Revisiting the enniatins: a review of their isolation, biosynthesis, structure determination, and biological activities

Abstract

Enniatins are cyclohexadepsipeptides isolated largely from Fusarium species of fungi, although they have been isolated from other genera, such as Verticillium and Halosarpheia. They were first described over 60 years ago, and their range of biological activities, including antiinsectan, antifungal, antibiotic, and cytotoxic, drives contemporary interest. To date, 29 enniatins have been isolated and characterized, either as a single compound or mixtures of inseparable homologs. Structurally, these depsipeptides are biosynthesized by a multifunctional enzyme, termed enniatin synthetase, and are composed of six residues that alternate between N-methyl amino acids and hydroxy acids. Their structure elucidation can be challenging, particularly for enniatins isolated as inseparable homologs; however, several strategies and tools have been utilized to solve these problems. Currently, there is one drug that has been developed from a mixture of enniatins, fusafungine, which is used as a topical treatment of upper respiratory tract infections by oral and/or nasal inhalation. Given the range of biological activities observed for this class of compounds, research on enniatins will likely continue. This review strives to digest the past studies, as well as, describe tools and techniques that can be utilized to overcome the challenges associated with the structure elucidation of mixtures of enniatin homologs.

A. Introduction and History of Enniatins

Enniatins are cyclohexadepsipeptides composed of alternating residues of three N-methyl amino acids, commonly valine (Val), leucine (Leu), and isoleucine (Ile), and three hydroxy acids, typically hydroxyisovaleric acid (Hiv). These secondary metabolites have been observed frequently from Fusarium species of fungi, although they have been isolated from a few other genera, as well (Table 1). The first enniatin [enniatin A (1)] was reported from the fungus Fusarium orthoceras var. enniatinum by Gaumann et al. in 1947,¹ and 1 was shown to be active against several mycobacteria in vitro. Shortly thereafter, another group working independently, isolated a compound they named lateritiin-I from Fusarium lateritium, which was found to be identical to 1.^2,3 In 1974, Audhya and Russel⁴ identified that previously reported enniatin A was actually a mixture of threo and erythro optical isomers. At that time only enniatin B had been observed pure from a natural source.⁵ Scientists worked to scale up the production of enniatin A, largely for the purpose of studying its ionophoric properties, which had started to emerge as the probable mechanism of action for enniatins.⁵ One isolation study^4,6 yielded a mixture ofenniatins A (1), A₁ (3), B (2), and B₁ (4), while an earlier insecticidal assay-guided study revealed that an enniatin A and A₁-rich complex was the active component in a fraction toxic to spruce budworm larvae.⁷ However, in 1992 the first report on the purification of enniatins A (1), A₁ (3), B (2), B₁ (4), and new analog A₂ (5),⁸ was published. From that point on, the isolation and characterization of enniatins expanded to include: enniatins B₂ (6), B₃ (7) and B₄ (8)⁹ [D (8), which has the same structure as B₄, was published in the same year]¹⁰; C (9);¹¹ E₁ (10), E₂ (11) and F (12);¹⁰ G (13);¹² H (14) and I (15) together with biosynthesized analogs G (13), C (9), and MK1688 (16);¹³ J₁ (17), J₂ (18), J₃ (19), and K₁ (20);¹⁴ L (21), M₁ (22), M₂ (23), and N (24);¹⁵ O₁ (25), O₂ (26), and O₃ (27);¹¹ and most recently, P₁ (28) and P₂ (29).¹⁶ Enniatin C (9) was reported first as a synthetic product,⁴ then as a biosynthetic product,¹³ before it was isolated as a natural product.¹¹

Table 1.

Sources and bioactivities of enniatins derived from the literature.

Enniatin	Biological Sourcea	Bioactivities
A (1)	Fusarium avenaceum;50 Fusarium sp. FO-130510	Cytotoxic against Hep G2, MRC-5; ACAT inhibition
A1 (3)	Fusarium avenaceum;50 Fusarium sp. FO-130510	Cytotoxic against Hep G2, MRC-5; ACAT inhibition
A2 (5)	Fusarium avenaceum50	Cytotoxic against Hep G2, MRC-5
B (2)	Fusarium avenaceum;50 Fusarium sp. FO-1305;10 Verticillium hemipterigenum13	Cytotoxic against Hep G2, MRC-5; ACAT inhibition
B1 (4)	Fusarium avenaceum;50 Fusarium sp. FO-130510	Cytotoxic against Hep G2, MRC-5; ACAT inhibition
B2 (6)	Fusarium acuminatum9 A, A1, B, B1, B2, B4	Cytotoxic against Hep G2, MRC-5;
B3 (7)	Fusarium acuminatum9	Cytotoxic against Hep G2, MRC-5
B4 = D (8) Because both were published in 1992	Fusarium acuminatum;9 Fusarium sp. FO-1305;10 Verticillium hemipterigenum13	ACAT inhibition
C (9)	Verticillium hemipterigenum13 B, B4, G, C	Antimalarial, antituberculous, cytotoxic
E1 (10)	Fusarium sp. FO-130510	ACAT inhibition
E2 (11)	Fusarium sp. FO-130510	ACAT inhibition
F (12)	Fusarium sp. FO-130510	ACAT inhibition
G (13)	Halosarpheia sp. (strain732)12 B, B4	Heps 7402
H (14)	Verticillium hemipterigenum13 B, B4, G, C	Antimalarial, antituberculous, cytotoxic
I (15)	Verticillium hemipterigenum13 B, B4, G, C	Antimalarial, antituberculous, Cytotoxic
MK1688 (16)	Verticillium hemipterigenum13 B, B4, G, C	Antimalarial, antituberculous, Cytotoxic
J1 (17)	Fusarium sp. F3114 B, B1, B2, B4	Against Botrytis cinerea
J2 (18)	Fusarium sp. F3114 B, B1, B2, B4	Against Botrytis cinerea
J3 (19)	Fusarium sp. F3114 B, B1, B2, B4	Against Botrytis cinerea
K1 (20)	Fusarium sp. F3114 B, B1, B2, B4	Against Botrytis cinerea
L (21)	Unidentified fungus BCC262915 B, B4, H, I	Antimalarial, antituberculous, cytotoxic
M1 (22)	Unidentified fungus BCC262915 B, B4, H, I	Antimalarial, antituberculous, cytotoxic
M2 (23)	Unidentified fungus BCC262915	Antimalarial, antituberculous, cytotoxic
N (24)	Unidentified fungus BCC262915 B, B4, H, I	Antimalarial, antituberculous, cytotoxic
O1 (25)	Verticillium hemipterigenum11 B, B4, G, C, H, I	Antimalarial, antituberculous, cytotoxic
O2 (26)	Verticillium hemipterigenum11 B, B4, G, C, H, I	Antimalarial, antituberculous, cytotoxic
O3 (27)	Verticillium hemipterigenum11 B, B4, G, C, H, I	Antimalarial, antituberculous, cytotoxic
P1 (28)	Unidentified Fusarium sp. VI 0344116	-
P2 (29)	Unidentified Fusarium sp. VI 0344116	-

^aOther enniatins that were isolated simultaneously are listed as well, where appropriate.

There are at least two other cyclohexadepsipeptides that resemble enniatins structurally, notably the destruxins¹⁷ and the beauvericins.¹⁸ What distinguishes these three is that the enniatins have alternating amino acid and hydroxy acid residues. Conversely, destruxins contain only one hydroxy acid and five amino acid residues, while beauvericins are comprised of alternating N-methylphenylalanine and hydroxy acid residues. Although all three types of cyclohexadepsipeptides have gained attention, likely due to their range of interesting biological activities,^19,20 the chemistry and the wide-ranging biological activities of the enniatins have not been reviewed.

B. Biosynthesis of Enniatins

The biosynthesis of enniatins was reviewed in 2002.²¹ Briefly, enniatins are biosynthesized by a 347-kDa multifunctional enzyme, termed enniatin synthetase (ESyn).^22,23 The initial biosynthetic step involves adenylation, which activates substrates as thioesters; this is followed by N-methylation and peptide bond formation. Once formed, the dipeptidol unit condenses with two other dipeptidols in an iterative fashion to form a linear hexadepsipeptide, which cyclizes in the final step to form the enniatin.²⁴ ESyn has low substrate specificity for L-amino acids, thus permitting incorporation of different amino acids and contributing to the chemical diversity of the enniatins (Figure 1). Despite variability in the amino acid units, distinct Fusarium strains seem to incorporate preferentially some amino acids over others, as evidenced by the fact that certain enniatins can be isolated from only a particular species or strain of fungus. For example, Fusarium sambucinum preferably produces enniatin A (1) while Fusarium scirpi preferably makes enniatin B (2). ²⁵ The D-hydroxy acid units are incorporated via D-2-hydroxyisovalerate dehydrogenase. This enzyme has high substrate specificity, as D-hydroxyisovaleric acid is conserved in most enniatins, except for enniatins H (14), I (15), L (21), M₁ (22), M₂ (23), and N (24) (Figure 1), and is present in most enniatin-producing fungi.^26,27 To date, most enniatins have been biosynthesized naturally from producing fungi; however, enniatins C (9), G (13), and MK1688 (16) were produced via precursor-directed biosynthesis.¹³ It is conceivable that as more advanced knowledge is generated on the biosynthesis of these compounds, other enniatins could be biosynthesized by altering the insertion of either (or both) the amino acid or hydroxy acid substrates.

Figure 1. Enniatins and their amino acid and hydroxy acid compositions.

enniatin	R1	R2	R3	R4	R5	R6
A (1)	sBu	sBu	sBu	iPr	iPr	iPr
A1 (3)	sBu	iPr	sBu	iPr	iPr	iPr
A2 (5)	sBu	iBu	sBu	iPr	iPr	iPr
B (2)	iPr	iPr	iPr	iPr	iPr	iPr
B1 (4)	iPr	sBu	iPr	iPr	iPr	iPr
a B2 (6)	iPr	iPr	iPr	iPr	iPr	iPr
a B3 (7)	iPr	iPr	iPr	iPr	iPr	iPr
B4 = D (8)	iPr	iPr	iBu	iPr	iPr	iPr
C (9)	iBu	iBu	iBu	iPr	iPr	iPr
* E1 (10)	iPr	iBu	sBu	iPr	iPr	iPr
* E2 (11)	iPr	sBu	iBu	iPr	iPr	iPr
F (12)	iBu	sBu	sBu	iPr	iPr	iPr
G (13)	iBu	iBu	iPr	iPr	iPr	iPr
H (14)	iPr	iPr	iPr	sBu	iPr	iPr
I (15)	iPr	iPr	iPr	sBu	sBu	iPr
MK1688 (16)	iPr	iPr	iPr	sBu	sBu	sBu
J1 (17)	iPr	iPr	Me	iPr	iPr	iPr
* J2 (18)	sBu	iPr	Me	iPr	iPr	iPr
* J3 (19)	Me	iPr	sBu	iPr	iPr	iPr
K1 (20)	iPr	iPr	Et	iPr	iPr	iPr
L (21)	iPr	iPr	iPr	iPr	iPr	Hy-sBu
* M1 (22)	iPr	iPr	iPr	iPr	sBu	Hy-sBu
* M2 (23)	iPr	iPr	iPr	iPr	Hy-sBu	sBu
N (24)	iPr	iPr	iPr	sBu	sBu	Hy-sBu
* O1 (25)	iBu	iPr	iPr	sBu	iPr	iPr
* O2 (26)	iBu	iPr	iPr	iPr	sBu	iPr
* O3 (27)	iBu	iPr	iPr	iPr	iPr	sBu
*b P1 (28)	iPr	iPr	Hy-Et	iPr	iPr	iPr
*b P2 (29)	iBu	iPr	Hy-Et	iPr	iPr	iPr

^a

B₂ and B₃ have one and two N-Methyls, respectively, in the R₁ and the R₁ and R₂ positions.

^b

The stereochemistry of the Hy-Et side chain was not determined.¹⁶

^*

Homologous among enniatins with the same “letter”.

C. Chemistry: Inseparable Mixtures of Enniatins

Eighteen out of the 29 known enniatins have been isolated as single compounds. The other 11 have been isolated as five mixtures of homologs [enniatins E [re-examined below as enniatin E₁ (10) and E₂ (11)]; J₂ (18), J₃ (19); M₁ (22), M₂ (23); O₁ (25), O₂ (26), O₃ (27); and P₁ (28), and P₂ (29)], despite the use of modern chromatographic techniques. Thus, their structures were characterized as mixtures.

Revisiting Enniatin E

In 1992, enniatins E₁(10) and E₂ (11) were the first reported inseparable mixture of enniatins (reported as enniatin E).¹⁰ Isolated with enniatins D (8) and F (12), the mixture of enniatins E₁ (10) and E₂ (11) was active as acyl-CoA: cholesterol acyltranferase (ACT) inhibitor.²⁸ These showed double the resonances in the NMR spectra when compared with enniatins D and F. The authors¹⁰ reported this as a mixture of two nonequivalent isomers after noting two key observations: (1) failing to see any inter-convertible spots on a silica gel TLC plate even at higher temperature, and (2) verifying by chemical degradation the amino acid and hydroxy acid components. While two homologs were proposed, a more efficient method is proposed below that serves to verify the identities based on mass spectrometry without the need for chemical degradation and illustrates, in general, some of the practical challenges of working with enniatin homologs. The difficulties in isolating and elucidating the individual components in the mixtures are oftentimes understated in the literature and, in our experience, daunting.

Recent studies on enniatins E₁ (10) and E₂ (11) emerged in our ongoing program to examine filamentous fungi for anticancer drug leads,^29,30 when promising cytotoxicity was observed in an extract of culture MSX 44957 [Fusarium sp., sequence deposited in Genbank (JQ809513)]. This extract yielded dipeptidols 30 and 31 (discussed below) and enniatins B₁ (4), D (8), and I (15). An additional compound, pseudo-enniatin E, was isolated as well that behaved like a dimer of enniatins, since it displayed twice the number of expected resonances in the ¹H and ¹³C NMR spectra when compared to compounds 4, 8, and 15 and generated a quasi molecular ion of m/z = 1335, essentially twice the mass of most enniatins. Importantly, this compound displayed a single peak when examined with at least eight different HPLC/UPLC columns, including both normal- and reverse-phase chemistries and a chiral column and via testing a suite of solvent conditions (examples are illustrated in Figure 2). These observations led to a tentative initial structure with double the number of amino acids and hydroxy acids of the known enniatins, although there were some challenges using 2D NMR experiments to finalize the structure. However, upon re-running and re-analyzing the mass spectral data, a consistent molecular ion was observed at m/z = 668 for [M+H]⁺. These data were acquired independently in two different mass spectrometry laboratories, and we suspect that the initial observation of m/z = 1335 was an artifact, likely representing [2M+H]⁺. The molecular ion matched seven known enniatins in literature: A₁ (3), E₁ and E₂ (10 and 11), G (13), I (15), and O₁ (25), O₂ (26), and O₃ (27). Of these, the enniatin O homologs could be ruled out, as they were described as a mixture of three inseparable homologs,¹¹ and our NMR data displayed patterns that were most similar to enniatins E₁ (10) and E₂ (11), although the original paper reported it as simply enniatin E.¹⁰

Figure 2.

Representative profile of enniatin E homologs when examined using an Acquity UPLC (Waters Corp., Milford, MA USA) and a solvent system of 20–100% MeOH-H₂O using four different columns (all from Waters Corp.): A) HSS-C18 (1.8 μm; 50 × 2.1 mm); B) BEH-C18 (1.7 μm; 50 × 2.1 mm); C) BEH-Phenyl (1.7 μm; 50 × 2.1 mm); D) HSS-T3 (1.8 μm; 100 × 2.1 mm).

To further characterize this isolated compound and compare and contrast differences with the literature on enniatin E₁ (10) and E₂ (11),¹⁰ a full suite of 1- and 2D NMR data were acquired in different solvents, including CD₃OD, DMSO-d₆, CDCl₃, and acetone-d₆, to see whether better resolution could be achieved. NMR data (Table 2) acquired in acetone-d₆ resulted in ¹H NMR signals that were more spread out and better resolved. However, ¹H NMR data in the literature¹⁰ were acquired in CDCl₃ and compared favorably to the data on “pseudo-enniatin E” acquired also in CDCl₃. ¹³C NMR data in acetone-d₆ also yielded the best data set, followed by DMSO-d₆, as most of the α-carbon resonances were individual peaks [12 peaks in acetone-d₆, as opposed to just six resonances when analyzed in CDCl₃ (Table 2)]. DMSO-d₆, with a boiling point of 189 °C, was utilized for VT NMR experiments, which permitted NMR spectra acquisition at higher temperatures, as compared to other more volatile NMR solvents. The VT NMR experiments probed whether the compounds were a result of conformers (if signals coalesce at higher temperature)³¹ or homologous (if coalescence was not observed at any point as the temperature was elevated). In the case of the pseudo-enniatin E, coalescence was not observed. In fact, at some higher temperatures, the peaks appeared even sharper than at room temperature (Figure 3), further evidence that pseudo-enniatin E was a mixture of homologs and not conformers.

Table 2.

NMR Data for Enniatin E₁ (10) and E₂ (11) (500 MHz in Acetone-d₆ and CDCl₃)

Position	Acetone-d6		CDCl3
	δH, Mult. (J in Hz)	δC	δH, Mult. (J in Hz)	δC
Hiv1
1	-	170.3	-	170.5
2	5.15, d (9.0)	75.0/74.9	4.95, d (9.0)	77.2
3	2.18–2.22, m	30.0	2.20–2.30, m	30.4
4	0.92–0.98, m	18.9	0.88–0.99, m	19.1
5	0.92–0.98, m	18.7	0.88–0.99, m	19.0
N-MeVal
6	-	169.0/168.6	-	170.2
7	4.88/4.51, d (10)	60.8/62.5	4.91, d (9.5)	61.3
8	2.18–2.22, m	31.9	2.02, m	33.6
9	0.92–0.98, m	20.6	0.88–0.99, m	20.2
10	0.92–0.98, m	19.4	0.88–0.99, m	19.6
N-Me	3.13, s	30.9/32.3	3.14, s	34.3
Hiv2
11	-	170.8/170.7	-	170.4
12	5.15, d (9.0)	74.7/74.4	5.10, d (8.0)	75.3
13	2.18–2.22, m	29.8	2.20–2.30, m	30.0
14	0.92–0.98, m	18.0	0.88–0.99, m	20.2
15	0.92–0.98, m	18.0	0.88–0.99, m	19.2
N-MeIle
16	-	169.1/168.7	-	168.7
17	5.01/4.71, d (10)	59.1/60.5	5.04, d (10)	59.8
18	2.03, m	33.9	2.20–2.30, m	30.0
19	0.92–0.98, m	17.7	0.88–0.99, m	16.0
20	2.18–2.22, m	24.9	2.20–2.30, m	27.8
21	0.84–0.89, m	9.8/10.2	0.83–0.87, m	11.1
N-Me	3.10/3.11, s	30.8/31.9	3.11, s	33.8
Hiv3
22	-	169.3/169.4	-	170.9
23	5.29/5.25, d (8.0)	74.5/74.4	5.20, d (9.0)	75.5
24	2.18–2.22, m	30.3	2.20–2.30, m	30.0
25	0.92–0.98, m	18.2	0.88–0.99, m	19.6
26	0.92–0.98, m	18.1	0.88–0.99, m	18.9
N-MeLeu
27	-	170.2/170.4	-	169.0
28	5.10, d (10)	54.6/54.4	4.57, m	57.3
29	1.85, m;1.64, m	37.5	1.80, m; 1.74, m	38.1
30	1.52, m	24.8	1.53, m	25.5
31	1.02, m	19.7	1.01, d (7.0)	20.5
32	1.45, m	25.2	1.35, m	25.0
N-Me	3.15/3.16, s	31.4/31.5	3.07, s	31.9

^aValues before and after slashes are chemical shifts for the individual homologs

Figure 3.

¹H NMR profile of enniatin E homologs (α-proton region) at three temperatures in DMSO-d₆ (500 MHz).

To determine the configuration of pseudo-enniatin E, acid hydrolysis was performed using methods described previously.³² The sample was extracted thrice with CH₂Cl₂, and both the aqueous and organic extracts were subjected to chiral HPLC analyses by comparing them to standards, which demonstrated that pseudo-enniatin E had the same composition of amino acids and hydroxy acids, all with the same configurations, as enniatins E₁ (10) and E₂ (11) (Hiv, N-MeIle and N-MeLeu).

Analysis of the mass fragmentation (Figure 4) through MSⁿ experiments for pseudo-enniatin E (m/z = 690 [M+Na]⁺) showed fragments consistent with the sequential loss of N-MeIle or N-MeLeu [the two amino acids being isomeric (m/z = 563)], Hiv (m/z = 463), N-MeVal (m/z = 350), and another Hiv (m/z = 250). Since enniatins have alternating amino acid and hydroxy acid residues, the initial sequential loss of residues revealed that N-MeVal was flanked by two Hiv units. The Hiv units had to be flanked by the two other amino acid residues - N-MeIle and N-MeLeu - which in turn were separated by one last Hiv unit. A second pattern emerged from the same mass spectrum that displayed sequential loss of N-MeVal (m/z = 577), followed by Hiv + NMeLeu (or NMeIle) (m/z = 350). Hence, in order for enniatin E to display double the resonances in the NMR spectra, the two homologs, which we now suggest should be termed enniatin E₁ (10) and E₂ (11) (Figure 1), had the following cyclic order: (1) N-MeIle-Hiv-N-MeVal-Hiv-N-MeLeu-Hiv (E₁) and (2) N-MeLeu-Hiv-N-MeVal-Hiv-N-MeIle-Hiv (E₂) (Figure 4).

Figure 4.

MSⁿ mass fragmentation of homologs of enniatin E. The boxes and circles highlight where N-MeIle and N-MeLeu were reversed in the two homologs.

Isolation of 1,4-perhydrooxazine-2,5-dione compounds 30 and 31

Two other compounds (30 and 31) were isolated together with the enniatins from the same Fusarium species (MSX 44957). These dipeptidols, characterized based on HRESIMS and NMR data (Table 3), may represent building blocks of enniatins in cyclized form.³³ Compound 30 showed m/z = 214.1452 [M+H]⁺, corresponding to a molecular ion of C₁₁H₁₉NO₃ (calculated for C₁₁H₂₀NO₃; 214.1443), while compound 31 showed m/z = 228.1599 [M+H]⁺, corresponding to a molecular ion of C₁₂H₂₁NO₃ (calculated for C₁₂H₂₂NO₃; 228.1608). To the best of our knowledge, there are no reports on the isolation of these as natural products, although compound 30 was obtained as a by-product upon hydrolysis and lactonization of N-α-bromoisovaleryl-N-MeVal in a synthesis paper.³³

Table 3.

NMR data for compounds 30 and 31 (500 MHz in CDCl₃)

Position	30		31
	δH, Mult. (J in Hz)	δC	δH, Mult. (J in Hz)	δC
1	-	170.4	-	170.3
2	4.47, d (9.0)	63.5	4.66, d (9.0)	61.8
3	-	169.4	-	169.2
4	5.11, d (8.0)	75.9	5.15, d (9.0)	75.6
5	2.27, m	28.1	2.08, m	34.0
6	0.86, d (7.0)	19.5	1.45, m; 1.11, m	25.3
7	1.03, d (7.0)	18.9	1.00, d (7.0)	16.3
8	-	-	0.89, t (8.0)	11.0
9	2.27, m	30.1	2.30, m	29.9
10	0.96, d (7.0)	20.7	1.02, d (7.0)	18.7
11	0.93, d (7.0)	18.7	0.97, d (7.0)	18.5
12	3.10, s	33.5	3.13, s	33.1

There are likely several plausible explanations for the existence of these dipeptidols in cultures of MSX 44957. Compounds 30 and 31 could be synthesized by a mutated enzyme, where instead of proceeding to an elongation step, the enzyme terminates the process and cyclizes at the dipeptidol stage. One could envisage them as the result of limited substrates, although we never limited nutrients intentionally. Alternatively, they may have formed spontaneously due to thermodynamically stable six-membered rings or could be a by-product of metabolism. Regardless, their identification concomitant with a series of enniatins may be valuable for future biosynthetic considerations.

Structure Elucidation: Techniques and Challenges

The structures of the enniatins have been determined by a variety of physical and chemical methods. With respect to X-ray crystallography, only enniatins B (2)³⁴ and N (24)¹⁵ have been reported in the literature. Alternatively, the use of NMR spectroscopy is more prominent. The 18-membered cyclic backbone of all enniatins consists of carbonyl carbons (three amides and three esters) and six α-carbons, three of which are oxygenated (δ_C 73 ~ 77 ppm) and three of which are connected to nitrogen atoms (δ_C 56 ~ 65 ppm). An α-proton (δ_H 4.44 ~ 5.32 ppm) is connected to each α-carbon in the backbone, while a methyl group is connected to each nitrogen in the amino acid portion (δ_H 3.08 ~ 3.18 ppm), except for enniatins B₂ and B₃, which have one and two less N-methyl moieties, respectively.⁹ The connectivities of each amino and hydroxy acid residue are established by HMBC data, which typically show correlations from the α-protons to two carbonyl carbons, and from the N-methyl protons to the ester carbonyl carbon that acylates that particular amino acid. The identity of each residue can be determined by COSY and TOCSY data and by multiplicities and J-values. Although signals for the methyl groups of Val, Leu, and Ile are typically overlapped in an enniatin ¹H NMR spectrum, a careful integration of the area will facilitate counting the number of methyl groups in the structure, and these are reported typically as interchangeable values. The chemical shifts of the β-protons for amino acids depend on whether the structure has a Val (δ_H 2.20 ~ 2.30 ppm), Leu (δ_H 1.70 ~ 1.80 ppm), or Ile (δ_H 2.09 ~ 2.11). The multiplicity of an α-proton of an Ile (triplet) would also appear different from that of a Leu (doublet) or Val (doublet). Moreover, the β-carbon of Ile differs in chemical shift from Leu (δ_C 33 ~ 35 ppm versus 37 ~ 38 ppm, respectively). Also, the DEPT-135 experiment can differentiate between Ile and Leu based upon the chemical shift of the carbon that changed phase. Biosynthetic considerations can also assist in the structure elucidation process. Although the enniatin synthases were mentioned to have low substrate specificity for L-amino acids, a study also showed that this enzyme preferentially incorporates only certain sizes of amino acid side chains.²⁴ Thus, alanine and α-aminobutyric acid are the only other amino acids shown to be incorporated into a naturally produced enniatin, specifically enniatins J₁ (17), J₂ (18), J₃ (19) and K₁ (20).¹⁴ In summary, if the enniatin will not form crystals suitable for X-ray analysis, NMR spectroscopy may be the best solution, although the spectrum will appear crowded and overlapped in certain diagnostic regions. Focusing on how resonances can be ascribed to distinct amino acid or hydroxy acid residues, particularly Val, Leu, Ile, and Hiv, should narrow the work involved considerably.

The modern challenge in the structure elucidation of enniatins arises when homologs are isolated. These enniatins co-elute and appear as a single peak in chromatographic separations (e.g. Figure 2), even using modern stationary phases, a suite of mobile phases, and even a chiral column. As inseparable mixtures, they display double (as exhibited by enniatins E₁/E₂, J₂/J₃ and M₁/M₂) or triple (as shown by enniatins O₁/O₂/O₃) the resonances in the NMR spectra.^10,11,14,15 In these situations, neither a change in solvent nor the use of variable temperature experiments results in simplification of resonances or coalescence of signals, thus ruling out the possibility of conformational isomers, as shown with compounds like bassianolide.³¹ Also, mass spectral data suggests a molecular weight in the range of typical enniatins (i.e. 611 to 683 a.m.u.). For certain enniatins, the structures of such homologs were elucidated in different ways, and a few cogent examples are discussed below.

Enniatin J₂/J₃¹⁴

In conjunction with NMR data, mass spectrometry played a key role in the structure elucidation of enniatins J₂ (18) and J₃ (19), particularly the use of ESIMSⁿ to determine the sequence of residues. Isolated together with enniatins J₁ (17) and K₁ (20) from an unidentified Fusarium strain, these homologs displayed a m/z of 626.3972 [M+H]⁺. NMR data and fragmentation information via ESIMS^2–4 indicated the presence of N-MeAla (m/z = 85), N-MeVal (m/z = 113), N-MeIle (m/z = 127), and Hiv (m/z = 100). The sequence was determined by subjecting the homologs to partial acid hydrolysis to obtain linear depsipeptides, which generated two major fractions that were subjected to further ESIMSⁿ. One fraction displayed the sequential loss of N-MeIle followed by N-MeAla, after the initial loss of Hiv. Keeping in mind that enniatins have alternating amino acid-hydroxy acid residues, these results indicated that N-MeIle and N-MeAla had to be on opposite ends of the linear depsipeptide and that N-MeVal separated both amino acids. In support of this finding, the other fraction showed a loss of N-MeVal followed by N-MeAla, after the initial loss of Hiv. Using the same logic, the N-MeIle had to be between N-MeVal and N-MeAla. Coupling these finding together distinguished the sequences of the two homologs (Figure 1).

Enniatin M₁/M₂¹⁵

Isolated from an unidentified fungus and reported as one of the first enniatins with a hydroxyl group in the side chain of the Hiv portion of the structure [the others being enniatins L (21) and N (24)], enniatins M₁ (22) and M₂ (23) were isolated as a single peak by reversed phase HPLC. However, this isolate displayed twice the expected resonances in both the ¹H- and ¹³C NMR spectra for a single enniatin molecule. Extensive 1- and 2D NMR analyses revealed that these homologs contained six units of N-MeVal, two units of Hiv, two units of 2,3-dihydroxy-3-methylpentanoic acid (Dmp), and two units of 2-hydroxy-3-methylpentanoic acid (Hmp). Even though there was no direct evidence for the exact sequence of residues for homologs M₁ (22) and M₂ (23), the authors suggested that there had to be a difference in their connectivities in order for these two homologs to display different NMR resonances. Indirectly, the NMR data from the related compound, enniatin N (24),¹⁵ from which a crystal structure was obtained, showing the presence of Hmp and Dmp residues in the hydroxy acid portion of the molecule, was used to match the resonances observed in enniatins M₁ (22) and M₂ (23) (Figure 1).

Enniatin O₁/O₂/O₃¹¹

Showing not just two but three sets of NMR resonances, enniatins O₁ (25), O₂ (26) and O₃ (27) were obtained from Verticillium hemipterigenum, whose fermentation conditions were optimized for enniatin production. NMR data (both ¹H and ¹³C spectra) revealed that these homologs consisted of six N-MeVal, three N-MeLeu, six Hiv, and three Hmp residues. Further structural analyses by NMR and mass spectrometry were performed on four amide fragments, produced from partial degradation using LiBH₄ followed by acetylation. The ¹H and ¹³C NMR data of the fragments matched similar data obtained from known enniatins [B (2), B₄ (8), and H (14)] subjected to the same reaction. In short, the absolute configuration and the sequences of the amino and hydroxy acid residues were derived by comparison with known compounds and by mass spectral analyses of the fragments, thus establishing the structures for enniatins O₁ (25), O₂ (26), and O₃ (27).

Identification of the configuration of the amino acid and hydroxy acid residues

The amino acid residues in all of the 29 enniatins described to date have the L-configuration, likely because the biosynthetic enzyme, ESyn, has a preference over the D-isomer. In a similar fashion, all of the hydroxy acids have the D-configuration, possibly due to the preference of the D-2-hydroxyisovalerate dehydrogenase. Although some of the enniatin literature is scant on experimental results, it may be a fair assumption, at least initially, that the L-amino acids and D-hydroxy acids are incorporated based largely on biosynthetic considerations. However, a simple acid hydrolysis can be performed on just a milligram quantity of the material and compared with standards using a chiral column [see discussion of enniatin E₁ (10) and E₂ (11)]. In this way, no derivatizations are required that would diminish the yield of the “product”. Moreover, in cases where inseparable homologs are present, acid hydrolysis can help determine the identity of the residues [see discussion of enniatin E₁ (10) and E₂ (11)]. Also, partial degradation using LiBH₄ followed by acetylation with acetic anhydride yielded fragments that were analyzed by NMR and mass spectrometry, as exemplified by enniatins O₁ (25), O₂ (26), and O₃ (27).¹⁰ In summary, although the biosynthetic considerations for the configuration of the residues have been consistent for 29 enniatins, to date, taking the research a step further and determining the configurations experimentally is straight forward and should be pursued; compounds with an unanticipated configuration could be uncovered in the future.

D. Biological Activities

Antimicrobial Activities

Enniatins have been shown to exhibit a wide array of biological activities, and one of the earliest reported was antimicrobial activity towards Mycobacterium spp., Staphylococcus spp., Bacillus spp., Pseudomonas aeruginosa and Escherichia coli.^1,35,36 Enniatins B (4), B₄ (8), H (14), I (15), G (13), C (9), MK1688 (16), O₁ (25), O₂ (26) and O₃ (27) were shown to inhibit the activity of Mycobacterium tuberculosis.^11,13 Furthermore, a recent study focused on the ability of enniatins to inhibit mycotoxigenic molds, such as Fusarium verticillioides, F. sporotrichioides, F. oxysporum, F. poae, F. tricinctum, F. proliferatum, Beauveria bassiana, Trichoderma harzianum, Aspergillus flavus, A. parasiticus, A. fumigatus, A. ochraceus, and Penicillium expansum.³⁷ Interestingly, of the four enniatins used in that study [A (1), A₁ (3), B (4), B₁ (4)], only enniatin B (2) displayed inhibition of some of the molds. Enniatins were also tested against Botrytis cinerea, a fungal phytopathogen that causes gray mold to many host species and infects seedlings stored in humid conditions. Of the series of enniatins tested, enniatins B (2), B₄ (8), and K₁ (20) were found to partially inhibit spore germination of B. cinerea.¹⁴

One mixture of enniatins, obtained from Fusarium lateritium WR, strain 437,³⁸ has been developed as an antimicrobial, fusafungine, and is administered as a topical treatment of upper respiratory tract infections by oral and/or nasal inhalation.³⁸ The drug is sold under the trade names Locabiotal (United Arab Emirates and Taiwan), Bioparox (Poland and Slovakia), Locabiosol (USA and Canada), and Fusaloyos (Spain) and is manufactured by Laboratoires Servier, a Frenchpharmaceutical company. Fusafungine showed bacteriostatic activity against a suite of microorganisms, such as Mycoplasma pneumoniae, Legionella pneumophila, Streptococcus pneumoniae, S. pyogenes, and Staphylococcus aureus, including methicillin-resistant S. aureus.

Enzyme Inhibitors

Enniatins A (1), A₁ (3), B (2), B₁ (4), D (8), E₁/E₂ (10 and 11) and F (12) have been probed, at least in a limited fashion, as enzyme inhibitors, particularly acyl-CoA: cholesterol acyltransferase (ACAT), although their activities were moderate compared to beauvericin, one of the most potent ACAT inhibitors of microbial origin.²⁸ Nevertheless, such inhibitions could be significant in the treatment and prevention of atherosclerosis and hypercholesterolemia. In a separate study, the hypolipidaemic action of enniatin B (2) could be ascribed both to ACAT inhibition and a reduction of triglyceride synthesis, thus diminishing the pool of the free fatty acid in cells.³⁹ Also, enniatin B (2) was found to inhibit 3′,5′-cyclo-nucleotide phosphodiesterase and was able to bind calmodulin.^40,41 This mode of action suggested an altering of signal transduction mechanisms of enzymes that are dependent on calmodulin, such as cyclic nucleotidase and protein kinases. In summary, although the number of studies of enniatins vs enzymes is somewhat limited, they seem to possess activities that warrant further exploration.

Cytotoxicity and Anticancer Properties

Towards cancer research, a study found enniatins B (2), B₁ (4), and D (8) to inhibit pleiotropic drug resistance 5 protein (Pdr5p), an ATP-binding cassette that is one of the major causes of drug resistance against a variety of compounds, including drugs like tamoxifen and doxorubicin.⁴² While many inhibitors of multidrug resistance (MDR) proteins have been described, these enniatins are important because they are found to be potent and specific for Pdr5p but are less toxic than other such leads.⁴² In another cancer-related study, enniatins A₁ (3) and B₁ (4) were found to induce apoptotic cell death and disrupt extracellular regulated protein kinase (ERK), a mitogen-activated protein kinase associated with cell proliferation,⁴³ while a mixture of enniatins A (1), A₁ (3), B (2), and B₁ (4) also induced apoptosis in several human cancer cell lines at low molar concentrations after 24 hours of treatment.⁴⁴

Our studies on enniatins were initiated based on cytotoxicity of an extract of the fungus Fusarium sp. (MSX 44957) tested against a panel of three human cell lines in culture (the methods and strategy for such studies have been outlined previously^32,45). As described earlier in this manuscript, a series of compounds were isolated and characterized, including enniatins E₁ (10), E₂ (11), B₁ (4), D (8), I (15), and compounds 30 and 31. Of these, the enniatins were more potent than the related building blocks, i.e. compounds 30 and 31, with the homolog mixture of enniatins E₁ (10) and E₂ (11) being the most potent (see Table 4). Despite these encouraging results, the activity of the enniatins was still approximately one to two orders of magnitude less potent than the positive control, camptothecin.

Table 4.

Cytotoxicity of enniatins and other compounds isolated from Fusarium sp. (MSX 44957) against a panel of human tumor cell lines.

Compound	IC50 values (in μM)a
	MCF-7	H460	SF268
Enniatin B1 (4)	1.4	1.2	2.2
Enniatin E1/E2 (10/11)	0.5	0.7	2.6
Enniatin D (8)	2.7	2.8	4.0
Enniatin I (15)	1.0	1.3	1.4
30	8.4	8.9	10
31	3.4	3.2	4.3
Camptothecinb	0.07	0.008	0.50

^aIC₅₀ values were determined as the concentration required to reduce cellular proliferation by 50% relative to untreated controls following 72 h or continuous exposure using methods described in detail previously.^51,52

^bPositive control.

Other Biological Activities

Not surprising for a series of compounds first described in the 1940s, the enniatins have shown a variety of other biological activities; a recent review highlights several others not covered in this manuscript.⁴⁶ In particular, enniatin A (1) was shown to possess anthelmintic properties against Nippostrongylus brasiliensis, Trichinella spiralis, and Heterakis spumosa.⁴⁷ Moreover, analogs of enniatin A (1) were synthesized, and two new derivatives were shown to possess strong in vivo activity against H. contortus in sheep.⁴⁸ A mixture of enniatins A (1), A₁ (3), B (2), B₁ (4) was cytotoxic against the Spodoptera frugiperda cell line, which is an insect cell line used to probe for in vivo cytotoxicity of fungal metabolites.⁴⁹ Several of the enniatins [B (2), B₄ (8), H (14), I (15), G (13), C (9), MK1688 (16), O₁ (25), O₂ (26) and O₃ (27)] were found to inhibit proliferation of the human malaria parasite Plasmodium falciparum.^11,13 Given this diverse range of activities, future researchers should be encouraged to examine the enniatins against a broad range of biological targets.

Enniatins as Ionophores

Given the pore shape of the cyclodepsipeptide core of enniatins, their ionophoric behavior has been suggested since the 1960s.⁵⁰ More recently, a study on the electrophysiological properties of an enniatin mixture containing A (1), A₁ (3), B (2), B₁ (4) revealed that they incorporate easily into the cell membrane as a passive channel and form cation selective pores.⁴⁰ By forming complexes with cations like K⁺, Na⁺, and Ca²⁺, enniatins evoke changes in intracellular ion concentration, disrupting cell function. Indeed, this may explain the broad range of biological activities observed for the enniatins.

E. Conclusion

Enniatins are secondary metabolites that have an interesting cyclohexadepsipeptide structure and that are produced, primarily, by Fusarium species of fungi. Although 29 of these have been described in the literature, it is anticipated that other analogs exist in nature, particularly given the diverse biosynthetic possibilities in the side chains of these compounds. Moreover, mixtures of homologs, where the position of amino acid units have been shuffled, will continue to challenge the skills and resolve of natural products chemists. Studies on these compounds have been, and likely will be in the future, driven by their broad range of biological activities. To date, one pharmaceutical product, fusafungine, has been derived from the enniatins; the future awaits other applications.