The mycete populations (yeasts and moulds) colonizing the surface layers of dry-cured meat products during maturation have been regarded as positively contributing to the chemosensory properties of final outcome, by means of oxidation prevention and generation of volatile compounds enhancing aroma [1,2]. In the case of dry-cured ham, yeasts form a film on ham surface during early and intermediate processing steps, while moulds can develop fungal mycelium even after yeast growth [3].

Among moulds isolated from dry-cured meat facilities and products, Penicillia were largely dominant, and the contamination levels have been found to be associated to the environmental conditions of ripening rooms [4,5], geographic area of production [6,7,8] and seasonality [9]. Among Penicillia, P. nordicum, a toxigenic contaminant of protein-rich food [10], was isolated from air and products of dry-curing plants [5,9]. Several P. nordicum isolates proved to be important producers of ochratoxin A (OTA), a mycotoxin classified by the International Agency for the Research on Cancer (IARC) as a possible human carcinogen, suspected to be involved in nephropathies and urothelial tract tumours [11]. OTA was detected in substrates contaminated by P. nordicum [12], including dry-cured pork samples [13,14]. Currently, the appropriate setting of environmental conditions (temperature, relative humidity and air circulation), is the only accepted tool to prevent the uncontrolled growth of P. nordicum inside dry-curing plants.

Biological control of P. nordicum contamination based on microbial antagonists can be investigated as a possible way to improve food safety without affecting sensory quality and properties of typical dry-cured meat products. As a rule, during manufacturing a wild yeast population spontaneously grows on dry-cured ham surface, achieving different counts according to processing stage and type [15,16]. Yeast species isolated from ham during processing mainly belonged to genera Debaryomyces, Candida and less frequently to Cryptococcus, Rhodotorula and Rhodosporidium [15,16,17]. Native yeasts more frequently isolated from dry-cured hams were regarded as non-pathogenic with the exception of Candida zeylanoydes [18] and proved to be able of growing to high populations in conditions of temperature, pH, moisture, a_w and salt typical of dry-cured ham surface [1,19]. Recently, selected yeasts were added to dry-cured meat products as starter cultures to enrich the volatile compound profile [2,20]. The antagonistic potential of yeasts against Penicillia has been considered for different food matrixes, such as cereals [21,22], coffee [23], fruits [24] or fermented food like wine [25], cheese and yogurt [26]. Recent studies report the use of antagonistic yeasts as surface starter cultures inhibiting fungal growth also in dry-cured meat products [8,27].

The aim of this work is to evaluate in vitro the ability of selected yeasts isolated from dry-cured ham surface to compete with P. nordicum and to inhibit OTA accumulation in the perspective of using them as surface starter cultures.

The microbial community growing on dehydrated surface of dry-cured ham during the maturing phase is dominated by several species of yeasts and moulds [15,34]; the former population, is present from early processing steps, while the latter becomes dominating during maturation [4]. The efficacy of native yeasts as biocontrol agents is questionable because moulds frequently colonize ham surface in the presence of yeasts [3].

Information on the antagonistic activities of yeasts isolated from dry-cured meat products against toxigenic fungi is scarce [35]: recent study differentiated yeasts isolated from dry-cured ham according to their inhibitory effect in vitro against an OTA-producing P. nordicum strain [19]. Although the mechanisms underlying the observed inhibition were not yet investigated, it is likely that the presence of a high yeast population (10⁸ cfu/mL) may restrict the availability of nutrients and sites for colonization, essential for the germination of mould spores [36]. Significant differences were found among assayed yeasts in terms of % inhibitory activity, being H. burtonii and C. zeylanoides the most, and D. hansenii the least effective in inhibiting P. nordicum growth.

D. hansenii is the predominant species in dry-cured ham isolates during processing [15,17], but this predominance becomes more remarkable in fully matured products. The high occurrence of D. hansenii in dry-cured meat products is most probably due to its moderately halophilic properties, accounting for its optimal growth at 3–5% salt [37]. A recent study demonstrated that a higher number of D. hansenii strains isolated from dry-cured meat products were able to growth up to 15% NaCl if compared to C. zeylanoides [16]. However, the lowest inhibitory activity exerted in the present study by D. hansenii against P. nordicum could be predictive of less effectiveness of this yeast as biocontrol agent.

The occurrence of C. zeylanoides was found to decrease from early to late processing steps [16,17]: in Italian dry-cured hams, though isolated up to final maturation, the incidence on muscle surface was lower during maturing phases than at the beginning of processing [15]. So, even if in vitro tested strains of C. zeylanoides highly inhibited P. nordicum, under dry-curing plant conditions (fluctuation of temperature, relative humidity, air circulation, decrease of water availability in surface muscle), C. zeylanoides, having been found to fail in maintaining high populations in matured dry-cured ham, could have less ecological fitness than D. hansenii. Furthermore, C. zeylanoides was included among emerging pathogenic yeasts [18] and considered unsuitable as biocontrol agent.

Hyphopichia burtonii, referred to as “yeast-like fungi”, was isolated from dry-cured hams after surface fat application during maturation. The yeast was probably introduced by the rice flour mixed with the spreadable pork fat (sugna) used at mid maturing to prevent excessive dehydration of ham surface [15]. It was reported as capable of inhibiting the spore germination of P. verrucosum and the growth of P. roquefortii [22,38], and the interest in this species is confirmed by the performance against P. nordicum showed in the present study. Otherwise, the attribution of H. burtonii to dry-cured ham native yeasts is questionable and its potential as biocontrol agent against P. nordicum should be tested in ripening plant conditions, before sugna application.

Debaryomyces maramus and C. famata (the anamorphic form of D. hansenii) were isolated throughout the processing of Parma ham and other traditional Greek and Spanish dry-cured products [15,39,40], even less frequently than D. hansenii and C. zeylanoides. In the case of orange fruits, a C. famata population (10⁸ cfu/mL) gave a successful biocontrol toward Penicillium digitatum [41] in the present study C. famata exerted lower inhibition against P. nordicum than H. burtonii and C. zeylanoides, and an higher inhibition than D. hansenii. Penicillium contamination in cured meat production facilities and products was found to be affected by a large variability [4,42]: accordingly, the biocontrol test was performed using inoculation levels of P. nordicum corresponding to 10², 10⁴, 10⁶ conidia/mL. Yeast inhibitory activity was affected by fungal concentration: at the lowest inoculum level, yeast biocontrol was more effective (P < 0.05), showing that the concentration of the mould is a key factor for yeast effectiveness. As a consequence, in future in vivo applications of antagonistic yeasts for dry-cured ham biocontrol, it will be important to keep the population of contaminating moulds as low as possible, in agreement with processing procedures for high quality dry-cured meat products aimed at this specific target [43]. Among yeast dominant species isolated from dry-cured hams, the tolerance to NaCl has been evidenced [16,19], a property enabling autochthonous yeasts to colonize and prevail in dry-cured ham surface microflora: this halotolerance is in agreement with the positive effect of 3% NaCl presence upon yeast inhibition activities. The same tolerance was recently demonstrated also for P. nordicum, capable of growing and producing OTA in dry-cured pork samples in the presence of 5% NaCl and 0.90 a_w [14]. A significant interaction “yeast × salt” was found and displayed in Figure 2. The inhibitory activity was significantly enhanced by 3% NaCl only for Debaryomyces spp., mostly in the case of D. hansenii.

The tested strains of D. hansenii gave, on average, the least inhibition among tested yeasts, but, thanks to its halophilic aptitude [37], may be that NaCl had a key role in inhibition expression. Some authors found that D. hansenii was able to synthesize toxic proteins with killer capacity against sensitive strains, but the lethal activity of these toxins was demonstrated or increased only in the presence of NaCl [44,45]. Other tested yeasts did not show any difference due to NaCl presence in the medium.

The biocontrol by means of antagonistic yeasts is a potential strategy both against toxigenic moulds and to prevent or minimize mycotoxin production. In this study, the production of OTA was examined when P. nordicum was in pure culture (control plates) or cultured with yeast strains (yeast-added plates). The temperature of the biocontrol test (20 °C) was previously found suitable for OTA high production in dry-cured pork model systems [14]. In the presence of yeasts, the growth inhibition of P. nordicum correlated with the sharp reduction of OTA (Table 3). In the above-mentioned test conditions, the competition for nutrients and space could be the mechanism inhibiting both P. nordicum growth and secondary metabolism, including OTA production [30]. In yeast-added plates, an average tendency of OTA production to increase according to P. nordicum inoculum concentration was found. In control plates, when the P. nordicum inoculum is low (10² conidia/mL), the nutrients not required for fungal growth could be used for OTA biosynthesis. Ochratoxin A frequency distribution into groups of equal frequency (33.3%) L, M and H changed according to yeast presence and species. Treated plates corresponded to low (L) or intermediate (M) OTA category production, while control plates fell mostly in high (H) category. In agreement with inhibitory activity values reported in Table 2, D. hansenii gave OTA value frequency different from H. burtonii, C. zeylanoides and D. maramus, C. famata differed from H. burtonii and C. zeylanoides and D. maramus from H. burtonii. These results suggested that the same mechanism of biocontrol was effective against both P. nordicum growth and OTA production; otherwise, other studies reported that although the growth of spoilage fungi is inhibited by biocontrol of antagonistic yeasts, their metabolic activities are not necessarily reduced [21,36]. The effect of 3% NaCl in control and yeast-added plates was examined with reference to OTA production (Figure 3): in control plates with NaCl, OTA reached higher values than in without-salt counterparts, suggesting a salt-stress related increase in mycotoxin biosynthesis. Otherwise, the presence of NaCl improved yeast biocontrol, enabling an almost total inhibition of OTA production. This finding is in agreement with the “NaCl” positive effect against P. nordicum growth in yeast-added plates (Table 2). Yeasts were also reported for decreasing OTA content by adsorbption on the external and internal part of the cells, sequestering the toxin or even degrading it [46]; the cell wall has a chemical composition that varies from yeast to yeast, regulating the adsorption/non-adsorption activity of yeasts [47]. The assay made to check OTA stability and recovery in control and yeast-added plates (Table 6), shows that OTA was stable in YPD medium with and without NaCl, and OTA recovery was not affected by yeast presence. In this respect, the sharp decrease of OTA in treated plates cannot be ascribed to degradation or adsorption by yeasts. Moreover, even in the case of adsorption, the usage of methanol for extraction, by denaturing yeast parietal protein, allowed OTA to be recovered (75–85% recovery range).

This study demonstrated that selected strains of yeasts isolated from dry-cured ham, exerted in vitro inhibition activity against P. nordicum cultured in the presence of high yeast population (10⁸ cfu/mL). The inhibition was effective against both fungal growth and OTA production, and was significantly influenced by yeast species, level of fungal contamination and NaCl presence in culture medium.

NaCl promoted a significant increase of biocontrol activity, favoring the most halophilic yeast strains such as D. hansenii, but other species exerted a more efficient control on P. nordicum growth and OTA synthesis. The antagonistic activity of tested yeasts was negatively affected by increasing levels of fungal contamination. Assayed yeasts could be effective biocontrol agents for dry-cured meat products, if capable of quickly growing and long lasting. In this respect, future works will be aimed to test the efficacy of different yeasts species in ripening plants, to optimize the inoculum distribution and to achieve suitable conditions for yeast stable and efficient colonization of product surface.