Aplicyanins – brominated natural marine products with superbasic character

1 Introduction

The chemistry of proton-transfer reactions and the interplay between acids and bases play a central role in almost all natural processes as well as in modern organic methods. A large majority of organic reactions is mediated by either acids or bases or involves their participation at some stage of the reaction pathway. Particular attention in organic synthesis has been given to so-called proton sponges – fused-ring organic diamines that are known for their exceptionally high proton affinity (PA) [1]. After the discovery of 1,8-bis(dimethylamino)naphthalene (DMAN) (Fig. 1) [2], which is the very first example of a proton sponge, many analogous compounds have been synthesized by replacing the molecular backbone with other aromatic structures or even more by introducing a “buttressing effect” by a variation of the basic nitrogen centers and their environment [3–11].

Fig. 1:

1,8-Bis(dimethylamino)naphthalene.

The high basicity of these compounds can be attributed to combinations of different factors, such as the destabilization of the base due to strong repulsion between lone pairs (causing extreme steric strain), the relief of strain in the protonated form due to the formation of a strong intramolecular hydrogen bond, and a difference in solvation energy between the base and its protonated form [12]. In the same time, the combination of steric and electronic features of these compounds is the reason for their low nucleophilicity. Fitted with this knowledge, one can now start to design superbasic proton sponges, e.g. based on metallosupramolecular chemistry [13], and to identify systems as proton sponges [14], which are already known in other contexts.

A complementary concept is the inspection of natural compounds. Natural products are a fertile and fascinating topic of chemical investigation for many years, driving the development of analytical and synthetic methods in chemistry. Parallel is the hope of mankind to find new substances with new and badly needed properties. In particular, natural products from the sea attract particular attention. Numerous nitrogen-containing metabolites belonging to unprecedented structural families and possessing interesting biological properties such as antitumor activity or cytotoxicity have been obtained from marine tunicates of the genus aplidium [15–29]. In their search for new antitumor agents from marine organisms, Reyes et al. [16] isolated from Aplidium cyaneum the cytotoxic and antimicotic alkaloids aplicyanins A–F (see Fig. 2).

Fig. 2:

Investigated aplicyanins.

Recently, we showed that the mycosporine-like amino acid porphyra-334 is a proton sponge, and to the best of our knowledge, it is the first example of a natural proton sponge [30]. Even more important, porphyra-334 is the first representative of an entire assembly of proton sponges within the mycosporine-like amino acids [31]. Although proximity effects are often responsible for high basicity (vide supra) in the synthetic proton sponges, the delocalization of the positive charge after protonation in the porphyra-334 is responsible for the high gas phase PA, as known from the proton sponges of Schwesinger et al. [32, 33]. As proton sponges based on guanidine moieties can also delocalize the positive charge [34, 35], we were especially interested in guanidine-based natural products. We investigated the aplicyanins A–F, as they are all fitted with a guanidine or guanidine-based moiety to find out if they are also natural proton sponges.

In this contribution, we wanted to test if the recently isolated aplicyanins are also superbasic proton sponges and if there is a correlation between the reported cytotoxicity [16] and the calculated gas phase PA.

2 Results and discussion

As a quantum chemical accessible descriptor for basicity, the gas phase PA, calculated according to eq. (1), is a well-established concept [36] (for successful applications of the concept of Raabe et al., see refs. [37–40]):

It is the benefit of this concept that published experimental gas phase PAs are available, and therefore we can evaluate our quantum chemical methods [Method A: B3LYP/6-311+G(2df,p)//B3LYP/6-31G(d)+ZPE(B3LYP/6-31G(d)); Method B: B3LYP/6-31G(d)//B3LYP/6-31G(d)+ZPE(B3LYP/6-31G(d)] against experimental values of some small bases and derive an equation to scale the calculated gas phase PAs (see Fig. 3 and Table 1).

Fig. 3:

Comparison of calculated and experimental gas phase PAs.

By using linear regression, we obtained the following two equations to scale the calculated proton affinity (PA) and to get values closer to the experimentally measured ones:

Method A:

Scaled calculated PA (6-311+G(2df,p)) =  (0.9378 × calculatedPA) – 13.4034 kcal mol−1 (r2 = 0.9925)

Method B:

Scaled calculated PA (6−31G(d)) =  (0.9603 × calculatedPA) – 4.2608 kcal mol−1 (r2 = 0.9837)

A comparison of the calculated gas phase PAs of the aplicyanins with the reference bases shows that aplicyanins A, C, and E can easily compete with 1,8-bis(dimethylamino)naphthalene and show an even higher PA of around –245 kcal mol⁻¹.

Taking the simplest molecule into account, which shows the basic moiety included in our aplicyanin structures, i.e. guanidine, we immediately see that aplicyanin B, aplicyanin D, and aplicyanin F show nearly identically or slightly lower (aplicyanin F) PAs as guanidine.

By contrast, aplicyanin A, aplicyanin C, and aplicyanin E exhibit a significantly higher basicity by approximately 10 kcal mol⁻¹. Because the structural variations of the indole system are clearly separated and far away from the imine-guanidine system, we can expect that they have no effect on the investigated PA, especially as we see in the case of guanidine and 2-acetylguanidine and the next larger references, tetrahydropyrimidin-2(1H)-imine and N-(tetrahydropyrimidin-2(1H)-ylidene)acetamide similar energy gaps. Closer inspection of tetrahydropyrimidin-2(1H)-imine and N-(tetrahydropyrimidin-2(1H)-ylidene)acetamide additionally confirms the expectation that the indole systems have no influence as tetrahydropyrimidin-2(1H)-imine and N-(tetrahydropyrimidin-2(1H)-ylidene)acetamide show a very similar PA as the corresponding aplicyanins without an indole substituent. The most pronounced difference in the structures of aplicyanin A, aplicyanin C, and aplicyanin E compared with aplicyanin B, aplicyanin D, and aplicyanin F, as well as in guanidine and 2-acetylguanidine or tetrahydropyrimidin-2(1H)-imine and N-(tetrahydropyrimidin-2(1H)-ylidene)acetamide, is the acetyl group bound to the imine nitrogen. Therefore, the first suggestion would be as follows: because acetyl groups are electron-withdrawing groups, the imine nitrogen atoms in aplicyanin B, aplicyanin D, and aplicyanin F have a reduced electron density and consequently are no longer as basic as in the other aplicyanins.

The reason for this reduced electron density can be easily found in the molecular structure of the N=C moiety of the guanidine building blocks (see Fig. 4 and Table 2). Throughout, aplicyanin A, aplicyanin C, aplicyanin E, and tetrahydropyrimidin-2(1H)-imine exhibit a short N=C imine double bond of 1.29 Å, nearly identical to the guanidine’s double bond length of 1.28 Å. After protonation, this bond is extended to 1.35 Å, again nearly identical to the value of protonated guanidine (1.35 Å). By contrast, aplicyanin B, aplicyanin D, aplicyanin F, and N-(tetrahydropyrimidin-2(1H)-ylidene)acetamide show in the neutral unprotonated form a somehow stretched NC-imine bond of 1.33 Å, very close to the value in 2-acetylguanidine (1.32 Å). This can be clearly attributed to the bond conjugation along the O=C–N=C moiety, giving extra stability to the unprotonated compound and consequently lowering the PA. This conjugation is severely hampered after protonation as it can be seen in the elongation of the OC–NC bond, leading to distances closer to the C–N single bonds. In the unprotonated and the protonated form, the conformation of the compound, including acetyl groups, is stabilized by a C=O···HN hydrogen bond (see Fig. 4).

Fig. 4:

Calculated (B3LYP/6-31G(d)) structures of unprotonated and protonated aplicyanin A and aplicyanin B.

An inspection of the aromaticity of the three ring systems showed no unusual feature. Although the two indole rings are clearly aromatic [nucleus-independent chemical shift (NICS)(1) values between –11 and –12], the third saturated cycle shows NICS values of around –1 and is therefore best considered as nonaromatic.

Comparing the experimental cytotoxicity tested on three human tumor cell lines and the antimicotic activity [16] against our calculated gas phase PAs, at first glance, some correlation is apparent. However, the explanation is simple: although the acetyl group at the guanidine nitrogen is suggested to have a key role in the biological activity, it also reduces the basicity, as is evident from the gas phase proton activity. Therefore, the gas phase PA cannot be taken as a direct indicator of the biological activity.

3 Conclusion

The structures of aplicyanins A–F have been investigated by quantum chemical methods with respect to their gas phase PAs. Although aplicyanin A, aplicyanin C, and aplicyanin E show a high PA of around –245 kcal mol⁻¹, aplicyanin B, aplicyanin D, and aplicyanin F are significantly weaker bases, as the O=C–bond in the acetyl group is in conjugation with the imine N=C group. The apparent correlation between the gas phase PAs and the investigated biological activity can be attributed to the presence or absence of an acetyl group at the guanidine nitrogen and therefore cannot be taken as a guide in the search for further specific compounds.

4 Experimental section

For comparability with our previous studies, we fully optimized all structures at the B3LYP/6-31G(d) density-functional theory (DFT) level [42–46] using the Gaussian 03 program [47]. Frequencies were computed at the same level to characterize stationary points and to obtain zero-point vibrational energies (ZPEs). All frequencies are unscaled. DFT, in particular B3LYP, has shown to provide accurate geometries and good harmonic vibrational frequencies for a broad range of molecules and ions [48–50]. As shown recently, the level of theory selected is well suited for nuclear magnetic resonance [51, 52] and NICS calculations [53–56]. The gas phase PAs were calculated for 0 K following the approach of Raabe et al. [36–40] at B3LYP/6-31G(d) and of Despotović et al. [57] applying B3LYP/6-311+G(2df,p)//B3LYP/6-31G(d) and always included the ZPE correction on the B3LYP/6-31G(d) level of theory.