**4. Discussion**

In the present study, six different molds belonging to *Aspergillus* spp., *Penicillium* spp., and *Fusarium culmorum* were used to artificially contaminate bread produced by fermentation of the dough with a commercial *S. cerevisiae*, or by its co-inoculation with *L. plantarum* UFG 121. The fungal strains were selected because they are representative molds of bread spoilage [30] and they have an ability to produce mycotoxins. In particular, *A. flavus* CECT 20802 produced aflatoxins B1, B2, M1, M2; *F. culmorum* CECT 2148 fusarine C; *P. roqueforti* CECT 20508 was able to synthesize PR toxin; while *P. expansum* CECT 2278 was responsible for the production of patulin and citrinin. In the present study, *L. plantarum* UFG 121, previously characterized for its broad antifungal activity against these mold strains [28], was investigated as a protective culture for bread production.

The employment of LAB strains with antagonistic activity has been widely proposed as an innovative green strategy to fight spoilage filamentous fungi in order to enhance the shelf life of bakery products [24,31–33]. In general, the antifungal effect has been attributed to the production of some secondary metabolites during the fermentation of sourdough [11]. Therefore, according to previous studies [24,34], in this work, a fermentation step of 18 h was performed in order to allow the UFG 121 strain to enrich the sourdoughs with bioactive antifungal compounds. However, differences between the control samples and the bread obtained by co-fermentation with *L. plantarum* UFG 121 were qualitatively detected as the bread was made that might have affected the production process. Nonetheless, as previous reported by Coda et al. [25], co-fermentation of the dough with yeast and selected LAB resulted in bread showing good chemical and textural features, including elasticity, color, and alveolation. Moreover, an improvement in terms of complexity of the aromatic profile was noted in the bread co-fermented with UFG 121. In agreement with this finding, Makhoul et al. [35], analyzing pro-technological microbe/matrix interactions during food fermentation, reported a greater impact of the microbial fraction on the volatile organic compounds of bread. However, the present work is only a preliminary study that should be further complemented by the analytical determination of the main physico-chemical parameters as well as the impact of a protective LAB culture on the organoleptic profile of the bread [36,37].

In our previous study, we found that the preservative potential of *L. plantarum* UFG 121 was mainly due to the production of lactic acid and phenyllactic acid [28]. Phenyllactic acid production by *L. plantarum* has been linked to the antifungal activity against fungal strains isolated from bakery products belonging to species of *Aspergillus*, *Penicillium*, and *Fusarium* [38]. Similarly, an increase in the shelf life of bread obtained by fermentation with *L. plantarum* CRL 778 and artificially contaminated with *Penicillium* spp. has been found to be related to the synthesis of acetic and phenyllactic acid as well as lactic acid [20]. Moreover, organic acids including phenyllactic acid from a strain of *L. amylovorus* have been found to be responsible for a higher shelf life in gluten-free breads [21]. Lactic acid, phenyllactic acid, and two cyclic dipeptides found in sourdoughs fermented by *L. plantarum* FST 1.7 have been identified as the main antifungal compounds able to retard the growth of *F. culmorum* and *Fusarium graminearum* on bread [18].

In a similar way, in the present study, *F. culmorum* CECT 2148 was completely inhibited in bread fermented by *L. plantarum* UFG 121. The in vitro assays performed suggest that CECT 2148 was the most sensitive mold when exposed to UFG 121 cell-free supernatant [28]. Moreover, this result has been confirmed in situ, since fermentation by UFG 121 and artificial contamination with CECT 2148 (after thermal stabilization) increased the shelf life of an oat-based formulation from less than one week to the second week of cold storage, indicating that a strong bioprotection could be provide by antifungal compounds produced in a 16 h fermentation step [28]. In contrast, fermentation with *L. plantarum* UFG 121 had no effect in countering the growth of *P. roqueforti* CECT 20508 and *A. niger* CECT 2805, but reduced bread contamination by the remaining tested molds to different extents. Interestingly, these results were only partially predicted by the in vitro assays [28], suggesting that interactions with the commercial yeast, process parameters, and/or the food matrix might modulate the antagonistic activity of selected LAB against fungal strains. Therefore, further investigations are

required to establish the effectiveness of the antifungal compounds, their synergistic interactions, and the complex microbial and physico-chemical relationships occurring in the food environment.
