Cyanophycin Production in a Phycoerythrin-Containing Marine Synechococcus Strain of Unusual Phylogenetic Affinity

Abstract

Thirty-two strains of phycoerythrin-containing marine picocyanobacteria were screened for the capacity to produce cyanophycin, a nitrogen storage compound synthesized by some, but not all, cyanobacteria. We found that one of these strains, Synechococcus sp. strain G2.1 from the Arabian Sea, was able to synthesize cyanophycin. The cyanophycin extracted from the cells was composed of roughly equimolar amounts of arginine and aspartate (29 and 35 mol%, respectively), as well as a small amount of glutamate (15 mol%). Phylogenetic analysis, based on partial 16S ribosomal DNA (rDNA) sequence data, showed that Synechococcus sp. strain G2.1 formed a well-supported clade with several strains of filamentous cyanobacteria. It was not closely related to several other well-studied marine picocyanobacteria, including Synechococcus strains PCC7002, WH7805, and WH8018 and Prochlorococcus sp. strain MIT9312. This is the first report of cyanophycin production in a phycoerythrin-containing strain of marine or halotolerant Synechococcus, and its discovery highlights the diversity of this ecologically important functional group.

Many forms of cyanobacteria store nitrogen as cyanophycin (CP; multi-l-arginyl-poly-[l-aspartic acid]), a polymer generally composed of equal amounts of arginine and aspartate (3, 22). CP is stored as intracellular granules, which have a characteristic fine structure (6, 15). In cells that produce CP, its synthesis is usually induced when growth is limited by some factor other than nitrogen availability (5, 27). Under phosphorus or light limitation, CP is degraded after nitrogen becomes limiting (4, 6). Furthermore, CP is not ribosomally synthesized, and the addition of ribosomal inhibitors like chloramphenicol can induce CP production (23).

The ability to store nitrogen as CP is unique to cyanobacteria, but not all cyanobacteria have this capability (3). CP production is common among nitrogen-fixing cyanobacteria and might be expected in taxa that experience a variable environment with respect to nitrogen availability. Marine picocyanobacteria in particular would appear to benefit from having the ability to store nitrogen in CP granules because they are ubiquitous in nutrient-poor water, where storms or other mixing events can lead to transient periods of enhanced nitrogen availability.

Picocyanobacteria in the form-genus Synechococcus represent a genetically diverse (28, 29, 39) and abundant (16, 32) component of the picoplankton community in the marine environment. Members of the group are found in estuaries, coastal waters, and the open ocean (11, 32). Nearly all cells in the open ocean and continental shelf produce phycoerythrin (PE), a pink or reddish pigment that enhances their ability to utilize the wavelengths of light that are available in these waters (11, 32, 37). Relatively little is known about the distribution of CP production among the marine Synechococcus. Of the five obligately marine or halotolerant strains of Synechococcus that have been tested, two are capable of CP production (13, 20, 27, 31). These strains were isolated from coastal plankton (PCC73109) or the coastal benthos (PCC7002); neither contains PE. In strains that contain PE, it is possible that CP production would not be selectively favored, since PE may play a specialized role as a nitrogen storage compound (40, 41).

In this study, we screened 32 clonal isolates of PE-containing marine Synechococcus by using a standard arginine assay for CP. We also sequenced 950 bp of 16S ribosomal DNA (rDNA) from the one strain that tested positive in order to determine its phylogenetic position among the cyanobacteria and to examine its relationship to other marine picocyanobacteria.

MATERIALS AND METHODS

Screening for CP production.

Thirty-two clonal isolates of PE-containing picocyanobacteria conforming to the form-genus Synechococcus (11) were used in this study. Fifteen isolates were isolated from a coastal station in the Black Sea near Trabzon, Turkey (30), and 17 were isolated from samples collected in the Arabian Sea during the Southwest monsoon along the track of cruise TN048 of the R/V Thomas G. Thompson (38). All strains synthesize a phycourobilin-lacking form of PE when grown in white light, and the majority of strains were cocci or short rods no more than 1.5 μm in their longest dimension (36). Cultures to be screened for CP production were grown in natural seawater medium with f/2 nutrient concentrations until they reached the late exponential phase. At this point, 15 mg of water-soluble chloramphenicol ml⁻¹ was added to the cultures. The cells were harvested by centrifugation about 48 h after addition of the chloramphenicol and assayed for the presence of CP by the arginine spectrophotometric assay (described below).

CP production in strain G2.1.

Three different methods were used to verify CP production in strain G2.1: the arginine spectrophotometric assay, amino acid analysis of the material in the CP fraction, and examination of the cells by transmission electron microscopy (TEM). Production of CP was induced by a nitrogen depletion and recovery experiment, which is known to yield high CP concentrations in cells that can synthesize CP (4, 17) (Fig. 1).

FIG. 1.

Experimental design and growth of each treatment showing culture condition at the time of key manipulations. The experiment was performed twice. Temperature and light were optimized for maximum growth (30°C, 115 microeinsteins m⁻² s⁻¹, 14 h:10 h light-dark), and the inoculum was acclimated to experimental conditions before time zero (36). Cultures were grown in ASPM artificial seawater with f/2 levels of nutrients (10), with the exception of nitrate, which was reduced to 75 mg liter⁻¹ in the exponential control. The nitrate cadmium reduction test (no. 957-26; standard methods, Environmental Protection Agency) (33) was used to determine nitrate concentrations. Fifteen milligrams of water-soluble chloramphenicol ml⁻¹ was added to the cultures as indicated on the graphs. Electron micrographs were taken of cells in the experimental treatment and the exponential control. CAP, chloramphenicol; N, nitrate.

Arginine assay.

Cells were harvested by centrifugation at 27,000 × g for 15 min. Pelleted cells were then broken by mortar and pestle in liquid nitrogen, and the resulting material was washed by the method of Simon (23). This procedure results in a 0.1 N HCl cell extract that is expected to be predominantly arginine and aspartate when CP is present in the cell sample, and a mixture containing relatively small amounts of amino acids when CP is not present (23). The amount of arginine in the washed material was then spectrophotometrically assayed as an indication of the amount of CP that was present by a modification of the Sakaguchi reaction (18).

Amino acid analysis.

Material was prepared for the amino acid analysis by breaking and washing the cells as done for the arginine assay. The samples were hydrolyzed at 115°C for 20 h. The amino acid analysis was conducted by AAA Service Laboratory, Boring, Oreg.

TEM.

Cells were prepared for TEM by a modification of the technique described by Stanier (26). Cells were prefixed in 0.5% (wt/vol) glutaraldehyde in 100 mM cacodylate buffer (pH 7) and then concentrated by centrifugation and mixed with 3% melted agar. The agar was cut into small blocks and left overnight in the glutaraldehyde cacodylate solution. Cells were rinsed in cacodylate buffer and fixed in 1% (wt/vol) osmium tetroxide in the same buffer for 1.5 h at room temperature. Sodium chloride was added to the glutaraldehyde, osmium tetroxide, and cacodylate solutions until the salinity of the solutions reached 3.5 g liter⁻¹ as measured on an ATAGO hand refractometer. Cells were dehydrated in an ethanol series and embedded in Spurrs resin (25). Thin sections were cut with a glass knife on an LKB Ultrotome and poststained with uranyl acetate and lead citrate. Micrographs were taken at 60 kV on a Philips CM12 electron microscope.

DNA isolation, amplification, and sequencing.

G2.1 genomic DNA was isolated by the method of Pitcher et al. (21), and a fragment of the 16S rRNA gene spanning Escherichia coli nucleotide positions 360 to 1335 was amplified as described by Miller and Castenholz (19). The amplification product was cleaned with the QIAquick-spin PCR purification kit (Qiagen) and then directly sequenced with an ABI Prism 377.

Sequence alignment and phylogenetic analyses.

A pairwise alignment of 16S rRNA gene sequence data for 30 cyanobacteria, including Synechococcus sp. strain G2.1 and three outgroup taxa (Bacillus subtilis, E. coli, and Aquifex pyrophilus) was obtained with MALIGN 2.77 (34) as described by Miller and Castenholz (19). This procedure incorporated secondary structural information for the mature 16S rRNA molecule into the alignment and imposed a gap cost-to-substitution cost of 3.

Phylogenies were reconstructed initially with the 30 aligned cyanobacterial sequences and the sequences from E. coli, B. subtilis, and A. pyrophilus; a simpler phylogeny was subsequently constructed based on 14 of the aligned cyanobacterial sequences and the B. subtilis sequence. A neighbor-joining phylogeny was inferred by using MEGA (version 1.01; S. Kumar et al., The Pennsylvania State University, University Park). It was constructed with a distance matrix calculated according to Kimura's (14) two-parameter model with a transition/transversion ratio of 2. Maximum-parsimony and maximum-likelihood analyses were performed with PAUP* (version 4b4; D. Swofford, Sinauer Associates, Sunderland, Mass.). The former employed a heuristic search by using the tree bisection-reconnection branch-swapping algorithm, with starting trees obtained by stepwise sequence addition and 10 replications of random sequence addition. The maximum-likelihood phylogeny was constructed by using Felsenstein's (9) two-parameter model for unequal base frequencies, a transition/transversion ratio of 2, and sequential addition of sequences. The neighbor-joining and maximum-parsimony phylogenies were bootstrap pseudoreplicated 1,000 times, whereas the maximum-likelihood phylogeny was bootstrap replicated 100 times. All phylogenies in the simpler analysis were rooted with B. subtilis.

Nucleotide sequence accession number.

The Synechococcus sp. strain G2.1 sequence has been submitted to GenBank and can be found under accession no. AY054298. The strain has been provided to the Culture Collection for Microorganisms from Extreme Environments at the University of Oregon (http://cultures.uoregon.edu) and can be ordered under strain identification no. CCMEE 5517.

RESULTS

Screening for CP production.

All 32 strains grew well in the experimental medium and yielded adequate cell material for analysis. Detectable amounts of arginine were only found in the CP cell fraction extracted from Synechococcus sp. strain G2.1; the same fraction extracted from all other strains contained no detectable amounts of arginine. Cells of Synechococcus sp. strain G2.1 are nearly isodiametric rods <2 μm in width; they are slightly larger than the cells of most other strains in the collection. Synechococcus sp. strain G2.1 was isolated from the Arabian Sea; the other strains that did not appear to make CP included two Arabian Sea strains very similar to G2.1 in size, which were isolated from the same original sample. Synechococcus sp. strain G2.1 does not fix nitrogen under aerobic growth conditions.

Experimental confirmation of CP production in Synechococcus sp. strain G2.1.

All inocula grew well through the first 100 h of the experiment, reaching the carrying capacity for the medium at ∼8 × 10⁶ cells ml⁻¹ after about 125 h (Fig. 1). Nitrate levels, measured at the beginning and end of the experiment, confirmed that nitrate was present in the cultures at the beginning of growth and depleted to undetectable levels by 150 h. Nitrate was readily detectable in the exponential control at the time of cell harvest. Arginine levels measured in the CP fraction of harvested cells were significantly higher in the experimental cultures than in any control cultures (P > 0.01 by Tukey-Kramer test for honestly significant difference) (Fig. 2).

FIG. 2.

Arginine content in the CP fraction extracted from cell pellets expressed on a per cell basis. Numbers indicate the mean of three cultures in each treatment; error bars represent 2 standard errors. Black bars, first replicate; gray bars, second replicate.

Based on quantitative amino acid analysis, the cell material recovered as the CP fraction from the experimental treatment was comprised of approximately equimolar quantities of arginine and aspartic acid (29 and 35 mol%, respectively). It also had an appreciable amount (15 mol%) of glutamic acid. The remaining 21 mol% was from background amounts of a wide variety of other amino acids. The exponential control showed only trace levels of amino acids.

The electron micrographs show that cells from the experimental culture, but not the exponential control, have prominent intracellular inclusions (Fig. 3). These inclusions have all of the characteristics of CP granules; they are spherical with irregular parameters, are not membrane bound, and are electron dense with a granular appearance (6, 15).

FIG. 3.

Electron micrographs of the experimental (A and B) and exponential (C and D) treatments. Arrows point to CP granules. Scale bars, 0.5 μm.

Phylogenetic analysis of strain G2.1.

Comparison of ca. 950 bp of the G2.1 16S rDNA sequence with the GenBank database by BLAST analysis yielded a closest match with the filamentous cyanobacterium Leptolyngbya foveolarum Komarek 1964/112 (BLAST score = 1,242; 91% sequence identity). A neighbor-joining phylogeny indicated that G2.1 formed a well-supported clade with this strain, along with Leptolyngbya sp. strain PCC73110, Phormidium sp. strain M-99, and Oscillatoria amphigranulata NZ-Concert (Fig. 4). This clade was also inferred for the maximum-likelihood analysis (ln likelihood score for the tree, −5,104) and for the two most equally parsimonious maximum-parsimony trees (length of 752 steps based on 209 parsimony-informative characters). A number of picocyanobacterial strains, including the widely studied Synechococcus sp. strains WH7805 and WH8103, also formed a well-supported clade in all three methods of phylogenetic reconstruction (Fig. 4). Similar results were also obtained for a larger phylogeny, which included 30 cyanobacteria and 3 outgroup taxa (http://evolution.uoregon.edu/mwood.htm).

FIG. 4.

Neighbor-joining phylogeny of the cyanobacteria inferred from ca. 950 bp of the 16S rRNA gene. A value at a node indicates the percentage of the time that the taxa to the right of the node formed a clade for 1,000 bootstrap pseudoreplicates. Only bootstrap values greater than 50% are indicated. GenBank accession numbers for sequence data are indicated in parentheses. Other sequences were obtained from the Ribosomal Database Project II (http://www.cme.msu.edu/RDP).

DISCUSSION

The results from the arginine assay, the amino acid analysis, and the electron micrographs definitively show that the Arabian Sea isolate Synechococcus sp. strain G2.1 is capable of producing CP granules. When nitrate-limited cells of Synechococcus sp. strain G2.1 in the experimental treatment were given nitrate and chloramphenicol, they produced appreciable quantities of CP; CP levels in the control treatments were undetectable. This result is consistent with the findings of Allen and Hutchison (4) and can be explained by their twofold hypothesis in which they postulate that in a nitrogen-starved cell, the enzymes for synthesizing CP are continuously active and the CP-degrading system is only synthesized after the cell starts to recover from nitrogen starvation. They further hypothesize that chloramphenicol inhibits synthesis of the protease that breaks down CP while the production mechanism is unaffected.

The amino acid analysis verified that the material analyzed in the arginine assay had the amino acid composition expected for CP; that is, it was composed of approximately equal molar quantities of arginine and aspartic acid with the possible addition of small amounts of glutamic acid. The glutamic acid found in the CP granules in this study confirms the finding by Merritt et al. (17) of glutamic acid in the CP granules of nitrogen-starved cells of Synechocystis sp. strain PCC6308. This is an important result, since efforts to determine the range of amino acids that can be incorporated into CP-like polypeptides have generated mixed results.

CP is synthesized by CP synthetase, which catalyzes an elongation reaction by using energy from ATP for peptide bond formation (24, 43). Berg et al. (7) recently expressed the gene for CP synthetase (cphA) from Anabaena sp. strain 29413 in E. coli and purified the enzyme. In their in vitro syntheses, aspartate could not be replaced by glutamate, but arginine could be replaced by lysine, citrulline, or ornithine, indicating that there is at least some substrate variability. CP synthetase from Synechocystis sp. strain PCC6308 has also been expressed in E. coli (1). When it was purified from the transformed E. coli and used for in vitro synthesis, radiolabeled canavanine and lysine were incorporated into CP, but glutamate was not (2). Previously, the same investigators found that glutamate is incorporated into CP in whole-cell extracts of the E. coli transformant (1). We found glutamate incorporated into CP produced by nitrogen-starved cells of Synechococcus sp. strain G2.1 after the addition of nitrate and chloramphenicol; this is perhaps more comparable to the conditions under which Merritt et al. (17) found glutamate incorporation into CP in nitrate-starved cells of Synechocystis sp. strain PCC6308 than the in vitro environment and highlights the importance of physiological conditions in determining the actual range of expression of a biosynthetic pathway. In both Synechococcus sp. strain G2.1 and Synechocystis sp. strain PCC6308, glutamate mobilization as a substrate for CP synthesis indicates that CP can be a dynamic part of nitrogen metabolism in these cyanobacteria.

The phylogenetic position of Synechococcus sp. strain G2.1 is interesting. A clade comprised of Synechococcus sp. strain G2.1, Leptolyngbya foveolarum, L. boryanum PCC73110, and Phormidium sp. strain M-99 was strongly supported by all three methods of phylogenetic reconstruction (Fig. 4). Leptolyngbya is a form-genus of filamentous cyanobacteria with flexible trichomes less than 3 μm in diameter (8). Our data provide the third example of a close phylogenetic relationship between a strain in the form-genus Synechococcus and one or more filamentous cyanobacteria. Previous reports have placed Synechococcus sp. strain PCC7335, a nitrogen-fixing strain capable of type III chromatic adaptation, and several strains of PE-containing Leptolyngbya from marine environments in the same clade (12, 35). Synechococcus sp. strain PCC7002, one of the marine strains of Synechococcus previously found to produce CP, and several other euryhaline strains of Synechococcus have also been placed in a clade of filamentous cyanobacteria that includes Oscillatoria rosea M-220 and Leptolyngbya fragile PCC7376 (11, 12, 35). Each of these examples represents a separate phylogenetic lineage, and none of the three Synechococcus strains that are related to filamentous cyanobacteria are close relatives of one another; taken together, these findings reflect the likely repeated evolution of filamentous growth forms from unicellular growth forms and vice-versa.

PE-containing picocyanobacteria in marine waters are often assumed to be members of the rapidly evolving cyanobacterial clade that includes Synechococcus sp. strains WH7803 and WH8103, Prochlorococcus strains MIT9303 and MIT9312, and all strains of Synechococcus known to be motile (28, 29). While Synechococcus sp. strain G2.1 is not a member of this clade, it is important to recognize that it would be included in estimates of Synechococcus abundance in many ecological studies because of its size and autofluorescence from PE. The same is also true for the unicellular nitrogen-fixing cyanobacteria recently described by Zehr et al. (42). It is increasingly apparent that few genetic or physiological properties can be automatically ascribed to small chroococcoid cyanobacteria in the sea merely because they contain PE.