The use of agro-industrial by-products has taken on considerable importance for animal feeding as a strategy to contribute both to solving the problem of the disposal of waste material and to reducing production costs for livestock feeding (Vasta et al., 2008).
In the Mediterranean area, the olive and olive oil industries have always played an important social and economic role (Molina-Alcaide and Nefzaoui, 1996). This industry produces substantial amounts of by-products and one of the most important is the olive cake (OC), representing approximately 50% of the conversion process (Servili et al., 2011). OC use for animal feeding is limited because of the seasonal availability and the low nutritive value, 0.32-0.49 UFL (Nefzaoui, 1991; Aguilera et al., 1992) due to the high lignin content and the high percentage of crude fibre (27-41%). However, in the sustainable agriculture the OC use in animal feeding can be useful mainly to reduce the environmental impact of this by-product, provided that the OC diet integration does not adversely affect the productive performances and the sensory, physico-chemical and aromatic properties of milk, cheese and meat.
The potential use of OC as a replacer of part of concentrate in diets for domestic animals has been explored. Different studies have been conducted on small ruminants as they are able to utilise the best types of feeding which are characterised by poor nutritive value (Lanzani et al., 1993).
Molina-Alcaide and Yáñez Ruiz (2008) pointed out the effect of OC in small ruminant diets on feed digestibility, milk yield, milk fat content, milk fatty acid quality, also responsible of an increase of milk protein in ewes.
The oxidative stability of lipids in lamb meat is increased by inclusion of destoned OC (DOC) in concentrate-based diets of lambs, specifically when DOC is used in combination with linseed (Luciano et al., 2013). Tufarelli et al. (2013) reported a significant effect of diet with different levels of partly destoned exhausted OC on growth performance and carcass traits of Gentile di Puglia breed lambs. OC administration for feeding ewes played positive effects on milk yield, whereas no effects on chemical composition and clotting properties were shown (Chiofalo et al., 2004). Moreover, OC administration in diet is responsible of an improvement of the dietetic-nutritional characteristics as shown by the increase of the unsaturated/saturated fatty acid ratio and by the decrease of the atherogenic and thrombogenic indices (Chiofalo et al., 2004).
Studies are currently in progress on the use of OC in diets for cattle. Use of dried OC in the diet of dairy buffaloes resulted in a high milk content of C18:0 and of C18:3n6 and in high amounts of secoiridoids, main phenolic compounds of olive (Terramoccia et al., 2013).
The fatty acid profile of Caciotta cheese was significantly influenced by nutritional integration of DOC for: saturated, monoinsatured, conjugated linoleic acid, n-3/n-6 content and nutritional indexes (Claps et al., 2012).
No study was found on the effect of DOC on the aromatic profile of milk and dairy products.
The aim of this study was to evaluate the aromatic profile of milk and dairy products of dairy cows supplemented with DOC and to compare the results with two techniques for the headspace volatile compounds analysis: thermal desorption by gas chromatography with a mass selective (GC/MS) detector and electronic nose.
Materials and methods
Production and composition of destoned olive cake
DOC was obtained by mechanical extraction of virgin olive oil performed using a Pieralisi Leopard 4 two-phases decanter. DOC has been transported and stored at room temperature, in suitable local of the farm. DOC composition is reported in Table 1.
Animals and diets
The experiment was performed into a farm of dairy cow in the South Italy area between the months of December and January. Ten dairy Friesan cows were divided into two groups (control and experimental), homogenous for weight (550 kg), distance from calving (60-70 d) and milk production (28 kg/d). Animal received the same diet reported in Table 2 and the feed was given as a mixed ration once a day. In the experimental group, the concentrate supplementation was replaced with DOC in a percentage of 15% dry matter. After 2 weeks of adaptation, animals were fed with the experimental diet for other 15 d.
Milk and cheese samples collection
Milk was collected and cumulatively processed in Caciotta cheese (a soft cheese, 25 d ripened) and in Semicotto cheese (hard cheese, 3 months ripened) for 13 times. Samples of milk and cheese were collected and stored at 20°C until GC/MS and electronic nose analysis.
Dynamic headspace-gas chromatography with a mass selective detector analysis
The volatile organic compounds (VOCs) were determined by dynamic headspace (DHS) analysis (Ciccioli et al., 2004) using capillary GC/MS. Fifty mL of milk sample and 5 g of cheese were transferred into a 100 mL and into a 15 mL glass container, respectively. Subsequently, samples were purged with 50 mL/min pure helium gas at 40°C in a water bath for 1h to isolate headspace volatiles. VOCs were trapped in a glass tube packed with Tenax TA, 60/80 mesh (Supelco, Bella Fonte, PA, USA), and Carbopack B, 60/80 mesh (Supelco) in a ratio of 2:1, respectively. The trapped compounds were thermally desorbed at 250°C during 15 min into the Gerstel TDS3 (Gerstel GmbH & Co. KG, Mellinghofen, Germany), Thermal Desorption System, and cryofocused at –70°C before being injected by heating at 250°C in the PTV inlet of CIS4 injector (Agilent Technologies Inc., Wilmington, DE, USA). VOCs were separated on Agilent HP-5MS fused silica capillary column (30 m×0.25 mm i.d., 0.25 µm film thickness) in Agilent GC 7890A at the following condition: helium flow rate 1 mL/min, interface open splitless, oven programme: 35°C for 5 min, then 5°C/min to 150°C for 5 min and 30°C/min to 270°C for 3 min. GC column was connected to the ion source (230°C) of a Agilent MSD 5975C quadrupole mass spectrometer (interface line 280°C), operating in the scan mode (40-450 amu). Ionization was done by electronic impact at 70eV; calibration was done by autotuning. The compounds were identified by comparing their mass spectra with those in the Wiley725 Mass Spectral Database and their retention time. Data were expressed as arbitrary unit (a.u.=peak area × 10–6).
Electronic nose used for this study was a Portable Electronic Nose PEN3 (AIRESENSE Analytics GmbH, Schwerin, Germany) with an array of 10 different metal oxide sensors positioned into a small chamber (V=1.8 mL). Detection limit of the hot sensors was in the range of 1 ppm. Sensors with good selectivity for sulphur organic compounds, methane, hydrogen, alcohol and hydrocarbons were used. The serial number, main applications and references of the 10 sensors are listed in Table 3 and their sensitive gases were applied according to the manufacture’s product manual.
To perform the assay, 1 g of cheese sample and 15 mL of milk sample were placed in a 50 mL vial. Measurements were conducted at a constant temperature (20°C for cheese samples and 30°C for milk samples). The sample run lasted 60 s and was followed by 300 s flushing time. Each measurement was carried out in triplicate. The set of signals of all sensors during measurement of a sample formed a pattern file. Patterns of multiple measurements dealing with the same problem were stored in a pattern file and acted as a training set. The pattern data were checked and analysed using WinMuster Version 22.214.171.124.
The data obtained with GC/MS were grouped according to their chemical nature in aldehyde, ketone, hydrocarbon, ester, alcohol and terpene classes. Compound class data were submitted to one-way analysis of variance using Systat Statistical Package (Systat 13, 2009) with the diet (control vs experimental diet) as treatment effect. Differences were considered significant at P≤0.05. All data collected of VOCs were processed with principal component analysis (PCA) using the previous version (2.16) of the software. PCA for electronic nose pattern data were performed by WinMuster Version 126.96.36.199 software to discriminate between the different samples.
The VOCs, identified either in milk and cheese, divided in aldehyde, ketone, hydrocarbon, ester, alcohol and terpene classes and expressed in a.u., are shown in Table 4. No significant differences are found between compound classes of control and experimental group, both in milk and in cheese. PCA of milk data (Figure 1A) explained 98.59% of the total variance, 81.19% by first principal component (PC1) and 17.40% by second principal component (PC2), and did not discriminate between the two groups, which overlapped. As regard VOCs of Caciotta cheese and Semicotto cheese (Figures 2A and 3A, respectively), PCA explained 93.96% (87.37% by PC1 and 6.59% by PC2) and 89.29% (76.31% by PC1 and 12.98% by PC2) of the total variance, respectively. No differences were found both in Caciotta cheese and in Semicotto cheese.
Also the results of the electronic nose have shown that the use of DOC has not affected footprint olfactory both the milk and cheese. PCA of electronic nose pattern data are presented in Figure 1B (95.37% of variance: 88.41% by PC1 and 6.96% by PC2) for milk, in Figure 2B for Caciotta cheese (98.37% of variance: 89.86% by PC1 and 8.51% by PC2) and in Figure 3B for Semicotto cheese (89.57% of variance: 70.31% by PC1 and 19.26% by PC2).
Our results, obtained both with GC/MS analysis and with electronic nose, show no effect of DOC on the aromatic profile of cows’ milk and dairy products. Moreover, these results support the possibility of DOC use for cow feeding, also because, according to literature (Molina-Alcaide et al., 2008; Tufarelli et al., 2013; Chiofalo et al., 2004), DOC has not negative effects on productive performances and on the physico-chemical properties of milk, cheese and meat. Furthermore Chiofalo et al. (2004) have not observed sensory peroxidative phenomena during cheese manufacturing and seasoning.
The second objective of this study was the comparison between the two different techniques used.
The headspace technique, in particular the DHS, is a very popular method for analysing volatile compounds in food (Canac-Arteaga, 2001). Many papers have investigated changes in volatile flavour compounds in milk and dairy products due to different feeding (Fedele et al., 2005), in fermented dairy products due to different starter cultures (Imhof and Bosset, 1994) and during cheese ripening (Thierry et al., 1999). However, this method is time consuming and costly, particularly when used for routine analysis purposes and it does not allow to establish whether such compounds are odour-active or not.
In recent years, electronic nose has been gaining attention as useful tool, especially for quality control in food and beverage industries. It is ideally suited for rapid screening capacity, but although it does not discriminate singular VOCs.
Electronic nose could discriminate different organoleptic properties (qualities, origins, defects and concentration of pollutant) of different samples. Many studies have investigated the use of electronic nose for aroma compound analysis of dairy products (Ampuero and Bosset, 2003), for the determination of the geographic origin of cheese (Pillonel et al., 2003), for the shelf-life determination of milk and for monitoring the cheese ripening process.
Comparative studies on the use of different techniques for the headspace aroma profile analysis were carried out on different food products like fish (Sun et al., 2013), while few comparative studies on milk and cheese were carried out (Pillonel et al., 2003). The use of the two techniques of investigation and the combined approach allow for a more complete understanding of the flavour profile.
The use of DOC, as an unconventional feed for livestock, has no effect on the aromatic profile of both milk and dairy products.
The results of two compared techniques seem to lead to the same conclusion. Here we have shown that data collected by electronic nose support its use as viable method for rapid detection of aromatic profile of milk and dairy products. The combination of the DHS-GC/MS and the electronic nose provides a great deal of information and a more complete picture of milk and dairy product complex flavour.
NDF, neutral detergent fibre; ADF, acid detergent fibre; ADL, acid detergent lignin.
|Sensor number||General description||Reference|
|S1||Aromatic compounds||Toluene, 10 ppm|
|S2||Very sensitive, broad range sensitivity, to nitrogen oxides||NO2, 1 ppm|
|S3||Ammonia, used as sensor for aromatic compounds||Benzene, 10 ppm|
|S4||Mainly hydrogen, selectively (breath gases)||H2, 100 ppb|
|S5||Alkenes, aromatic compounds, less polar compounds||Propane, 1 ppm|
|S6||Sensitive to methane broad range||CH3, 100 ppm|
|S7||Reacts on sulphur compounds||H2S, 1 ppm|
|S8||Detects alcohols, partially aromatic compounds||CO, 100 ppm|
|S9||Aromatics compounds, sulphur organic compounds||H2S, 1 ppm|
|S10||Reacts on high concentrations||CH3, 100 ppm|
|Compounds classes*||Milk||Caciotta cheese||Semicotto cheese|
|Control group||DOC group||SE||P||Control group||DOC group||SE||P||Control group||DOC group||SE||P|
*Arbitrary units (a.u.=peak area × 10–6). DOC, destoned olive cake; SE, standard error; ns, not significant.