M. Longissimus dorsi from beef carcasses were used in the production. M. Longissmus dorsi muscles (right and left parts) were cut across the centre into two pieces. In total, four pieces were obtained for production of pastırma. Production was carried out twice. One carcass was used for each experiment.

First, fat and connective tissue from the surface of meat were removed. Then, deep incisions were made on meat strips. After this, production stages were performed as given in Table 1. Four groups of pastırma were produced by using the curing mixtures: 100 g NaCl, 0.3 g sucrose, and 150 ppm KNO₃ or 300 ppm KNO₃ or 450 ppm KNO₃ or 600 ppm KNO₃ for 1 kg meat.

Ten grams of sample were homogenized with 100 mL distilled water (1:10 w/v), and pH value was measured using a pH meter. Water activity (a_w) was determined by using a TH-500 a_w Sprint (Novasina, Pfäffikon, Switzerland).

In residual nitrite analysis of pastırma samples, the method given by Tauchmann [11] was used. Residual nitrite amount of samples was determined as NaNO₂ ppm units based on coefficient and absorbance values which were calculated using sample weight, dilution factor and standard curve.

The TBARS values of samples were determined according to the method of Lemon [12] and were given as μmol malondialdehyde/kg. Non-protein nitrogenous matter (NPN-M) content was determined according to Anonymous [13]. The results were expressed as g/100 g of samples.

Cross-sectional color values (L*, a*, and b*) of samples were measured using a colorimeter (CR-200, Minolta Co, Osaka, Japan). Color value measurements were carried out according to the criteria defined by Commission Internationale de I’E Clairage based on three-dimensional color measurement. Analysis was performed on sliced samples as three replicates.

The extraction of volatile compounds was done using a CAR/PDMS (Supelco 75 μm, Bellefonte, PA, USA) fibre according to Kaban [8]. After the extraction, the compounds adsorbed by fibre were desorbed from the injection port of the gas chromatography (Agilent Technologies 6890N, Santa Clara, CA, USA) for 6 min at 250°C, and the compounds were determined by a mass selective detector (Agilent 5973, USA). Volatile compounds were separated in a DB-624 (Agilent J&W Scientific, Santa Clara, CA, USA; 60 m, 0.25 mm i.d., 1.4 μm film) capillary column. Helium was used as carrier gas. The temperature programme was started when the fibre was inserted and held at 40°C for 6min and subsequently programmed from 40°C to 110°C at 3°C/min and at a rate of 4°C/min to 150°C, then, at a rate of 10°C/min to 210°C where it was held for another 12 min. The gas chromatography/mass spectrometry (GC-MS) interface was maintained at 280°C. Mass spectra were obtained by electron impact at 70 eV, and data were acquired across the range 30 to 400 amu. The results were evaluated by comparing with mass spectra from a database developed by the National Institute of Standards and Technology (NIST) and WILEY, or standard molecules (for calculating kovats indices, Supelco 44585-U, Bellefonte, PA, USA) and by matching their retention indices with those in the literature. Each sample was analysed as three replicates.

In the study, nitrate level (150, 300, 450, and 600 ppm KNO₃) was taken as factor. The experiments were carried out according to randomized complete block design as two replicates. The data was tested by variance analysis and differences between means were evaluated by Duncan’s multiple range tests using SPSS 20 statistics software (Armonk, NY, USA, 2011). The relationship between the various nitrate levels and volatile compounds detected with solid phase microextraction (SPME) GC/MS analysis was evaluated by principal component analysis (PCA) using Unscrambler v.10.01 (Camo Process AS., Oslo, Norway).

Nitrate level had significant effect (p<0.05) on pH value of pastırma samples. While the lowest mean pH value was determined in the group containing 150 ppm nitrate, pH value increased with increasing of nitrate ratio, and the highest mean pH value was found in the group with added 600 ppm nitrate (p<0.05) (Table 2). The pH value was found above 5.5 in all of groups. This finding was in agreement with the results of other researchers [8,10]. The highest pH value of pastırma should be 6.0 according to the Turkish Food Codex Communiqué on Meat and Meat Products [14].

Nitrate level had no statistically significant effect on a_w value (p>0.05). Water activity values were determined below 0.90 in all groups, and the a_w value ranged from 0.841 to 0.863 (Table 2). In pastırma, water activity values are affected by the salt concentration and drying processes. Therefore, pastırma belongs to the group of intermediate moisture food [15,16]. Water activity shows a decrease depending on the progress of production time [8]. In pastırma, water activity is the most important hurdle factor for microbiological stability, it is stated that it should be below 0.90 [8,15].

Nitrate and nitrite are important additives in pastırma production in terms of color formation, ensuring of curing flavor, prevention of rancidity and antimicrobial activity. However, reduction of nitrate to nitrite is necessary in the processes using nitrate for formation of these effects [1]. In the reduction of nitrate to nitrite, microorganisms having nitrate reductase activity play an important role. On the other hand, the residual nitrate and nitrite must be under certain values in the final product because of negative effects on human health.

Nitrate level had no statistically significant effect on residual nitrite amounts of pastırma samples produced with different nitrate levels (p>0.05), and amount of residual nitrite varied from 3.35 to 15.05 ppm (Table 2). According to this result, residual nitrite amount was less than 50 ppm given as the limit value [17], even if 600 ppm nitrate was used. In another study about cured dried whole processed meat products, residual amount of nitrite was determined as 16.8 ppm [18].

Nitrate level had a significant effect on TBARS value of pastırma (p<0.01). The highest TBARS value was determined in the group produced using 450 ppm KNO₃ (p<0.05). Statistically, there were not any differences between 300 ppm and 600 ppm potasium nitrate levels (p>0.05) (Table 2). The lowest TBARS value was found in the group containing 150 ppm KNO₃ and this value was statistically different from mean TBARS values of other groups (p<0.05). Thiobarbituric acid reactive substances, secondary products of lipid oxidation are used as indicators of lipid oxidation. Kaban [8] reported that lipid oxidation continues during the production of pastırma and as a result of that, TBARS value increases.

In dry-cured meat products, proteolysis is one of the most important reactions. Amounts of non-protein nitrogenous substances such as peptides and free amino acids increase as a result of proteolysis. It was determined that non-protein nitrogenous matter content as an indicator of proteolysis increase during pastırma production process [8]. In the present study, non-protein nitrogenous matter content varied from 4.32 to 4.69 g/100 g in samples (Table 2). However, there were no statistical differences among groups in terms of non-protein nitrogenous matter content (p<0.05).

The factors such as the use of curing agent (nitrate, nitrite, and nitrate/nitrite), pH value and myoglobin content of raw material and manufacturing process are effective in color formation of cured meat products [1,19]. Nitrate is the most important curing agent used in the production of pastırma [8,10]. This curing agent, which is generally used as potassium nitrate, can usually be added to salt in various amounts before the production. Nitrate level had no significant effect on L*, a*, and b* values of pastırma samples (p>0.05) (Table 2). According to these result, use of 150 ppm nitrate in pastırma production was sufficient for color formation.

Many volatile compounds belonging to various chemical groups occur as a result of many reactions during pastırma production [1,8,16]. A total of 46 compounds consisting of 11 different chemical such as aliphatic hydrocarbons, sulfur compounds, aromatic hydrocarbons, aldehydes, terpenes, ketones, esters, furans, acids, alcohols, and nitrogen compounds were identified in pastırma samples produced using different nitrate levels (Table 3).

From 11 aldehydes identified, nitrate level had a significant effect on the 2-pentanal, hexanal, heptanal, octanal, and propanal- 2-methyl-3-phenyl at levels of p<0.05 or p<0.01. 2-Methyl-3- phenyl propanal from statistically significant aldehydes showed the highest value at 150 ppm level (Table 3). Aldehydes have an important share in volatile profile of pastırma [8,16]. In a market research conducted by Kaban and Kaya [16], it was reported that total peak area of aldehydes ranged from 17.21 to 43.80, and hexenal was dominant compound.

3-Methyl thiopene, 10 sulphur compounds identified from pastırma samples, was only found significantly at p<0.05 level (Table 3). Sulphur compounds have been found in the other studies conducted on pastırma [8,16]. In pastırma, sulphur compounds were concluded to be resulting from çemen used in final stage [8]. Garlic also found in the composition of çemen has been shown as source of many sulphur compounds by Ramirez and Cava [20].

Three ketones as 1,3-pentadiene, 6-methyl-5-hepten-2-one, and 3,5-octadien-2-one were determined in pastırma samples (Table 3). None of these compounds were statistically significant (p>0.05). Similarly, it was also determined that numbers and percentage values of ketones are low in studies of volatile compounds on pastırma [8,16]. Kaban [8] reported that ketones were not found in raw material used in pastırma production, but ketones were generated during the production.

In pastırma samples, six different aromatic hydrocarbons; toluene, benzene- 2-methyl-1-propenyl, benzene-1-methoxy-4-(1-propenyl), 1-methylene-1H-Indene, benzene-1,3-bis (1,1- dimethylethyl), and benzene-1,2-dimethoxy-4-(2-propenyl), were determined (Table 3). The sources of aromatic hydrocarbons can show differences considerably. Of these compounds, toluen can come from lipid degradation as well as animal feeds [21]. On the other hand, Marco et al [22] reported that this compound may also occur as a result of amino acid catabolism. As can be understood from the average toluen levels that was found statistically significant, when nitrate level increased, toluen amount decreased, and this compound was not observed at 600 ppm level. Kaban [8] also determined that toluen in the final product was not found.

Aliphatic hydrocarbons are compounds that have a limited effect on aroma of dry-cured meat products due to having high threshold values [20]. In the present study, two different aliphatic hydrocarbons; hexane and tridecane, were identified. Nitrate level had a significant effect on both of them (p<0.05). Aliphatic hydrocarbons, secondary products of lipid oxidation, were determined at different levels in other studies related to pastırma [8,16]. A decrease in the amounts of these compounds was observed depending on increasing of nitrate level.

Esters occur usually as a result of esterification of carboxylic acids and alcohols in meat products. Low molecular weight esters such as ethyl ester, may be come from carbohydrate metabolism [23]. On the other hand, Olesen et al [24] reported that ethyl acetate level in fermented meat products produced by using nitrate was higher than the samples produced by using nitrite. In the present study, nitrate level showed significant effects on butyl propionate (p<0.05), 2,4-Hexadienoic acid methyl ester and propyl hexanoate (p<0.01) at different significance levels. Level of other esters determined except 2,4-hexadienoic acid methyl ester decreased with increasing nitrate levels. The highest level of 2,4-hexadienoic acid methyl ester was determined in pastırma samples containing 300 ppm nitrate. In esters, as in aromatic and aliphatic hydrocarbons, the decrease was usually observed with increasing level of nitrate.

Terpenes are generally considered to be sourced off spices, especially from pepper [25]. However, it was also indicated that some terpenes were found in meat as a result of their existence in animal feed [26]. In a study on pastırma, five terpenes including α-pinene, β-pinene, 3-carene, D-limonene and o-cymol were determined in raw materials and it was concluded that terpenes were related to animal feed [8]. In this study, five terpenes were identified as β-myrcene, D-limonene, linalool, eugenol, and caryophyllene (Table 3). The levels of β-myrcene (p<0.05) and D-limonene (p<0.01) compounds found statistically significantly decreased depending on increasing nitrate level and even could not be determined at the level of 600 ppm nitrate.

In pastırma samples produced with different levels of nitrate, the nitrate level showed statistically significant effect on 1-methyl- 1H-pyrrole (p<0.05). Soto et al [27] also determined this compound in dry cured loin. Only one alcohol (3,5-octadiene-2-ol) was determined in pastırma produced using different levels of nitrate. However, nitrate level had no significant effect on 3,5-octadiene-2-ol (p>0.05). Five different alcohols in the production of pastırma were detected by Kaban [8]. Alcohols are significant aroma compounds in dry fermented meat products because of low odor threshold, and their most important sources are lipid oxidation, carbohydrate metabolism and amino acid catabolism [28].

2-Pentyl-furan identified in this study was also been determined in a previous study on pastırma by Kaban [8]. This compound was affected very significantly (p<0.01) from nitrate level and its level decreased as nitrate level increased. Additionally, this compound was also determined in raw-cured whole processed meat products such as cured loin [7], Iberian ham [29], dry cured ham [30]. It was known that this compound occurs during the heat treatment, also might occur as a result of the oxidation of linoleic acid [29].

In pastırma samples, pentanoic acid, 2-methyl and hexanoic acid as acids were found (Table 3). In this research, nitrate level, the main cause of variation, had significant effect at on 2-methyl-pentanoic acid (p<0.05) and on hexanoic acid (p<0.01). The lowest average level of hexanoic acid and pentanoic acid, 2-methyl was determined in the group containing 600 ppm nitrate. In a study carried out on pastırma, Kaban and Kaya [16] identified two different acids as butanoic acid, 3-methyl, and hexanoic acid.

PCA was applied to study the relationships between nitrate level and volatile compounds of pastırma. Two principal components explained 81% of the total variance. The first component explained 57% of the variation in measured properties and the second component 24% of the variation. Figure 1 shows a Biplot with principal component 1 (PC1) plotted against principal component 2 (PC2). The biplot obtained from this analysis showed that all of the volatile compounds evaluated except 1,3-pentadiene were highly related with PC1 (Figure 1). PC1 differentiated the 600 ppm nitrate group showing a negative correlation from the groups containing 150, 300, and 450 ppm nitrate. The 300, 450, and 600 ppm nitrate groups were placed on the positive axis of PC2, however, 150 ppm nitrate group showed only negative correlation with PC2.