse A/S (Denmark) using MS peptide mapping (MALDI-TOF/MS) and sequence analysis (MALDI-TOF/TOF).

Protein spots of interest were digested with trypsin and the resulting peptides were concentrated on ZipTip micropurification columns. Samples were eluted onto an anchorchip target for analysis on a Bruker Autoflex III MALDI-TOF/TOF instrument. The peptide mixtures were analyzed in the positive reflector mode and 5–10 of the peptides in each sample were selected for analysis by MS/MS fragmentation for partial peptide sequencing.

The MS and MS/MS spectra for each sample were combined and used for database search (NRDB1, 7155275 protein sequences) using the mascot software version 2.2.03 (http://www.matrixscience.com). Search was performed using the following criteria: 1 missed trypsin cleavage allowed, 60 p.p.m. mass tolerance, cysteine carbamidomethylation and methionine oxidation as variable modifications (see also Supporting Information) (Vintila & El-Shehawy, 2010).

Results and discussion

Samples of N. spumigena strain AV1 (hereafter termed Nodularia) were collected every fourth hour along the diel cycle. As can be seen in Fig. 1a, there was a rapid increase in nitrogenase activity at the onset of the light period and the highest activity was found about 8 h into the light cycle, whereas the lowest activity was recorded about 4 h into the dark cycle. This is in agreement with previous findings for nitrogen-fixing cyanobacteria from the Baltic Sea (Evans et al., 2000; Stal et al., 2003; Moisander et al., 2007). The expression of nifH in Nodularia was light dependent and strongly induced during the first light hours (Fig. 1b), with the peak in expression preceding that of the nitrogenase activity.

Figure 1.

 Diel patterns related to nitrogen fixation in Nodularia spumigena strain AV1. (a) Nitrogenase activity measured by the acetylene reduction (AR) assay; (b) nifH expression measured by real-time reverse transcription-PCR; (c) Western blot against the NifH protein; (d, e) Western blot of small (7 cm) 2D SDS-PAGE gels against NifH at peak activity (16:00 hours) (d) and lowest activity (04:00 hours) (e). Error bars denote SE (n=3).

To investigate whether the NifH protein was inactivated or degraded in darkness, samples were collected for one-dimensional (1D) SDS-PAGE and subsequent Western blot analysis. Two forms of NifH were clearly apparent, but their relative protein levels did not vary along the diel cycle (Fig. 1c). Next, samples collected at the lowest and the highest nitrogenase activities, respectively, were subjected to 2D SDS-PAGE (7 cm gels) and subsequent Western blot analysis (Fig. 1d and e). This revealed that two protein spots were recognized by the anti-NifH antibody, being of similar masses (∼32.5 kDa), but with different isoelectric (pI) points, and that these were differentially abundant under light and darkness regimes. However, there was no detection of the heavier NifH form of ∼35 kDa. To resolve this discrepancy, a larger (18 cm) gel was immunoblotted after a 2D SDS-PAGE separation of the Nodularia protein extract collected at the beginning of the light phase (see Fig. 1). As can be seen in Fig. 2a, three forms of NifH were now detected, likely due to the better resolution capacity of the proteins on the larger gel and the higher protein amount. The identity of the three spots was confirmed as NifH by MALDI-TOF/TOF, corresponding to the same gene. Two of the NifH spots were of a similar molecular weight (∼32.5 kDa), but with different pI (as on the smaller gel), while a third spot had a higher molecular mass (∼35 kDa). It is therefore apparent that the single, lighter protein band (32.5 kDa) observed using 1D SDS-PAGE (Fig. 1c) is actually composed of two forms with a minor difference in molecular weight, but different pI, whereas the heavier NifH protein band in the 1D-gel (35 kDa) consists of a single NifH form and does not respond to differences in light regimes (data not shown).

Figure 2.

 NifH forms and localization in Nodularia spumigena strain AV1. (a) Western blot of a large (18 cm) 2D SDS-PAGE gel against NifH. A sample was taken just after the onset of light. (b) Western blot of 1D SDS-PAGE gel against NifH after protein fractionation. Lane 1, soluble fraction; lane 2, membrane fraction; and lane 3, total protein. (c) Western blot against D2 protein of PSII. Lane 1, soluble fraction; lane 2, membrane fraction; and lane 3, total protein.

Efforts were also made to identify the post-translational modification that occurs on NifH and even though the NifH protein modification was not identified, we were able to rule out protein phosphorylation using three different methods, including MS (data not shown).

The slower migrating and heavier form of NifH in the 1D-gels was less abundant in the 2D SDS-PAGE gels and therefore we investigated the possibility that it may be membrane associated. For this purpose, a total protein fraction, collected at the highest nitrogenase activity, was separated into membrane-bound and soluble fractions. As can be seen in Fig. 2b, the lighter NifH form was predominantly found in the soluble fraction (lane 1) while the heavier form of NifH dominated in the membrane fraction (lane 2). The purity of the fractions was verified using Western blot analysis and an antibody raised against the photosystem II/D2 protein (Fig. 2c). In addition, anaerobic extraction of NifH was used to exclude the possibility of degradation (data not shown).

These results could either imply that two of the NifH forms of the N. spumigena nitrogenase are modified post-translationally, one of the subunits being constitutively modified to be membrane associated, or that the expression of the nifH gene may result in two different NifH forms, of which only one is post-translationally modified. Membrane association of the nitrogenase Fe-protein may be debatable since previous studies, using electron microscopy, have shown a random distribution of NifH throughout the heterocyst cell compartment, although localization close to the numerous membranes filling up the compartment cannot be excluded (Stal & Bergman, 1990; Colón-López et al., 1997). It is possible that the function of the observed modification is, as opposed to the effect in other bacteria (Nordlund & Ludden, 2004), to protect nitrogenase from inactivation/degradation. The activity of nitrogenase in heterocystous cyanobacteria is dependent on light for electron transfer and, to a certain extent, ATP production. Exposure to darkness limits the electron flow from ferredoxin to nitrogenase, lowering the nitrogenase activity (Evans et al., 2000). Therefore, theoretically, there would be no need for inactivation of nitrogenase and modification of NifH might serve another function. It has been speculated that the modification of nitrogenase in cyanobacterial heterocysts may serve as a target signal for degradation or to prevent oxygen-induced damage of the protein (Smith et al., 1987; Chow & Tabita, 1994).

To investigate whether any post-translational modification of NifH occurs upon nitrogen supplementation, ammonium ions (1 mM) were added to the Nodularia cultures. Figure 3 shows that nitrogenase activity is reduced by ammonium within 6 days after the addition, probably not due to post-translational modification, but rather due to the decrease in nifH transcription detected (Fig. 3a and b). 1D- and 2D-gel electrophoresis (7 cm gel) and Western blot analysis of proteins from ammonium-supplemented cultures showed a gradual decrease in the total NifH protein abundance rather than a switch to a modified form (Fig. 3d and f). This is also in agreement with previous studies, indicating that there is no ‘switch-off’ effect of nitrogenase upon addition of combined nitrogen (Yoch & Gotto, 1982; Ramos & Guerrero, 1983). All forms of NifH decreased to the same extent, suggesting that these forms are all part of the active nitrogenase.

Figure 3.

 Characterization of nitrogenase expression during ammonium supplementation. (a) Nitrogenase activity measured by the acetylene reduction (AR) assay, (b) nifH1 expression pattern measured by real-time reverse transcription (RT)-PCR, (c) nifH2 expression pattern measured by real-time RT-PCR, (d) Western blot against NifH protein, (e, f) Western blot of small (7 cm) 2D SDS-PAGE gels against NifH before (e) and 1 day (f) after ammonium supplementation. Error bars denote SE (n=3).

The N. spumigena genome has been sequenced recently (available at http://img.jgi.doe.gov), but a database search based on our MS data only resulted in partial NifH sequences from N. spumigena. Gene information from Integrated Microbial Genomes (http://img.jgi.doe.gov) revealed the occurrence of two nifH genes in N. spumigena. One of the nifH genes (from here on called nifH1) is localized upstream of a complete nif-gene cluster that includes nifDK and appears to be interrupted by a 5210 bp DNA insertion element containing three ORFs (Fig. 4a), whereas the other nifH copy (from here on called nifH2) is intact, but exclusive and does not belong to a nif-gene cluster (Fig. 4b). It therefore appears that the interrupted nifH1 (annotated on the genome draft as two genes gi:119509285; gi:119509281) is the gene being transcribed and active under nitrogen-fixing conditions. Real-time reverse transcription-PCR verified this hypothesis and showed that nifH2 (gi:119509973) transcription levels were about 250 times lower (using 1 μg of cDNA as a starting quantity for the PCR reactions) than those of nifH1. Moreover, the transcription of nifH2 did not respond to nitrogen supplementation or removal (Fig. 3c) as clearly as nifH1 (Fig. 3b). It is known that some heterocystous cyanobacteria contain an alternative Mo-nitrogenase (encoded by the nif2 operon) that is transcribed in both vegetative cells and heterocysts under anaerobic conditions and/or an alternative V-type nitrogenase (encoded by the vnf operon) that is transcribed under conditions of molybdenum deficiency in heterocysts (Thiel, 1993; Thiel et al., 1995, 1997; El-Shehawy & Bergman, 2003). However, the second copy of nifH of N. spumigena (nifH2) does not encode an alternative nitrogenase similar to nif2 or vnf, as it is not localized in a nifHDK cluster, and is transcribed at a very low concentration regardless of the external availability of nitrogen. Whether or not nifH2 is involved in nitrogen fixation under other growth conditions does, however, require further investigation. It is worth mentioning that because nifH1 is interrupted by a 5210 bp insertion element, primers targeting N. spumigena nifH in environmental samples will also hybridize to nifH2 if the primers are complementary to the nifH1 sequences overlapping the insertion element. A blastn search showed that some of the nifH sequences of N. spumigena deposited in the GenBank actually belong to nifH2 and not nifH1 of the complete nifHDK cluster (data not shown).

Figure 4.

 Gene organization of the two nifH copies found in Nodularia spumigena strain CCY9414. (a) Organization of the interrupted nifH gene (nifH1) and (b) the second nifH gene (nifH2) in N. spumigena CCY9414 as reported by Integrated Microbial Genomes (http://img.jgi.doe.gov).

Analysis using blastp revealed that the DNA insertion element interrupting nifH1 is similar in organization to the genetic elements present in nifD genes of other cyanobacteria (Golden et al., 1985). To investigate the possibility that a DNA rearrangement occurs during heterocyst differentiation also in Nodularia, semi-quantitative PCR was performed on DNA from isolated heterocysts, total DNA from intact filaments (Fig. 5a) and on cDNA (Fig. 5b). The data obtained indicate that the intervening DNA element is absent in the heterocyst DNA, suggesting a DNA rearrangement during differentiation.

Figure 5.

 Semi-quantitative PCR of nifH1 from genomic DNA and cDNA of Nodularia spumigena strain AV1 and genomic DNA of seven additional Nodularia spumigena strains. (a) Semi-quantitative PCR of nifH1 with primers spanning the element. Lane 1, DNA ladder; lane 2, genomic DNA from isolated heterocysts; and lane 3, genomic DNA from whole filaments from N. spumigena strain AV1. (b) Semi-quantitative PCR of nifH1 with primers spanning the element. Lane 1, DNA ladder; lane 2, cDNA from nitrogen-fixing filaments; and lane 3, cDNA from cells grown 6 days on 1 mM ammonium from N. spumigena strain AV1. (c) PCR of genomic DNA from different N. spumigena strains. Lanes 1 and 2, whole filament and vegetative cell DNA, respectively, from strain NSOR10; lanes 3 and 4, whole filament and vegetative cell DNA, respectively, from strain NSGG-1; lane 5, whole filament DNA of strain KAC7; lane 6, whole filament DNA of strain KAC66; lane 7, whole filament DNA of strain KAC71; lane 8, whole filament DNA of strain NSZG0205; lane 9, whole filament DNA of strain NSBL-5; lane 10, negative control; and lane 11, DNA ladder.

One of the ORF within this DNA element is annotated as a phage integrase (gi:119509282). Sequence homology search using blastp revealed that this protein is most similar in sequence to XisA (accession number: YP_001864116) from N. punctiforme PCC 73102. The XisA protein is a site-specific recombinase responsible for the excision of the nifD DNA element in Nostoc and Anabaena (Lammers et al., 1986; Brusca et al., 1989; Brusca et al., 1990). Therefore, the product of the gene annotated as phage integrase is likely responsible for the excision of the nifH DNA insertion element in N. spumigena during the differentiation of heterocysts.

As a support for a gene rearrangement event, the PCR product obtained for nifH1 was about 40 bp longer than it should have been fusing the annotated gene sequences for nifH of N. spumigena CCY9414. We sequenced the nifH1 gene product from N. spumigena strain AV1, translated the sequence and aligned it with NifH protein sequences from N. spumigena CCY9414, Nostoc, Anabaena and Cyanothece (Fig. 6). From the alignment, it seems that the N. spumigena CCY9414 NifH lacks 13 amino acids in a conserved part of the protein (marked in red). However, when the N. spumigena CCY9414 insertion element is translated in silico, these 13 amino acids are actually present towards the C-terminal part of the insertion element and may thus be an annotation error in the genome draft (data not shown).

Figure 6.

 Alignment of the NifH protein sequence from different cyanobacterial species. 1, Cyanothece sp. PCC 7425; 2, Nodularia spumigena AV1; 3, Nostoc sp. PCC 7120; 4, Nostoc punctiforme PCC 73102; 5, N. spumigena CCY9414; 6, Anabaena variabilis ATCC 29413. The NifH1 protein sequence of N. spumigena strain CCY9414 was obtained by combining the two separately annotated NifH protein sequences (gi:119509285 marked in blue; gi:119509281 marked in purple) while the NifH1 protein sequence of N. spumigena strain AV1 was obtained by translating the sequenced gene product of nifH1.

To further investigate the presence of the DNA element in N. spumigena, we performed PCR using genomic DNA isolated from whole filaments of seven other N. spumigena strains (Table 1). As can be seen in Fig. 5c, the nifH1 insertion element is not present in all N. spumigena strains investigated. Nodularia spumigena strains NSOR10 and NSGG-1 lost heterocysts when cultures were supplemented with ammonium and therefore DNA from filaments comprised solely of vegetative cells could be subjected to PCR. Both strains were found to lack the nifH1 DNA insertion element. Interestingly, the nifH1 DNA insertion element was found to be present in all the tested strains that exhibited the uncoupling behavior between nitrogen fixation and heterocyst differentiation characteristic of N. spumigena strain AV1, as reported previously (Vintila & El-Shehawy, 2007; Vintila & El-Shehawy, 2010). It is therefore tempting to speculate that there might be a possible connection between the presence of the nifH1 insertion element and the observed uncoupling between nitrogen fixation and heterocyst differentiation in N. spumigena. Numerous reports indicate that blooms of N. spumigena form when the Baltic Sea is nitrogen limited and seem to be more influenced by temperature, salinity and phosphorus rather than the availability of combined nitrogen (Granéli et al., 1990; Kononen, 1992; Pliński & Jóźwiak, 1999; Kiirikki et al., 2001; Moisander et al., 2003; Stal et al., 2003; Wasmund & Uhlig, 2003; Vurio et al., 2005).

In summary, the data clearly show that (1) the Baltic Sea N. spumigena strain AV1, as well as other Baltic Sea strains, possesses two copies of nifH, but that only one copy is actively expressed under nitrogen-fixing conditions and that its expression level was modulated by the addition of an external nitrogen source. (2) Nodularia spumigena expresses three forms of NifH; one form is heavier in mass, while the two others are similar in mass, but differ in their isoelectric focusing point. The modification of the lighter form is regulated by a light/dark signal, but not by nitrogen supplementation, while the heavier form appears to be constantly modified and potentially membrane bound. (3) nifH in N. spumigena strain AV1, as well as in the other Baltic Sea strains tested, is interrupted by a DNA insertion element that is excised during heterocyst development. Our findings clearly demonstrate that N. spumigena adapts a distinct regulatory circuit related to nitrogen fixation. Understanding of such a circuit is required to understand the physiological traits of N. spumigena in terms of growth and nitrogen fixation. Such an understanding, combined with our existing knowledge of the environmental factors that may promote bloom formation, is important for the development of efficient and sustainable plans to manage the Baltic Sea summer blooms.

Acknowledgements

This work was supported financially by the Swedish Research Council FORMAS, the Royal Swedish Academy of Science, StockholmMarine Research Centre (SMF) and Knut and Alice Wallenberg's Foundation.

Ancillary