Highly Diverse Cyanobactins in Strains of the Genus Anabaena

Abstract

Cyanobactins are small, cyclic peptides recently found in cyanobacteria. They are formed through proteolytic cleavage and posttranslational modification of short precursor proteins and exhibit antitumor, cytotoxic, or multi-drug-reversing activities. Using genome project data, bioinformatics, stable isotope labeling, and mass spectrometry, we discovered novel cyclic peptides, anacyclamides, in 27 Anabaena strains. The lengths of the anacylamides varied greatly, from 7 to 20 amino acids. Pronounced sequence variation was also detected, and only one amino acid, proline, was present in all anacyclamides. The anacyclamides identified included unmodified proteinogenic or prenylated amino acids. We identified an 11-kb gene cluster in the genome of Anabaena sp. 90, and heterologous expression in Escherichia coli confirmed that this cluster was responsible for anacyclamide production. The discovery of anacyclamides greatly increases the structural diversity of cyanobactins.

Cyanobacteria are a prolific source of secondary metabolites with potential as drug leads or useful probes for cell biology studies (23). They include biomedically interesting compounds, such as the anticancer drug lead cryptophycin (15), and environmentally problematic hepatotoxic peptides, such as microcystins and nodularins produced by bloom-forming cyanobacteria (23). Many of these compounds contain nonproteinogenic amino acids and modified peptides and are produced by nonribosomal peptide synthesis (23, 26).

The cyanobactins are a new group of cyclic peptides recently found in cyanobacteria (4). They are assembled through posttranslational proteolytic cleavage and head-to-tail macrocyclization of short precursor proteins. The cyanobactins are low-molecular-weight cyclic peptides that contain heterocyclized amino acids and can be prenylated or contain d-amino acids (3, 4). The cyanobactins that contain heterocyclized amino acids include patellamides, ulithiacyclamides, trichamide, tenuecyclamides, trunkamides, patellins, and microcyclamides and are synthesized in this manner (3, 4, 20, 24, 28). They possess antitumor, cytotoxic, and multi-drug-reversing activities and have potential as drug leads (4, 18, 20).

Cyanobactins containing heterocyclized amino acids are found in a variety of cyanobacteria (4). A recent study demonstrated that the cyanobactin biosynthetic pathway is prevalent in planktonic bloom-forming cyanobacteria (14). However, the products of these gene clusters encoding new cyanobactins are unknown. Here we report discovery of a novel family of low-molecular-weight cyanobactins and show that these compounds are common in strains of the genus Anabaena. These anacyclamides exhibit pronounced length and sequence variation and contain unmodified or prenylated amino acids.

MATERIALS AND METHODS

Cyanobacterial strains and cultivation.

The Anabaena strains were grown in Z8 (12) medium lacking a source of combined nitrogen, except for a few cultures for which nitrogen was included in the medium (Table 1). Strains were grown in 40-ml cultures under continuous light with a photon irradiance of 5 to 12 μmol m⁻² s⁻¹ at 20 to 25°C. Z8 medium containing ³⁴S was used to detect sulfur-containing peptides. The MgSO₄·7 H₂O of Z8 medium was replaced by a stable isotope of MgSO₄ (catalog no. IS7080; 90 atom% ³⁴S; Icon).

TABLE 1.

Anacyclamides and their detection by LC-MS in Anabaena strains^a

Anabaena strain	Origin	Amino acid sequence	Length (amino acids)	Unmodified peptide			Modified peptide (one group)			Modified peptide (two groups)			Accession no.
				Designation	Retention time (min)	[M+H]+	Designation	Retention time (min)	[M+H]+	Designation	Retention time (min)	[M+H]+
90b	Lake Vesijärvi, Finland	TSQIWGSPVP	10	A10	26.48	1053.5		NDf					FJ461733
0TU33S16b	Lake Tuusulanjärvi, Finland	TSQIWGSPVP	10	A10	26.44	1053.5		ND					GU183825
0TU39S15B	Lake Tuusulanjärvi, Finland	TSQIWGSPVP	10	A10	26.46	1053.5		NAg					GU183826
SYKE971/6c	Lake Kotojärvi, Finland	TSQIWGSPVP	10	A10	26.44	1053.5		ND					GU183827
299Bb,c	Lake Vesijärvi, Finland	TSQIWGSPVP	10	A10	26.45	1053.5		NA					GU183828
1TU31S9	Lake Tuusulanjärvi, Finland	SSVIWGSPVP	10	B10	30.20	1010.5		NA					GU183829
SYKE763A	Lake Tuusulanjärvi, Finland	SSVIWGSPVP	10	B10	29.08	1010.5		ND					GU183830
299Ab	Lake Vesijärvi, Finland	SAQWQNFGVP	10	C10	27.51	1115.5		ND					GU183831
SYKE748A	Lake Tuusulanjärvi, Finland	SAQWQNFGVP	10	C10	27.50	1115.5		ND					GU183832
1TU44S9	Lake Tuusulanjärvi, Finland	NAHWQNFGVP	10	D10	23.17	1151.5		ND					GU183833
PH256	Lake Knud, Denmark	YAPLQNFGVPd	10	E10	33.05	1087.6	E10P	40.98	1155.6				GU183834
BIR260	Baltic Sea, Gulf of Finland	DNWLGEWIGIP	11	A11	40.47	1281.6		ND					GU183835
BIR250B	Baltic Sea, Gulf of Finland	DNWLGEWIGIP	11	A11	40.49	1281.6		ND					GU183836
BIR257	Baltic Sea, Gulf of Finland	DNWLGEWIGIP	11	A11	40.37	1281.6		ND					GU183837
37b	Lake Sääskjärvi, Finland	HAFIGYDQDPTGKYP	15	A15	19.48	1690.8	A15G	35.12	1826.9e				GU183838
141b	Lake Vesijärvi, Finland	HAFIGYDQDPTGKYP	15		ND	1690.8	A15G	35.08	1826.9e				GU183839
1TU39S8	Lake Tuusulanjärvi, Finland	RERFVYP	7	A7	17.12	948.5		ND					GU183840
1TU32S11c	Lake Tuusulanjärvi, Finland	YSNKPSDFSP	10		ND	1123.5	F10P	30.84	1191.6				GU183841
1TU33S10b	Lake Tuusulanjärvi, Finland	YDDKLNLSPd	9		ND	1046.5	A9P	33.96	1114.6				GU183842
202A1/35b	Lake Vesijärvi, Finland	WGNGTGLDWKLLTGGISASP	20		ND	2012.0	A20P	35.80	2081	A20PP	41.32	2149	GU183843
202A1b	Lake Vesijärvi, Finland	WGNGTGLDWKLLTGGISASP	20		ND	2012.0	A20P	35.75	2081	A20PP	41.21	2149	GU183844
SYKE844Cc	Lake Lohjanjärvi, Finland	WGNGTGLDWKLLTGGISASP	20		ND	2012.0	A20P	35.79	2081	A20PP	41.24	2149	GU183845
1TU28S13b,c	Lake Tuusulanjärvi, Finland	WGNGTGLDWKLLTGGISASP	20		ND	2012.0	A20P	35.76	2081	A20PP	41.27	2149	GU183846
TR232	Baltic Sea	HQPWHAAPd	8		ND	925.4	A8P	23.57	993.5				GU183847
SYKE816	Baltic Sea, Gulf of Finland	FSPDWRAP	8	B8	26.93	957.5		ND					GU183848
PH262	Lake Knud, Denmark	VIQHYLFP	8	C8	32.35	998.5		ND					GU183849
SYKE844B	Lake Lohjanjärvi, Finland	LIGIMHP	7	B7	30.56	762.4		ND					GU183850

^aThe strains, their origins, and the predicted amino acid sequences and masses, as well as retention times, for the anacyclamides detected by LC-MS are shown. We detected prenylated (68-mass-unit increase) and geranylated (136-mass-unit increase) anacyclamides in 10 strains. The numbers of amino acids detected and derivatizations are indicated in the designations of the peptides (P, prenyl; G, geranyl). The five underlined peptides were validated by comparison to a synthetic peptide (retention time for A10, 26.46 min; retention time for C10, 27.54 min; retention time for A15, 19.48 min; retention time for A7, 17.04 min; retention time for B7, 30.53 min).

^bAxenic.

^cThe Z8 medium included added nitrogen.

^dBold type indicates a prenylated amino acid.

^e[M+H]⁺ indicates geranylation.

^fND, not detected.

^gNA, not analyzed.

Annotation of the acy gene cluster.

We obtained the complete genome sequence of the toxic bloom-forming freshwater cyanobacterium Anabaena sp. 90. Open reading frame (ORF) predictions were made with Glimmer, and the acy gene cluster was annotated using Artemis (Sanger Institute). Functional analysis of proteins was performed using BLASTp, and the predictions for start sites of the genes were checked manually. A putative peptide precursor gene was identified using BLASTp searches and homology of the N-terminal sequence of the 49-amino-acid AcyE protein and the precursor proteins of patellamide, tenuecyclamide, trichamide, and microcyclamide (4, 20, 24, 28).

DNA extraction, amplification, and sequencing.

Genomic DNA was extracted from the strains of Anabaena using an E.Z.N.A. plant DNA mini kit (Omega Bio-Tek, Doraville, GA). We prescreened 74 Anabaena strains for the presence of cyanobactin biosynthetic genes using a PCR approach based on the conserved subtilisin-like protease gene, as previously described (14). In order to identify an Anabaena strain producing a cyanobactin containing a sulfur amino acid, the acyE peptide precursor gene was amplified using primers preRNAF (5′GAAGAACATCCGCCCCCAACAAGTTG3′) and preRNAR (5′CTCCGCGTCGTCGCCTGCAAAAGG-3′) and primers PreF (5′-GCCTTCACCAAACCAGTCTTCTTCAT-3′) and PreR (5′-CATCGAGGCGAACCGTGCGCCAAGGGAT-3′) from the genomic DNA of Anabaena strains that were selected based on the presence of a cyanobactin biosynthetic gene fragment. PCRs were performed as previously described (14). Sequencing was carried out with an ABI 3730xl automated sequencer (Applied Biosystems) by Macrogen (Seoul, Korea). Sequences were checked and edited by using the Chromas 2.2 program (Technelysium Pty. Ltd.) and were aligned by using the Bioedit sequence alignment editor.

Cloning the acy biosynthetic gene cluster.

The entire 11-kb acy gene cluster was amplified from genomic DNA of Anabaena sp. 90 by PCR using primers patex2f (5′-ATGGATCCTGATGGACTGTAGTGTGAG-3′) and patex5r (5′-TACTCGAGAGGTTTTGGGACTCTTTAG-3′) in three 60-μl reaction mixtures containing 1× PCR buffer for Super Taq Plus (HT Biotechnology Ltd.), 200 μmol of each nucleotide (Finnzymes), 0.75 μM of each primer, 0.8 U Super Taq Plus proofreading polymerase (HT Biotechnology Ltd.), and 100 ng of Anabaena sp. 90 template DNA. The thermocycling conditions were 94°C for 2 min, followed by 30 cycles of 94°C for 30 s, 56.4°C for 30 s, and 68°C for 9 min and then a final extension at 68°C for 20 min. The PCR products were separated on a 0.6% agarose gel containing 0.5× Tris-acetate-EDTA and run for 30 min at 100 V. The gel was stained using SYBR Safe DNA gel stain (Invitrogen) and was visualized using a Dark reader (Clare Chemical Research Inc.) to avoid DNA damage from UV light during gel extraction. The 11-kb PCR product was gel extracted with a MinElute gel extraction kit (Qiagen) and cloned into the pCR 2.1-TOPO vector using a TOPO TA cloning kit (Invitrogen) with an insert-to-vector molar ratio of 3:1. The vector was used to transform chemically competent Escherichia coli One Shot TOP10 cells according to the manufacturer's instructions. The resultant pNL1901 and pNL1902 plasmids were analyzed by PCR and restriction digestion to ensure that the integrity of the insert was maintained during the cloning and amplification in E. coli. The transformants containing the 11-kb insert in the plasmid were grown overnight with shaking at 120 rpm at 28°C in 50 ml of LB medium containing 100 μg ml⁻¹ of ampicillin (sodium salt; Sigma-Aldrich) for liquid chromatography (LC)-mass spectrometry (MS) analysis. The cells were harvested by centrifugation at 10,000 × g for 10 min (Eppendorf 5804R centrifuge) and dried with a Heto vacuum centrifuge (Heto-Holten A/S), which yielded ca. 43 mg (dry weight).

Chemical analysis.

Cells of Anabaena strains were collected from the 40-ml cultures by centrifugation at 7,000 × g for 7 min. The collected cells were freeze-dried with Supermodulyo (Edwards High Vacuum International) or dried with a Heto vacuum centrifuge, which yielded 5 to 12 mg (dry weight). The dried Anabaena cells, as well as E. coli transformants, were extracted with 1 ml of methanol (HiperSolv for high-performance liquid chromatography [HPLC]; BDH Laboratory Supplies) in 2-ml plastic tubes containing glass beads (cell disruption medium; 0.5-mm-diameter glass beads; Scientific Industries Inc.). Each mixture was homogenized by shaking it with a FastPrep cell disrupter (Bio 101, Thermo Electron Corporation, Qbiogene, Inc.) for 30 s at a speed of 5 m s⁻¹. The mixture was centrifuged at 10,000 × g for 5 min, and the supernatant was used for LC-MS analysis.

The methanol extracts were analyzed with a high-performance liquid chromatograph combined with a mass spectrometer (Agilent 1100 series LC/MSD with Ion Trap XCT Plus and an electrospray ion source) in order to detect low-molecular-weight peptides. Peptides were separated from the extracts by HPLC using a Phenomenex C₁₈ (2) column (2.0 mm by 150 mm; particle size, 5 μm). The mobile phase gradient consisted of 0.1% aqueous (water purified with Milli-Q Plus; Millipore) formic acid (50% solution in water; Fluka, Sigma-Aldrich) (solvent A) and 0.1% formic acid in isopropyl alcohol (Sigma-Aldrich) (solvent B). Two different settings were used; one setting was used for screening methanolic extracts of Anabaena cells, and the other setting (values in parentheses below) was used for further structural characterization of natural and synthetic peptides. The percentage of solvent B was increased from 5 to 50% in 60 min. A flow rate of 0.15 ml min⁻¹ was used, and the columns were heated to 40°C during separation. The positive-ion mode of electrospray ionization was used. The pressure of the nebulizer gas (N₂) was 30 lb/in² (35 lb/in²), the drying gas flow rate was 8 liters min⁻¹, and the temperature was 350°C. The capillary voltage was set at 5,000 V (4,500 V), and the capillary offset value was 300 V. A skimmer potential of 85 V (100 V) and a trap drive value of 144 (111) were used. Spectra were recorded using a scan range from m/z 100 to m/z 2,200. Identification of the anacyclamides was based on the molecular weights calculated from the predicted peptide AcyE precursor amino acid sequences and the assigned fragment ion patterns of MSⁿ (n = 1 to 3) spectra. A comparison with the retention time and MS data for a synthetic reference was used in five cases (Beijing SBS Genetech Co., Ltd., China; anacyclamide B7 from JPT Peptide Technologies GmbH, Germany). The level of biosynthetic anacyclamide A10 in Anabaena sp. 90 was determined by LC-MS using synthetic peptide that was >98% pure and the m/z 1053.5 (MH⁺) signal.

Nucleotide sequence accession numbers.

The sequence of the acy gene cluster has been deposited in the GenBank database under accession number FJ461733. All acyE gene sequences have been deposited in the GenBank database under accession numbers GU183825 to GU183850.

RESULTS

Characterization of acy genes in Anabaena sp. 90.

We identified an 11-kb gene cluster encoding two subtilisin-like proteases during preliminary assembly and annotation of the Anabaena sp. 90 genome. The 11 genes in the acy gene cluster are organized in two operons with a putative bidirectional promoter (Fig. 1A and Table 2). The acy gene cluster is bordered by genes encoding a signal recognition particle GTPase and ribosomal protein S16. AcyA and AcyG have homology to the proteases involved in the biosynthesis of patellamides, trunkamide, trichamide, tenuecyclamide, patellin, and microcyclamides (Table 2). AcyE is a 49-amino-acid protein with N-terminal homology to the peptide precursor proteins in the other cyanobactin pathways (Fig. 2). Eight amino acids are conserved in the N-terminal leader sequence in all eight precursor proteins (Fig. 1B and 2). In addition, AcyB, AcyC, and AcyF are homologous to proteins that are encoded by almost all cyanobactin gene clusters but have unknown functions (Table 2). There are five ORFs between acyE and acyG that encode hypothetical and conserved proteins. It is not clear if these ORFs encode proteins that participate in the biosynthesis of an anacyclamide. Interestingly, the acy gene cluster encoded no homologs of the PatD, McaD, TriA, and TenD proteins, which are postulated to catalyze the heterocyclization of serine, threonine, and cysteine to thiazolines and oxazolines. Furthermore, the AcyG protein lacks the oxidase domains of PatG homologs, which are hypothesized to catalyze the oxidation of thiazolines and oxazolines to thiazoles and oxazoles in cyanobactin biosynthesis.

FIG. 1.

Anacyclamide biosynthetic pathway in Anabaena sp. 90. (A) acy gene cluster from Anabaena sp. 90. The genes indicated by black and gray arrows have homology to genes present in other cyanobactin gene clusters. The acyE gene encoding the precursor peptide is indicated by a gray arrow and is enclosed in a box. The genes indicated by open arrows encode conserved and hypothetical proteins without known or predicted functions. (B) AcyE peptide precursor from Anabaena sp. 90 with the hypervariable region of the 49-amino-acid protein encoding the mature anacyclamide shaded, indicating the position of cleavage and macrocyclization. The eight conserved amino acids in the AcyE leader peptide and other known cyanobactin precursor leader peptides are indicated arrowheads. (C) Structure of the decapeptide anacyclamide A10 from Anabaena sp. 90.

TABLE 2.

Functional assignment of the predicted proteins encoded by the acy gene cluster of Anabaena sp. 90 (accession no. FJ461733) and homology with proteins encoded by the microcyclamide (accession no. AM774406 and AM931579), trichamide (accession no. CP000393), tenuecyclamide (accession no. EU290741), patellamide (accession no. AY986476), and patellin (accession no. EU290742) biosynthetic gene clusters

Anabaena sp. 90			Proteins encoded by cyanobactin biosynthetic gene clusters in other cyanobacteria
Protein	Length (amino acids)	Predicted function	Microcyclamide	Trichamide	Tenuecyclamide	Patellamide	Patellin
AcyC	62	Unknown	McaC		TenC	PatC	TruC
AcyB	67	Unknown	McaB	TriJ	TenB	PatB	TruB
AcyA	658	Subtilisin-like protease	McaA	TriH	TenA	PatA	TruA
AcyE	49	Precursor protein	McaE	TriG	TenE	PatE	TruE
ORF5	277	Transposase-like protein		TriE
AcyF	296	Unknown	McaF		TenF	PatF	TruF1/TruF2
ORF7	131	Unknown
ORF8	227	Unknown
ORF9	47	Unknown
ORF10	201	Unknown
AcyG	710	Subtilisin-like protease	McaG	TriK	TenG	PatG	TruG

FIG. 2.

Peptide precursor proteins from Nostoc spongiaeforme var. tenue (TenE), Lyngbya aesturarii PCC 8106 (LynE) (4), Prochloron (PatE) (20), Microcystis aeruginosa PCC7806 (McaE1), Microcystis aeruginosa NIES 298 (McaE2) (28), Prochloron (TruE) (4), Trichodesmium erythraeum IMS 101 (TriG) (24), and Anabaena sp. 90 (AcyE). The homologous amino acids are indicated by asterisks. The areas of the precursor protein which form the mature cyanobactin are shaded.

AcyD peptide precursor diversity in strains of Anabaena.

The absence of the conserved cleavage sites in the AcyE precursor protein in Anabaena sp. 90 made direct prediction of the peptide product of the acy gene cluster impossible (Fig. 2). We used a novel approach to identify the product of the acy gene cluster. Anabaena strains were screened using PCR primers specific for the acyA protease gene in order to identify which Anabaena strains might produce cyanobactins. Subsequently, the acyE gene, encoding the precursor protein, was amplified from these Anabaena strains and sequenced (Table 1). The entire acyE gene was successfully amplified from 26 strains and sequenced. The N-terminal leader sequences and the 8 amino acids at the C terminus were very similar in all peptide precursors (data not shown). We identified a 7- to 20-amino-acid region that exhibited pronounced sequence variation between the conserved N- and C-terminal regions of the AcyE precursor protein. We postulated that the mature cyanobactin is formed from this region of the precursor peptide. A single strain (Anabaena SYKE 844B) was identified in which methionine, a sulfur-containing amino acid, was encoded in this hypervariable region of the acyE gene. We hypothesized that this strain would produce a cyanobactin that could be detected using ³⁴S stable isotope labeling without knowing the exact amino acid composition of the peptide.

LC-MS identification of anacyclamides.

In order to test this hypothesis, Anabaena sp. SYKE 844B was cultivated in medium containing either inorganic ³⁴S or ³²S as the principal source of sulfur. Cell extracts from both cultures were analyzed by LC-MS, and the sum mass spectra, containing all of the ions of the chromatogram, were compared (Fig. 3). We detected shifts of 2 mass units in mass spectra of the labeled (³⁴S) and unlabeled (³²S) extracts. A 2-mass-unit shift corresponding to one sulfur atom and one methionine was found. The m/z 762 (MH⁺) and 784 (MNa⁺) ions in the unlabeled extract were shifted to the m/z 764 and 786 ions in the ³⁴S-labeled extract. The MS² spectra of the m/z 762 and 764 ions were identical, with the exception of a shift originating from the ³⁴S atom. An ion mass of m/z 762 corresponds to a protonated anacyclamide B7 peptide (Fig. 3 and Table 1). This was confirmed by comparison to synthetic anacyclamide B7. The synthetic and biosynthetic peptides had the same retention time (Fig. 4), and the product ion spectra for positive (m/z 762) and negative (m/z 760) precursor ions of the peptides were identical (Fig. 5). Identification of the product of the acy gene cluster from Anabaena sp. SYKE 844B made it possible to predict the sites of proteolytic cleavage in all AcyE peptide precursors and calculate a predicted mass based on assumptions concerning cyclization and the amino acid content. We predicted that the product of the Anabaena sp. 90 acy gene cluster is a decapeptide with a monoisotopic molecular mass of 1,052 mass units, anacyclamide A10 (Fig. 1C). The corresponding m/z 1053 (MH⁺) and m/z 1075 (MNa⁺) ions were detected in Anabaena sp. 90 cell extracts by LC-MS. The biosynthetic and synthetic protonated anacyclamide A10 (MH⁺ and MH⁺-H₂O) produced very similar spectra when tandem mass spectrometry was used (Fig. 6). The biosynthetic and chemically synthesized anacyclamides had the same retention time, indicating that all amino acids were in the l configuration.

FIG. 3.

Shift in mass detected from comparison of mass spectra for extracts of Anabaena sp. strain SYKE 844B cultivated on normal medium (³²S) (spectrum below the line) and on medium containing ³⁴S (spectrum above the line), presented as mirror images. The m/z 764 peak is the protonated ion from ³⁴S-labeled anacyclamide B7, and the m/z 786 peak is its Na adduct; the corresponding ³²S ions are the m/z 762 and 784 ions, respectively.

FIG. 4.

Ion chromatograms for the protonated biosynthetic and synthetic reference anacyclamides A10, C10, A15, and B7. Ion masses and the anacyclamides are shown in Table 1. Chromatogram 1, Anabaena strain 90; chromatogram 2, Anabaena strain 0TU33S16; chromatogram 3, Anabaena strain SYKE971/6; chromatogram 4, Anabaena strain 299A; chromatogram 5, Anabaena strain SYKE748A; chromatogram 6, Anabaena strain 37; chromatogram 7, Anabaena strain SYKE 844B. The peak eluting later in the ion chromatogram of strain 0TU33S16 originated from a cyclic peptide belonging to another peptide family.

FIG. 5.

Comparison of the product ion spectra for the m/z 762.4 (MH⁺) and m/z 760.4 (M-H⁻) molecular ions of synthetic and biosynthetic (Anabaena SYKE 844B) anacyclamide B7.

FIG. 6.

Comparison of biosynthetic anacyclamide A10 of Anabaena sp. 90 to a synthetic standard and anacyclamide A10 produced in E. coli. (A to C) Mass spectra of (A) protonated biosynthetic anacyclamide produced by Anabaena sp. 90, (B) synthetic anacyclamide, and (C) biosynthetic anacyclamide produced by E. coli(pNLA90). (D and E) Mass spectra of (D) sodiated anacyclamide A10 from Anabaena sp. 90 and (E) a synthetic standard.

The predicted anacyclamides were detected in cell extracts of the Anabaena strains studied (Table 1). Altogether, 18 anacyclamides were unique (Table 1). Three more synthetic anacyclamides (anacyclamides C10, A15, and A7) were analyzed to ensure that the structural interpretation was correct. The retention times of the biosynthetic and synthetic anacyclamides were the same (Fig. 4), and the product ion spectra of MH⁺ were practically identical (data not shown). Two arginine residues in anacyclamide A7 induced the formation of an intense doubly charged M+2H²⁺ ion and a highly characteristic product ion spectrum, which unequivocally identified the structure.

At first we were able to detect 11 of the 18 different types of predicted anacyclamides in the strains of Anabaena (Table 2). Heterocyclized variants of anacyclamides were not detected in any of the strains, suggesting that the products of the acy gene cluster are unmodified cyclic peptides. However, it became apparent that seven of the anayclamides were produced in a derivatized form. Anacyclamides F10P, A9P, A20P, and A8P with masses that were 68 mass units greater than the masses predicted from the AcyE precursor peptide sequence were detected, while anacyclamide A20PP with a mass that was 136 mass units (2 × 68 mass units) greater than the predicted mass was also detected. Unmodified anacyclamide was not produced together with the derivatized peptide. Anacyclamide E10P was 68 mass units larger than predicted, and anacyclamide A15G was 136 mass units larger than predicted. Loss of 68 mass units was not observed for anacyclamide A15G. The unmodified anacyclamides eluted earlier than the derivatized anacyclamides, which demonstrated the hydrophobic nature of the structural units (R) with masses of 68 and 136 mass units (Table 1). Another characteristic describing these groups is the weak bonds to the peptidic structure in anacyclamides E10P, A15G, F10P, A9P, and A8P as the groups were partially lost in the ion source. Coelution of the M+H⁺ and M-R+H⁺ ions showed that M-R+H⁺ is probably formed from M+H⁺ via loss of 68 mass units during ionization in the ion source (Fig. 7A). The product ion spectrum for the M+H⁺ ion contained an intense M-R+H⁺ ion peak, and this is evidence that the M-R+H⁺ ion present in the ion source originated from the M+H⁺ ion (Fig. 7B). Finally, the evidence indicating that the structure of the M-R+H⁺ ion present in the ion source and the structure of the M-R+H⁺ ion formed via fragmentation in the trap from the M+H⁺ ion are identical is that the two M-R+H⁺ ions produced nearly identical product ion spectra (Fig. 7C and D). Data obtained for other derivatized anacyclamides unambiguously proved that anacyclamides E10P, A15G, F10P, A9P, and A8P contained loosely bound groups with masses of 68 or 136 mass units (Fig. 7). Anacyclamides A20P and A20PP lost the 68-mass-unit groups only when they were fragmented in the ion trap. Detailed analysis of MS² and MS³ fragmentation patterns suggested that the 68-mass-unit group is located in tyrosine or tryptophan in anacyclamides E10P, A9P, and A8P (Table 1).

FIG. 7.

Chromatographic and mass spectrometric data showing the presence of a structural unit with a mass of 68 mass units in anacyclamide A9P. (A) Ion chromatograms of m/z 1114.6 and m/z 1046.5 ions. (B) Product ion spectrum for protonated derivatized anacyclamide A9P m/z 1114.6 ion. (C and D) Product ion spectra for the m/z 1046.5 protonated ion formed (C) in the ion source and (D) in the ion trap.

Heterologous expression.

Cyanobacteria are difficult to manipulate genetically, and Anabaena sp. 90 is recalcitrant to genetic manipulation. In order to demonstrate that the acy gene cluster is responsible for the production of anacylamide, we amplified the 11-kb acy gene cluster by PCR, cloned the entire cluster into the pCR2.1 plasmid, and used the plasmid to transform E. coli TOP10. Clones containing the acy gene cluster in both orientations (pNLA901 and pNLA902) were identified by restriction analysis and used for heterologous expression of anacyclamide. We cultivated E. coli TOP10(pNLA901) and E. coli TOP10(pNLA902) overnight at 28°C in 50 ml of LB medium containing 100 μg ml⁻¹ ampicillin. We detected the predicted anacyclamide A10 in cell extracts of E. coli TOP10 carrying either plasmid. The ion masses of anacyclamide A10 in Anabaena sp. 90 were m/z 1053 (MH⁺) and m/z 1075 (MNa⁺), and the corresponding ions were found in E. coli TOP10(pNLA901) and E. coli TOP10(pNLA902) but not in the negative control (Fig. 6). The retention time of anacyclamide A10 produced in E. coli was also identical (±0.04 min) to the retention time of the synthetic and biosynthetic peptides from Anabaena sp. 90. The successful heterologous expression of anacyclamide A10 in E. coli confirmed that the acy genes are responsible for anacyclamide biosynthesis. The amount of anacyclamide A10 produced by Anabaena sp. 90 was 100 ng mg⁻¹ (dry weight) of biomass, and the amount produced by E. coli TOP10(pNLA901) 5 ng mg⁻¹ (dry weight) of biomass.

DISCUSSION

Cyanobactins are recently described low-molecular-weight cyclic peptides that contain heterocyclized amino acids and are formed through proteolytic cleavage of short precursor proteins (4). Some cyanobactins also contain disulfide bridges, prenylated amino acids, or d-amino acids (3, 4). Here we describe the discovery of anacyclamides, a new family of cyanobactins produced by strains of the genus Anabaena. The cyclic anacyclamides are composed of 7 to 20 amino acids, and the sequence variation is so pronounced that just a single conserved amino acid is present in all anacyclamides (Table 1). However, we found no evidence that anacyclamides contain heterocyclized amino acids. In keeping with this, we did not find genes encoding the proteins predicted to catalyze heterocyclization or oxidation of the heterocycles in the acy gene cluster of Anabaena sp. 90 (Table 2).

We found anacyclamides comprised of just unmodified proteinogenic amino acids and also a number of anacyclamides comprised of amino acids derivatized with structural units having masses of 68 or 136 (2 × 68) mass units. Mass spectrometric and chromatographic elution data presented here suggest that these structural units are prenyl (C₅H₈) and geranyl (C₁₀H₁₆) groups attached to the side chain heteroatom of the amino acid. Neutral loss of 68 mass units is characteristic of O-prenyl-substituted flavonoids (8), as detected for derivatized anacyclamides E10P, A9P, F10P, A20P, A20PP, and A8P. O-Prenyl-, N-prenyl-, and N-geranyl-substituted peptides have been reported for cyanobacteria and other marine organisms. Myriastramides and prenylagaramides contain O-prenylated tyrosine (6, 16), mollamides contain O-prenylated serine or threonine (5), cyclomarins contain N-prenylated tryptophan (21), microguanidine contains N-geranylated arginine (9), and aeruginoguanidines contain N-prenylated and geranylated arginine (10). Our findings broaden the known structural variation of cyanobactins to include unmodified and prenylated homodetic cyclic peptides lacking heterocyclized amino acids and exhibiting pronounced length and sequence variation.

The pronounced chemical diversity of anacyclamides that we observed is achieved through hypervariation at the nucleotide level in the region of the acyE gene encoding the mature peptide. At present it is not clear how the length and sequence variation is achieved at the genetic level in nature. The hypervariation suggests that there is strong selection for rapid evolution, maintenance, and diversification of anacyclamides in Anabaena. The biosynthetic machinery for assembling cyanobactins resembles that of bacteriocin protein toxins found in a wide range of bacteria (13). It is not clear why Anabaena produces anacyclamides, but, analogous to bacteriocins, they may be produced for chemical warfare (1). The anacyclamides might participate in interstrain competition or act as antimicrobial agents against bacteria. Anabaena produces a range of toxins and enzyme inhibitors widely believed to act as grazing deterrents (23).

Cyanobacteria are a rich source of low-molecular-weight peptides with a broad range of structural variations. Strains of the cyanobacterial genus Planktothrix produce cyclic peptides containing only unmodified proteinogenic amino acids, including planktocyclin (2), oscillapeptin (7), and agardhipeptin (22). Strains of the genus Planktothrix also produce prenylated cyclic peptides called prenylagaramides (16). Microcystis aeruginosa produces unmodified and prenylated versions of the undecapeptide kawaguchipeptin that exhibit antimicrobial activity (11). Interestingly, all of these compounds contain a proline residue (2, 7, 11, 16, 22). Anacyclamides exhibit considerable sequence variation, in contrast to other cyanobactins, and the terminal proline is the only amino acid conserved in all anacyclamides (Table 1). The presence of proline may be necessary for cleavage and/or macrocyclization of the anacyclamide precursor peptide by the AcyA and AcyG proteases. In an analogous situation the presence of a conserved asparagine residue is necessary for both cleavage and macrocyclization of the precursor in plant cyclopetide biosynthesis (19). The biosynthetic origins of the unmodified and prenylated cyclic peptides produced by Planktothrix and Microcystis are unclear, but the chemical similarities suggest that such peptides may be synthesized in a fashion similar to anacyclamides.

Low-molecular-weight cyclic peptides consisting of unmodified proteinogenic amino acids have been found in a broad range of organisms, including cyanobacteria (2, 22) and plants (25). Unmodified cyclic peptides have a range of interesting pharmaceutical activities, including antimicrobial, antiplatelet, antimalarial, immunosuppressive, antileukemic, and enzyme-inhibiting activities (2, 17, 22, 25). Thus, it has been established that cyclic peptides such as anacyclamides can exhibit interesting bioactivities and warrant further investigation. Research on the chemical diversity and biosynthetic potential of anacyclamides may lead to useful biotechnological and industrial applications. Cyanobacteria are notoriously difficult to manipulate, and strains of the bloom-forming genus Anabaena are recalcitrant to genetic manipulation. To date there have been just a few examples of expression of a cyanobacterial metabolite in a heterologous host (3, 4, 20, 27). The heterologous expression of the anacyclamide gene cluster in an E. coli surrogate host opens the possibility of determining the role of each gene in the acy gene cluster in the assembly of the cyclic peptide, as well as the production of a library of recombinant cyclic anacyclamides.

In conclusion, we detected 18 anacyclamide variants in 27 strains of Anabaena originating from different habitats using mass spectrometry. The anacyclamides identified contained unmodified proteinogenic or prenylated amino acids. In contrast to previously described cyanobactins, the length of the anacyclamides varied greatly, from 7 to 20 amino acids, and there was pronounced sequence diversity with conservation of just a single amino acid, proline. Such great length variation or sequence diversity of cyanobactins has not been reported previously for cyanobacteria. In addition, heterologous expression of the gene cluster in E. coli provides further avenues for detailed characterization of the machinery, as well as its utilization in production of novel compounds belonging to this biomedically interesting class of cyclic peptides. This work also emphasizes the power of genome mining, culture collections, and labeling experiments, as well as a combination of molecular and chemical tools, for discovering new families of bioactive compounds.