Low-Fat and High-Quality Fermented Sausages

The present study, considering for the first time microbiological concerns due to the use of lemon albedo as a fat replacer, aimed at the selection of an anti-Listeria strain to be used as protective culture in low-fat southern Italian fermented sausages. In fact, these kinds of products require appropriate bio-protective strategies to avoid risks due to Listeria monocytogenes. Sixty-seven Lactiplantibacillus plantarum strains isolated from diverse sources were screened for their antimicrobial activity and their interaction with starter strains (Latilactobacillus sakei 152 and Staphylococcus xylosus MVS9). Lactiplantibacillus plantarum Lpls100, highlighting both listericidal activity and the ability to promote Staphylococcus xylosus MVS9 growth, was used as a protective strain in low-fat fermented sausages prepared with lemon albedo as a fat replacer. The effect of the albedo and the protective strain on the fermentation process and the final quality was ascertained. Results highlighted that the use of the albedo did not affect the growth of starter strains and enhanced some quality features, such as fatty acid profiles and certain sensory attributes. However, the albedo also produced a slow decrease in water activity, compromising the microbial quality. The anti-Listeria strain, enhancing coagulase negative cocci growth and exerting antimicrobial activity, avoided the inconveniences caused by the use of the albedo. Moreover, the anti-Listeria effectiveness was assessed through a challenge test using a Listeria cocktail. The study revealed that Lactiplantibacillus plantarum Lpls100, regardless of the presence of the albedo, assures a prompt inhibition of Listeria spp. Therefore, its use could be an important contribution to the quality of low-fat fermented sausages.


Introduction
The consumption of saturated fatty acids (SFAs) and their replacement with unsaturated fatty acids (UFAs) is widely debated [1]. Health organizations all over the world promote the choice of a diet low in saturated fat. The Health Assembly approved, in 2018, the 13th General Programme of Work, which guides the work of the World Health Organization (WHO) in 2019-2023 [2]. Conversely, recent meta-analyses studies found no evidence of a relationship between saturated fat intake and coronary heart disease or cardiovascular disease [3,4]. Despite this evidence, the American Heart Association released a Presidential Advisory on dietary fats with a meta-analysis that provided evidence that re-established their advice to reduce SFAs and replace it with polyunsaturated fatty acids (PUFAs) [5]. Reduction of fat intake and the elimination of industrially-produced trans-fats from the food supply are identified in GPW13 as part of WHO's priority actions to achieve the aims

Lemon Albedo
Lemon albedo was obtained directly from by-products of a local Limoncello (Italian liqueur) manufacturer as reported by Fernandez-Gines et al. [33]. The powder obtained was vacuum packed until use.

Starter Culture
Latilactobacillus sakei 152 (Lt. sakei 152) and Staphylococcus xylosus MVS9 (S. xylosus MVS9), both belonging to the Collection of the Department of Agricultural, Environmental, and Food Sciences (DiAAA) of the University of Molise (Italy), were used in this study as a multi-strain starter culture. Both strains were previously isolated from traditional-fermented sausage and selected for their technological features [12,[34][35][36].

Screening of Anti-Listeria Activity of Lactiplantibacillus plantarum Strains
The inhibitory effect produced by sixty-seven Lb. plantarum strains against L. monocytogenes and four strains of L. innocua was evaluated. Trials were performed in BHI broth (Oxoid). Growing cells of each indicator strain were inoculated in BHI in order to obtain a final concentration of about 5 Log colony-forming units/mL (CFU/mL) and were combined with cell free supernatant (CFS) of each producer strain. Tubes of BHI without cell free supernatant of producer strains and inoculated with Listeria strains as described above, were used as a control. The pH of each tube was adjusted to values of about 5.5 using HCl 1N or NaOH 1N. Tubes were incubated at 30 • C for 30 h because this time corresponds to the advanced stationary phase of L. innocua when cultured under control conditions. At the end of the incubation time, 1 mL of sample was withdrawn from each tube, and Listeria cells were counted by plate counts performed in BHI agar [28].
The antagonism effect was expressed using an arbitrary index calculated as follows: where Y end_exp represents the count levels of Listeria cells after 30 h (t end ) of incubation in combination with Lb. plantarum CFS; Y end_c represents the count levels of Listeria cells after 37 h (t end ) of incubation under control conditions (without combination); and Y 0_c represents the count levels of Listeria cells in the batches at the beginning of the incubation (t 0 ) without combination with Lb. plantarum CFS. The equation takes into account the effect of producers' CFS on the final count levels of L. innocua Y end_exp − Y end_c , and normalizing these values with respect to the behaviour under control conditions (Y end_c − Y 0 c ), listeria inhibitory and listericidic actions were discerned. In fact, different values of the antimicrobial index indicated diverse anti-Listeria effects: Listeria inhibitory action (0 < Anti-Listeria_score ≤ 1) or Listericidic action (Anti-Listeria_score > 1).

Detection of Interactions between Lactobacillus plantarum Strains and Conventional Starter Culture Strains
Sixty-seven Lb. plantarum strains were used as producers, while Lt. sakei 152 and S. xylosus MVS9 were used as indicators. The effect of Lb. plantarum strains on the growth of indicators was assessed in sarcoplasmic protein extract (SPE), prepared as described by Basso et al. [47] and inoculated with growing cells of indicator strains combined with cell-free supernatants of producer strains as reported by Tremonte et al. [35]. In all cases, the growing cell inoculum was adjusted to obtain an OD600 nm value of 0.2. The microbial growth of Lt. sakei 152 and S. xylosus MVS9 after incubation at 28 • C for 30 h was ascertained by plate counts performed in MRS agar and in Mannitol salt agar (MSA), respectively [35,36]. Based on these values, an arbitrary index was estimated and calculated as follows: where Y end_ic and Y end_iexp represents the count levels of indicators cells at the end of incubation alone or in combination with Lb. plantarum CFS, respectively, and Y 0_ic represents the count levels of indicator cells in the batches at the beginning of the incubation (t 0 ) without Lb. plantarum CFS. Values lower than 0 indicate inhibitory action against the indicators, whereas positive values correspond to a stimulation effect of producers on indicator strains. Values close to zero correspond to a neutral interaction. Specifically, considering previous results [35,36], the following breakpoints were defined: - Salami, southern Italian style, was produced in a small factory by mixing minced pork meat (longissimus dorsi and psoas) refrigerated at 0 ± 1 • C. During mixing, NaCl (2.5%), glucose (0.3%), black pepper (0.2%), KNO 3 (0.02%), sterile physiological solution (1% v/w), and the starter culture, containing Lt. sakei 152 and S. xylosus MVS9 (6 Log CFU/g each), were added. The mixture was divided into four batches: C, control batch produced with minced pork fat addition (14%); A, experimental batch produced with lemon albedo (4%) and without pork fat; CPc, experimental batch produced as C and inoculated with Lb. plantarum Lpls100 as a protective culture addition (about 6 Log CFU/g); APc, experimental batch prepared as A but with Lb. plantarum Lpls100 as a protective culture addition (about 6 Log CFU/g). After stuffing, the products were left undisturbed for 12 h at about 2-4 • C and were then were dried for seven days (relative humidity increasing from 60 to 80%; temperature decreasing from 18 to 11 • C). Subsequent ripening stages were carried out in store rooms at 80% relative humidity and 11 ± 1 • C temperature for 42 days.

Inoculation of Salami with Listeria innocua
In order to simulate Listeria contamination during the process, a challenge test on fermented sausage, prepared as reported above, was carried out in parallel. For this purpose, the experimental design proposed by Paratata et al. [48] with some modifications was adopted. In detail, aliquots of each mixture batch described above were used and inoculated (1% v/w) with 3 Log CFU/g of the multi-strain Listeria cocktail, a mixture of four strains of L. innocua: the strain ATCC 33090 and strains Li03, Li12, and Li13. Prior to use, the microbial cultures were revitalized in BHI (Oxoid, Milan, Italy) and sub-cultured two times in SPE at 18 • C. At the exponential phase, the strains were inoculated in the mixture for the challenge test, obtaining four experimental batches: CL, AL, CPcL, and APcL, which were prepared as the corresponding batches C, A, CPc, and APc, but all inoculated with the multi-strain Listeria cocktail.
As reported above, the mixture was stuffed into natural pig casings, and subsequent drying and ripening were carried out. The mixture batches (C, A, CPc, APc) described in the above paragraph were used as control batches. For each treatment (C, A, Cpc, APc, CL, AL, CPcL and APcL), three independent batches were prepared, and two samples from each batch were taken at different time intervals.

Physicochemical and Microbiological Analyses
Analyses were performed on samples from each batch at time 0 and after 3, 7, 14, 21, 28, 35, and 42 days of ripening. The pH was determined in three different points of two samples from each batch, using a pH-meter with a Mettler Toledo spear probe (Novate Milanese, Italy). Water activity was measured in triplicate using a Water Activity Meter CR2 (AQUALAB Instrument, Washington, USA) on two different samples from each batch. Microbiological analyses were performed in duplicate on two samples from each batch as reported in the literature [49,50]. Briefly, about 10 g of samples were decimal diluted in a sterile solution of 0.1% peptone water, homogenised in a Stomacher 400 Lab-blender (Seward Medical, London, UK) for 3 min and serially diluted under the same sterile solution. Counts of lactic acid bacteria (LAB), micrococci and staphylococci coagulase-negative cocci (CNC), enterococci, Brochothrix thermosphacta, Listeria spp., total and faecal coliforms, Pseudomonas spp., and Eumycetes were detected after appropriate incubation in proper media and conditions as described in a previous study [30,46,50,51]. Results were expressed as the mean of measurements reported as Log CFU/g.
To verify the presence of the starter culture during the ripening period, five colonies were randomly picked up from the MRS and MSA plates. The cultures isolated from MRS (120 cultures) and MSA (150 cultures) agar plates were presumptively identified and typed by RAPD-PCR [12,30,52].

Listeria Detection
To detect the behaviour of Listeria in the batches intentionally inoculated with the multi-strain Listeria cocktail, at time 0 and after 3, 7, 14, 21, 28, 35, and 42 days of ripening, three samples were collected from each batch inoculated with Listeria strains (CL; AL; CPcL; and APcL) and from control batches (C, A, CPc, and APc). Listeria spp. were enumerated on Oxford agar (Oxoid) prepared with selective supplements (SR140, Oxoid) after incubation at 35 ± 1 • C for 48 ± 2 h. Analyses were performed in duplicate, and the results were expressed as Log CFU/mL.

Fatty-Acid Determination of the Neutral Lipid Fraction
The fatty-acid (FA) compositions were determined by high-resolution gas chromatography, analysing the fatty acid methyl esters obtained by trans-esterification of about 25 mg of each fat sample, which were previously dissolved in 2 mL petroleum ether, and using 2 mL BF 3 -methanol reagent [53]. A Fisons model MFC800 gas chromatograph (Fisons, Milan, Italy) equipped with a 60 m × 0.32 mm i.d. and 0.5 µm film thickness Stabilwax fused silica capillary column (Restek, Bellefonte, PA, USA) was used. The following conditions were applied: (i) oven: 5 min at 170 • C, followed by heating (1 • C/min) up to 220 • C, where the temperature was held for 30 min; (ii) carrier: helium 20 cm/s at 170 • C; (iii) injector: 250 • C, 1 µL; split 40:1; (iv) detector: FID, 250 • C. The FFA composition was determined according to Tremonte et al. [12].

Sensory Analysis
The sensory analysis was performed on salami samples from batches without Listeria inoculum (from batches C, A, CPc, and APc) at 42 days of ripening. To define the sensory profile, a descriptive panel of 50 judges, habitual consumers of dry-fermented meat products, was engaged. The judges were trained in preliminary sessions to produce a list of attributes useful to define the sensory profile of low-fat fermented sausage. On the basis of the frequency of citations (>60%), nine descriptors were selected: juiciness, softness, greasiness, rancidity, smell, faecal smell, bitter taste, sour taste, and salty taste. A multiple comparison test was used to determine whether there was a significant difference among the samples from the different batches. A nine-point scale anchored by 1 (extremely low) and 9 (extremely high) was used. Salami samples from each batch were sliced into 2 mm thick slices and equilibrated for 1 h at room temperature according to ISO 8589-1 [54]. Water and unsalted crackers were provided between samples as palate cleansers and to remove any residual flavours.

Statistical Analysis
Statistical analyses were performed following the approach used by Tremonte et al. [40] and by Sorrentino et al. [30]. Similarity in the Anti-Listeria profiles was calculated using multivariate analysis. Results obtained from microbiological counts, pH, and a w were analysed by a General Linear Model based on ANOVA considering the effect of batches and ripening time. In all cases, statistical significance was attributed to p values of < 0.05. Statistical data were expressed as the mean ± standard error or standard deviation. IBM SPSS Statistics 21 and RStudio (v3.5.0) were used for data analyses and graphical representation.

Anti-Listeria Lactiplantibacillus plantarum Strain Selection
The anti-Listeria ability expressed by Lb. plantarum strains from diverse isolation sources was investigated. In a previous study, a relationship between the isolation sources characterized by harsh environmental conditions (antimicrobial substances and low pH) and the presence of lactic acid bacteria able to produce a strong antimicrobial activity was highlighted [39,55]. To date, more extensive literature is available on protective or anti-Listeria microbial cultures and their advantages and concerns in meat products [27,56]. Selection criteria of protective and starter cultures were also elucidated [57]. However, too little attention has been paid to isolation sources, and the protective strains were isolated and selected only from meat products [58][59][60]. In the present study, Lb. plantarum strains from fermented sausages as well as from other and harsher environments, such as wine, sourdough, and honey, were evaluated. Figure 1 shows the heat-map regarding the anti-Listeria action expressed by the CFS of sixty-seven Lb. plantarum strains against L. monocytogens ATCC 15313 and versus the four strains of L. innocua. The results are reported as Anti-Listeria_score (calculated as reported in Section 2.3.3) assuming values from −1 to +2. Negative values correspond to stimulatory activity of Listeria strains, while positive values indicate an inhibition or listericidal action.
The results showed that a response substantially similar was exhibited by the five indicator strains. This finding confirms that L. innocua could be considered a surrogate of the pathogen L. monocytogenes [61]. On the other hand, significant differences were detected among the producer strains, confirming, as also reported in literature [62], that the antimicrobial activity is a character that is strain-dependent. In particular, the producer strains were grouped into four main clusters (CL1-CL4, Figure 1) depending on their actions against the indicator strains. In detail, sixteen strains, grouped in CL4, did not produce an anti-Listeria effect, and eleven of them stimulated Listeria growth. It is noteworthy that all strains grouped in CL4 were isolated from fermented sausages. The other assayed strains, all isolated from sources (fermented olives, sourdough, wine, and honeybee guts) characterized by harsher features, exhibited a positive Anti-Listeria_score. However, differences in the antagonistic effects were recognized among the producers. In detail, thirty-seven strains (CL3) produced a low-moderate inhibition action, and seven strains (CL2) exhibited a strong inhibitory action. In addition, seven strains (CL1) were distinguished from all others for their listericidal action. In fact, these strains highlighted Anti-Listeria_score values higher than 1, indicating a reduction in the Listeria count over the incubation time. These last strains were isolated from wine, honeybee guts, and sourdough, all matrices characterized by low pH and other environmental stressors [39,52]. The results confirmed that certain environments with harsh features harbor and select specific strains able to produce antimicrobial activity [39].

266
The results showed that a response substantially similar was exhibited by the five indicator 267 strains. This finding confirms that L. innocua could be considered a surrogate of the pathogen L.
268 monocytogenes [61]. On the other hand, significant differences were detected among the producer 269 strains, confirming, as also reported in literature [62], that the antimicrobial activity is a character that 270 is strain-dependent. In particular, the producer strains were grouped into four main clusters (CL1-271 CL4, Figure 1) depending on their actions against the indicator strains. In detail, sixteen strains, 272 grouped in CL4, did not produce an anti-Listeria effect, and eleven of them stimulated Listeria growth.

273
It is noteworthy that all strains grouped in CL4 were isolated from fermented sausages. The other 274 assayed strains, all isolated from sources (fermented olives, sourdough, wine, and honeybee guts)

279
In fact, these strains highlighted Anti-Listeria_score values higher than 1, indicating a reduction in 280 the Listeria count over the incubation time. These last strains were isolated from wine, honeybee guts,

281
and sourdough, all matrices characterized by low pH and other environmental stressors [39,52]. The 282 results confirmed that certain environments with harsh features harbor and select specific strains able 283 to produce antimicrobial activity [39].  Cluster analysis shown as a heat-map with boxplot annotation concerning the anti-Listeria activity (Anti-Listeria_score) produced by sixty-seven Lactiplantibacillus plantarum strains isolated from different sources (origin). The anti-Listeria activity was assayed against Listeria monocytogenes (Lm_ATCC_15313) and four strains of Listeria innocua (Li_ATCC_33090, Li03, Li12, Li13).

Culture Starter Compatibility
To evaluate the effect of Lb. plantarum strains on the growth starter, growing cells of starter strains (Lt. sakei 152 and S. xylosus MVS9) were combined with cell-free supernatants of the assayed Lb. plantarum strains. By analyzing the values of the "compatibility index", different situations emerged. In detail, the combination of Lt. sakei 152 with the CFS from each Lb. plantarum strain showed substantial neutralism, shown by very similar count values between the control (when Lt. sakei was cultured singly) and experimental batch (when Lt. sakei was cultured in combination). In fact, for all combinations, the values of the "compatibility index", labelled "sakei compatibility", were approximately zero (data not shown). Conversely, between S. xylosus MVS9 growing cells and Lb. plantarum CFSs, several interactions were detected. In fact, as reported in Figure 2, the compatibility index, labelled "CNC compatibility", showed values ranging about from −0.60 to +0. 50. showed substantial neutralism, shown by very similar count values between the control (when Lt. 290 sakei was cultured singly) and experimental batch (when Lt. sakei was cultured in combination). In 291 fact, for all combinations, the values of the "compatibility index", labelled "sakei compatibility", were 292 approximately zero (data not shown). Conversely, between S. xylosus MVS9 growing cells and Lb.

293
plantarum CFSs, several interactions were detected. In fact, as reported in Figure 2, the compatibility 294 index, labelled "CNC compatibility", showed values ranging about from -0.60 to + 0.50.

307
Considering the extensive literature on the CNC role in fermented sausage and particularly on the 308 role of these bacteria in lipolytic, proteolytic, and aromatic behavior [63,64], approaches to promote 309 the growth of these strains should be strongly recommended.

311
In order to select the most appropriate protective strain for low-fat and high-quality fermented 312 sausage, the interactions with strains of conventional starter cultures were assessed. This aspect, in 313 the last decade, has been adopted in the selection of fermentative starter cultures [35,36]; however, 314 there is still little consideration of the selection of protective cultures. To date, only Corbo et al. [65], 315 in the selection of antagonistic cultures for sausages, evaluated the interactions with starter strains 316 highlighting substantial neutralism. In our study, several positive or negative interactions between and from the interactions with S. xylosus strains were combined and analysed together. In Figure 3, A moderate interaction was found for 14 combinations between S. xylosus MVS9 and the CFS of 14 Lb. plantarum strains. Twenty-seven combinations highlighted negative values which indicate the inhibitory action of the related Lb. plantarum strains against the indicator. Conversely, twenty-six combinations showed positive values of the CNC compatibility index, denoting a stimulatory action of Lb. plantarum strains on the growth of S. xylosus MVS9. All these findings emphasized that the choice of a specific Lb. plantarum strain, even as a protective culture, could significantly affect the growth of the starter strain. In addition, several previous studies highlighted a strong relationship between the interaction and the expression of important technological activities [12,35,40]. Considering the extensive literature on the CNC role in fermented sausage and particularly on the role of these bacteria in lipolytic, proteolytic, and aromatic behavior [63,64], approaches to promote the growth of these strains should be strongly recommended.

Protective Strain Choice
In order to select the most appropriate protective strain for low-fat and high-quality fermented sausage, the interactions with strains of conventional starter cultures were assessed. This aspect, in the last decade, has been adopted in the selection of fermentative starter cultures [35,36]; however, there is still little consideration of the selection of protective cultures. To date, only Corbo et al. [65], in the selection of antagonistic cultures for sausages, evaluated the interactions with starter strains highlighting substantial neutralism. In our study, several positive or negative interactions between Lb. plantarum strains and S. xylosus strains were detected. Therefore, data from anti-Listeria activity and from the interactions with S. xylosus strains were combined and analysed together. In Figure 3, the scatterplot highlights the relationship between "CNC-compatibility" and the "Anti-Listeria_score" for each Lb. plantarum strain. the scatterplot highlights the relationship between "CNC-compatibility" and the "Anti-    On the basis of the results reported in Figure 3, the number of strains useful as protective cultures was strongly reduced. In detail, about half of the strains exhibiting anti-Listeria activity also had an inhibitory effect on the growth of S. xylosus MVS9. Moreover, only three strains among those with listericidal activities were compatible with the starter strain. In fact, the strains TB38, Lpla8, and Lpls100, isolated from wine, honey guts, and sourdough, respectively, were collocated in the comfortable zone. Lb. plantarum Lpls100 was chosen as the highest performing strain and was subsequently used as a protective culture.

Quality Features of Low-Fat Fermented Sausages
In order to evaluate the effects produced by lemon albedo, used as a fat replacement, and by the potential protective effect of Lb. plantarum Lpls100, conventional and experimental batches of fermented sausages were prepared and monitored during ripening in order to investigate the principal quality features. Moreover, the presence of the bacterial strains used as starter strains or protective strains was ascertained. For this purpose, typing was performed by RAPD-PCR on both lactobacilli and CNC isolated from all batches (C, A, CPc, APc, CL, AL, CPcL, and APcL) at 0, 21, and 42 days of ripening. The results highlighted the effectiveness of the inoculation and sustained ability of the strains used as starters to survive during the ripening of salami. In fact, more than 90% of presumptive CNC isolates from the four batches had RAPD-PCR profiles with a coefficient of similarity higher than 85% with S. xylosus MVS9. Similar results were obtained for lactobacilli isolated from MRS plates; in fact, about 90% of isolates, phenotypically identified as Lt. sakei and Lb. plantarum, showed a coefficient of similarity higher than 85% with the strains Lt. sakei 152 and Lb. plantarum Lpls100, used as starters and protective culture (data not shown).

Physicochemical Features of Conventional and Low-Fat Sausages
The behaviours of pH and a w of each batch, during ripening, are shown in Figure 4A,B respectively. The pH values resulted in agreement with those commonly found in other fermented sausages from southern Italy [40,66,67]. However, ripening time and batches affected pH values (p < 0.05). In all batches, a significant decrease in pH values was detected during the ripening time, falling from 5.80 ± 0.09 to values ranging between 5.03 ± 0.02 and 5.12 ± 0.02 after 7 days of ripening. This behaviour was consistent with those revealed in Italian sausage prepared with starter cultures [68][69][70]. This last finding suggests the effectiveness of the starter. As regard the different batches, during the entire period time, in the samples from batch A were detected pH values lower than those observed in the other batches. Also, other authors found that the use of citrus fibre at 4% produced a decrease in the pH values [16,71]. Moreover, in the samples from batch A, the pH values decrease until the 7th day, a behaviour substantially constant were detecting. A similar trend, also with values slightly higher, characterized also the samples from batch C.   Conversely, the samples from batches CPc and APc highlighted a significant increase from the 14th day of ripening. Moreover, in these samples, during the entire ripening period, similar values were detected. Therefore, the presence of the anti-Listeria strain would seem to remove the effect, even if slight, due to albedo. However, this aspect deserves further attention, also considering the behaviour in microbial populations. In detail, the protective strain Lb. plantarum Lpls100, due to its ability to promote the growth of the proteolytic strain S. xylosus MVS9, could produce an increase in pH values affecting the sensory features of the product.
The a w ( Figure 4B) decreased from initial values of about 0.97 to about 0.89 at the end of ripening. However, regardless of the presence of Lb. plantarum Lpls100, significant differences emerged among the batches with or without albedo. In fact, the a w recorded in A and APc from the 14th day of ripening was about 0.02 units higher than that recorded in relative batches without albedo (batches C and CPc). These findings are in agreement with the literature [16] and could be attributable to the properties of lemon albedo that, due to its high pectin content (25%), produces a water content increase in the product [20,72]. All these properties, in our opinion, could also affect other quality features, such as microbiological and sensory properties. In fact, the lower decrease in a w values, as also evidenced in a recent study [48], could affect the microbial behaviour, compromising the more prompt inhibition of undesirable microbial groups.

Microbiological Features of Conventional and Low-Fat sausages
The behaviour of the most relevant microorganisms is reported in Figure 5. Except for yeast and moulds which showed a substantially constant trend during ripening (data not shown), for all microbial groups, significant differences (p < 0.05) in microbial load were found depending on the different sampling time and the diverse batches. During the ripening period, a significant increase in LAB and CNC count levels was observed, whereas a decrease in levels of undesirable microbial populations was detected. This is consistent with literature highlighting, once again, that useful lactic acid bacteria and CNC effectively contribute to the ripening process [73][74][75]. Moreover, data also showed that diverse spoilage bacteria, such as Brochothrix thermosphacta, Pseudomonas spp., and Coliforms persisted and were detectable until the advanced stage of ripening. These microorganisms are widely described as contaminants of fresh meat as well as the most common spoilage microorganisms of fermented meat products [76][77][78][79]. features, such as microbiological and sensory properties. In fact, the lower decrease in aw values, as 382 also evidenced in a recent study [48], could affect the microbial behaviour, compromising the more 383 prompt inhibition of undesirable microbial groups.  However, several differences were found among the different batches. For LAB counts, similar count levels were detected in all batches. This finding indicates that the use of the albedo did not affect the growth of useful strain starters. Interesting differences among the batches were noticed in CNC behaviour. Specifically, even if the same trend was detected in all batches, the samples prepared with protective culture (APc and CPc) addition showed count levels higher than those observed in samples from batches A and C. This difference could be related to the ability of selected Lb. plantarum Lpls100 to promote the growth of S. xylosus MVS9. Moreover, the higher levels of CNC would also justify the pH increase during the second step of ripening, which was higher in APc and CPc than in batches A and C. In fact, as already reported in a previous study, CNC plays a significant role in proteolytic activity and consequently an increase in pH values [35,36]. Particular attention should be paid to the differences among the batches concerning the undesirable microorganisms. A greater persistence of certain Enterobacteriaceae (total and faecal coliforms), of Pseudomonas spp. as well as of Brochothrix thermosphacta was found in the samples from batch A. In particular, B. thermosphacta was detected in batch A until the end of ripening. This bacterium belonging to the Listeriaceae family, is phylogenetically related to L. monocytogenes, and is considered a strict meat spoiler [80]. A similar situation was also detected for Listeria spp. (data not shown) which, even if found since time 0 of ripening at negligible values, persisted in batch A until the end of ripening. The greater persistence of undesirable bacteria in sample from batch A could be related to their a w values being higher than those detected in the samples from batches C and CPc. However, for the samples from batch APc, even if characterized by a w values similar to those detected in batch A, the prompt decrease in undesirable bacteria was assured by the presence of protective Lb. plantarum Lpls100. In detail, the effect of Lpls100 on Listeria spp. was elucidated in Section 3.3.

Final Fatty Acid Composition and Sensory Features
Fatty acids (FAs) represent an important component in fermented sausages, affecting their sensory and nutritional features [40,81]. The percentages of unsaturated fractions and of saturated fatty acids are reported in Figure 6A. Results showed that unsaturated fatty acids (MUFA + PUFA) were higher than saturated fatty acids (SFA) in all batches. This finding is consistent with other studies reported in the literature [82,83]. However, as highlighted in Figure 6A, differences among the batches were noticed. In particular, considering values of PUFA and SFA levels, significant differences among the batches could be revealed. For this purpose, the PUFA/SFA ratio was considered ( Figure 6B). This ratio has been adopted as an index to describe the quality of fermented sausages, and the nutritional recommendations highlighted that the PFA/SFA ratio should be higher than 0.45 [84]. Interesting results were shown by the samples from all experimental batches. In detail, ratio values close to the threshold of 0.45 were exhibited by the samples from batch C. All others samples highlighted PUFA/SFA ratios significantly higher than 0.45. Specifically, a higher ratio value (0.53) was shown by the samples from batch APc. Therefore, on the basis of this last evidence, a role of both albedo and protective strain Lpls100 could be assumed.
In addition to the nutritional aspects, the fat replacement could affect the sensory features since fats have an important role in appearance, softness, flavour, smell, juiciness, and rancidity. As reported in the literature, the addition of fat substitutes in fermented sausages influences the final texture [85]. However, to date, the effect of lemon albedo is still debated [20]. Some authors showed that the addition of lemon albedo to fermented sausages did not affect sensory properties such as smell, or salty and acidic taste, and increased juiciness [19], while others assured that the addition of citrus fibre in fermented sausages decreased the smell, consistency, and juiciness [33]. On these bases, the effect of lemon albedo and of protective strain Lpls100 on sensory attributes was also ascertained by a panel of trained judges [54]. The mean values of sensory attributes are reported in Figure 7. The results showed that fat replacement and protective strain Lpls100 did not negatively affect the sensory features and they positively influenced some attributes. In detail, no significant differences among the batches with albedo lemon (A, APc) or without (C and CPc) were found in juiciness, salty taste, and rancidity. This last finding is interesting considering that so far the rancidity and the lipid oxidation represent the main critical issues in the preparation of fermented sausage. The samples from batch A showed lower softness, faecal smell, bitter and sour taste than those from the other batches. These attributes and in particular the faecal smell could be due to the persistence of several spoilage microorganisms in samples prepared with albedo addition (batch A). However, the combination of the albedo with protective strains (APc) enhanced the sensory quality, preventing the low softness, bitter, and sour taste perception. This finding could be due to the antimicrobial effect of protective stain Lpls100 as well as to its ability to promote CNC growth that, as reported in previous studies, could positively contribute to the texture and sensory features [12,35]. texture [85]. However, to date, the effect of lemon albedo is still debated [20]. Some authors showed 454 that the addition of lemon albedo to fermented sausages did not affect sensory properties such as 455 smell, or salty and acidic taste, and increased juiciness [19], while others assured that the addition of 456 citrus fibre in fermented sausages decreased the smell, consistency, and juiciness [33]. On these bases,

Anti-Listeria Action in Low-Fat Fermented Sausages
Based on the results reported in the previous paragraph, it clearly appeared that the main concern of low-fat fermented sausages is the microbiological quality. In this context, the evaluation of the effect expressed by the albedo and the protective strain Lb. plantarum Lpls100 on the behaviour of a multi-strain Listeria cocktail, intentionally added, could offer useful knowledge. For this purpose, four experimental batches (CL, AL, CPcL, and APcL) were prepared as the corresponding batches C, A, CPc, and APc and inoculated with the multi-strain Listeria cocktail.
protective stain Lpls100 as well as to its ability to promote CNC growth that, as reported in previous 470 studies, could positively contribute to the texture and sensory features [12,35].     During the ripening time, the samples from the batches inoculated with the Listeria cocktail showed values of pH and a w closely comparable to those detected in the samples from the relative control batches (C, A, CPc, and APc)-data not shown. Specifically, the samples from the batches AL and APcL highlighted a w values about 0.02 units higher than those detected in the samples from batches CL and CPcL. Moreover, for the pH values, differences among the batches were detected during the second step of ripening. In detail, in the samples from batches CPcL and APcL, higher pH values were detected than those registered in samples from CL and AL. Moreover, the effective presence of the strains used as starter or protective culture was determined by RAPD-PCR and as reported above, their occurrences were confirmed at 0, 21, and 42 days of ripening. Therefore, the starter and protective strains represented the main dominant microbial population during the fermentative process. The effectiveness on the protective strain Lpls100 was clearly shown by Figure 8, which reports the behaviours of Listeria spp. in the different batches. Specifically, the behaviours of Listeria spp. in the four batches showed two different trends.
physicochemical parameters, promote the growth of Listeria spp. In detail, based on our results and 519 considering the literature, the aw parameter seems to be crucial in the behaviour of Listeria spp. [86].

520
These assertions are consistent with other evidence in the literature reporting that at fermentation 521 temperatures below or equal to 20 °C, aw reduction became more important than pH decreases and In the batches without protective strain addition, an increase from 3.2 Log CFU/g to 4.9 Log CFU/g (batch CL) or 6.6 Log CFU/g (batch AL) was detected. Conversely, for the batches prepared with protective culture addition, a decrease was highlighted. This last evidence clearly highlights the protective and anti-Listeria effectiveness of Lb. plantarum Lpls100. However, to better understand the role of the albedo, the significant differences between batches AL and CL in the increasing behaviour deserve to be considered. Specifically, in these batches, the Listeria count levels appeared substantially similar until the 18th day of ripening when significant differences were observed, and, in samples from batch AL, count levels almost 2 Log CFU/g higher than those revealed in the samples from batch CL were observed. Therefore, the albedo and the related environmental features, such as physicochemical parameters, promote the growth of Listeria spp. In detail, based on our results and considering the literature, the a w parameter seems to be crucial in the behaviour of Listeria spp. [86]. These assertions are consistent with other evidence in the literature reporting that at fermentation temperatures below or equal to 20 • C, a w reduction became more important than pH decreases and if this reduction is slow, Listeria may survive longer [87]. Moreover, other authors showed that although the pH decrease was rapid during the first 48 h of fermentation, the slow dehydration of the sausages and the small change of a w did not contribute to rapid inactivation of the pathogens.

Conclusions
Lemon albedo could represent a suitable substitute for total replacement of pork fat in the preparation of fermented sausages. However, for its use to be feasible, careful consideration is needed. In fact, on the one hand, the albedo did not affect the behaviour of useful fermentative bacteria, such as lactic acid bacteria and CNC, and positively influenced nutritional features and some sensory attributes such as juiciness, greasiness, and salty taste. On the other hand, because of its ability to retain water, the albedo produced a slow decrease in a w values. Actually, the a w decrease seems to be crucial for the behaviour of Listeria that harbours the pathogen species L. monocytogenes, as well as for the behaviour of spoilage bacteria, such as B. thermosphacta. Considering the significant Listeria risk for fermented sausage, the use of additional and bio-protective barriers in the production of low-fat fermented sausages is essential. Moreover, an appropriate selection of protective culture is fundamental. In fact, the antagonistic action against the undesirable bacterial target as well as the synergic or compatible relationship with starter strains must be considered. Lb plantarum Lpls100, highlighting high compatibility with starter strains and a strong anti-Listeria action, could be effectively used to assure the high microbiological quality of low-fat fermented sausages.