Abstract

An assay based on an electronic olfactory system was set to evaluate tomato fruits by sensing the aromatic volatiles during postharvest storage of 21 days at C in darkness. Olfactory system measurements were coupled with colour values. Odour profile and senescence parameters were carried out at 7-day intervals. Discriminant function analysis applied to electronic nose data showed three components, accounting for 99.2% of the total variance. In the present assay, separation among groups according to storage time (0, 7, and 14 days) was observed for wildtype. Overexpressed (Money Maker) lines/plants of tomato showed difference between odour profile for day 0 and day 21, even tough a no clear discrimination between 7 and 14 days was observed. Fruit lost weight almost linearly with shelf life () presenting an averaged loss of 21% () for over-expressed (Money Maker) lines/plants, 13% () for silenced (Money Maker), and 14% () for wild type during 21 days of storage. Colour values , , and data showed that colour properties changed during storage for all the lines considered. Correlations between odour profiles and colour parameter were obtained showing that the electronic nose is a useful technique for monitoring short-term storage of tomato.

1. Introduction

Flavour is defined as the aroma and taste perceived by the human senses and as such is an important food quality attribute [1]. The flavour of tomato results mainly from a combination of volatile compounds for aroma and of sugars and acids for taste. The aroma composition of fresh tomatoes has been studied and over 400 components have been identified, but only a limited number are useful to explain the global fresh tomato aroma. Several studies report aroma composition by cultivars [2], stages of ripeness [3], different culture conditions [4], and treatments [5] suggesting that these parameters influence the aroma composition of tomato.

Many efforts are made to maintain optimal visual quality (e.g., uniform colour, absence of decay, etc.) to attract customers. As a consequence, internal quality attributes, such as flavour, texture, and nutritive value, which are not readily detectable during sorting operations, receive less attention.

Visual appearance is a critical factor driving the initial choice for purchase, but subsequent purchases are influenced greatly by eating quality. Colour in tomato is the most important external characteristic to assess ripeness and postharvest life. Degree of ripening is usually estimated by colour charts. Colorimeters, on the other hand, express colour in numerical terms along the , , and axes. However, most of the tomato literature, mainly express colour changes in terms of different mathematical combinations of and on the chromatic equatorial plane. As referred by Lรณpez Camelo and Gรณmez [6], different colours are present during tomato ripening simultaneously. Chlorophyll is degraded from green to colourless compounds at the same time that carotenoids are synthesized from colourless precursor (phytoene) to ฮพ-carotene (pale yellow), lycopene (red), ฮฒ-carotene (orange), and xanthophylls and hydroxylated carotenoids (yellow) in a kind of parallel biosynthetic pathway.

On the other hands, the odour of a food product is detected when its volatiles enter the nasal passages at the back of the throat and are perceived by receptors of the olfactory system [7]. Currently, the most common methods for measuring tomato flavour include sensory and instrumental studies. In sensory analysis, taste and aroma aspects of food products are evaluated by panels of specially trained people. Consumer studies provide unique information about the acceptance levels of a food, which is also widely used for the determination of overall quality.

The most important problems affecting sensory analysis include standardisation of measurements, correctness of training, stability, accuracy, and reliability.

The introduction of the electronic nose (E-nose) approach that employs an array of chemical sensors based on conducting polymers, metal oxides, surface acoustic wave devices, quartz crystal microbalances, or combination of these devices has provided an alternative to classical instrumental analysis [8]. Basically, the sensor elements give a signal pattern characteristic of the mixture of volatiles in the headspace of the sample.

This signal pattern is then evaluated using pattern recognition techniques such as neural networks and multivariate statistical techniques [9]. In horticulture, the electronic nose has been successful in monitoring pears [10], apples [11], and other fruits and vegetables [12].

The aim of this work was to study the organoleptic maturation of different transgenic lines of tomato plants using an electronic nose composed of metal oxide sensors and senescence parameters techniques. Short term of storage was analyzed using multivariate techniques to monitor quality of the fruit.

2. Materials and Methods

2.1. Fruit Material

Wild-type tomato plants cv. Money Maker and tomato plants overexpressing and silencing Asr1 gene under the control of promoters 35S and B33 were grown under controlled conditions in a greenhouse (200โ€‰ฮผmol PAR s-1โ€‰m-2, 60% RH, 23ยฐC).

Fruits were harvested manually from plants grown in the National Institute of Agropecuary Technology, during the summer at the ripening stage 5 (light red) (USDA colour chart, 1975). Fruits of uniform shape and size and free from fungal infection were selected. After harvest, fruits were washed with a solution of hypochlorite (150โ€‰ppm de Cl2 as hypochlorite of sodium), air-dried at atmospheric temperature, and individually labelled and weighed. Samples were kept at ยฐCโ€‰and 85% RH and analyzed weekly (7 days) for three weeks (21 days).

2.2. Measurement of Senescence Parameters

The loss in weight (Scout-Pro OHAUS, USA) of individual fruit was determined at weekly intervals as a percentage of initial weight at harvest. A mean of four fruits was used for each sampling period. Skin fruit colour was monitored using a ByK Gardner Spectro guide 45/0 Gloss. Colour values were measured at four points of the Ecuador line of the fruit, and CIELab system was used.

Among the several existing colour scales, CIELAB colour space is a three-dimensional spherical system defined by three colorimetric coordinates. The coordinate is called the lightness. The coordinates and form a plane perpendicular to the lightness. The coordinate defines the deviation from the achromatic point corresponding to lightness, to red when it is positive and to green if negative. Similarly, the coordinate defines the turning to yellow if positive and to blue if negative.

Colour index (CI) was calculated according to

2.3. Electronic Nose

An electronic nose (EN) comprising 18 semiconductor oxide metallic sensors pure and doped semiconductor (MOS), coupled with a mass spectrometer system (NE-MS, Alpha Prometheus, Alpha MOS) was used to discriminate odours of the fruits.

The used device is equipped with two types of sensors: P and T sensors and LY ones. P and T are metal oxide sensors based on tin dioxide SnO2 (n-type semiconductor), the difference between them resides in the geometry of the sensors. The LY sensors are metal oxide ones based on chromium titanium oxide (p-type semiconductor) and on tungsten oxide (n-type semiconductor). Table 1 presents an overview of the sensors of the electronic nose, as well as the chemical compounds to which they are sensitive. In the presence of a reducing gas, there is absorption with an electronic exchange of gas towards the sensors: the conductance of the n-type increase while for the p-type the resistance will increase, due that n-type are based on tin dioxide SnO2 and p-type are based on chromium titanium oxide [13].

Doping with different elements increases SnO2 selectivity for different gases. The adopted configuration results are very flexible for general purposes and convenient for a wide range of applications. Sensors are relatively nonspecific and can combine the signals of all the sensors in a unique signal (Figure 1). Each curve represents a different sensor. The curves represent the sensor conductivity (y-axis) over time (x-axis) when the volatiles from the fruit reach the measurement chamber, with respect to its value measure when carrier gas reaches the sensor.

Electronic nose data is analyzed by multivariate methods like principal component analysis and discriminate function analysis. The result obtained using these method are bi-dimensional plots, were axes are determinate by the sensors that contribute most to discriminate odour. On the other hand, similar odours tend to be grouped in clusters.

2.3.1. Samples for Electronic Nose

Each individual fruit belonging to wild-type (nontransgenic) plants and to transgenic plants was macerated in a stomacher machine for 30โ€‰s and 60โ€‰g pulp was mixed with 15โ€‰mL of saturated CaCl2 solution (added all at once) in the stomacher for another 5โ€‰s [14]. For electronic nose measurement, samples of โ€‰g were placed in five 10โ€‰mL glass vials equipped with a screw cap and silicon septum.

The experimental part was divided into two steps. In the first step, sensors response of electronic nose was evaluated and experimental conditions of electronic nose were optimised using wild-type tomato. Once the experimental conditions and methodology of electronic nose was established, sensory evaluation was performed using overexpressing tomato plants. It was not possible to analyse silenced plants due to the fact that the amount of fruit was not enough.

2.3.2. Parameters Used for Electronic Nose Analysis

Samples were stabilised at 40ยฐC for 10โ€‰min (incubation time) and shaked (500โ€‰rpm). Then, 1โ€‰mL of headspace sample was injected, the acquisition time being 120โ€‰s with a frequency of 0.5โ€‰s. Synthetic air was employed as carrier gas with a flow of 30โ€‰mLโ€‰minโˆ’1. Samples were analyzed thrice.

2.4. Statistical Analysis

In this work, statistic analysis was done under two approaches: univariate analysis with a completely randomized design and Pearson correlation; and multivariate discriminant and principal components analysis. The statistical software used was SPSS v. 12 (Illinois, USA).

3. Results and Discussion

3.1. General Senescence Parameters
3.1.1. Loss of Weight

With increasing the storage time fruit lost weight almost linearly (), and averaged loss of 21% () for overexpressed (Money Maker), 13% () for silenced (Money Maker), and 14% () for wild type was obtained after 21 days of storage. Maharaj et al. [5] reported that mature green tomato fruit (var. Capello), stored at 16ยฐC and under high relative humidity for a period of 35 days, represented a loss of weight of 16% during 21 days of storage. Similar results reported Maharaj were observed, but in different mature stages.

3.1.2. Colour

Initially all fruits were light red (rating 5) in colour (Figure 2). The effects of processing and storage time on lightness , , and coordinates and colour index for each sample are shown in Tables 2, 3 and 4.

Significant differences in and values were obtained for tomato wild-type samples. A decrease in values due storage was observed. Short term storage demonstrated that tomatoes colour were darker and less yellow than fresh samples. On the other hand, values showed an increase, reaching the highest value at day 14 and then a decrease in red colour. CI increased during storage, having the highest value at day 14 (Table 2).

For silenced (Money Maker) tomato, significant differences in and values due storage were observed. Fresh tomato had the highest and values. The value averaged after 7 days in storage indicated a significant loss in green colour. Colour index showed an increase during storage (Table 3).

and values decrease during storage, reaching the lowest values at day 14 for over- expressed samples. Parameter increases during storage. CI shows the highest difference between initial and day 7 (Table 4).

Colour development in tomato is sensitive to temperature, having a better plastid conversion when temperature is above 12ยฐC and below 30ยฐC [6]. Tijskens and Evelo [15] demonstrated that suffered big changes if tomatoes were ripened at high temperatures (over 30ยฐC) and yellowing took place due to the inhibition of lycopene synthesis and the accumulation of yellow/orange carotenoids. On the other hand, at low temperatures (below 12ยฐC), chlorophyll is not degraded and lycopene accumulation does not take place.

When red colour pigments started to be synthesized, a decreasing value indicated the darkening of the red colour. This behaviour was observed in all samples between 7 and 14 days of storage.

In this research a significant decrease in parameter is observed after day 7. On the other hand, Brunink et al. [8] reported that values changed very little during ripening. This could be related to the fact that ฮถ-carotenes (pale-yellow colour) reach their highest concentration before full ripening, where lycopene (red colour) and ฮฒ-carotene (orange colour) achieve their peaks [16, 17].

3.2. Electronic Nose
3.2.1. Wild-Type and Overexpressed Samples

E-nose data was analyzed applying discriminant function (DFA); analysis was performed using Wilksโ€™ lambda stepwise method for variable selection.

DFA was chosen because it considers the relation of data points for the specified classes. On the other hand, DFA takes into account the distribution within classes and the distances between them. Therefore, it allows us to collect information from all sensors in order to improve the resolution of classes.

The criterion used was the significance of with a maximum of 0.05 to enter and a minimum of 0.10 to exit. The sensors that allow the classification of odour profiles over time were LY2/LG, LY2/G, LY2/AA, LY2/gCTl, LY2/Gct, P10/2, and T40/1.

Three discriminant functions (DFs) were found for the wild-type and overexpressed (Money Maker) lines/plants samples, accounting for 99.2% of the total variation (Figure 3). For wild-type samples three groups were obtained according to storage time (0, 7, and 14 days). On the other hand, three groups were obtained according to storage time (0, 7, 14, and 21 days) for the overexpressed tomato samples. Storage at days 7 and 14 did not show discrimination in odour for overexpressed samples. After 21 days of storage, the overexpressed samples showed difference in odour. These results can be attributed to differences in the volatile-fraction composition during storage that impacts on their odour profiles. Berna et al. [18] reported similar results with tomato (L. esculentum Mill.).

3.2.2. Correlation between E-Nose Data and Colour

In order to observe the electronic nose performance for monitoring the behaviour of fruit quality during storage time, olfactory measurements were related with colour parameters. Principal component analysis was applied to colour and electronic nose data considering only the sensors selected by DFA (LY2/LG, LY2/G, LY2/AA, LY2/gCTl, LY2/Gct, P10/2, and T40/1). Two components were obtained that explained 88.2% of the total variance (Figure 4). PC1 is correlated positively with colour parameters and and sensors LY2/G, LY2/AA, LY2/gCTl, and LY2/Gct. Only samples at storage time T0 (initial time) were correlated positively with PC1. On the other hand, PC1 was correlated negatively with colour parameters and sensors LY2/LG, P10/2, and T40/1. Samples at storage time T21 were correlated negatively with PC1. This result suggests that the different storage time of tomatoes could be monitored by means of the electronic nose.

4. Conclusions

The study of organoleptic mature plants using transgenic lines of tomato Money Maker with the gene ASR1 overexpressed and silenced under the constitutive 35S promoter and the patatin B33 promoter of potato showed changes in colour during storage.

Electronic nose showed differences in odour profiles during short-term storage for either overexpressed or wild-type (Money maker) tomatoes. Future research is needed in order to compare tomato lines response, focusing the attention on 7 and 14 days of storage.

In the last decade, odour research was focused principally on the identification of potent odorants, the determination of their odour relevance, and their release in different foods. Nowadays, the development of the electronic nose methodology, with a chemical sensory array, provides a powerful tool to analyze odour as a set of odorants present within a given sample. Sensory analysis, as a branch of the food industry, will be benefited with the adoption of this methodology.

Acknowledgment

The authors would like to thank Mrs. Mรณnica Pecile for her collaboration in this project.