Aflatoxins are a group of acutely toxic metabolites produced by toxigenic strains of fungi such as Aspergillus flavus Link and Aspergillus parasiticus Speare (Asao et al. 1963). These toxins have closely similar structures and form a unique group of highly oxygenated, naturally occurring heterocyclic compounds. Four principal aflatoxins (AFs) are produced by this fungus: AFB1, AFB2, AFG1, and AFG2. AFB1 is considered among the most powerful hepatocarcinogenic agents known. In 1987 the International Agency for the Investigation in Cancer stated that aflatoxins represent a high potential risk for cancer in humans (IARC 1987). Contamination of foods and feedstuffs with aflatoxins is a serious problem for human and animal health. Many experiments have therefore been performed to reduce the level of aflatoxins in contaminated crops (Mercado et al. 1991; Samarajeewa et al. 1991). The aim of these methods is either to remove or to destroy the toxin from food or feed, and they can be classified into chemical, biological, and physical methods.

A large number of chemicals can react with aflatoxins and convert them to less toxic and mutagenic compounds. These chemicals include acids (Buchi et al. 1967; Dutton and Heathcote 1968), bases (Dollear et al. 1968; Mann et al. 1970; Park et al. 1981, 1984, 1988), oxidizing agents (Cater et al. 1974; Applebaum and Marth 1982; McKenzie et al. 1997, 1998), bisulfites (Doyle and Marth 1978a, b; Moerck et al. 1980; Hagler et al. 1982, 1983), and gases (Brekke et al. 1978; Samarajeewa et al. 1991). However, most of the chemical processes that have been investigated are impractical (required to be carried out under extreme conditions of temperature and pressure), are unsafe (due to the formation of toxic residues), and compromise the nutritional, sensory, and functional properties of the product. Currently, ammoniation and treatment with sodium bisulfite are the major industrial processes most widely used to inactivate aflatoxins in peanut meal, maize, and cottonseed destined for animal feeding.

Our previous studies have shown that certain organic acids perform a detoxification function in treating aflatoxin-contaminated commodities, providing certain protection to ducklings from aflatoxin toxicity and reducing the mutagenic activity of aflatoxins in the Salmonella microsomal screening system (Méndez-Albores et al. 2005, 2007).

These results suggest that detoxification of AFB1 initially involves the formation of a β-keto acid structure (catalyzed by the acidic medium), followed by hydrolysis of the lactone ring, yielding aflatoxin D1 (AFD1), derived from decarboxylation of the lactone ring-opened form AFB1, which is 450 times less mutagenic than AFB1 and presents an 18-fold toxicity decrease (Méndez-Albores et al. 2008; L. Lee et al. 1981). Although x-ray and nuclear magnetic resonance spectra of the AFB1 molecule have been reported (van Soest and Peerdeman 1970a, b), there is no specific information available concerning the structural and electronic parameters of the lactonic ring in the aflatoxin molecule. As far as we know, there are few theoretical studies of this compound (Pachter and Steyn 1985; Pavão et al. 1995; Okajima and Hashikawa 2000; Billes et al. 2006; Yiannikouris et al. 2006; Bren et al. 2007; Ramírez-Galicia et al. 2007), and they do not address the importance of the lactonic ring in the reactivity properties of AFB1. Consequently, the purpose of the present research deals with determination of the structural and electronic properties of the lactonic ring of the aflatoxin B1 compound, particularly when the lactonic ring suffers a rupture induced by the aqueous acid medium (Scheme 1). This study is realized by means of theoretical calculations, using density functional theory (DFT)-based methods and supercomputer facilities.

Scheme 1
scheme 1

Rupture of the lactonic ring, promoted by the aqueous acid medium

Full size image

Computational Methodology

In the first step, the target molecule was built with standard bond lengths and bond angles, using the PC Spartan 02 program (Deppmeier et al. 2002). The connectivity frame of this compound suggests four possible stereoisomers (Fig. 1a), because both H9a and H6a of the C9a and C6a centers can be on the same side, either the front, yielding structure 1, or back and front, yielding structure 2. In addition, there are alternate configurations: H6a in front and H9a in back (3) and the inverse structure (4). Therefore, the first task was to establish the conformation of maximum stability. Thus, the geometry for each stereoisomer was fully optimized using the AM1 (Dewar et al. 1985) semiempirical method implemented in Spartan 02. Furthermore, DFT calculations were carried out using the Gaussian 03 (Frisch et al. 2003) program at the level of theory defined by the hybrid Becke’s three-parameter hybrid functional, B3LYP (Becke 1988, 1993; Lee et al. 1988), which includes a mixture of Hartree–Fock exchange with DFT exchange correlation, has been employed. The basis set used includes the split-valence basis sets of 6-31G(d,p) (Hariharan and Pople 1973). Also, a diffuse function 6-311+G(d,p) (Clark et al. 1983; Frisch et al. 1984) was used. The default convergence criteria were also employed.

Fig. 1
figure 1

a Four possible stereoisomers of the aflatoxin molecule, both H9a and H6a of the C9a and C6a centers, can be on the same side, either the front, yielding structure 1, or the back, yielding structure 2. Or they can lie in an alternate pattern: H6a in front and H9a in back (3) or the inverse (4). b Selected geometrical parameters of the B3LYP/6-311+G(d,p) optimized structures, 1 and 3. Bond lengths are given as angstroms, and the atomic charges for some selected atoms (electron units; boldface) determined in the natural population analysis are indicated

Full size image

Then the optimized geometries for the ground and higher-energy states were determined with the B3LYP/6-31G(d,p) and B3LYP/6-311+G(d,p) methods. The minima were verified by performing a vibrational analysis with these methods. The structural parameters (bond lengths, bond angles, dihedral angles), natural atomic charges, electrostatic potential molecular surfaces, HOMO (lowest occupied molecular orbital), and LUMO (lowest unoccupied molecular orbital) were used to analyze the electrophilic or nucleophilic character, which can be involved in the reactivity of the aflatoxin B1 compound. These properties were analyzed at the B3LYP/6-311+G(d,p) level.

The bond orders were calculated for the B3LYP/6-31G(d,p) optimized structure of stereoisomer 1. We have proposed a mechanism for B1 (1) with an acidic treatment, to explain the formation of a compound that contains both hydroxyl and carboxylic groups. Also, we performed gas-phase geometry optimization employing the 6-311+G(d,p) basis set, for intermediates and transition states (TSs) of 1, which were confirmed by frequency calculations at the same level of theory. In all cases intrinsic reaction coordinate (IRC) calculations were performed to test whether the determined TSs connect with the proper reactants and products.

Absorption spectra determining vertical electronic excitations from the ground state using the TD-DFT method implemented in Gaussian 03, B3LYP/6-31G(d,p), were used (Stratmann et al. 1998; Maitra and Tempel 2006).

Results and Discussion

Ground State of Aflatoxin B1

As stated above, in the first phase of this study the four stereoisomers 1–4 of aflatoxin B1 were fully optimized using semiempirical calculations at the AM1 level, with the observation that the hydrogen atoms can lie on the same side, originating structures 1 and 2, with the same estimation for the formation energy of 138.3 kcal/mol. As shown in Fig. 1a, when these hydrogen atoms are in alternating fashion structures 3 and 4 are formed, with a considerably lower formation energy, 89.8 kcal/mol (Supplementary Table 1S). That is, 3 and 4, are located, with respect to 1 and 2, at a significant higher energy, 48.5 kcal/mol, also indicating the strong structural and energetic effects produced by the H atoms lying alternately or on the same side. As shown in Fig. 1a, in the ground-state structure the A, B, C, and D rings are in-plane, whereas in structure 3, all the rings tend to be on-plane, but the E ring is strongly distorted or stressed. These results indicate that the presence of 3 and 4 in a given sample may be disregarded.

Although the geometric parameters can be reasonably predicted by the semiempirical AM1 method, frequently this approach yields an incorrect electronic structure. For these reasons, the already located 1–4 structures were reoptimized at a higher level of theory, using DFT-basedmethods. In particular, the B3LYP functional and the 6-31G(d,p) and 6-311+G(d,p) basis sets were used. DFT calculations were done with the Gaussian 03 program.

The results obtained confirmed that 1 and 2 are the lowest-energy states, since 3 and 4 were located 31.6 kcal/mol higher in energy, at the B3LYP/6-311+G(d,p) level, or 31.9 kcal/mol using B3LYP/6-31G(d,p) (see Fig. 1b). Moreover, the vibrational analysis performed indicated that 1 and 2 are true minima on the potential energy surface, since real frequencies were obtained for these structures.

On the other hand, and according to the literature, the x-ray results (van Soest and Peerdeman 1970a, b) have revealed that the ground-state structure of the aflatoxin B1 compound corresponds when the hydrogen atoms are located on the same side of the molecular plane, in agreement with our theoretical results. The lowest-energy structure of aflatoxin, 1, is shown in Fig. 1b. It is worth mentioning that AM1 produces an average deviation of 0.018 Å for the equilibrium bond lengths, while this deviation is 0.015 and 0.014 Å for the B3LYP/6-31G(d,p) and B3LYP/6-311+G(d,p) methods. At the latter level of description, the two carbonyl groups, C1−O12 and C11−O14, in rings A and B have negative charges on the oxygen centers, implying repulsion between these oxygen centers (see structure 1 in Fig. 1b), which are separated by 3.10 Å; note that the A and B rings are rigid. Smaller repulsion is expected for the O6 and O7 centers, which have negligible negative charges and a separation of 2.35 Å. Considering the large number of their geometric parameters we present only their bond lengths (Table 1), their valence angles (Table 2), and some dihedral angles (Table 3). Then the average deviations of the bond lengths obtained with split-valence basis sets using the B3LYP functional are closer to the experimental values; see Table 1. Note that the equilibrium bond lengths for the C ring, 1.39 Å, fall in the regime of double-bond formation, suggesting an aromatic behavior. Moreover, C is conjugated with the B cycle, which, together with their planarity, as stated above, originates fluorescence. This phenomenon is discussed below.

Table 1 Optimized bond lengths (Å) for aflatoxin B1 (1a–1h and 3) molecules at the levels of theory B3LYP/6-311+G(d,p) and the higher-energy triplet state of 1 (1aT)a
Full size table
Table 2 Optimized bond angles (degrees) for aflatoxin B1 (1a–1h and 3) molecules at the levels of theory B3LYP/6-311+G(d,p) and the higher-energy triplet state of 1 (1aT)a
Full size table
Table 3 Optimized dihedral angles (degrees) for aflatoxin B1 (1a–1h and 3) molecules at the levels of theory B3LYP/6-311+G(d,p) and the higher-energy triplet state of 1 (1aT)a
Full size table

Considering the almost-planar geometry of the aflatoxin molecule, as well as the rupture of the lactonic ring under acidic conditions; it is necessary to realize other theoretical studies supporting this behavior. Therefore, bond orders, charges, map of electrostatic potential, HOMO, and LUMO, as well as some key steps in the hydrolysis reaction and fluorescence, were also determined.

Bond Orders

The triplet state is promoted by light in the 425-nm region (Purchase 1974) and plays an important role in fluorescence behavior and in the reactivity properties of the aflatoxin compound. Therefore, we proceed with the estimation of the bond orders (Mayer 1986a, b) for both the ground state and the higher-energy triplet state. The results are reported in Table 4; the differences between these states are also indicated. The objective is to examine whether the excited triplet state presents significant differentiation in terms of bond length and bond order. The C4−C3b, C3b−C9c, and C9b−C5a distances are clearly lengthened with respect to the ground state. A similar behavior occurs for the C3a−C11a, C11−O14, and O10−C9c bonds on the B ring, which is the main active region of aflatoxin B1. Note also the observed shortening of the C4−C5, C9b−C9c, C11a−C11, C11−O10, and C3a−C3b bond lengths. The bond lengths and bond orders are reported in Tables 1 and 4, for both singlet and triplet species; they clearly indicate a conjugation on benzene ring C and a considerable relaxation in the lactonic group. These results also suggest some conjugation on the lactonic ring B, contributing to the stability of aflatoxin B1. The data also show that the main difference lies in the bonds involving the central benzene ring C of aflatoxin B1. Consequently, the bond order is smaller for the C11−O10 bond, which is important for the breaking of this bond during the acidic hydrolysis. In fact, this C–O bond suffers a rupture in several types of hydrolysis. On the other hand, the calculated bond length for O10−C9c is relatively small, with a big bond order. Therefore, these bond order and bond lengths results suggest that bond breaking is more likely to occur at the C11−O10 bond.

Table 4 Bond orders for aflatoxin B1 (1a) molecules at the level of theory B3LYP/6-31G(d,p) and the higher-energy triplet state of 1 (1aT)
Full size table

Charges

Figure 1b and Table 5 report the natural population analysis (Reed et al. 1985) of the atomic charges for some selected atoms of aflatoxin B1, obtained with B3LYP/6-311+G(d,p). These values indicate that the O10 atom is slightly more negative than the O14 center. Thus, the addition of a proton may occur on the O10 or O14 atoms. The total energy results indicate that the addition at O14 is favored, because it is 25.3 kcal/mol (relative electronic energy), more stable than the addition at the O10 atom. The charges for the excited molecule, listed in Table 5, also were calculated with the B3LYP/6-311+G(d,p) method. The changes in charge distribution between the ground and the excited triplet state gave an indication of the fluorescent ability of the molecule.

Table 5 Calculated B3LYP/6-311+G(d,p) natural population analysis of the atomic charges (e −) of some atoms for aflatoxin B1 (1a–1h and 3) molecules and the higher-energy triplet state of 1 (1aT), as well as their difference (excited state [1aT] − ground state [1a])
Full size table

The charge difference between the excited and the ground state is also tabulated in Table 5. It is evident that an increase in electron charge is observed for the C1, C3a, O10, and C11 atoms, whereas a decrease was found for O12 and O14, with the excitation. These results indicate that the charge at both carbon, C11, and oxygen, O14, atoms must participate in the C=O bond, showing a small, but significant, increase in C=O bond length, from 1.194 Å in the ground state up to 1.200 Å in the excited state. Up to this point, these results allow a characterization of the structural and electronic details of the lactonic region, which is important in the hydrolysis reaction.

On the other hand, with respect to the excited state, the C4 and C5a atoms of the benzene ring show an increase in their atomic charges, whereas the atomic populations in O13 and C5a are decreased; these changes were reflected in the slight shortening of the C4−C5 and C5−C5a bond lengths. Most of the electron charge in the excited state is localized in the lactonic ring, as well as in some carbon atoms of the benzene ring. These results are in agreement with the experimental observation that the phenomenon of fluorescence disappeared when aflatoxin B1 was submitted to an acid hydrolysis process (Méndez-Albores et al. 2008).

Molecular Electrostatic Potential Surfaces

As mentioned before, the electrostatic potential has been primarily used to predict sites and relative reactivities toward electrophilic or nucleophilic attacks (Murray and Politzer 1998, Murray et al. 1998, 1999). From these molecular electrostatic potential surfaces (MEPSs) it was possible to identify that aflatoxin B1 has several possible sites for an electrophilic attack. The most negative values for MEPSs were found between O12, O14, and O10; in this sense the O14 position is more favorable for electrophilic attack (see Fig. 2), since the carbonylic oxygen atom of the lactonic moiety, O14, is the site with the largest negative potential, indicating the highest nucleophilic character of this atom. The sites with a positive potential are scattered among the different hydrogen atoms, C6a, C8, and the hydrogen atoms of the methyl group, CH3O13. According to the reaction conditions, these positive centers are not feasible for a nucleophilic attack. On the other hand, the natural population analysis indicated a strong deficiency of electron density on the C11 carbon atom, due to the oxygen atom O14, which polarizes the C11 carbon atom.

Fig. 2
figure 2

Electrostatic potential molecular surface of 1a

Full size image

In general, nucleophilic attacks on target molecules could occur at positive sites, and an electrophilic procedure should occur at the negative sites of aflatoxin B1. The electrostatic potential is a useful expression of this idea.

HOMO and LUMO Localizations

The contour plots of the HOMO and LUMO frontier orbitals were obtained with the B3LYP/6-311+G(d,p) method. As shown in Fig. 3, HOMO is mainly localized on the O12 and O14 sites; also, some π-bond contributions are recorded on ring C. In other words, HOMO has strong contributions from “p” orbitals, as can be appreciated in the carbon atoms of the title molecule; this charge distribution is additional to the lone electron pairs of the named oxygen atoms. Therefore, HOMO indicates the regions with a more Lewis-base character of aflatoxin B1. Due to the low reactivity of a π-bond with a proton under the experimental conditions previously stated, the major Lewis-basic sites correspond to O12 and O14. A similar pattern was observed for the HOMO-1 and HOMO-2 orbitals, which are located 0.28 and 0.40 eV below HOMO. Moreover, in agreement with the proposed AAc2 mechanism, the most reactive site belongs to O14, and O12 makes more contributions to LUMO (Fig. 3).

Fig. 3
figure 3

LUMO, HOMO, HOMO-1, and HOMO-2 contour plots for the aflatoxin molecule (1) at the B3LYP/6-311+G(d,p)

Full size image

On the other hand, the charge distribution and the energy of LUMO can be indicative of the carcinogenic activity of a substance, according to Leão and Pavão (1997); they indicated that the carcinogens have a low LUMO energy. In the case of the aflatoxin B1 compound, the LUMO has a considerable lower absolute energy, −2.38 eV, and it shows the main contributions around the C11 atom, which is the region of interest.

Hydrolysis of Aflatoxin B1

Coming back to the aflatoxin B1 ground-state structure (see Fig. 1b), the relatively longer O10−C11 distance, 1.423 Å, as compared with the O10−C9c bond length, 1.354 Å, is consistent with the feature that the O10 atom may cause rupture of the O10−C11 bond, when the protonation is done just at the O10 site, as indicated in structure 1a in Figs. 4 and 5 (see also Scheme 2 and Table 6). However, protonation may also occur at O14, as shown in structure 1b in Fig. 4. In both cases, the acidic medium produces hydrolysis of the lactone ring, yielding a compound that contains both hydroxylic and carboxylic groups, 1h. In more detail, in the first step of the reaction mechanism (stereoisomer 1: a → b) (see Fig. 4), our calculations indicate that proton addition occurs between the O10 and the O14 atoms, but closer to O10, yielding a TS which has an imaginary frequency at 1489 cm−1), with vector displacements around the O10−H−O14 atoms; 1b. The O14−H and O10−H contacts in 1b show distances of 1.441 and 1.174 Å, respectively, which indicate a stronger bonding interaction for O10−H. The bond lengths for C11−O10 (1.462 Å) and O10−C9c (1.407 Å) are rather longer than those of 1a. Besides, for the TS 1b the C11−O14−H and C11−O10−H angles were found to be 76.8º and 77.5º, whereas the O14−C11−O10 angle was found to be 100.8º (slightly smaller than sp 3 hybridization). In 1a this angle is 116.6º, revealing that the hybridization is of the sp 2 type. Furthermore, the TS 1b may produce two intermediates, 1c and 1d, with the proton lying at O14 or at O10, respectively. The results obtained suggest that the H−O14 (1c) structure may originate the more favorable pathway, because it is considerably more stable, by about 25.3 kcal/mol in electronic energy, than 1d. In fact, the reaction pathways originating from 1d may be disregarded. The reaction pathways from TS to O14−H and from TS to O10−H structures were verified by an IRC calculation.

Fig. 4
figure 4

Intermediates and transition states for the reaction pathway 1a → 1h at the B3LYP/6-311+G(d,p) level of theory for molecule 1

Full size image
Fig. 5
figure 5

Key points of the potential energy surface, at B3LYP/6-311+G(d,p), for the reaction pathway of aflatoxin B1 (1a → 1h), of system I

Full size image
Scheme 2
scheme 2

Steps proposed for the hydrolysis of aflatoxin B1 under acidic conditions

Full size image
Table 6 Electronic energy (in the gas phase) for the reaction pathway of aflatoxin B1 (1a → 1h) of system 1, at the levels of theory B3LYP/6-311+G(d,p)
Full size table

Then the addition of one proton produces the intermediate 1c. The C11−O14 bond length was calculated to be 1.293 Å, which is a rather long C–O bond length, i.e., 0.099 Å longer than that of the 1a structure of aflatoxin. The interaction of 1a with a proton enhances the electronic delocalization of the B and C rings, as can be seen by the shortening of the C–C distances (see structure 1c in Fig. 4). As can be observed, the C11−C11a, C11−C10, C3a−C3b, C9b−O9c, C4−O5, C4−O13, and C5a−O6 bond lengths are clearly shortened, whereas the O10−C9c, C9c−C3b, C3a−C11a, C4−C3b, and C9b−C5a distances are lengthened with respect to the ground state.

The interaction of a proton with O14 affects, first, the net charge distribution of the C11−O14 bond. The O14 atom becomes more negative, while the charge on C11 also becomes positively charged. This interaction enhances the electrophilicity of the carbon atom C11. The C11 charge in 1c is more positive, by 0.035 e −, than that of 1a. Another notorious phenomenon in 1c is the charge distribution for some carbon atoms of the aromatic ring B. Two of them, O10 and C3b, increase their positivite character by 0.110 and 0.047 e − respectively, while C9c and C11a become more negative (0.103 and 0.037 atomic charge unit).

Further, our results also indicate that the lactonic ring may also be broken at the O10−C11 bond, producing a C11≡O14 acylium ion as an intermediate, which is recognized in structure 1d, located 25.3 kcal/mol above 1c (relative electronic energy). It is worth noting that this intermediate can be stabilized by resonance (Boer 1968; Olah and Westerman 1973), which is in agreement with the theoretical calculations for the C11≡O14 distance; the result obtained, 1.138 Å, indicates a strong bond formation.

Also, a TS called TS2 (imaginary frequency, 1497 cm−1), 1e, is suggested before the final product, 1h. The addition of one water molecule produced the intermediate 1f. The C11−O14 distance was calculated to be 1.309 Å, indicating a rather long C–O bond. Similarly, the C11−O14 bond is longer, by 0.115 Å, than 1a. Also, the interaction of 1c with the H2O molecule produces an increase in the electronic delocalization in ring C, which includes the contributions of the C3a, C11a, C1, and O12 sites of 1f. As shown in Fig. 4, the C11−O10, C11a−C3a, C3b−C4, C5a−C9b, and C9c−O10 distances are lengthened. To a minor extent, the C3a−C3b, C4−C5, C5−C5a, C11−C11a, and C9b−C9c bond lengths are shortened, with respect to 1a. The interaction of one water molecule with the C11 atom affects, first, the net charges of the C11−O14 bond. The O14 atom becomes more negative, and the charge on C11 is more positive, by 0.049 e −, than the corresponding value for 1a. Besides, others atoms of the B ring suffer changes in the charge distribution. Two of them, C3a (0.106 e −) and C3b (0.009 e −; 0.589) increases their positive charges with respect to 1a, while C9c, C11a, and O10 become more negative (0.033 and 0.123 atomic charge units).

A 1g TS, TS3 (imaginary frequency, 282.6 cm−1), is suggested before the final product, 1h, is formed. Abstraction of the proton from the oxygen atom provokes changes in the O16−C11, C11−O14, and C3a−C3b distances that are clearly reflected in the lengthening of these bond lengths. In addition, C3a−C11a showed a shortening of the distance (by 0.013 Å). All bond lengths and bond angles in the ring B of structure 1g change considerably and affects those of rings A and C, with respect to 1f. At the same time, planarity of ring A is absent, since the C3C3aC3bC9c and C11aC3aC3bC4 dihedral angles are −55.4o and 61.5o, respectively. As for the electronic details, the O14 and C11 atoms of 1g become positively and negatively charged, by 0.046 e − and 0.003 e −, respectively, compared to the corresponding value for 1f.

The hydrolysis of the aflatoxin B1 molecule originates two conformers as final products. Figure 4 shows the final product, 1h. These two species can occur with or without the formation of an intramolecular hydrogen bond. In fact, in 1h an O10−H−O14 bonding is formed, with the relatively short H−O14 distance of 1.733 Å. Around the H−bond region, the C11a−C3a, C11−O16, and C9c−O10 distances are shortened, and the C11−O14 and O10−H10 bonds are lengthened.

It should be noted that the H-bond formation stabilizes structure 1h, by about 7.4 kcal/mol (relative electronic energy). This result is indicative of considerable bond strength for this bond and exemplifies the importance of this type ofn bonding on the stability of this conformer. It is important to mention that the generality of this analysis is in complete agreement with a hydrolytic mechanism for the lactone (a cyclic ester) compound, belonging to the class of acid-catalyzed, acyl cleavage, second-order reactions (AAC2) (Zimmermann and Rudolph 1965), as proposed by Ingold.

In this sense, it is convenient to note that, to classify the catalytic hydrolysis of esters, eight possible mechanisms have been proposed, the most common being BAC2 for basic conditions and AAC2 for acid catalysis. The above notations (Sykes 1986) indicate that the O–R bond of the acyl group is commonly broken by a bimolecular procedure. Consequently, the lactone ring of aflatoxin B1 must be broken by means of an AAC2 mechanism under diluted acid conditions. In this mechanism, the ester first accepts a proton at the carbonyl oxygen and this change enhances the positive charge on the carbonyl carbon. This protonation facilitates the successive addition of water at that position to form a tetrahedral intermediate, consisting of the protonated cyclic ester and water.

Fluorescence of Aflatoxin B1

The molecular structure is one of the most important features in fluorescent species. The dihedral angles for the ground-state structure of aflatoxin B1 are reported in Table 3. These results indicate that the A, B, C, and D rings show a planar conformation, with all dihedral angles varying by less than 1° from planarity, whereas the E ring is located outside that plane. Particularly, the coplanarity of the phenyl C ring with the B one promotes some degree of conjugation, enabling the formation of an extended π system on these rings, which in turn yields an efficient absorption of relatively low-energy photons.

Due to planarity, the degree of rotational freedom is notably restricted and thus the rotational-vibrational relaxation mechanism does not contribute to the loss of the absorbed energy. As a result, the singlet–triplet excitation, with the associated emission for returning to the ground state, is the main fluorescence mechanism for relieving the energy excess, yielding a molecule with high fluorescence quantum yields.

In the singlet state, the B3LYP/6-31G(d,p) time-dependent results (see Table 7) indicate that the lowest excitation occurs at 377 nm but with a zero value for the oscillator strength (f). On the contrary, the singlet excitation occurring at 309 nm (4.01 eV) shows the highest f value (0.24) and represents a HOMO(-2)-LUMO transition; these results are in agreement with the reported values (Guedes and Ericsson 2006). Note that the use of B3LYP/6-31+G(d,p) for this single excitation renders a value of 312 nm (3.96 eV), with an oscillator strength of 0.23, which is in agreement with a previously reported value, (Bauernschmitt and Ahlrichs 1996).

Table 7 Absorption spectra determining vertical electronic excitations from the ground state using the TD-DFT with B3LYP/6-31G(d,p)
Full size table

For the triplet state the B3LYP/6-31G(d,p) results indicate 408 nm for the main transition, with an f value of 0.32 and representing predominantly a HOMO-LUMO(+2) event, which is in reasonable agreement with the experimental value of 425 nm (Purchase 1974). Note that this wavelength falls in the visible region. It is to be mentioned that use of the larger 6-311+G(d,p) basis set yields 433 nm for this transition, in better concordance with the experiment. A similar value, 431 nm was obtained using the 6-31+G(d,p) basis set for the transition estimation, which takes as input a 631G(d,p) optimized geometry.

Conclusions

The aflatoxin B1 compound was studied by means of DFT; the determined structural and electronic properties allow the characterization of its reactivity behavior in acidic hydrolysis. It was suggested that the reaction pathway that involves the lactone ring is the main one responsible of this process. The results obtained agree with the experimental findings for this type of hydrolysis. It was found that the B ring is broken completely during the reaction mechanism, producing transference of charge for the ground state structure 1, and the excited singlet indicated the ability of the molecule to display fluorescence. These fluorescence properties are lost when the planarity of the molecule is broken, namely, by hydrolysis under acidic conditions and opening of the lactonic ring. The atomic charges, electrostatic potential, and HOMO and LUMO indicate that the oxygen atom O14 is the more fitting site of the aflatoxin moiety for interaction with protons, generated from the aqueous citric acid conditions. Thus, it was shown that the lactonic ring is hydrolyzed under acidic conditions by an acid-acyl bimolecular mechanism (AAC2), implying the deletion of its carcinogenic properties.