Isolation and Structural Elucidation of Euryjanicins B–D, Proline-Containing Cycloheptapeptides from the Caribbean Marine Sponge Prosuberites laughlini

Abstract

Three new cyclic peptides, euryjanicins B (2), C (3), and D (4), have been isolated from the Puerto Rican marine sponge Prosuberites laughlini, and the structures were elucidated by chemical degradation, ESIMS/MS, and extensive 2D NMR methods. When tested against the National Cancer Institute 60 tumor cell line panel, all of the purified isolates displayed weak cytotoxicity.

Cyclic peptides are a frequently encountered class of natural products that often exhibits a wide variety of essential biological functions. A large number of cyclic peptides with unique structures and diverse pharmacological activities have been reported from invertebrate animals and microorganisms typically associated with marine habitats.¹ Additionally, a smaller number of cyclic peptides have been isolated from higher plants.² Because of greater resistance to in vivo enzymatic degradation and higher bioavailability than their noncyclic counterparts, cyclic peptides often possess superior therapeutic potential.³ A noteworthy class of marine cyclopeptides is represented by proline-rich compounds, usually containing seven or eight amino acid residues. The role of proline in these molecules, which often occur as complex mixtures of structurally related derivatives, has been associated with control of conformation of the molecule in solution because of the restricted dihedral angle of proline.⁴ Examples of this kind of bioactive natural products are the axinellins,⁵ axinastatins,⁶ phakellistatins,⁷ hymenamides,⁸ stylopeptides,⁹ stylostatin 1,¹⁰ wainunuamide,¹¹ dominicin,¹² and others.¹³

Within the context of a long-term effort to discover new metabolites with antitumor and anti-infective properties from marine invertebrates from the Caribbean region, we recently had the opportunity to chemically scrutinize the bioactive extract of the marine sponge Prosuberites laughlini (order Hadromerida, family Suberitidae) collected in the waters off Puerto Rico.¹⁴ As a result, we recently reported the structure of euryjanicin A (1), a crystalline cyclopeptide whose basic structural motif consists of a 21-membered ring constructed from seven amino acid residues, two of which are prolines.¹⁵ In this paper we report the isolation, structure characterization, and bioactivity of three additional members of the family, named euryjanicins B–D (2–4), from the above marine sponge. All three compounds are proline-containing cycloheptapeptides, which, unlike euryjanicin A (1), generate well-resolved NMR spectra in CDCl₃. From these spectra it was obvious that for each cyclopeptide one conformation strongly predominated in this solvent.

Results and Discussion

The lyophilized sponge was extracted with MeOH, and the residue obtained was partitioned according to a modified Kupchan procedure, affording two extracts of increasing polarity: EtOAc and nBuOH.¹⁶ Our initial biological screenings showed that the EtOAc fraction retained most of the original antitumor activity ascribed to the extract but showed no antituberculosis activity. However, the nBuOH extract was strongly antitubercular but weakly cytotoxic. Preliminary chromatographic and ¹H and ¹³C NMR analyses revealed that the cytotoxic EtOAc fraction contained substances of a peptidic nature, whereas the strongly anti-infective nBuOH extract was mainly comprised of a mixture of the known pyrrole alkaloids monobromoisophakellin¹⁷ and hymenidin.¹⁸ Therefore, the EtOAc-soluble material was selected for further isolation work. This extract was subjected to normal-phase Si gel column chromatography (CC) using a mixture of CHCl₃ and MeOH that had been previously saturated with NH₃ (80/20). Similar fractions were pooled together on the basis of their TLC, NMR, and biological activity profiles. Subsequent evaluation of the first fraction by ¹H and ¹³C NMR indicated that this fraction contained a mixture of small peptides. Further fractionation was performed by normal-phase Si gel (100 g) CC with a mixture of hexane and EtOAc (95/ 5). Close inspection by NMR indicated that subfractions 10 and 11 appeared to contain most of the peptides observed in the initial NMR analysis. Purification of subfraction 10 was achieved by C18 Si gel reversed-phase CC; this yielded pure euryjanicin A (1),¹⁵ euryjanicin B (2), and the known cyclic octapeptide dominicin (5).¹² Subsequent purification of fraction 11 by reversed-phase HPLC afforded pure euryjanicin C (3) and euryjanicin D (4).

Structure Elucidation of Euryjanicin B

Compound 2 was obtained as an optically active, colorless oil that showed a pseudomolecular ion peak at m/z 710 in the positive ion ESIMS spectrum. The molecular formula of 2 was determined to be C₃₆H₅₁N₇O₈ by the HRESIMS (m/z 710.3871 [M+H]⁺), requiring 15 sites of unsaturation. The ¹³C NMR spectrum revealed resonances consistent with seven amide carbonyls (δ 173.3, 171.4, 171.3, 171.0, 170.7, 169.8, 169.5), seven α-methine carbons (δ 62.2, 60.7, 59.3, 56.5, 55.3, 53.6, 52.4), a secondary carbinol (δ 69.0), and a monosubstituted phenyl [δ 136.9 (C), 129.7 (CH), 128.6 (CH), 127.0 (CH)] ring system, suggesting a heptapeptide with phenylalanine and threonine units. The seven amino acids were unambiguously identified by 2D NMR techniques. Three independent spin systems of the type X–CH–CH₂–CH₂–CH₂–X′ were defined using TOCSY, COSY-GPQF, and HSQC (Table 1), indicating the presence of three proline units. This contention was further substantiated by the observation of only four peptide-bond NH proton signals in the ¹H NMR spectrum. The spin systems X–CH–CH(CH₃)₂, X–CHCH₃, and X–CH(OH)CH₃ were identified, suggesting the existence of valine, alanine, and threonine residues. The remaining independent spin system of the type X–CH–CH₂–X′ was attributed to phenylalanine by the HSQC correlations: δ 3.13 and 3.07 [βH₂(Phe⁴)]/δ 37.7 [βC(Phe⁴)]. The amino acid composition was confirmed by HPLC analyses of the acid hydrolysate of 2 after derivatization with the Marfey reagent (FDAA), allowing the absolute configurations at the α-carbons to be assigned as the L configuration for all residues.¹⁹ Because only 14 of the calculated 15 degrees of unsaturation could be accounted for by the functionalities in the seven individual amino acids, it became obvious that euryjanicin B was a cyclic peptide. Indeed, the cyclic nature of 2 was evident also by the high degree of chemical shift dispersion observed for the peptide-bond amide proton signals resonating between δ 7.72 and 6.14, its solubility in organic solvents, and the fact that it was negative to a ninhydrin test. The sequence of amino acids was assigned on the basis of a combined approach of 2D-NMR and electrospray tandem mass spectrometry techniques (ESIMS/MS).

Table 1.

¹H (500 MHz), ¹³C NMR (125 MHz), TOCSY, HMBC, and NOESY Correlations for Euryjanicin B (2)^a

amino acid	position	δC (mult)b	δH, mult (J in Hz)	TOCSYc	HMBCd	NOESYe
L-alanine1	CO	173.3, qC			α, β, αThr, NHThr2
	α	52.4, CH	4.28, dd (7.0, 7.5)	β, NH	β, NH	β
	β	18.7, CH3	1.54, d (7.5)	α, NH	α, NH	NH
	NH		7.72, d (7.0)	α, β		α, β, δPro7, NHThr2, αPro6
L-threonine2	CO	169.8, qC			α
	α	55.3, CH	4.75, dd (3.0, 8.0)	β, NH	γ, NH	β, γ, δPro3
	β	69.0, CH	4.02, br s	γ, NH	α, γ	γ, δPro3
	γ	20.0, CH3	1.13, d (6.0)	α, β, NH		α, β
	NH		7.54, d (8.0)	α, β, γ		α, αPro3
L-proline3 trans	CO	171.3, qC			α, NHPhe4
	α	62.2, CH	4.15, dd (8.5, 8.0)	β	δ	γ, βThr2
	β	29.5, CH2	2.35, m	α, γ, δ	α, δ	α
	γ	25.7, CH2	1.55, m	δ
			1.95, m	α, β, δ	δ	δ
	δ	47.7, CH2	3.68, m	γ, α	γ	γ
			3.42, m	γ		γ
L-phenylalanine4	CO	171.0, qC			α, β, NHVal5
	α	53.6, CH	4.80, dd (9.5, 15.5)	β, NH	β, NH	β
	β	37.7, CH2	3.13, dd (5.5, 14.0)	NH	α, ortho, NH	ortho, NH
			3.07, dd (7.5, 14.0)
	γ	136.9, qC			α, β, meta
	ortho	129.7, CH	7.16, d (7.0)		β, meta, para	α,β, NH
	meta	128.6, CH	7.34, m	ortho	para
	para	127.0, CH	7.29, m	ortho	ortho	ortho
	NH		6.14, d (9.5)	α, β		α, β, ortho, NH
L-valine5	CO	169.5, qC		α	α
	α	56.5, CH	4.36, dd (8.5, 9.0)	β, NH	β, γ, NH	γ, δPro6
	β	30.3, CH	2.06, m	γ	α, γ	α
	γ	19.8, CH3	0.97, d (6.5)	α, β	α
		17.9, CH3	0.94, d (6.5)	α, NH		α, NH
	NH		6.60, d (8.5)	α, β, γ		α, β, γ, NHPhe4
L-proline6 trans	CO	170.7, qC			α, β
	α	59.3, CH	4.59, dd (5.0, 8.5)	β, γ, δ	δ	αPro7
	β	28.6, CH2	2.14, m	δ	α, δ	α
			1.75, m	δ
	γ	24.8, CH2	2.11, m		α
			2.06, m
	δ	47.6, CH2	3.63, m	α, γ	β, γ	γ
			3.42, m	γ
L-proline7 cis	CO	171.4, qC			α, β, γ, NHAla1
	α	60.7, CH	4.35, d (7.0)	β, δ	γ, δ	βPro6, β, γ
	β	31.8, CH2	2.44, dd (7.0, 12.3)	α, γ, δ	α, δ
	γ	22.0, CH2	1.95, m	α	α, δ
			1.75, m	δ		α, δ
	δ	46.6, CH2	3.58, m	α, γ	α	γ

^aSpectra were recorded in CDCl₃ at 25 °C. Chemical shift values are in ppm relative to the residual CHCl₃ (7.25 ppm) or CDCl₃ (77.0 ppm) signals.

^b13C NMR multiplicities were obtained from APT experiments.

^cNMR data recorded in a Bruker (400 Mz) spectrometer.

^dHMBC mixing time = 50 ms.

^eNOESY mixing time = 200 ms.

The carbonyl carbons within each residue were assigned from HMBC correlations between the C=O and their respective α-protons (Table 1), whereas each peptide-bond amide proton was identified from ¹H–¹H COSY correlation to its adjacent α-proton and by reference to known values for these residues. The two fragments Pro⁷-Ala¹-Thr² and Pro³-Phe⁴-Val⁵ were assigned by two-bond ¹H–¹³C correlation as follows: NH (Ala¹)/CO (Pro⁷), NH (Thr²)/CO (Ala¹) and NH(Phe⁴)/CO (Pro³), NH(Val⁵)/CO (Phe⁴). The complete amino acid sequence of euryjanicin B (2) was established as cyclo-(–Ala-Thr-Pro-Phe-Val-Pro-Pro–) by the following pivotal NOE correlations: αH(Pro⁶)/αH(Pro⁷); αH(Val⁵)/δH₂(Pro⁶); βH(Thr²)/δH₂(Pro³) (Figure 1). Although the complete sequence was not unequivocally solved using the available NMR data, it was confirmed on the basis of the results of ESIMS/MS experiments. The mass spectrum of euryjanicin B displayed [M + H]⁺ at m/z 710, the fragmentation of which was followed by MSⁿ. The ESIMS/MS spectrum of euryjanicin B showed preferential opening of the macrocycle at the Val⁵-Pro⁶ amide bond, followed by a series of major fragmentation pathways. One ion series started with loss of 99 amu due to Val⁵, leaving m/z 611 (Pro⁶-Pro⁷-Ala¹-Thr²-Pro³-Phe⁴ plus H), which then lost 244 amu (Pro³-Phe⁴), affording m/z 367 (Pro⁶-Pro⁷-Ala¹-Thr² plus H). The latter fragment lost 101 amu (Thr²), yielding m/z 266 (Pro⁶-Pro⁷-Ala¹ plus H) (Figure 2). Another pathway left the major fragment m/z 593 [M+H–H₂O–Val⁵]⁺ after loss of 99 amu (Val⁵) from m/z 664 [M+H–H₂O]⁺, which then lost sequentially 147 amu (Phe⁴) and 97 amu (Pro³), leaving m/z 446 [M+H–H₂O–Val⁵-Phe⁴]⁺ and 349 [M+H–H₂O–Val⁵-Phe⁴-Pro³]⁺, respectively. The daughter ion spectrum also contained additional fragment ions of lower abundance ascribable to the y ion series at m/z 613 [M+H–Pro⁶]⁺, 516 [M+H–Pro⁶-Pro⁷]⁺, and 445 [M+H–Pro⁶-Pro⁷-Ala¹)]⁺.

Figure 1.

HMBC (H→C) and NOE correlations (dashed arrows) for euryjanicin B (2) in CDCl₃.

Figure 2.

Major fragmentation pathways of 2 following opening of the protonated cyclic structure during ESIMS/MSⁿ.

The structure of euryjanicin B (2) was thus confirmed as cyclo-(–Ala-Thr-Pro-Phe-Val-Pro-Pro–). The geometry of the peptidic linkages was assigned on the basis of the differences in ¹³C chemical shifts of the Cβ and Cγ of the proline residues.²⁰ The ¹³C NMR data of euryjanicin B indicated that two proline peptide bonds were trans, as shown by the small ¹³C NMR chemical shift difference of Pro³ΔδCβ–Cγ = 3.8 and Pro⁶ΔδCβ–Cγ = 3.8, and one cis, as indicated by the large ¹³C chemical shift difference of Pro⁷ΔδCβ–Cγ = 9.8 (Table 1).^21,22 Adjacent cis and trans proline residues, which have been previously found in cyclic peptides such as wainunuamide¹¹ and phakellistatin 8,²³ are known to be powerful β-turn inducers.^4a

Structure Elucidation of Euryjanicin C

Euryjanicin C (3), a white semisolid, showed a high-resolution ESIMS spectral quasimolecular ion peak at m/z 768.4662 ([M+H]⁺) consistent with the molecular formula C₄₀H₆₂N₇O₈, requiring 14 degrees of unsaturation. The IR absorption bands at 3324 and 1651 cm⁻¹ were attributed to amino and amide carbonyl groups, respectively. The heptapeptide nature of 3 was evident from the molecular formula and ¹³C NMR spectra, which showed seven carbonyl signals (δ 172.9, 171.8, 171.5, 171.0, 171.0, 170.8, 169.5) and seven α-methine carbons (δ 61.4, 61.1, 59.6, 55.3, 54.4, 54.1, 50.5) (Table 2). Only five amide NH signals were detected in the ¹H NMR spectrum, suggesting a heptapeptide with two proline units. As in 2, the amino acid residues were identified by extensive NMR analyses—COSY-GPQF, TOCSY, HSQC, HMBC, and NOESY—which allowed us to establish the identity of the seven residues: 2 × Pro, 2 × Ile, Phe, Ser, and Leu. These residues accounted for 13 degrees of unsaturation out of the 14, requiring that 3 is a cyclopeptide, too. The absolute configurations at the α-carbon for these amino acids were based on Marfey’s method, and we also found in this case that all residues belong to the L series. The amino acid sequence analysis was conducted in a similar manner to that for 2, by 2D NMR analysis, e.g., HMBC and NOESY spectra (Table 2). From the results of the useful HMBC correlations, as shown in Figure 3, the structure was established to be cyclo-(–Ser-Ile-Pro-Leu-Phe-Pro-Ile–), which was also confirmed by the correlations observed during NOESY NMR experiments (Figure 3).

Table 2.

¹H NMR (500 MHz), ¹³C NMR (125 MHz), TOCSY, HMBC, and NOESY Correlations for Euryjanicin C (3)^a

amino acid	position	δC (mult)b	δH, mult (J in Hz)	TOCSY	HMBCc	NOESYd
L-serine1	CO	169.5, qC			α, NHIle2
	α	54.4, CH	4.50, m	β		β, γIle2, δIle2
	β	63.5, CH2	3.80, d (2.0)
	NH		6.54, d (8.5)	α, β		NHIle7, α
L-isoleucine2	CO	172.9, qC			α
	α	55.3, CH	4.31, dd (9.3, 10.5)	β, ε	β	β, δ, ε
	β	37.2, CH	1.79, m	γ, ε	α
	γ	24.6, CH2	1.63, m	δ, ε	γ
			1.20, m
	δ	10.6, CH3	0.94, t (7.5)
	ε	14.9, CH3	0.86, d (7.0)		α
	NH		6.87, d (9.0)	α	COSer1	αSer1, βSer1, βIle2
L-proline3 cis	CO	171.8, qC			NHLeu4, α
	α	61.4, CH	4.46, d (7.0)	β, γ, δ	β, γ, δ	αIle2, εIle2, β
	β	31.7, CH2	2.40, dd (7.0, 12.0)	γ	α, δ
			2.05, m
	γ	22.1, CH2	1.97, m	δ	α, δ	δ
			1.71, m	δ
	δ	46.4, CH2	3.62, m	β, γ	α, β	β, γ
			3.49, m	β, γ
L-leucine4	CO	171.5, qC			α, β, NHPhe5
	α	50.5, CH	4.51, m	β, δ	β, γ	β, δ
	β	36.6, CH2	1.61, m	γ, δ	α	γ
			1.42, m	γ, δ
	γ	24.8, CH	1.46, m	δ	β, δ
	δ	16.0, CH3	0.89, d (3.5)
		21.8, CH3	0.81, d (7.0)		β, δ
	NH		7.60, d (6.0)	α	COPro3, δ	α, β
L-phenylalanine5	CO	171.0, qC			β
	α	54.1, CH	4.62, dt (3.5, 8.3)	β	β, NHPhe5	β, αPro6
	β	38.0, CH2	3.05, dd (5.0, 8.3)	γ	ortho
	γ	134.9, qC			β, meta
	ortho	129.7, CH	7.24, d (7.5)		β, para	α, β
	meta	128.9, CH	7.31, m		ortho
	para	127.5, CH	7.28, m		ortho
	NH		7.55, d (3.5)	α		α, β, αLeu4, NHLeu4
L-proline6 cis	CO	171.0, qC			α, NHIle7
	α	61.1, CH	3.89, d (7.0)	β, γ	β, γ	β, αPhe5, βPhe5
	β	30.9, CH2	2.18, dd (7.0, 12.0)	γ, δ	α, δ
			1.16, m
	γ	21.7, CH2	1.77, m	β	α, β, δ
			1.57, m	δ	β
	δ	46.5, CH2	3.59, m	β, γ	α, β	γ
			3.45, m			γ
L-isoleucine7	CO	170.8, qC			α, NHSer1
	α	59.6, CH	4.28, dd (5.5, 9.0)	β, δ, ε	NHIle7, β, γ	β, ε
	β	36.1, CH	2.09, m	ε	α
	γ	24.8, CH2	1.42, m	δ	α
			1.21, m	δ	α, γ
	δ	11.4, CH3	0.86, t (7.5)
	ε	22.9, CH3	0.90, d (4.0)
	NH		7.50, d (9.0)	δ		α, β

^aSpectra were recorded in CDCl₃ at 25 °C. Chemical shift values are in ppm relative to the residual CHCl₃ (7.25 ppm) or CDCl₃ (77.0 ppm) signals.

^b13C NMR multiplicities were obtained from APT experiments.

^cHMBC mixing time = 50 ms.

^dNOESY mixing time = 200 ms.

Figure 3.

HMBC (H→C) and NOE correlations (dashed arrows) for euryjanicin C (3) in CDCl₃.

Analogously to the preceding cyclopeptide, the protonated molecular ion of euryjanicin C at m/z 768 was fragmented by ESIMS/MS. A complex spectrum was recorded due to simultaneous fragmentation of two isomeric linear peptides, following the opening of the cyclic structure through preferential fragmentation at the Ser¹-Ile⁷ and Pro⁶-Ile⁷ amide bonds. After loss of H₂O the major linear peptide arising upon fragmentation at Ser¹-Ile⁷ left a major fragment at m/z 750 [M+H–H₂O]⁺, which in turn underwent N-terminal fragmentation, leaving m/z 637 [M+H–H₂O–Ile⁷]⁺, 540 [M+H–H₂O–Ile⁷-Pro⁶]⁺, 393 [M+H–H₂O–Ile⁷-Pro⁶-Phe⁵]⁺, and 280 [M+H–H₂O–Ile⁷-Pro⁶-Phe⁵-Leu⁴]⁺. On the other hand, fragmentation of the lesser abundant linear peptide generated after cleavage of the Pro⁶-Ile⁷ peptide bond left a major fragment at m/z 655, corresponding to [M+H–Ile⁷]⁺, which underwent sequential losses of 87 amu (Ser¹), 113 amu (Ile²), and 97 amu (Pro³), leaving m/z 568 [M+H–Ile⁷-Ser¹]⁺, 455 [M+H–Ile⁷-Ser¹-Ile²]⁺, and 358 [M+H–Ile⁷-Ser¹-Ile²-Pro³]⁺, respectively. These related patterns of b fragments could be clearly distinguished in the daughter ion mass spectrum, leading to the reconstruction of the sequences inferred previously from 2D-NMR experiments.

The solution (CDCl₃) conformation about the proline peptide linkages of 3 could be assigned from the chemical shift differences between the proline β- and γ-carbons. Thus, Pro³ and Pro⁶ showed ΔδCβ–Cγ = 9.6 and 9.2, respectively, indicating that these proline peptide bonds are both cis. Confirmation for these assignments came from the NOESY spectrum. Cross-peaks between αH(Pro³)/αH(Ile²) established the cis geometry for the Pro³ peptide bond, while cross-peaks between αH(Pro⁶)/αH(Phe⁵) confirmed the cis-peptide link for the Pro⁶ peptide bond. In point of fact, stylopeptide 1, isolated by the Pettit group from the South and Western Pacific Ocean sponges Stylotella sp. and Phakellia costata and whose structure was established as cyclo-(–Pro-Leu-Ile-Phe-Ser-Pro-Ile–), is a structural isomer of euryjanicin C (3).^9a However, while both cyclopeptides are comprised of the same amino acid residues, there is no sequential homology among them.

Structure Elucidation of Euryjanicin D

Euryjanicin D (4) was obtained as an optically white semisolid that gave a [M + H]⁺ ion in the HRESIMS at m/z 802.4514, appropriate for a molecular formula of C₄₃H₅₉N₇O₈, requiring 18 sites of unsaturation. General features of both the ¹H and ¹³C NMR spectra (Table 3) suggested 4 was a peptide. Seven ¹³C NMR resonances had chemical shifts appropriate for amide carbonyls, and seven resonances had chemical shifts appropriate for amino acid α-carbons, consistent with a heptapeptide. Detailed analyses of the COSY-GPQF, TOCSY, HSQC, and HMBC data (Table 3) revealed that euryjanicin D (4) contained one serine and two each of proline, isoleucine, and phenylalanine residues. Hydrolysis of peptide 4 with 6 N HCl followed by Marfey’s derivatization and HPLC analyses of the amino acids in the hydrolysate confirmed the presence of these four amino acids and established that all seven amino acid residues had the L configuration. The seven amino acid residues accounted for all of the atoms in the molecular formula of euryjanicin D (4) and 17 of the sites of unsaturation. Furthermore, there was no evidence for terminal amino or carboxylic acid functionalities, and therefore, euryjanicin D (4) was presumed to be a cyclic peptide.

Table 3.

¹H NMR (500 MHz), ¹³C NMR (125 MHz), TOCSY, HMBC, and NOESY Correlations for Euryjanicin D (4)^a

amino acid	position	δC (mult)b	δH, mult (J in Hz)	TOCSY	HMBCc	NOESYd
L-phenylalanine1	CO	170.8, qC			α, β, NHSer2
	α	53.8, CH	4.51, m	β	β	β, ortho
	β	38.6, CH2	3.05, m		ortho	ortho
	γ	135.0, qC			β, meta
	ortho	129.2, CH	7.22, d (7.0)		β, para
	meta	128.9, CH	7.33, m
	para	127.5, CH	7.27, m		ortho
	NH		7.26, d (5.0)	α, β	β	α, βIle7
L-serine2	CO	170.0, qC			β
	α	54.9, CH	4.51, m	β	β	β, αPro7, NH
	β	63.2, CH2	3.89, br d (14.5)		α
			3.71, dd (5.0, 11.0)			βPhe1
	NH		7.47, d (6.0)	α, β	α, COPhe1	α
L-proline3 cis	CO	171.7, qC			α, NHIle4
	α	61.4, CH	3.59, d (7.5)	β, γ, δ	β, γ, δ	β
	β	30.9, CH2	1.96, m	γ	γ, δ
			0.95, m	γ
	γ	21.6, CH2	1.68, m	β	α, β, δ
			1.44, m	β
	δ	46.1, CH2	3.38, d (9.5)	β, γ	α, β
L-isoleucine4	CO	171.4, qC			α, NHPhe5
	α	60.4, CH	4.15, d (8.8)	β, ε	ε, NH	δ, ε
	β	32.3, CH	1.91, m	α, δ, ε	α, δ, ε	γ, δ, ε
	γ	24.4, CH2	1.40, m	δ	α, δ, ε	δ
			0.95, m
	δ	9.89, CH3	0.77, d (7.0)
	ε	15.8, CH3	0.84, d (6.5)		α, β, γ
	NH		7.11, d (8.8)	α, ε	β, γ, δ, ε
L-phenylalanine5	CO	171.6, qC			β
	α	53.2, CH	4.51, m	β	β, ortho, NH	β, αPro6, NH
	β	37.8, CH2	3.08, m		ortho, meta, NH	ortho, NH
	γ	135.4, qC			α, meta
	ortho	129.6, CH	7.26, d (5.0)		para
	meta	128.9, CH	7.33, m
	para	127.5, CH	7.27, m		ortho
	NH		7.71, d (7.0)	α, β		β, γ, ε
L-proline6 cis	CO	170.6, qC			α, NHIle7
	α	61.1, CH	3.92, d (7.5)	β, γ, δ	β, δ	β, orthoPhe5
	β	30.7, CH2	2.18, dd (6.0, 12.0)	γ	α, γ
			1.04, m	γ		γ
	γ	21.7, CH2	1.45, m	β	α, β	β
			1.74, m	β
	δ	46.4, CH2	3.54, dt (8.0, 11.8)	β, γ	α, β
			3.41, m			γ
L-isoleucine7	CO	171.6, qC			α, NHPhe1
	α	56.4, CH	4.12, d (9.3)	β	γ, δ	NHPhe1
	β	36.2, CH	2.09, m	ε	α, δ, ε	γ, δ, ε
	γ	25.0, CH2	1.44, m	δ	ε, α, δ
			1.12, m			δ
	δ	11.1, CH3	0.85, d (7.0)	α, γ	γ
	ε	15.8, CH3	0.90, d (7.0)		α, γ
	NH		7.67, d (9.3)	α, β, δ		β, γ, ε

^aSpectra were recorded in CDCl₃ at 25 °C. Chemical shift values are in ppm relative to the residual CHCl₃ (7.25 ppm) or CDCl₃ (77.0 ppm) signals.

^b13C NMR multiplicities were obtained from APT experiments.

^cHMBC mixing time = 50 ms.

^dNOESY mixing time = 200 ms.

The sequence analysis was conducted in a similar fashion to those in 2 and 3 by 2D-NMR analysis, e.g., HMBC and NOESY spectra. From the HMBC correlations as shown in Figure 4, the structure was established to be cyclo-(–Phe-Ser-Pro-Ile-Phe-Pro-Ile–), which was also confirmed by NOE correlations observed in the NOESY spectrum. Finally, this sequence was confirmed by tandem MS fragmentation analysis of a linear acylium ion at m/z 784 generated after loss of H₂O following preferential opening of the macrocycle at the Phe¹-Ile⁷ amide bond (Figure 5). Once the amino acid sequence of euryjanicin D (4) was determined, the final structural feature to be elucidated was the geometry of the peptidic linkages at the Pro residues. In this instance, the NMR data of 4 pointed to a cis geometry for both the Ser²-Pro³ and Phe⁵-Pro⁶ peptide bonds (ΔδCβ–Cγ = 9.3 and 9.0, respectively). Subsequent confirmation for these assignments followed from the NOESY spectrum. Strong cross-peaks between αH(Ser²)/αH(Pro³) and αH(Phe⁵)/αH(Pro⁶) confirmed that indeed the proline peptide bonds were cis.

Figure 4.

HMBC (H→C) and NOE correlations (dashed arrows) for euryjanicin D (4) in CDCl₃.

Figure 5.

N-Terminal and C-terminal MSⁿ fragmentations of the acyclic acylium ion at m/z 784 generated from cyclopeptide 4 after in-source protonation, ring-opening, and loss of H₂O.

Our overall chemical investigation of the cell growth inhibitory sponge extract yielded four new cyclic heptapeptides, euryjanicin A–D (1–4), and the previously known cyclic octapeptide dominicin (5). These molecules are structurally related to several other bioactive proline-rich cyclopeptides reported from marine sponges.^5–13 Yet, the isolated peptides 1–5 were found to be only marginally active to inactive against the National Cancer Institute (NCI) 60 tumor cell line panel (Table 4). The problem of lost activity associated with proline-rich cyclic peptides has been observed by others.^9a,13b This phenomenon may be due to changes in the conformation of the cyclic peptide during isolation or to its ability to bind to a potent antineoplastic substance present in very low concentration and therefore detectable only during biological screenings.^7,24 Peptides 1–5 have been tested for their antimalarial (Plasmodium falciparum W2 clone), anti-Mtb, antiviral [HBV, HSV-1, HSV-2, FluA (H5N1), FluB, RSV A, WNV, Rift Valley fever], and anti-inflammatory activity in rat neonatal microglia and were inactive in all of the assays.

Table 4.

Percent Growth of Selected Human Cancer Cell Lines Treated with Euryjanicins A–D (1–4) and Dominicin (5)^a

cell line	1	2	3	4	5
NCI-H522b	72.2	83.9	95.1	94.6	100
LOX IMVIc	89.0	88.6	88.0	87.6	100
IGROV1d	82.6	91.3	100	100	100
UO-31e	73.6	76.3	83.4	82.7	32.8

^aInhibitory activity tested at a single high dose (10⁻⁵ M) against the full NCI 60 cell line panel.

^bNon-small cell lung cancer.

^cMelanoma.

^dOvarian cancer.

^eRenal cancer.

Cycloheptapeptides 1–4 and cyclooctapeptide 5 are characterized by the presence of two or three prolines, an array of apolar residues such as Leu, Ile, and Val, and (in the case of 1–4) one or two aromatic residues such as Phe or Trp. Euryjanicins A (1), C (3), and D (4) each contains a serine residue, whereas euryjanicin B (2) and dominicin (5) possess one threonine unit. Except for metabolites 1 and 5, one set of resonances was observed for each of the amino acid residues in compounds 2–4, indicating that in each instance one rigid solution conformation dominates in CDCl₃. On the other hand, the ¹H and ¹³C NMR spectra of 1 and 5 in CDCl₃ gave broad signals resulting from the equilibria of slow exchanging conformers about the peptidic ring.^12,15 The remarkable analogy in structure and in amino acid content between cyclic peptides isolated from marine sponges and those stemming from marine-derived cultured bacteria or fungi suggests a possible microbial or symbiotic origin for sponge-derived cyclic peptides such as 1–5.²⁵

Experimental Section

General Experimental Procedures

Optical rotations were recorded with a Rudolph Autopol IV polarimeter. The UV data were recorded with a Shimadzu UV-2401PC spectrometer, and the IR analyses were performed with a Nicolet Magna IR 750 spectrometer. 1D and 2D NMR data were recorded on a Bruker DRX-500 or Bruker AV-500 FT-NMR spectrometer. ¹H NMR chemical shifts were recorded with respect to the residual CHCl₃ signal (7.26 ppm), and ¹³C NMR chemical shifts were reported in ppm relative to the CDCl₃ signal (77.0 ppm). Mass spectrometric measurements were generated at the Mass Spectrometry Laboratory of the University of Illinois at Urbana–Champaign. Column chromatography was performed on Analtech Si gel (35–75 mesh) and monitored with TLC analyses carried out on Analtech glass precoated Si gel plates and visualized using UV light and I₂ vapors. All solvents were distilled from glass prior to use.

Animal Material

The Caribbean sponge Prosuberites laughlini (Diaz, Alvarez & van Soest, 1987)²⁶ (phylum: Porifera; class: Demospongiae; order: Hadromerida; family: Suberitidae) was collected at a depth of 50 feet by scuba off Aguadilla, Puerto Rico, in April 2006. A voucher specimen (No. PLAG-01) is stored at the Chemistry Department of the University of Puerto Rico. The encrusting (0.3–4.0 cm thick) sponge was dull orange to yellow externally, lighter internally with surface visually smooth in thin specimens, rugose on thicker ones. The specimen collected possessed oscules with transparent membranes (1–6 mm in diameter) with thin (1–2 mm) canals departing radially and was soft and compressible, easy to tear.

Extraction and Isolation

The freshly collected sponge was freeze-dried for 5 days, and the dried organism (622.7 g) was repeatedly extracted with MeOH (16 L). The combined MeOH extracts were evaporated to dryness, and the resulting brown oil (83.1 g) was partitioned between EtOAc (3 × 1 L), nBuOH (3 × 1 L), and H₂O (1 × 1.5 L). The combined EtOAc extracts were concentrated in vacuo to give 6.1 g of a dark brown oil that was chromatographed over Si gel (201 g) with 20% MeOH • NH₃ in CHCl₃ as eluent. The first fraction (2.79 g) was subsequently chromatographed over Si gel (100 g) with 5% EtOAc in hexane as eluent. Subfractions 10 (182.4 mg) and 11 (155.3 mg) contained all of the peptides. Fraction 10 was purified by reversed-phase C18 column chromatography (5 g) using 15% H₂O in MeOH, yielding pure euryjanicin A (1) (43.6 mg; 0.007%), euryjanicin B (2) (21 mg; 0.003%), and dominicin (5) (32.3 mg; 0.005%). Purification of fraction 11 via C18 reversed-phase HPLC using a 10 mm × 25 cm Ultrasphere ODS column, 5 μm, with 30% H₂O in MeOH yielded pure euryjanicin C (3) (15.3 mg; 0.002%) and euryjanicin D (4) (39.9 mg; 0.006%).

Euryjanicin B (2)

colorless oil; [α]²⁰_D −55 (c 0.8, CHCl₃); IR (film) ν_max 3345, 3087, 3064, 2973, 1667, 1518, 1454, 753 cm⁻¹; UV (MeOH) λ_max (log ε) 204 (4.2) nm; ¹H NMR (500 MHz, CDCl₃) and ¹³C NMR (125 MHz, CDCl₃) (see Table 1); HRESIMS m/z [M + H]⁺ 710.3871 (calcd for C₃₆H₅₂N₇O₈, 710.3877).

Euryjanicin C (3)

white semisolid; [α]_D²⁰ −60 (c 1.1, CHCl₃); IR (film) ν_max 3324, 3063, 3028, 2961, 1651, 1530, 1449, 754 cm⁻¹; UV (MeOH) λ_max (log ε) 204 (3.9) nm; ¹H NMR (500 MHz, CDCl₃) and ¹³C NMR (125 MHz, CDCl₃) (see Table 2); HRESIMS m/z [M +H]⁺ 768.4662 (calcd for C₄₀H₆₂N₇O₈, 768.4660).

Euryjanicin D (4)

white semisolid; [α]_D²⁰ −123 (c 0.85, CHCl₃); IR (film) ν_max 3307, 3063, 3029, 2963, 1667, 1516, 1454, 1347, 752, 702 cm⁻¹; UV (MeOH) λ_max (log ε) 205 (4.0); ¹H NMR (500 MHz, CDCl₃) and ¹³C NMR (125 MHz, CDCl₃) (see Table 3); HRESIMS m/z [M + H]⁺ 802.4514 (calcd for C₄₃H₆₀N₇O₈, 802.4503).

Acid Hydrolysis of Euryjanicins B–D (2–4)

Pure euryjanicins B–D (0.5 mg) were hydrolyzed in 0.5 mL of 6 N HCl at 110 °C for 12 h in a 1.0 mL reaction vial. The cooled reaction mixture was evaporated to dryness, and traces of HCl were removed from the residual hydrolysate by repeated evaporation from H₂O (3 × 0.5 mL) using N₂ gas.

Absolute Configuration of Amino Acids

To a 4 mL vial containing 1 μmol of pure amino acid standards in 200 μL of H₂O was added 1 μmol of N-α-(2,4-dinitro-5-fluorophenyl)-L-alanine amide (L-FDAA) in 400 μL of acetone followed by 100 μL of 1 N NaHCO₃. The mixture was heated for 1 h at 40 °C. After cooling to rt, 100 μL of 2 N HCl was added and the resulting solution was filtered through a small 4.5 mm filter and stored in the freezer until HPLC analysis. Half of each peptide hydrolysate mixture was dissolved in 200 μL of H₂O, and to this was added 1.5 μmol of L-FDAA in 400 μL of acetone followed by 100 μL of 1 N NaHCO₃. The derivatization reaction was carried out and worked up as described above. An 8 μL aliquot of the resulting mixture of L-FDAA derivatives was analyzed by reversed-phase HPLC. A 5 mm × 250 mm Spheri-5 C18 column, 5 μm, with a linear gradient of (A) 9:1 triethylammonium phosphate (50 mM, pH 3.0)/CH₃CN and (B) CH₃CN with 0% B at start to 40% B over 55 min (flow rate = 1 mL/min) was used to separate the L-FDAA derivatives with UV detection at 340 nm. Each chromatographic peak was identified by comparing its retention time with the L-FDAA derivative of the pure L-amino acid standard and by co-injection. In all cases a peak at 39.2 min was observed, which was attributed to excess L-FDAA. Retention times (min) are given in parentheses: L-Pro (32.11), L-Ile (46.85), L-Leu (47.40), L-Thr (26.15), L-Ala (31.74), L-Val (40.70), L-Phe (39.23), L-Ser (23.2), L-Trp (46.0).

Electrospray Ionization Mass Spectroscopic (ESIMS/MS) Analyses

ESIMS/MS analyses were performed with a LTQ ion trap mass spectrometer (Thermo Fisher). Aliquots of the peptide (100 μL) in solution (50% CH₃CN(aq)/0.1% formic acid) were injected in the ion source at a flow rate of 1 μL/min, the capillary temperature was 200 °C, and the applied voltage was 1 kV. Fragmentation experiments were carried out using a collision energy of dissociation of 20%. All ESIMS are reported as averaged mass.

Biological Assays

Additional experimental details for our primary in vitro antimicrobial assays against Mycobacterium tuberculosis and Plasmodium falciparum have been previously described.^27,28 All of the in vitro antiviral, anti-inflammatory, and cancer cell cytotoxicity assays for cyclopeptides 1–5 were used as indicated.^29–31

Chart 1.