Baranjski kulen and kulenova seka were purchased from the Matijević family farm (Suza, Croatia) and from the Croatian meat industry Belje plus Ltd. (Beli Manastir, Croatia) in 2016. The production process of baranjski kulen (PGI) is precisely specified and very similar to the production process of kulenova seka (9). A brief description of the production of both products with an emphasis on the differences between baranjski kulen and kulenova seka as well as the differences in the process between traditionally and industrially produced sausages is given below.

Raw materials for production of these sausages originated from T2 (pigs with higher final mass) and K (sows obtained by breeding) category pigs. For traditional sausages, the raw material was obtained only from the T2 category, while both T2 and K category pigs were used for the industrial version. At least 80% of the meat part must come from the first category meat (ham and loin), while the use of third category meat is not allowed. According to a specified process, prior to preparation of the filling, the raw material was tempered, meat at between -2 and 4 °C and back fat at between 
-8 and 4 °C. The tempered and pre-minced back fat was then mixed with the meat in a required meat/fat ratio (in kg/kg: 90:10 for baranjski kulen and 80:20 for kulenova seka) with the addition of sweet red pepper, red hot pepper, ground white pepper, garlic and salt. To the best of our knowledge, in traditional sausages fresh garlic and sucrose were added without other additives. On the other hand, in industrial sausages, besides dry garlic and dextrose, starter cultures (Staphylococcus equorum and Lactobacillus curvatus) were added for fermentation control along with antioxidants, nitrites and nitrates. At this point, the mixture was minced to a granulation of 6 or 8 mm and filled into pork intestines. Baranjski kulen is filled in a natural casing, derived from pig’s caecum, while kulenova seka is filled in a natural casing derived from pig’s rectum. Prior to fermentation, conditioning was carried out for 24 h. Sausages were then transferred into the fermentation chamber for at least 7 days. During fermentation, the products were periodically smoked, provided that the smoke temperature did not exceed 22 °C. Fermentation conditions such as temperature (5-24 °C), air humidity (75-92%) and air circulation were monitored only in industrial processing. The ripening lasted for at least three months at a temperature up to 15 °C and air humidity from 60 to 85%. The main differences from traditional products were in the duration and control of fermentation and ripening conditions. Consequently, the total processing time for industrial baranjski kulen was 90 days, industrial kulenova seka 50 days, traditional baranjski kulen 150 days, and for traditional kulenova seka 65 days. For the purpose of preliminary investigation, one of each industrially and traditionally produced baranjski kulen and kulenova seka were analysed.

A representative portion of each sausage was homogenized in distilled water for 5 min (1:4, g/g). The supernatant was decanted, filtered and then centrifuged (5415R; Eppendorf, Hamburg, Germany) at 10 000×g and 4 °C for 15 min. The supernatant was placed in a clean tube and used for soluble protein fraction, peptide and amino acid analyses. One of each industrially and traditionally produced baranjski kulen and kulenova seka were analysed. All samples were prepared in duplicate.

Two-dimensional gel electrophoresis (2-DE) was performed as described by Butorac et al. (10). Isoelectric focusing (IEF) was performed using Flatbed IEF System (Cleaver Scientific, Rugby, UK). The second dimension, sodium dodecyl sulphate–polyacrylamide gel electrophoresis (SDS-PAGE) was run with OmniPAGE System (Cleaver Scientific) through 1-mm thick 12% polyacrylamide gel. The proteins were visualized by Coomassie Brilliant blue staining. 2-DE gels were made in duplicates. All chemicals and materials were purchased from Sigma-Aldrich Co., Merck, Darmstadt, Germany. Densitometry analysis was performed with image analysis software (Discovery Series PDQuest 2-DE analysis software, v. 7.4.0) (11) integrated with a VersaDoc 4 000 Imaging System (Bio-Rad, Hercules, CA, USA). Master gel was created automatically based on the differentially expressed spots identified across all the gels. Differential analysis was obtained by overlapping images of gels based on Gaussian distribution with a cut-off>3.0.

Differentially displayed spots were excised from gels and subjected to tryptic in-gel digestion according to the procedure described by Shevchenko et al. (12). After digestion, the extracted peptides were purified with reversed-phase columns (ZipTip C4; Millipore, Darmstadt, Germany) and 0.5 µL of purified peptides was spotted onto a metal matrix-assisted laser desorption/ionization (MALDI) plate. Sample droplets were dried, and then 0.5 µL of matrix solution was added. The matrix contained 1.4 mg of α-cyano-4-hydroxycinnamic acid (CHCA; Sigma-Aldrich, Merck) dissolved in 1 mL of solvent mixture containing 85% acetonitrile (ACN) and 15% aqueous solution of 0.1% trifluoroacetic acid (TFA).

Mass spectrometry (MS) was acquired with MALDI time-of-flight/time-of-flight (TOF/TOF) (Autoflex Speed; Bruker, Bremen, Germany) in a positive ion reflection mode. The instrument parameters were set using Flex Control v. 3.4 software (Bruker) (13). Mass spectra were obtained by averaging 1000 laser shots covering a mass range m/z=700-4000. After the recording of MS spectra, ten of the most intense precursor signals were selected for tandem mass spectrometry analysis (MS/MS). Protein identification and database search were performed by ProteinScape v. 3.1.0 software (Bruker) (14). Combined MS and MS/MS data were matched against the SwissProt database (15) by MASCOT search engine. The parameters were set as follows: one trypsin miscleavage, oxidation of methionine and 100 ppm precursor mass tolerance. The Gene Ontology (GO) Consortium’s online tool (http://www.geneontology.org/) was used for functional characterization of differentially expressed proteins (16).

The volume of 500 µL of each sample prepared as described in the section Protein, peptide and amino acid isolation from sausages was pipetted in Microcon filter 10 kDa (Millipore, Bedford, MA, USA) and was several times recentrifuged at 5000×g (5415R; Eppendorf) until 200 µL of the filtrate were collected.

The NanoLC system Dionex Ultimate 3 000 RSLCnano (Thermo Fisher Scientific, Waltham, MA, USA) equipped with a UV/Vis detector coupled to Proteineer fcII (Bruker) was used for peptide separation and collection directly onto the MALDI plate. Chromatographic separation was performed on a column Acclaim PepMap 100 C18 3 µm, 100 Å, 75 µm i.d.×15 cm (Thermo Fisher Scientific) at 40 °C. The flow rate was 0.3 µL/min, and the injection volume was 1 µL. Mobile phase A consisted of 0.1% aqueous solution of TFA (V/V) and mobile phase B consisted of 0.1% TFA in ACN. The 75-minute gradient elution was programmed to increase over 70 min with solvent B from 2 to 90% and then to condition the column back to the initial conditions until the completion of the run. The spotter make-up flow was set to 100 µL/h (1.4 mg CHCA matrix dissolved in 1 mL of 50% ACN aqueous solution). The total number of collected fractions was 192.

MS acquisition was performed using MALDI-TOF/TOF (Autoflex Speed; Bruker). Acquisition was performed in positive ion reflection mode. The instrument parameters were set using Flex Control v. 3.4 software (Bruker) (13). Mass spectra were obtained by averaging 1000 laser shots covering a mass range m/z=700-4 000. MS/MS analysis was done with the following parameters: signal-to-noise (S/N) threshold 10, 100 ppm mass tolerance between compounds, merging compounds separated by less than 6 fractions, 5.0 Da as minimal mass distance to co-eluting compounds. Using the SwissProt protein database (15), mammal taxonomy was searched with ProteinScape v. 3.1.0 (Bruker) (14). A search using MS/MS data was matched against the Swiss-Prot database (15) by the MASCOT search engine. The parameters were set as follows: unspecified cleavage, oxidation on methionine, asparagine and glutamine deamination as variable modifications and 100 ppm mass tolerance. Functional characterization (GO) of identified proteins was performed by the above-mentioned online tool (16).

The volume of 500 µL of each sample was prepared as described above for peptide analysis, the difference being in pipetting in Microcon filter 3 kDa (Millipore, Bedford). Then it was recentrifuged several times at 5000×g (Eppendorf) until 200 µL of filtrate were collected. Filtrate was diluted 5 times in water, and 10 µL of diluted sample were derivatized with AccQ-Tag reagent (Waters, Milford, MA, USA) according to the manufacturer’s instructions.

Amino acid analysis was performed as described in Waters AccQ-Tag Chemistry Package instruction manual using a Waters liquid chromatography system with fluorescent detector. Individual amino acids were identified by comparison of their retention times with those of calibration standards.

Volatile extraction by headspace solid phase microextraction (HS-SPME) was carried out according to the modified method of Ma et al. (17). Representative sausage samples were minced in saturated sodium chloride solution (1:5, g/g). The solution was mixed at room temperature for 20 min. An aliquot (10 mL) of homogenized sample was placed in 20-mL flat bottom headspace vial along with the magnet and then capped with a crimp cap fitted with a polytetrafluoroethylene (PTFE)/silicone septum (Supelco, Bellefonte, PA, USA). The headspace vial was heated at 40 °C for 30 min in order to achieve equilibration. SPME carboxen/polydimethylsiloxane (PDMS) fibre (Agilent Technologies, Santa Clara, CA, USA) was preconditioned prior to analysis at 250 °C for 10 min in a GC injection port. Then, the volatile components in the samples were adsorbed onto the fibre for 180 min. One of each industrially and traditionally produced baranjski kulen and kulenova seka were analysed. All samples were prepared in duplicate.

After extraction, the SPME fibre was inserted directly into the injection port of the gas chromatograph GC 7890B equipped with 7693A ALS interface and 240 Ion Trap GC/MS detector (Agilent Technologies). The volatile compounds were immediately thermally desorbed for 2 min at 250 °C in the SPME-specific liner in the injector port of GC-MS. A DB-35MS capillary column (35%-phenyl)-methylpolysiloxane phase, 30 m×0.25 mm×0.25 µm (Agilent Technologies) was used for separation. Helium 6.0 (purity 99.9999%) was used as a carrier gas with a constant flow rate of 1.5 mL/min. Chromatographic conditions were as follows: the oven was held for 2 min at 40 °C, heated to 240 °C at 5 °C/min, and held for 3 min at this temperature. The total run time was 45 min. The mass spectrometer was operated in the electron impact mode with a source temperature of 200 °C, an ionizing voltage of 70 eV, the transfer line temperature of 240 °C, manifold temperature of 40 °C and ion trap temperature of 150 °C. The mass spectrometer scanned masses from 30 to 350 m/z at a scan time of 2 µscan/s and maximum ionization time 65 000 µs. Peak identification was carried out by comparing the obtained mass spectra with those in the NIST 2014/EPA 7/NIH Mass Spectral Database (18). The data were processed by the analysis of variance (one-way ANOVA) and the least significant difference (LSD) post hoc test was used to compare the mean values of concentrations at p<0.05. Statistica v. 13.2 software (19) was used. All four products were compared with each other based on the obtained results.

A total of 435 protein spots were detected on 2-DE gels of industrial baranjski kulen and 450 on 2-DE gels of the traditional sample. A total of 446 protein spots were detected on 2-DE gels of industrial kulenova seka and 450 on 2-DE gels of the traditional sample. In comparative analysis of all 2-DE gels, 51 spots were detected as differentially displayed, of which 46 were successfully identified. Fig. 1 shows an image of the 2-DE master gel. Table 1 presents the complete list of significantly altered proteins when comparing all four products. Based on the data obtained from the MASCOT database, 11 proteins were identified in more than one spot. Generally, a protein shift on 2-DE gel can be associated with posttranslational modification that leads to a change in overall protein charge and/or molecular mass or can be due to limited proteolysis (20).

Peptidomic analysis revealed that traditional kulenova seka contained the lowest number of peptides (in total 43), whereas the highest number of peptides was identified in industrial kulenova seka (in total 107). A total of 72 peptides were identified in traditional baranjski kulen and 96 in industrial baranjski kulen. Table 2 shows the list of proteins from which the identified peptide sequences originated. Most of the peptides in all samples arise from actin.

With respect to the observed proteomic and peptidomic results, according to GO (16), most of the identified proteins and peptides can be grouped into: proteins related to muscle contraction, small molecule metabolic process, cellular response to stress and vesicle-mediated transport.

Quantitative comparison of the concentration of the amino acids between traditional and industrial baranjski kulen and kulenova seka is shown in Fig. 2. From the amino acid profile, it can be observed that significantly lower concentrations of almost all amino acids were measured in traditional kulenova seka than in other products. These results are in correlation with the peptide analysis and could be related to lower protein degradation. However, the low amount of amino acids after fermentation could also indicate their metabolisation by bacteria (2). The predominant amino acids in all samples were arginine and glutamate/glutamine. A similar profile was found in Spanish sausage, where the highest amounts of glutamine, arginine and tyrosine were measured (22). A high concentration of glutamic acid was also detected in the sausages after the addition of starter cultures (2). During drying and ripening of other types of sausages, an increase in the concentration of phenylalanine was observed (29). On the contrary, in traditional baranjski kulen, very low concentrations of phenylalanine were measured compared to other products. This may be due to the drying and ripening processes (29), but also due to amino acid degradation pathway (30).

Relative amounts of volatile aroma compounds detected in the headspace of baranjski kulen and kulenova seka produced by traditional and industrial technologies are reported in Table 3. In total, 43 compounds were (tentatively) identified, including four terpenes, two organosulphur compounds, three miscellaneous compounds, 11 benzenoids, 19 phenols and 4 furanoid and pyranoid compounds.

Only a few terpenes were detected in relatively low amounts (Table 3). Terpenes in dried meat products are found mainly as ingredients of added spices (17). Since the production of baranjski kulen and kulenova seka did not involve spices abundant in monoterpenes, the obtained lower amounts were expected. Similar results were obtained previously for slavonski kulen (31). The identified terpenes possibly originated from black and ground red pepper used in the production, since these spices contain moderate amounts of terpenes (17). The identified terpenes may also originate from the animal feed (32). In this study, no significant differences in terpene amounts were found among the investigated products.

Organosulphur compounds identified in this study (Table 3) most certainly originated from garlic, used as a spice in the production of baranjski kulen and kulenova seka. These compounds are produced after disruption of garlic bulb tissues, which activates the allinase enzyme to produce allicin from alliin. Allicin is then further decomposed to finally yield volatile sulphur compounds with a general formula R₁–(S)_n–R₂. Among them, diallyl disulphide, probably a direct product of allicin degradation (33), is the most abundant volatile compound in garlic (34). Consequently, it is the most representative organosulphur compound in the volatile profile of slavonski kulen, as shown previously by Jerković et al. (31) and confirmed in this study with baranjski kulen and kulenova seka. No significant difference in the amount of diallyl disulphide was determined between industrial and traditional baranjski kulen and kulenova seka, respectively (Table 3). However, another organosulphur compound, methyl ethyl sulphide, was detected only in industrial products. In addition, volatiles derived from spices, especially compounds originating from garlic, can be subjected to oxidative degradation due to increased amounts of unsaturated fatty acids. Also, the differences between samples could originate from the differences in raw material (35). Since these parameters were not controlled in this study, their effects should be investigated in detail in the future.

Interestingly, compounds generated in the lipid degradation processes including autooxidation, mostly represented by short-chain saturated and unsaturated aldehydes, ketones and alcohols that contribute to the aroma of a sausage with rancid or oxidised sensory attributes, were barely detected. Such a result corresponded to that obtained by Jerković et al. (31), who found relatively low percentages of these compounds in the volatile profile of slavonski kulen. The authors related this to the occurrence of high concentrations of phenols and furans, which are derived from the smoking applied during curing (36, 37). They act as antioxidants and therefore limit autoxidation. In this study, a higher level of heptanal was found in both industrial than in both traditional sausages. On the contrary, higher amount of n-hexadecanoic acid was found in traditional baranjski kulen and kulenova seka. It is possible that both compounds derived from lipolysis, with the increased amount of n-hexadecanoic acid in traditional sausage, probably as a direct result of increased triglyceride hydrolysis (38), and the increased amount of heptanal in industrial sausage as a consequence of the greater activity of other enzymes.

Compounds from the amino acid degradation pathway are mainly represented by phenylalanine degradation products (30), such as phenylmethanol, phenylmethanal, phenylacetaldehyde, phenyl cyanide and phenylethanol. Industrial baranjski kulen was the most abundant in phenylacetaldehyde, while industrial kulenova seka contained the highest amount of phenylmethanol. Methoxylated benzenoids identified in this study were also possibly partly derived from the oxidative and/or non-oxidative degradation of l-phenylalanine (39), however, no significant differences were observed among the products regarding their amounts. The occurrence of indane-4-carboxaldehyde was specific for industrial baranjski kulen. A tendency towards higher amounts of benzene-based compounds, such as 3,4-dimethoxytoluene, 1,4-dimethoxy-2-methylbenzene, 1,2,3-trimethoxybenzene, although without statistical significance, was noted in traditional products. These compounds are known to originate from smoke used in the production of smoke-cured meat products (40).

Benzonitrile was found only in industrial products. It is known that this and similar compounds, such as phenylacetonitrile, derive from the reaction of acids and sodium nitrite. As production of industrial baranjski kulen and kulenova seka included the addition of nitrites, preservatives added to inhibit the growth of bacteria (41), benzonitrile identification was not surprising.

Qualitatively, the most abundant group of compounds identified in the industrial and traditional baranjski kulen and kulenova seka investigated in this study were smoke-derived volatile phenols (Table 3). Among them, methoxyphenols, such as 2-methoxyphenol and 4-ethyl-2-methoxyphenol, dominated the profile. This was slightly in disagreement with the results of Jerković et al. (31), who found higher concentrations of phenol and certain methyl-substituted phenols than methoxyphenols in the volatile profile of slavonski kulen. However, the amounts of the former compounds in this study were also high. The highest amounts of phenol, 2-methylphenol, 3-methylphenol, 2-methoxyphenol, 4-methoxy-3-methylphenol and 2-methoxy-4-methylphenol were found in traditional kulenova seka. The 2,4,5-trimethylphenol was detected only in this product. Besides, a tendency towards higher amounts of particular ethyl, vinyl and propyl phenols was observed in traditional products, although without a significant difference (Table 3). It is possible that the traditional process included a higher degree of smoking than the industrial one or that the differences derived from different types of wood used. Another source of volatile phenols is enzymatic degradation by microorganisms (42), so higher amounts in traditional products could be connected to the use of smaller amounts of preservatives in their production, as suggested above. According to other authors (31, 42), smoke-derived phenols exhibit low sensory threshold values that make them important contributors to slavonski kulen flavour with their woody, pungent and smoky odours.

Traditionally produced kulenova seka was characterized by the highest amount of 2,4-dimethylfuran. Benzofuran also exhibited a tendency towards the highest values in this product, while two other compounds from this group, 3-ethenylpyridine and 2-acetyl-5-methylfuran, were detected only in traditional kulenova seka (Table 3). Similarly to phenols, furans are also characteristic of wood smoke derived from cellulose, hemicellulose and lignin pyrolysis, originating mainly from the Maillard reaction (43). During smoking, these compounds are adsorbed by the sausage surface and a part of them interacts with other sausage components. A fraction that remains adsorbed on the surface but does not change chemically by the mentioned interactions is present in its headspace and contributes to the smoked sausage flavour (31). It is possible that a higher degree of smoking in the production of traditional sausage in this study was the main cause of the observed differences.