Patulin is a polyketide metabolite, like several other major mycotoxins, e.g., aflatoxins, fumonisins and ochratoxins, although this last toxin is a polyketide/amino acid hybrid compound. The study of its biosynthesis is historically significant for two reasons. Firstly, the acetated hypothesis of Birch [11] was based upon the incorporation of radiolabeled acetate into 6-methylsalicylic acid (6MSA), and led to the recognition of a major class of natural products, the polyketides, previously suggested by Collie. The second reason is that the enzyme involved in the first patulin biosynthesis step was the first polyketide synthase to be studied and characterized in vitro [12]. The biosynthetic pathway of patulin consists of about 10 steps as suggested by several biochemical studies and by the identification of several mutants that are blocked at various steps in the patulin biosynthetic pathway [13] (Figure 2).

The first step in the production of patulin is the formation of 6MSA by the condensation of one acetyl-CoA and three malonyl-CoA units. This formation is carried out by a single multifunctional enzyme that has several enzymatic activities: acetyl and malonyl transferase, ketoacyl synthase, ketoreductase and dehydratase [12]. This enzyme also possesses an acyl carrier protein function [14,15] and consists of four identical 176 kDa polypeptidic chains [16]. Historically, Bu’Lock and Tanenbaum showed that radiolabeled 6MSA was converted into patulin [17,18].

The studies using ¹⁴C and ³H radioisotopes and ¹³C and ²H stable isotopes showed that 6MSA is then modified extensively to form patulin. The products generated from m-cresol and gentisylaldehyde are structurally similar to 6-methylsalicylic acid [19,20,21]. 6MSA is modified to m-cresol by 6MSA decarboxylase, then the methyl group of m-cresol is oxidized to form an aldehyde group. This step is followed by a hydroxylation reaction that leads to gentisaldehyde formation. However, at this point, it seems that for over a decade the sequence of the different intermediates was much less clear. The conversion of gentisaldehyde to a two ring structure such as patulin needs the opening of a ring by a mechanism mediated either by a monooxygenase or by a dioxygenase. The isolation of several patulin-minus mutants of P. griseofulvum led to the identification of four post aromatic precursors: isoepoxydon [22] phyllostine [23], neopatulin [24] and ascladiol [25]. The discoveries of these compounds led also to the hypothesis that an epoxidation step occurs after gentisaldehyde synthesis. The nature of the epoxidation has remained a matter of speculation since a point of contention existed as to whether the substrate of the epoxidation reaction was gentisyl alcohol or gentisaldehyde [26]. Indeed, although a kinetic pulse labeling study performed by Forrester and Gaucher [19] showed that only the following co-metabolites are readily converted into patulin: acetate, 6MSA, m-cresol, m-hydroxybenzyl alcohol, m-hydroxybenzaldehyde and gentisaldehyde, a crude extract which catalyzed the epoxidation of gentisyl alcohol to phyllostine was isolated [26].

Several enzyme activities related to patulin biosynthesis have been characterized. Among them, the ring hydroxylation of m-hydroxybenzyl alcohol to gentisyl alcohol requires NADPH and molecular oxygen for activity. Inhibition in a reversible manner by carbon monoxide suggests the involvement of cytochrome P450 [27].

Five of the enzymes involved in patulin biosynthesis have been partially or fully purified: the first, 6-methylsalicylic acid synthase, the second, 6-methylsalicylic acid decarboxylase [28], m-hydroxybenzyl alcohol dehydrogenase, the seventh, isoepoxydon dehydrogenase (IDH) [29] and the eighth, neopatulin synthase [29].

In filamentous fungi, the genes encoding the enzymes involved in the synthesis of secondary metabolites are usually contained together in clusters on chromosomes [30,31]. A large number of gene clusters related to secondary metabolite production has been discovered and particularly those responsible for the biosynthesis of several mycotoxins, e.g., aflatoxins [32], fumonisins [33], trichothecenes [34], ergot alkaloids and zearalenone [35]. In contrast, a gene cluster involved in patulin biosynthesis has only recently been isolated and only some genes related to patulin biosynthesis have been identified. The first gene was the 6-methylsalicylic acid synthase (6MSAS) gene from P. griseofulvum [14,15]. The idh gene encoding the seventh enzyme, isoepoxydon dehydrogenase (IDH), was originally isolated from P. griseofulvum [29], and then from several other Penicillium species [36,37,38] and from B. nivea [8,39]. Recently, a gene located downstream of the idh gene has been isolated, which encodes a protein with a high homology to isoamyl alcohol oxidase [40]. Finally, genes encoding cytochrome P450 enzymes—involved in two steps of the patulin biosynthesis pathway—have been isolated from A. clavatus, and characterized by heterologous expression in yeast [41]. The first cytochrome P450 (CYP619C3) transforms m-cresol to yield m-hydroxybenzyl alcohol whereas the second cytochrome P450 (CYP619C2) catalyses the hydroxylation of m-hydroxybenzyl alcohol to gentisyl alcohol. This last monooxygenase can also hydroxylate m-cresol to 2.5 dihydroxytoluene (toluquinol). This result confirms an earlier experiment showing that the same enzyme preparation could convert m-cresol and m-hydroxybenzyl alcohol respectively to toluquinol and gentisyl alcohol [26,27]. Toluquinol is a well-known compound produced by P. griseofulvum. Its role as patulin precursor was discussed in earlier studies but the use of a pulse–chase radiolabeling technique or a culture of force-fed deuterated toluquinol has shown that toluquinol is metabolized to deoxyepoxydon and not to patulin [19,20]. On the basis of these studies, toluquinol is a co-metabolite of patulin, but not an intermediate in the patulin biosynthetic pathway.

Recently, a cluster of 15 genes involved in patulin biosynthesis has been identified in the A. clavatus genome [41]. All the genes are located in a 40 kb region. The genes encode the enzymes necessary for the biosynthesis of the toxin, but also the specific regulatory factor and transporters. This cluster contains three transporter genes: one ABC (ATP binding cassette) transporter, one MFS (Major Facilitator Superfamily) transporter and one acetate transporter. The cluster also contains genes for enzymes: one putative carboxyl esterase (PatB), one putative Zn-dependent alcohol dehydrogenase (PatD), one GMC (Glucose-Methanol-Choline) oxidoreductase (PatE), one gene that probably encodes a putative decarboxylase displaying an amido hydroxylase conserved domain (PatG), two genes (PatH and PatI) encoding cytochromes P450 (responsible for the hydroxylation of m-cresol to m-hydroxybenzyl alcohol and of m-hydroxybenzyl alcohol to gentisyl alcohol, respectively) and a gene encoding a putative dioxygenase (PatJ).

The cluster contains also the 6msas gene (PatK), one putative transcription factor gene (PatL), the isoepoxydon dehydrogenase gene (PatN) previously isolated as idh gene, a putative isoamyl alcohol oxidase described recently from P. griseofulvum (PatO), and finally a gene which does not have any obvious function (PatF). Related clusters are also present in the genome of some species that do not produce patulin such as P. chrysogenum, Talaromyces stipitatus and Aspergillus terreus. They are also shown in Figure 3. The fact that these latter species do not produce patulin can be explained by the lack of some key genes such as the 6msas gene in T. stipitatus or idh gene in A. terreus or P. chrysogenum. Although Varga, et al. [42] confirmed that A. terreus was not a patulin producing species, the isolation of the AtX gene encoding a 6MSAS has been previously reported. The gene encoded a functional enzyme that led to 6MSA formation after heterologous expression in A. orizae [43]. In A. terreus, another biosynthesis pathway, that of terreic acid, needs 6-methylsalicylic acid synthase activity [44]. Since there is only one 6msas gene in the A. terreus genome, it is very likely that this gene encodes an enzyme involved in terreic acid biosynthesis in this species.

Although their roles are not proven, the involvement of some genes can be predicted with regard to the steps identified biochemically in the synthesis of patulin. PatG contains the amido hydroxylase superfamily signature sequence motifs shared by γ-resorcylate decarboxylase and 5-carboxyvanillic acid decarboxylase [45]. Pat G is most likely involved in the decarboxylation of 6-methylsalicylic acid to m-cresol. On the basis of the high homology with VBS (59%), which catalyses the transformation of versiconal into versicolorin B in the aflatoxin pathway, and the similarity between their chemical structures, PatE could catalyze the last step of the patulin pathway, to yield patulin from ascladiol. The very recent demonstration that vesicles play a key role in aflatoxin biosynthesis and the hypothesis of the role of subcellular compartmentalization could be explain the presence of three transporter genes in the cluster [46].

The level of nutrient nitrogen in the culture medium determines when the patulin pathway is expressed. In their study, Grootwassink and Gaucher [47] showed that the age at which a P. griseofulvum culture produced the key enzyme m-hydroxybenzyl alcohol dehydrogenase increased when the nutrient nitrogen content of the medium increased or when the size of the inoculums decreased. Resuspension of the mycelium in a nitrogen-free 4% glucose solution immediately induces the pathway enzymes and patulin production. Distinct metabolic phasing can also be observed in cultures with ammonium as the sole nitrogen source. When ammonium ions were added to cultures actively producing patulin, a rapid loss of secondary metabolism occurred [48]. The effect of ammonium repression has been determined to be at the transcription level. The addition of 30 mM ammonium chloride to cultures producing idh and 6msas transcripts caused a considerable decrease in the amounts of both transcripts, suggesting that a rapid turnover of patulin mRNA occurred and not enzyme inhibition. This down-regulation has been also observed for various mycotoxins such as sterigmatocystin and ochratoxin A [49,50]. In P. griseofulvum, the upstream regions of both the 6msas and idh genes have several putative GATA sites that could interact with a global nitrogen regulatory factor encoded by the nrfa gene, an AreA orthologous gene in P. griseofulvum. In a reporter gene study, Summerer [51] showed that reporter gene activity is observed only in cultures lacking ammonium and that deletion of regions containing the proposed GATA binding sites resulted in a loss of expression activation. DNA mobility shift assays demonstrated that the upstream regions of the 6MSA synthase and isoepoxydon dehydrogenase genes bind strongly to the NRFA DNA binding domain [52].

In contrast, there is no carbon metabolite repression of patulin biosynthesis. Like aflatoxin, glucose is an optimal carbohydrate source for patulin production but it is unclear if the carbon source directly induces patulin production or modulates its biosynthesis through general metabolism.

Manganese had been shown to be an essential requirement for patulin biosynthesis in P. griseofulvum [53]. Inhibitor studies using actinomycin D and cycloheximide showed that manganese exercised its affect on patulin biosynthesis at the level of transcription and influenced the coordinated appearance of pathway enzymes [54]. This was confirmed by Northern Blot analysis [29]. Manganese has a substantial effect on the expression of the idh gene, but only a limited effect on the 6msas gene, resulting in 6-methylsalicylic acid accumulation but very little patulin production.

In P. griseofulvum shaken cultures, the 6MSAS enzyme seems to appear four hours before the 6MSA decarboxylase, m-hydroxybenzylalcohol dehydrogenase, and isoepoxydon dehydrogenase. Induction of later enzymes in the pathway by certain patulin pathway intermediates has been shown [55]. Indeed, the addition of early pathway metabolites i.e., 6MSA or m-hydroxybenzyl alcohol to 16 hour old shaking cultures shifted the appearance of the fourth enzyme forward by about two hours. When various concentrations of 6MSA or m-hydroxybenzyl alcohol were added to growing cultures of a mutant strain unable to produce patulin and 6MSA, the second enzyme, 6MSA decarboxylase appeared 1.5 hours earlier, with maximum levels about seven-times higher than the uninduced control. Induction was greatest with m-hydroxybenzyl alcohol. This regulation has been shown to occur at the transcriptional level [29].

Finally, using a reporter gene strategy, Summerer [51] showed that the pH of the culture medium affected the transcription of the idh gene. In P. griseofulvum, optimal transcription occurs when the initial culture pH is 5.0, and decreases when the pH is either raised or lowered.

Patulin has a strong affinity for sulfhydryl groups. Patulin adducts formed with cysteine are less toxic than the unmodified compound in acute toxicity, teratogenicity, and mutagenicity studies. Its affinity for SH-groups explains its inhibition of many enzymes.

In rodents, the oral LD50 of patulin ranges between 29 and 55 mg/kg body weight (b.w.) [56,57,58]. Poultry seem less sensitive, with an oral LD50 of 170 mg/kg b.w. When administered by the intravenous, intraperitoneal or subcutaneous routes, patulin is 3-6-times more toxic. Toxic signs consistently reported in all studies were agitation, in some cases convulsions, dyspnea, pulmonary congestion, edema, and ulceration, hyperemia and distension of the gastro intestinal tract [59].

Some compounds were able to modulate the toxicity of patulin. For example, inhibitors of cytochrome P450 such as proadifen (SKF 525A) considerably increase the toxicity of this mycotoxin, while P450 inducers do not modify its toxicity [60]. Similarly, when a patulin/cysteine adduct was administered to mice intraperitoneally, no acute toxicity was observed at levels up to 150 mg of patulin/mouse [61].

The sub-acute administration of patulin has been mainly studied in rats, where it was shown to induce weight loss, gastric and intestinal changes and alterations in renal function [62,63]. Repetitive doses lead to signs of neurotoxicity (tremors, convulsions) as well as an inhibition of several enzymes (ATPase) in the intestine [64] and the brain, in particular, with consequences in terms of lipid metabolism [65].

Similar clinical signs were observed in mice, hamsters and chickens. In monkeys, no sign of toxicity was observed after daily treatments with 5 to 500 µg/kg b.w. for four weeks. Only monkeys receiving 5 mg of patulin/kg b.w./day for two weeks rejected their food during the last three days of the experiment [66].

Selmanoglu and Kockaya [67] measured thyroid and testicular hormones in rats receiving 0.1 mg patulin/kg b.w./day patulin by the oral route for 60 or 90 days. A 60-day exposure increased the plasma level of testosterone and decreased T₄ hormone while there was no change in T₃, TSH, LH and GH. When the exposure lasted for 90 days, there was an increase in testosterone and in LH without any other clinical signs. Histological examination of the thyroid showed lymphoid cell infiltration and enlargement of interstitial tissue. At the tested level, edema, fibrosis, local Leydig cell hyperplasia and disorganization of the seminiferous tubule epithelium were observed. This was associated with a decreased sperm count [68].

As summarized in Table 1, patulin is recognized as mainly inducing gastrointestinal disorders with ulceration, distension and bleeding, and at higher doses, alterations in renal function.

Patulin did not increase revertant frequency in the Ames test using several strains of Salmonella typhimurium, but some studies have shown mutagenic activity in Saccharomyces cerevisiae strains and in Bacillus subtilis [69,70]. Patulin gave a negative result for genotoxicity by the SOS microplate assay [71], however, it was clearly positive in the initiator tRNA acceptance assay for carcinogens [72].

Several studies have demonstrated that patulin acts as a clastogen in mammalian cells, e.g., inducing micronuclei without kinetochores and structural chromosomal aberrations in cultured Chinese hamster lung fibroblast V79 cells [73,74,75]. Chromosome aberrations and gene mutations were induced by patulin in FM3A cells, a C3H mouse mammary carcinoma cell line [76], in V79 cells [77], and in mouse lymphoma L5178Y cells [78]. Chromosome and chromatid gaps and breaks were also induced by patulin in Chinese hamster ovary cells but not in human peripheral blood lymphocytes [79]. Similar differences in the susceptibility of different cell systems to the genotoxic activity of patulin were reported with respect to sister chromatid exchanges, which were induced in Chinese hamster ovary cells and human peripheral blood lymphocytes [79] but not in Chinese hamster V79 cells [73]. Evidence for oxidative DNA damage was obtained in human embryonic kidney cells after treatment with patulin [79,80] and it is also implied by the observation that ascorbic acid decreases the induction of micronuclei and chromosomal aberrations in PAT-treated V79 cells [75].

In conclusion, although the data on genotoxicity were variable, most assays carried out with mammalian cells were positive while assays with bacteria were mainly negative. In addition, some studies indicated that patulin impaired DNA synthesis. These genotoxic effects might be related to its ability to react with sulfhydryl groups and to induce oxidative damage [80,81]. Nevertheless, the WHO [59] concluded from the available data that patulin is genotoxic.

The few studies of long-term toxicity on patulin showed an absence of tumors in rats orally exposed to 0.1 to 2.5 mg patulin/kg b.w./day for 74 to 104 weeks [82,83]. According to the International Agency for Research on Cancer (IARC), patulin is classified in the group 3 as ‘‘not classifiable as to its carcinogenicity to humans” [70].

As part of a two-generation reproduction study, the offspring of Sprague Dawley rats were exposed to 1.5 mg patulin/kg b.w./day [84]. Patulin caused an increased resorbtion in F1 litters, but this effect was not observed in the F2 generation. In this generation only a decrease in the weight of the foetuses without observable deformations was noticed. Similar observations were made by Reddy, et al. [85] in rats receiving 1.5 mg patulin/kg b.w. intraperitoneally; a dose of 2 mg toxin/kg b.w. induced the abortion of all embryos.

In mice, the same oral dose provoked the death of the offspring with brain, lung and cutaneous haemorrhages [82]. A higher oral dose (3.8 mg/kg b.w./day for 12 days) did not modify the number of implantations and delivered fetuses without inducing any deformation. By contrast, intraperitoneal injection increased cleft palate as well as malformations of the kidneys [86].

When injected into the air cell of chick eggs, patulin was reported to be embryotoxic at levels of 2.35–68.7 µg/egg depending on the age of the embryo, and teratogenic at levels of 1–2 µg/egg. Patulin/cysteine adducts showed the same toxic effects, but at much higher doses: 15–150 µg of patulin equivalents [61]. Rat embryos were exposed to patulin-treated (0.00–62 µM) rat serum for 45 h. The embryos that were exposed to the highest dose did not survive beyond 40 h of incubation. Patulin induced a significant reduction in the protein and DNA content, yolk sac diameter, crown rump length, and somite number count. Patulin treatment also increased the frequency of defective embryos. Anomalies included growth retardation, hypoplasia of the mesencephalon and telencephalon, and hyperplasia and/or blisters of the mandibular process [87].

As described for other mycotoxins, patulin can alter the immune response of the host [88]. Numerous in vitro studies have demonstrated that patulin inhibits several macrophage functions. Indeed, Sorenson et al. [89] showed that in vitro exposure of alveolar rat macrophages to patulin inhibited protein synthesis and altered membrane functions. Patulin also significantly decreased the production of O₂⁻, phagosome-lysosome fusion, phagocythosis, as well as lysosomal enzyme and microbiological activity in mouse macrophages [90].

More recently, patulin was found to reduce the cytokine secretion of IFN- γ and IL-4 by human macrophages [91] and of IL-4, IL-13, IFN-γ, and IL-10 by human peripheral blood mononuclear cells and human T cells [92]. However, this mycotoxin was less potent than gliotoxin [91]. Similarly, Marin et al. [93], using the thymoma cell line EL-4, observed a reduction in the production of interleukins 2 and 5, with concentrations of 500 ng patulin/ mL. Similarly, in this study, the inhibitory effect of patulin was lower than that observed with T-2 toxin. This decrease in cytokine secretion was not due to the cytotoxic effects of patulin but to the depletion of intracellular glutathione [92].

In vivo studies using mice indicate variable effects of patulin on the immune system. These effects include an increased number of splenic T lymphocytes and depressed serum immunoglobulin concentrations [94], depressed delayed hypersensitivity responses [95] and increased neutrophil numbers and resistance to Candida albicans infection [96].

A thorough battery of immunology endpoints was examined in mice administered patulin daily by gavage for 28 days at doses of 0.08 to 2.56 mg patulin/kg b.w. [97]. These doses were calculated to be approximately equivalent to estimated human exposure levels. Changes in immune cell numbers included depressed peripheral blood leukocytes and lymphocytes numbers (observed at 1.28 and 2.56 mg toxin/kg b.w./day), an increased number of splenic monocytes and Natural Killer cells (from 0.08 mg toxin/ kg b.w./day), an increased number of cytotoxic T lymphocytes (at 2.56 mg patulin/kg b.w./day) and changes in the percentage of immunoglobulin (Ig)⁺, CD3⁺, CD4⁺/CD8⁻ and CD4⁻/CD8⁺ lymphocytes in the spleen. These changes in cell number did not reflect functional changes. There were no measurable changes in immune function in patulin-treated mice using the following endpoints: antibody response to sheep red blood cells (SRBC), mixed leukocyte responses and Natural Killer cell functions. The authors concluded that exposure to patulin, at levels consistent with potential human exposure in food would not be likely to alter immune responses [97].

In rats, as in other species, patulin reduces the oxidative burst in peritoneal macrophages [93] and the production of reactive oxygen species in HL-60, human human promyelocytic leukemia cells [80]. Mc Kinley et al. [58] suggested that the increase in neutrophils was due to gastro-intestinal inflammation induced by the toxin.