Scytodecamide from the Cultured Scytonema sp. UIC 10036 Expands the Chemical and Genetic Diversity of Cyanobactins

Abstract

Cyanobactins are a large family of cyanobacterial ribosomally synthesized and post-translationally modified peptides (RiPPs) often associated with biological activities, such as cytotoxicity, antiviral, and antimalarial activities. They are traditionally described as cyclic molecules containing heterocyclized amino acids. However, this definition has been recently challenged by the discovery of short, linear cyanobactins containing three to five amino acids as well as cyanobactins containing no heterocyclized residues. Here, we report the discovery of scytodecamide (1) from the freshwater cyanobacterium Scytonema sp. UIC 10036. Structural elucidation based on mass spectrometry, 1D and 2D NMR spectroscopy, and Marfey’s method revealed 1 to be a linear decapeptide with an N-terminal N-methylation and a C-terminal amidation. The genome of Scytonema sp. UIC 10036 was sequenced and bioinformatic analysis revealed a cyanobactin-like biosynthetic gene cluster consistent with the structure of 1. The discovery of 1 as a novel linear peptide containing an N-terminal N-methylation and a C-terminal amidation expands the chemical and genetic diversity of the cyanobactin family of compounds.

Graphical Abstract

Scytodecamide is a novel N- and C-protected peptide. Isolated from the freshwater cyanobacterium Scytonema sp. UIC 10036, it is a linear decapeptide featuring N-terminal N-methylation, and a C-terminal amidation. Genome sequencing and bioinformatics analysis revealed a cyanobactin biosynthetic gene cluster consistent with the structure of scytodecamide.

Introduction

Cyanobacteria are photosynthetic prokaryotes that produce chemically diverse and biotechnologically relevant natural products, such as polyketides, non-ribosomal peptides, and a wide range of ribosomally synthesized and post-translationally modified peptides (RiPPs).^[1–3] Cyanobactins encompass a large family of RiPPs, often associated with biological activities, such as cytotoxicity, multidrug reversing, antiviral, and antimalarial activities.^[4–6] They were traditionally described as cyclic peptides of ribosomal origin that contain heterocyclized amino acid (aa) residues.^[6,7] However, this definition has been recently challenged. Genome mining of cyanobacterial strains has resulted in novel cyanobactins that did not fit the classic definition. Anacyclamides and prenylagaramides are cyclic peptides containing no heterocyclized residues.^[8,9] In 2013, Leikoski and colleagues discovered the first linear cyanobactins, aeruginosamides B and C and viridisamide.^[10] These short peptides (three to five aa residues) contain N-terminal N-prenylations and methylated C-termini connected to thiazole rings.^[10] Enzymatic modifications of peptides, such as cyclizations, methylations, prenylations, epimerizations, and amidations, have been shown to increase the functional diversity of peptidic natural products.^[11,12]

The canonical biosynthetic pathway for cyanobactins, as recently reviewed,^[13,14] is ribosomal synthesis of the precursor peptide (E gene), followed by heterocyclization (D gene, if the product contains heterocycles), proteolytic cleavage of the leader peptide (A gene), and macrocyclization/cleavage of the C-terminal follower peptide (G gene). Heterocyclizing enzymes (D protein) recognize the precursor peptide at recognition sequence I (RSI), whereas proteases (A protein) cleave the leader peptide at RSII, and macrocyclases/proteases (G protein) recognize RSIII. Tailoring enzymes such as oxidases, methyltransferases and prenyltransferases may also be involved in cyanobactin pathways usually leading to late-stage modifications.^[15–18]

In a previous study, strain Scytonema sp. UIC 10036 was found to produce scytophycin C and tolytoxin.^[19] Scytophycins and tolytoxin were first characterized by the Patterson group. ^[20,21] Tolytoxin was later shown to be a potent actin inhibitor.^[22] Potential uses of these compounds as chemical probes or as anticancer agents inspired further studies into their biosynthesis,^[23] total chemical synthesis,^[24] and production under different culture conditions.^[25,26] The chemical potential of strain Scytonema sp. UIC 10036 encouraged us to perform media studies to challenge its growth and production of secondary metabolites.^[19] In short, strain UIC 10036 was cultured under various levels of nitrate and phosphate. Untargeted comparative metabolomics was applied to assess the influence of these media components on its metabolomic profile. Results obtained for strain Scytonema sp. UIC 10036 revealed a) changes in tolytoxin production under the different conditions, b) increased heterocyst glycolipid production under nitrate starvation, and c) increased production of a potentially novel compound of m/z [M - H]^- 994 in low nitrate and high phosphate in comparison to the standard Z medium.^[19] The latter was investigated in this work.

Herein we present the isolation and complete structural elucidation of scytodecamide (1), a N- and C-protected linear decapeptide. Moreover, we report the genome sequence of the producing strain Scytonema sp. UIC 10036 and the bioinformatics identification of a cyanobactin-like biosynthetic gene cluster consistent with the structure of 1.

Results and Discussion

In order to isolate and elucidate the structure of the potentially novel compound, cell mass from large-scale cultures (18 L) of Scytonema sp. UIC 10036 was harvested, lyophilized, and extracted with MeOH:DCM 1:1 (v/v). Dry extract was subjected to fractionation using Diaion^® HP-20SS as solid phase and a step gradient of water and isopropyl alcohol (IPA). The target compound was found to be enriched in the fraction eluting at 40:60 IPA:H₂O (F3) by LC-MS analysis. Further purification by semi-preparative HPLC afforded scytodecamide (1).

Scytodecamide (1) was obtained as a white, amorphous powder. HRMS data (m/z 996.6450 [M + H]⁺) was consistent with the formula C₄₇H₈₅N₁₁O₁₂, requiring 11 degrees of unsaturation. MS/MS fragmentation of 1 showed a sequential loss of ten aa residues (Supplementary Information S3). The peptidic nature of 1 was further confirmed by the inspection of DEPTQ and ¹H NMR spectra (Table 1). The DEPTQ spectrum of 1 in methanol-d₄ contained ten amide carbonyl signals (δ_C 172–178) as well as resonances typical of α-carbons in aa (δ_C 53–70) and aliphatic methyl, methylene and methine groups (δ_C 12–42). Methylene signals at δ_C 62.2 and 63.0 ppm suggested the presence of two Ser residues. The ¹H NMR spectrum of 1 in methanol-d₄ exhibited resonances characteristic of aa α-protons (δ_H 3.8–4.4), aliphatic methylene and methine groups (δ_H 1.2–2.5), and 13 aliphatic methyl signals (δ_H 0.8–1.5). The methyl doublets at δ_H 1.41 and 1.48 ppm suggested the presence of two Ala residues. A deshielded methyl singlet at δ_H 2.34 ppm was also observed. Comparison of ¹H NMR spectra of 1 in methanol-d₄ and -d₃ showed only six exchangeable amide signals (δ_H 7.5–8.2). Additional ¹H NMR spectra acquired in pyridine-d₅ with TFA vapor revealed a total of 11 exchangeable signals between δ_H 8.1–9.3 ppm.

Table 1.

NMR spectroscopic data of scytodecamide (1)^a

		Methanol-d3b and -d4				Pyridine-d5 with TFA vapor
	Position	δC, mult	δH, mult (J in Hz)	COSY	HMBC	δC, mult	δH, mult (J in Hz)	COSY	HMBC
Val amide	1	176.7, C				175.4, C
	2	60.9, CH	4.17, m	2-NH, 3	1, 3	60.4, CH	5.05, m	2-NH, 3	1, 3, 4
	3	31.4, CH	2.23, o (6.8)	2, 3-Me, 4	2, 4	31.2, CH	2.8, m	2, 3-Me, 4	2, 4
	3-Me	19.8, CH3	0.98, d (6.8)	3	3	20.5, CH3	1.25, d (6.8)	3	2, 3
	4	18.8, CH3	1.01, d (6.8)	3	3	19.1, CH3	1.38, d (6.8)	3	2, 3, 3-Me
	1-NHa		nd				8.08, br		1
	1-NHb		nd				8.14, br		2
	2-NH		7.56, d (8.8)	2	1Leu1		8.23, d (8.8)	2	2, 3, 1Leu1
Leu1	1	175.7, C				174.4, C
	2	54.8, CH	4.33, dd (11.0, 3.8)	NH, 3	1, 3	54.7, CH	5.05, m	3	1, 2
	3	41.5, CH2	1.58, m	2	2, 4	41.6, CH2	2.12, ddd (13.8, 9.8, 4.0)	2	2, 4
			1.93, m				2.39, ddd (13.8, 11.2, 4.3)
	4	25.4, CH	1.93, m	4-Me, 5	5	25.3, CH	2.34, m	4-Me, 5	3
	4-Me	22.7, CH3	0.91, d (6.3)	4	5	24.1, CH3	1.04, d (6.6)	4	3, 5
	5	21.2, CH3	0.89, d (6.3)	4		21.7, CH3	0.92, d (6.6)	4	3
	NH		7.69, d (6.4)	2	1Ser1		8.30, d (7.4)	2	2, 3, 1Ser1
Ser1	1	173.2, C				172.4, C
	2	59.7, CH	4.22, dd (6.6, 3.6)	NH, 3	1	59.9, CH	4.95, m	NH, 3
	3	62.2, CH2	3.94, dd (11.9, 3.6)	2	1, 2	62.7, CH2	4.47, dd (11.6, 3.8)	2	1
			4.04, dd (11.9,6.6)				4.61, dd (11.6, 6.4)
	OH		nd				nd
	NH		7.69, d (6.4)	2	1Ala1		8.32, d (4.4)	2	2, 3, 1Ala1
Ala1	1	177.56, C				176.2, C
	2	53.5, CH	4.09, q (7.3)	2-Me	1, 2-Me	52.9, CH	4.64, m	NH, 2-Me	2-Me
	2-Me	16.8, CH3	1.41, d (7.3)	2	1, 2	17.3, CH3	1.71, d (7.3)	2	1, 2
	NH		8.24, d (4.14)	2			8.66, br	2	1Leu2
Leu2	1	176.9, C				176.0, C
	2	56.6, CH	4.06, m	NH, 3	1, 3, 4	55.9, CH	4.62, m	NH, 3	3, 4
	3	40.5, CH2	1.69, m	2	1, 2, 4	40.4, CH2	2.01, m	2	1, 2, 4
			1.77, m				2.08, m
	4	25.7, CH	1.74 m	4-Me, 5	3	25.5, CH	2.02, m	4-Me, 5	3
	4-Me	23.9, CH3	0.92, d (6.3)	4	3	23.3, CH3	1.01, d (6.1)	4	3
	5	22.3, CH3	0.88, d (6.3)	4	3	22.4, CH3	0.91, d (6.1)	4
	NH		7.98, d (4.83)	2			8.39, br	2	2, 3, 1Leu3
Leu3	1	177.2, C				176.7, C
	2	55.9, CH	4.19, t (7.9)	NH, 3	1, 3, 4	55.5, CH	4.63, m	NH, 3	3, 4
	3	40.5, CH2	1.75, m	2	1, 2, 4	40.3, CH2	1.95, m	2	1, 2, 4
			1.81, m				1.99, m
	4	26.3, CH	1.70, m	4-Me, 5	5	25.8, CH	1.88, m	4-Me, 5	4-Me
	4-Me	23.0, CH3	1.02, d (6.7)	4	3	23.0, CH3	0.95, d (6.5)	4	3
	5	22.9, CH3	0.97, d (6.7)	4	3	22.5, CH3	0.84, d (6.5)	4	3
	NH		7.62, d (6.4)	2	1Ala2		8.15, br	2	1Ala2
Ala2	1	177.8, C				176.8, C
	2	53.3, CH	4.06, m	NH, 2-Me	1, 2-Me	53.2, CH	4.43, m	NH, 2-Me	1, 2-Me
	2-Me	16.5, CH3	1.48, d (7.3)	2	1, 2	17.2, CH3	1.52, d (7.3)	2	1, 2
	NH		nd				8.78, d (4.4)	2	1Pro
Pro	1	176.0, C				175.2, C
	2	64.1, CH	4.29, t (8.2)	3	1, 3	63.4, CH	4.68, m	3	1, 3
	3	30.6, CH2	2.01, m	2, 4	1, 2, 5	30.3, CH2	2.13, m	2, 4	2, 5
			2.43, m				2.35, m
	4	26.4, CH2	2.06, m	3, 5	5, 3	25.9, CH2	1.77, m	3, 5	3, 5
			2.18, m				1.98, m
	5	49.3, CH2	4, dt (9.2, 6.8)	4	4, 2Ser2	49.0, CH2	4.11, m	4	3, 4
			4.07, m				4.14, m
Ser2	1	172.7, C				172.7, C
	2	53.5, CH	4.94, dd (6.4, 4.9)	3	1, 3	53.7, CH	5.51, m	NH, 3	1, 3
	3	63.0, CH2	3.83, dd (10.2, 6.4)	2	1, 2	63.1, CH2	4.25, m	2	1, 2
			4.14, dd (10.2, 4.7)				4.54, dd (10.4, 5.4)
	OH		nd				nd
	NH		nd				9.3, br	2
N-Me-Ile	1	175.2, C				nd
	2	70.1, CH	2.94, d (5.2)	3	1, 3, N-Me, 4	69.1, CH	3.63, br	3
	3	39.5, CH	1.72, m	2, 3-Me, 4		38.5, CH	2.11, m	2, 3-Me, 4
	3-Me	15.9, CH3	0.94, d (6.9)	3	3, 4, 2	16.1, CH3	1.13, d (6.8)	3	2, 3, 4
	4	26.7, CH2	1.22, m	3, 5	3, 3-Me, 5	26.5, CH2	1.44, m	3, 5	3, 3-Me, 5
			1.55, m				1.83, m
	5	12.1, CH3	0.91, t (7.6)	4		12.2, CH3	0.88, d (7.4)	4	3, 4
	N-Me	35.6, CH3	2.34, s		2	35.0, CH3	2.67, s		2
	NH		nd				nd

^aFrequency of 900 MHz for ¹H and 226 MHz for ¹³C.

^bMethanol-d₃ was used for the assignment of NH protons as well as their COSY correlations to α-protons and HMBC correlations to carbonyl carbons.

Interpretation of COSY and TOCSY spectra allowed for the assignment of ten aa residues: two Ser, two Ala, three Leu, one Ile, one Pro, and one Val. The methyl singlet at δ_H 2.34 (methanol-d₄) showed a single HMBC correlation with the α-carbon of the Ile residue (δ_C 69.1), revealing this to be an N-Me-Ile. HBMC correlations between the broad exchangeable signals at δ_H 8.08 and 8.14 ppm (pyridine-d₅ with TFA vapor) with the Val carbonyl (δ_C 175.4) and α-carbon (δ_C 60.4), respectively, suggested this residue to be a terminal Val amide.

The aa sequence in 1 was established with the combined analysis of the HMBC spectra in methanol-d₃ and pyridine-d₅ with TFA vapor and HRMS/MS fragmentation (Figure 1, Table 1). In methanol-d₃, HMBC correlations from Val amide NH (δ_H 7.56) to Leu1 C-1 (δ_C 175.7), from Leu1 NH (δ_H 7.69) to Ser1 C-1 (δ_C 173.2), from Ser1 NH (δ_H 7.69) to Ala1 C-1 (δ_C 177.6), from Leu3 NH (δ_H 7.62) to Ala2 C-1 (δ_C 177.8), and from Pro H-4 (δ_H 4.00 and 4.07) to Ser2 C-2 (δ_C 53.5) were observed. In pyridine-d₅ with TFA vapor, additional correlations from Ala1 NH (δ_H 8.66) to Leu2 C-1 (δ_C 176.0), from Leu2 NH (δ_H 8.39) to Leu3 C-1(δ_C 176.7), and from Ala2 NH (δ_H 8.78) to Pro C-1 (δ_C 175.2) supported the sequence of Ser2-Pro-Ala2-Leu3-Leu2-Ala1-Ser1-Leu1-Val-CONH₂. Despite extensive efforts, no NMR correlations were observed between Ser2 and the terminal N-Me-Ile. The placement of N-Me-Ile as the N-terminal aa was confirmed by HRMS/MS fragmentation of 1 (Figure 1).

Figure 1.

HRMS/MS fragmentation and key NMR correlations of 1

The absolute configuration of the aa in 1 was determined by acid hydrolysis followed by the advanced Marfey’s method. All residues were assigned the l configuration based on the comparison with aa standards (Supplementary Information S20). The final structure of 1 was established as l-N-Me-Ile-l-Ser-l-Pro-l-Ala-l-Leu-l-Leu-l-Ala-l-Ser-l-Leu-l-Val-CONH₂.

Scytodecamide (1) features two terminal modifications. N-methylation and C-terminal amidation have been associated with improved membrane permeability and resistance to peptidases, respectively.^[12,27,28] The unusual structural features of 1 increased our interest in investigating its biosynthetic pathway. Either a nonribosomal peptide synthetase (NRPS) or a RiPP pathway can be envisioned for biosynthesis of 1.

The genome of Scytonema sp. UIC 10036 was sequenced with short-read Illumina and long-read PacBio technologies. None of the NRPS gene clusters detected were in agreement with the structure of 1. Instead, bioinformatic analysis revealed a putative cyanobactin gene cluster. Open reading frames (ORFs) within the gene cluster were annotated and subsequently analyzed for sequence similarity to known proteins (Supplementary Information S21). The putative scd biosynthetic gene cluster (BGC) consisted of six genes spanning approx. 8.3 kb (Figure 2, Table 2). It was positively identified by the presence of the precursor gene scdE encoding the exact peptide sequence found in scytodecamide (1) (Figure 2). ScdE as shown in Figure 2 is composed of 70 aa including the leader peptide, recognition site II (RSII), and the ten aa core sequence. Although there are alternative methionines in the leader peptide sequence that could function as start codons, the one shown in Figure 2 and leading to the 60-aa leader sequence has the best ribosome biding site consensus (GGAG), followed by the first internal methionine leading to a 40-aa leader sequence (GAG). A comparison of known cyanobactin precursor peptides and ScdE is shown in Supplementary Information S22. A second core peptide can be tentatively predicted, however we did not detect this peptide in cultures of strain UIC 10036. A potential RSI sequence can also be predicted, although the sequence is not very well conserved (Figure 2 and Supplementary Information S22).

Figure 2.

The putative scd biosynthetic gene cluster from Scytonema sp. UIC 10036 and the hypothetical biosynthetic pathway of 1, with an expansion showing the amino acid sequence of the precursor peptide ScdE, the substrate of the biosynthetic enzymes. The N-terminal protease ScdA would cleave off the leader peptide of ScdE at recognition sequence II (RSII), and ScdMT and ScdTA would catalyze methylation and transamidation, respectively. The substrates and timing of reactions catalyzed by Scd enzymes remain to be elucidated. For example, it is conceivable that transamidation by ScdTA could happen either before or after cleavage of the leader sequence by ScdA. Genes of the scd BGC are named according to Gu et al.^[13]

Table 2.

Predicted function of ORFs in the scd BGC.

Protein	Length (aa)	Predicted function
ORF-2	188	hypothetical protein
ORF-1	119	hypothetical protein
ScdC	56	unknown (PatC homolog)
ScdB	65	unknown (PatB homolog)
ScdMT	587	didomain protein methyltransferase/DUF
ScdA	697	didomain protein N-terminal protease/DUF (PatA homolog)
ScdE	70	precursor protein
ScdTA	627	transamidase
ORF+1	82	hypothetical protein
ORF+2	415	tRNA nucleotidyltransferase

Overall, the structure of 1 conforms to the current definition of the cyanobactins. However, as a decapeptide, it is larger than the linear cyanobactins reported to date. Moreover, N-terminal N-methylation and C-terminal amidation have been previously described in RiPPs, but not for a cyanobactin.^[29]

The scd BGC encodes a homolog of the N-terminal protease PatA from the patellamide biosynthetic pathway (ScdA, 83% similarity at the aa level; Figure 2, Table 2).^[30,31] PatA catalyzes the cleavage of the N-terminal protease site RSII of the precursor peptide, removing the leader peptide sequence.^[30] PatA contains two domains, an N-terminal serine protease domain and a C-terminal domain of unknown function (DUF). Besides an A protease, cyanobactin gene clusters may encode a macrocyclase/C-terminal protease which is related to PatG/AgeG.^[4,10,30] PatG catalyzes macrocyclization^[30] while AgeG from the aeruginosamide pathway cleaves off a C-terminal peptide sequence leading to linear products.^[16] G enzymes also contain a DUF domain^[32] and recognize the precursor peptide C-terminal protease signature RSIII.^[33] No homologs for G enzymes were found in the putative scd gene cluster, which is consistent with the lack of an RSIII in the precursor protein sequence (Figure 2 and Supplementary Information S22). Lack of RSIII is an unusual feature of precursor peptide ScdE since reported cyanobactin precursor peptides all contain RSIII (Supplementary Information S22).

Furthermore, genes with sequence similarity to the short conserved proteins PatB and PatC, were identified (ScdB and ScdC; Figure 2, Table 2).^[5,6] Despite being conserved among nearly all cyanobactin gene clusters, their function remains unknown, and studies have shown that these genes are non-essential for in vivo production of patellamides.^[34] Consistent with the structure of 1, no homologs of PatD (heterocyclase) or PatF (prenyltransferase) were found.^[30,31]

The scd BGC contains one gene, scdMT, with a N-terminal S-adenosylmethionine-dependent methyltransferase domain and a C-terminal region that shows sequence similarity to the DUF domain of proteases and macrocyclases ^[32] (Supplementary Information S23-S25). Genes containing methyltransferase domains were previously found in six out of 30 cyanobactin BGCs identified by genome mining, of which four encode unknown compounds.^[10] The two known compounds are aeruginosamide and viridisamide, featuring N-terminal N-prenylations and C-terminal O-methylations. The only cyanobactin featuring N-methylation is the cyclic cyanobactin microcyclamide in which the nitrogen of a histidine residue is methylated. However, a methyltransferase could not be identified in the microcyclamide BGC.^[35] In five out of six cases, methyltransferases identified by genome mining^[10] are fused with other functions such as prenyltransferase, and the DUF domain of proteases and macrocyclases. The only characterized cyanobactin methyltransferase is AgeMTPT, which contains methyltransferase and prenyltransferase domains.^[16] AgeMTPT was shown to N-prenylate a peptide precursor containing RSIII, after which AgeG cleaves off the RSIII and AgeMTPT O-methylates the C-terminal carboxyl group towards biosynthesis of aeruginosamides. The scdMT gene potentially expands the genetic diversity for N-protection in cyanobactins by putatively encoding a N-terminal N-methyltransferase. The substrate and timing of methylation remains to be elucidated.

Finally, the structure of scytodecamide features an unusual C-terminal amide. A protein BLAST analysis of the scd gene cluster revealed the gene scdTA to encode a putative transamidase (= amidotransferase, Supplementary Information S21). Li et al. recently showed that TtmN, an amidotransferase from Streptomyces afghaniensis NRRL 5621, catalyzes the amidation of the carboxylated intermediate prethiotetraamide C to yield thiotetroamide C.^[36] ScdTA shows 29% identity and 57% similarity, respectively, to TtmN at the aa level. We, therefore, hypothesize that ScdTA is responsible for the C-terminal amidation of scytodecamide.

Scytodecamide (1) was tested for cytotoxicity against MDA-MB-435 human melanoma, MDA-MB-231 human breast cancer, and OVCAR3 human ovarian cancer cells, but no activity was detected up to 25 μM.

Conclusion

Media studies analyzed by untargeted comparative metabolomics led to the discovery of scytodecamide (1), a new linear decapeptide featuring an N-terminal N-methylation and a C-terminal amidation. The genome of Scytonema sp. UIC 10036 was sequenced using Illumina and PacBio technologies. Bioinformatic analysis revealed an 8.3-kb cyanobactin-like biosynthetic gene cluster consistent with the structure of 1. Novel modes of N- and C-terminal protection – N-methylation and C-terminal amidation, respectively – expand the chemical and genetic diversity of cyanobactins.

Experimental Section

General Experimental Procedures:

Optical rotation was recorded using a PerkinElmer 241 polarimeter. UV spectrum was acquired on a Varian Cary 5000 spectrophotometer. IR spectrum was obtained using a Thermo-Nicolet 6700 with Smart iTR™ accessory. 1D and 2D NMR spectra were acquired on Bruker AVII 900 MHz spectrometer equipped with a 5 mm TCI cryoprobe (Supplementary Information S4–19). NMR chemical shifts were referenced to residual solvent peaks (CD₃OD/CD₃OH δ_H 3.31 and δ_C 49.0; Pyr-d₅ δ_H 8.74 and δ_C 150.35). LC-UV-HRMS experiments were performed on a Shimadzu LCMS-IT-TOF equipped with a UPLC C18 column (2.1 × 50 mm x 1.7 μm). Mobile phases consisted of H₂O (A) and CH₃CN (B), both acidified with 0.1% formic acid, flowing at a total rate of 0.5 mL/min. The gradient program was set as follows: 5–100% B for 7 minutes,1-min wash, and 2-min re-equilibration. PDA acquisition ranged from 190 to 450 nm at 4.17 Hz sampling rate. HRMS data were recorded in positive and negative modes from 150 to 3000 m/z. Additional HRMS/MS spectra were acquired at the University of North Carolina at Greensboro using a Thermo QExactive Plus MS (Thermo-Fisher). HRMS/MS spectra were obtained by automatic data-dependent fragmentation with a collision energy of 35 eV.

Biological Material and Culture Conditions:

Scytonema sp. UIC 10036 was isolated from a sample collected in December of 2006 at Homestead, Florida (N 25°24.2′, W 80°33.4′). Unialgal strain isolation was accomplished by micropipette isolation techniques.^[37] Scytonema sp. UIC 10036 was cultured for 6 weeks in 18 L of liquid Z medium^[19] (one 13 L flask containing 10 L of medium and four 2.8 L Fernbach flasks, with 2 L of medium each). Cultures were grown at 22 ^oC under fluorescent lights (2.6 klx) in an 18/6 h light/dark cycle. Sterile air was continuously passed into cultures. Cells were harvested by centrifugation and ~500 mg of wet cell mass was set aside for genomic DNA isolation. The remaining cells were lyophilized to allow for the isolation of 1.

Morphological Identification:

Cultured UIC 10036 was analyzed under a Zeiss Axiostar Plus light microscope with a Moticam 5 microscope camera (Supplementary Information S1). Observed morphological features were: cylindrical and isopolar trichomes, monoseriate, homogeneous width along the filament, thin sheaths, common false branching, cells shorter than wide, rounded apical cells, solitary heterocysts of quadrangular shape, no akinetes, hormogonia arising terminally by trichome disintegration. Preliminary morphological identification of UIC 10036 was based on the taxonomic systems by Komarek et al.^[38,39]

Phylogenetic Analysis and Genome Sequencing:

For genomic DNA isolation, ~250 mg of wet cell mass was resuspended in 2.5 mL of lysis buffer (10mM Tris, 0.1 M EDTA, and 0.5% w/v SDS) and lysozyme (Sigma-Aldrich, final concentration of 1 mg/mL). The mixture was incubated in a water bath at 37 ^oC for 30 min. Subsequently, proteinase K (Thermo Fisher Scientific) was added at a final concentration of 100 μg/mL. After incubating the mixture in a water bath for 1 h at 50 ^oC, the sample was centrifuged and the supernatant was removed. Subsequently, the Nucleospin^® soil kit (Macherey-Nagel) was used to extract genomic DNA from the pretreated cells according to the manufacturers’ instructions. The 16S rRNA gene was amplified by PCR on an Applied Biosystems 2720 Thermal Cycler using the cyanobacteria specific primers 109F (5’-CGGACGGGTGAGTAACGCGTGA-3’) and 1509R (5’-GGTTACCTTGTTACGACTT-3’). The 50 μL PCR reaction consisted of 100 ng of genomic DNA, 0.4 μM of each primer, 0.2 mM of each dNTP, and 0.05 U/μL Taq DNA polymerase (New England Biolabs) in the buffer supplied with the enzyme. The following thermal cycling conditions were used: 2 min at 95 ^oC, 35 cycles of 95 ^oC for 30 s, 49 ^oC for 30 s, 72 ^oC for 120 s, and a final hold for 5 min at 72 ^oC. PCR products were submitted for Sanger sequencing using the primers 109F, 359F (5’-GGGGAATTTTCCGCAATGGG-3’), and 1509R. The partial 16S rRNA sequence of Scytonema sp. UIC 10036 has been deposited at GenBank under accession no. MH780925.

Phylogenetic analysis of Scytonema sp. UIC 10036 was performed using Mega 7.0.^[40] Muscle was used to create a multiple sequence alignment containing strains from the Bergey’s Manual of Systematic Bacteriology^[41] as well as other strains available on NCBI. The evolutionary history was inferred using the Neighbor-Joining tree building method and 1000 bootstrap replicates were performed to determine the consistency of the tree. The bootstrap values are shown next to the branches of the tree and any branches with bootstrap values less than 75% were removed from the tree. Evolutionary distances were computed using the Kimura 2-parameter method with a gamma distribution of 0.55 (Supplementary Information S2).

For genomic DNA isolation for genome sequencing, ~250 mg of cell mass was washed with nuclease-free water and resuspended in 500 μL cell lysis solution (Promega) and lysozyme solution (Sigma-Aldrich, final concentration of 1 mg/mL). After incubating the mixture in a water bath at 37 ^oC for 30 min, the sample was centrifuged and the supernatant was removed. The pellet was washed with nuclease-free water, centrifuged, and the supernatant was removed. Subsequently, the Nucleospin^® soil kit (Macherey-Nagel) was used to extract genomic DNA from the pretreated cells according to the manufacturers’ instructions. Genomic DNA was prepared for sequencing using a Nextera XT DNA library preparation kit (Illumina, San Diego, CA, USA). Sequencing was performed on an Illumina Miseq system. Genomic DNA was also subjected to one RSII single-molecule real-time (SMRT) DNA sequencing on the Pacific Biosciences (PacBio) system with a 6-hour movie capture using P6-PC4 chemistry. Samples were prepped using standard PacBio templated protocol for 10kb libraries and size selected using Blue Pippin (Sage Science) instrument starting at 4kb. De novo hybrid assembly of Illumina and PacBio reads was performed using SPAdes v3.9.^[42,43]

Bioinformatic Analysis:

The genome of UIC 10036 was annotated using RAST.^[44–46] Biosynthetic gene clusters were roughly predicted with antiSMASH version 4.^[47] The identified putative scytodecamide (scd) biosynthetic gene cluster was further annotated using FgenesB (http:/www.softberry.com) and BLAST^[48] followed by manual curation. The putative scd biosynthetic gene cluster has been deposited at Genbank under accession no. MN265868, and at MiBIG under accession code BGC0001880. The genome of Scytonema sp. UIC 10036 has been deposited at GenBank under accession code WJFC00000000.

Extraction and Isolation:

After lyophilization, dry cell mass of UIC 10036 was extracted with 1 L CH₂Cl₂/MeOH 1:1 three times (24 h each time) at 4 ^oC with occasional agitation. Cell material was filtered and the extraction solution was concentrated in a rotatory evaporator to yield 1.3 g of dry extract. Extract was fractionated using a chromatographic column containing Diaion^® HP-20SS and applying a step gradient of IPA/H₂O (0:100, 20:80, 40:60, 70:30, 90:10, and 100:0). LC-MS analysis detected 1 in the fraction eluting at 40:60 IPA/H₂O (F3). Compound isolation was accomplished by semipreparative HPLC using an Agilent C18 column (250 × 10.0 mm), with a flow rate of 4 mL/min, and a gradient of H₂O (A) and ACN (B), both acidified with 0.1% formic acid. Gradient program was set as follows: 20% B for 5 min, 20–100% B for 35 min, 100% B for 10 min, 100–20% B for 5 min, 20% B for 4 min. Compound 1 (~1.1 mg) eluted at rt 16.1 min.

Scytodecamide (1): white amorphous powder; [α]²⁵_D none detected (c 0.09, MeOH); UV (MeOH) λ_max (log ε) 213 (3.65), 262 (3.58) nm; IR (neat) ʋ_max 3303 (br), 2960, 2931, 2873, 1663, 1543, 1436, 1387, 1369, 1311, 1205, 1185, 1138, 1058, 840, 801, 723 cm⁻¹; ¹H and ¹³C NMR (CD₃OD/CD₃OH and Pyr-d₅ with TFA vapour) see Table 1; HRESIMS m/z 996.6450 [M + H]⁺, calcd for C₄₇H₈₆N₁₁O₁₂, 996.6457.

Acid Hydrolysis and Marfey’s Analysis:

Approximately 0.2 mg of 1 was subjected to hydrolysis with 6 N HCl (1 mL) at 110 ^oC for about 18h. Residual HCl was removed by resuspending the sample with water (0.5 mL) and applying repeated evaporation. The hydrolysate of 1 was split into two equal portions (~0.1 mg each) prior to derivatization. Hydrolysate as well as amino acid standards were derivatized with l- and dl-Advanced Marfey’s reagent (l-FDLA and dl-FDLA). Approximately 0.1 mg of the hydrolysate or standards were mixed with 110 μL of acetone, 50 μL of H₂O, 20 μL of 1 N NaHCO₃, and 20 μL of l-FDLA or dl-FDLA solution (10 mg/mL in acetone). The mixtures were stirred at 40 ^oC for 1h. After 1h, the mixtures were removed from heat and quenched with 20 μL of 1 N HCl. Reaction mixtures were evaporated and resuspended in CH₃CN for LC-MS analysis. LC-MS analyses were performed on a Shimadzu LCMS-IT-TOF with a Phenomenex Kinetex C18 column (50 × 2.1 mm, 1.7 μm) and a gradient of H₂O (A) and CH₃CN (B) both with 0.1% formic acid. Flow rate was kept at 0.5 mL/min and gradient program went from 25 to 65 %B over 9 min. Extracted ion chromatograms and retention times were compared for each Marfey’s derivative for the assignment of absolute configuration (Supplementary Information S20).

Antiproliferative Assay:

Cytotoxicity assays were performed following established protocols.^[49] Tested cell lines included MDA-MB-435 (human melanoma), MDA-MB-231 (human breast cancer), and OVCAR3 (human ovarian cancer).