Journal of Applied Botany and Food Quality 94, 182 - 191 (2021), DOI:10.5073/JABFQ.2021.094.022 1Department of Crop Sciences, Division Quality of Plant Products, University of Goettingen, Goettingen, Germany 2 Institute for Ecological Chemistry, Plant Analysis and Stored Product Protection, Julius Kühn-Institute (JKI), Federal Research Centre for Cultivated Plants, Quedlinburg, Germany Assessment of sensory profile and instrumental analyzed attributes influenced by different potassium fertilization levels in three tomato cultivars Bashar Daoud1*, Marcel Naumann1, Detlef Ulrich2, Elke Pawelzik1, Inga Smit1 (Submitted: August 1, 2021; Accepted: October 17, 2021) * Corresponding author Summary Sensory properties are an essential quality aspect when the con- sumption of fresh tomato is under consideration. The flavor of tomato is defined as a combination of taste sensations (sweetness, sourness), aroma (volatile compounds), and texture (firmness, mealiness), some of which are proven to be affected by insufficient nutrient supply − especially potassium (K). This study intends to undertake a holis- tic assessment of the K fertilization effect on the flavor of tomato by connecting the use of sensorial and instrumental methods. An optimal K supply significantly increased the sensory descriptors sweetness, sourness, and aroma as well as the instrumental estimated color, firmness, total soluble solids (TSS), titratable acids (TA), and dry matter (DM) in a cultivar-specific manner. The volatile organic compounds (VOCs) were not significantly affected by K fertilization. The evaluation by the panelists confirmed the results of the instru- mental analyses, by which an increment in the fruit quality with ris- ing K supply could be detected. An optimal K supply of 3.66 g/plant could be suggested to increase tomato flavor in the cocktail cultivars studied: Primavera and Yellow Submarine. Cultivar effects should, therefore, be considered for defining the optimal K fertilizer dose that favors high tomato fruit quality and, hence, better flavor. Keywords: Solanum lycopersicum L., potassium, sensory evaluation, instrumental analyses, volatile organic compounds Introduction The tomato is one of the most important vegetables in the world. In 2019, around 181 million tons of tomatoes were produced (FAO- STAT, 2021). The increasing annual demand for tomato can be at- tributed to its versatility and suitability for several dishes (AdegbOlA et al., 2019), as well as its fruitfulness in nutrients like minerals and antioxidants (AFzAl et al., 2015). In the European Union, 40% of tomatoes are consumed fresh and 60% are processed for different products (eurOpeAn, 2020). Fresh fruits can be described by their extrinsic (e.g. color, shape, and firmness) as well as the intrinsic cha- racteristics (e.g. taste and aroma) (OlTmAn et al., 2014). The flavor is a complex attribute and derived from the interaction between the volatile compounds, such as hexanal and 2-isobutylthiazole, and non- volatile components like sugar, acids, and minerals (beckleS, 2012). The flavor of tomato is frequently described as a sweet-sour taste accompanying special aromatic aspects like ‘fruity’ and ‘floral’ (bAldwin et al., 2008). However, consumers have often complained about the poor flavor of fresh tomato (TiemAn et al., 2012). Therefore, the flavor of the tomato needs to be comprehensively considered, and not only for the consumers, but also for the producers (piOmbinO et al., 2013). Moreover, the extrinsic and intrinsic characteristics of the flavor are remarkably influenced by many factors like weather conditions and the nutrient status of the plant and soil (mATTheiS and FellmAn, 1999). This study focused on the effect of potassium (K) nutrition on the flavor of tomato. Being an essential macronutrient, K is involved in many physiological and biochemical processes in plants (hAwkeSFOrd et al., 2012). Cellular K plays a role in catalyzing many enzymes, in addition to having major functions in osmotic pressure adjustment (zörb et al., 2014). Furthermore, sufficient K nutrition reinforces the resistance of the plants against biotic stresses like diseases and insects (bidAri and hebSur, 2011). K is also involved in the relocation of photosynthetic assimilates to sink organs, resulting in an increment in the sugar content in the cytosol (leSTer et al., 2006). Consequently, increasing the K fertilizer dose has demonstrated a positive enhancement on total soluble solids (TSS), titratable acids (TA) (SOnnTAg et al., 2019), dry matter (DM) (JAvAriA et al., 2012), and firmness (leSTer et al., 2010). SeriO et al. (2007) and TAber et al. (2008) could state a significant influence of K supply on the lycopene content and, hence, on skin color. The positive effect of K fertilization on increasing yields and fruit quality has been pointed out by many studies (AFzAl et al., 2015; AmJAd et al., 2014). Volatile organic compounds (VOCs) have been considered as sensory indications for flavor preferences (gOFF and klee, 2006). Though around 400 volatile compounds have been detected in tomatoes, only 15-20 compounds, such as hexanal, 2-isobutylthiazole, and 6-methyl- 5-hepten-2-one, have been found to characterize the flavor of the tomato (kAnSki et al., 2020). Most volatile compounds are derived from essential nutrient precursors like amino acids, carotenoids, and fatty acids (rAmblA et al., 2014). Flavor can be measured by instrumental analyses, e.g. TSS, TA, and color, and by sensory evaluation. Several studies investigated the interaction between sensory evaluation and instrumental analyses in tomatoes (kAnSki et al., 2020; TAndOn et al., 2003). To the best of our knowledge, very few studies (AFZAL et al., 2015; WANG et al., 2009) have been measuring the effect of K nutrition on the instrumental and sensory characteristics and their interactions in tomato. Our work attempts to investigate the effect of K fertilization on sensory and physicochemical traits. It also aims to verify whether the results obtained by instrumental methods can be confirmed by the human senses. We hypothesize that: (I) increasing the K supply modifies the values of instrumental analyzed traits and the intensity of sensory quality; (II) the effect of K fertilization on the sensory quality can be recognized by human senses; (III) instrumental analyzed traits will distinctly correlate with sensory quality; and (IV) the effect of K fertilization will be cultivar-dependent. Materials and methods Experimental set-up In summer 2016, an outdoor experiment was conducted with three tomato cultivars. Two cocktail tomato cultivars − Primavera and Yellow Submarine (Kiepenkerl, Everswinkel, Germany) − and one Potassium influences sensory and instrumental analyzed traits in tomato fruits 183 salad tomato cultivar − Lyterno F1 (Rijk Zwaan, De Lier, Nether- lands) − were chosen. The cocktail cultivars were used in previous experiments and showed a good response to K fertilization (SOnnTAg et al., 2019). The salad cultivar was chosen based on the breeders’ description highlighting this cultivar as being high in lycopene. Therefore, it was expected that Lyterno F1 would respond well to varying K supply as regards its color, which has been shown for high lycopene cultivars by TAber et al. (2008) and SeriO et al. (2007). All cultivars were sown on March 30 in planting trays with capa- cities of 0.1 L. After three weeks, the seedlings were transplanted into 11 cm pots with capacities of 1 L in a greenhouse. Greenhouse conditions comprised 16 hours of daylight, with a mean temperature of 22 °C during the day and 18 °C at night. The soil in the trays and pots was a mixed peat (‘A 400’, Stender, Schermbeck, Germany). The final transplantation to the outdoor location took place after seven weeks of sowing on May 25. The seedlings were planted into 20 cm Mitscherlich vessels with capacities of 6.2 L filled with peat substrate (Gartentorf, Naturana, Vechta, Germany). Three different concen- trations of K − K1 low with 0.5 g K/plant; K2 medium with 2.19 g K/ plant; and K3 optimal with 3.66 g K/plant − in the form of liquid K2SO4 were applied weekly during the growing season. Nitrogen (N) was applied on a weekly basis along with K − as a mixture of NH4NO3 and Ca(NO3)2 · 3H2O − for K3 treatment and every two weeks for K1 and K2 treatments. Another N solution − (NH4)2SO4 − was applied for K1 and K2 treatments, alternating with the previous mixture every two weeks to balance the sulfate supply. Other plant macro- and micronutrients were applied at the final transplantation and two more times during the season (Tab. A1). The plants were irrigated with distilled water when required and were pruned to one shoot weekly. They were arranged in a randomized design, with four blocks representing four replicates per cultivar and K level. During harvest, the fruits of each sample were split into three subsamples. One sample set was used for the sensory evaluation by the panelists; the second subsample was used for extraction of VOCs; and the third for instrumental analyses. The number of fruits used for each type of quality analysis is given in the Table A2. Instrumental analyses The K concentration, color, firmness, TA, TSS, DM, and volatile compounds were estimated at fruit maturity. Based on the method of kOch et al. (2019), the K concentration was determined by digesting 100 mg fine powder of lyophilized tomato fruits in 4 mL of 65% nitric acid and 2 mL of 30% hydrogen peroxide for 75 min at 200 °C and 40 bar in a microwave (Ethos terminal 660, Milestone, Sorisole, Italy). Subsequently, the samples were analyzed using inductively coupled plasma-optical emission spectrometry (ICP-OES; Vista-RL ICP-OES, Varian, Palo Alto, USA). Fruit color was determined by Minolta Chroma Meter CR-400 (Konica Minolta Optics, Japan) at the two equatorial sides of each fruit in the Lab modus, where the a value represents the red color intensity of Lyterno and Primavera fruits, while the b value represents the yellow color intensity of Yellow Submarine fruits. Afterwards, the firmness was estimated by a penetration test (5 mm staple micro cylinder, speed: 6 mm/s, distance: 6 mm) on the equatorial side of these fruits with a texture analyzer (TA.XT2, Stable Micro System, Surrey, UK). TSS, TA, and DM were estimated for the same fruits. The fruits were mixed for two minutes with a kitchen blender (MQ 5000 Soup, Braun, Neu-Isenburg, Germany) to achieve a homogenized puree. An amount of 10 g of this puree was dried for estimating DM, and the rest of the puree was centrifuged for 20 minutes at room temperature and at 5000 g (Centrifuge 5804 R, Eppendorf, Hamburg, Germany) to estimate TSS and TA based on SOnnTAg et al. (2019). Immediately after harvest, VOCs were extracted from fresh fruits, as described by ulrich and OlbrichT (2013). The fruits were rinsed with distilled water, cut into quarters, and homogenized in a solution with 20% (m/v) NaCl by a kitchen blender (MQ 5000 Soup, Braun, Neu-Isenburg, Germany). The homogenate was centrifuged for 30 minutes at 4 °C and 3000 g (Centrifuge 5804 R, Eppendorf, Hamburg, Germany). To 8 mL of the supernatant and 4 g of NaCl, 16 μL of the internal standard (5 μL octanol + 10 mL ethanol) were added. The samples were vortexed and stored at -20 °C until analysis by gas chromatography − FID, as previously described by ulrich and OlbrichT (2013). Sensory evaluation A group of 12 panelists had been trained weekly over two months, resulting in eight training sessions in accordance with the iSO 13299 (2016) sensory analysis guidance (iSO 8586, 2014), by focusing especially on the quantitative descriptive analysis of the type of tomato fruits used in this study. The panel performance was checked after each training session and the result was used for improving the training. The sensory descriptors color and odor intensity, juiciness, sweetness, sourness, bitterness, skin strength, aroma, and aftertaste were elaborated with the sensory panel (Table A3). The scale from 0% (minimum intensity) to 100% (maximum intensity) was used to determine the intensity of all descriptors that were studied. The final sensory evaluation was performed during the second week of August on fully ripe fruits for three consecutive days, with a single cultivar being evaluated each day with respect to sensory fatigue of the panelists. The experimental set-up of the outdoor trial was retained during the sensory evaluation. The replicate samples deriving from the four blocks were evaluated separately by the sensory panel. Overall, each panelist evaluated 12 samples that derived from four field plots multiplied by three fertilization levels. The samples were provided to the panel in a randomized design generated by the EyeQuestion software. The evaluation was accomplished in a sensory laboratory that provided separated booths, in accordance with iSO 8589 (2007). The fruits of cocktail cultivars were cut into halves, while those of the salad cultivar were cut into quarters immediately before being served in transparent plates that were coded with three- digit numbers. The panelists performed the evaluation and the data were acquired digitally using the EyeQuestion software. Between the served samples, the panelists were directed to consume a piece of bread and tap water to naturalize the basic tastes. Statistical analyses Statistical analyses were performed mainly by using the SPSS Software, Version 22 (IBM Corporation, New York, United States). Data were proven to be normally distributed with the Shapiro-Wilk test (p<0.05), and the variance homogeneity was verified with Welch’s test. General fertilizer effects were tested at the significance level of p<0.05 with one-way ANOVA before separating the means of each fertilization treatment within the cultivars by using Tukey’s post-hoc test. In order to connect sensory and physicochemical traits, Pearson’s correlation analysis was calculated with SPSS and a principal component analysis (PCA) was calculated with the Statistica Software, Version 13.3 (TIBCO Statistica, Tulsa, United States). The panel performance was calculated by a 2-way ANOVA with assessor and sample as main effects with the Software PanelCheck V1.4.0. Results Fruits’ K concentration The fruits’ K concentration was significantly influenced by the fer- tilization level (Tab. 1). As anticipated, the level K1 significantly dis- played the lowest values. The supply of the fertilizer level K3 could 184 B. Daoud, M. Naumann, D. Ulrich, E. Pawelzik, I. Smit only significantly raise the K concentrations in the two cocktail to- mato cultivars compared to level K2. Instrumental and sensory determined color Instrumental analyzed color values increased significantly with rising K fertilization only in the cocktail cultivars, where the color-b value (yellow) of Yellow Submarine fruits and the color-a value (red) of Primavera fruits were more intense in K3 (Tab. 1). Based on the panelists’ evaluation, color intensity was increased significantly only in Primavera (Tab. 3). Consequently, the color values affected by K fertilizations depended remarkably on the cultivar. The principal component analysis (PCA) in three cultivars confirmed the ANOVA results. In the PCA, color intensity and instrumental determined color were located closely to each other in Primavera (Fig. 2) but distanced from each other in Lyterno F1 (Fig. 1) and Yellow Submarine (Fig. 3). Additionally, the correlations of color intensity with the instrumental determined color were low and nonsignificant: color-a (r = 0.23) and color-b (r = 0.45) (Tab. A5). Volatile organic compounds and odor intensity Around 16 known volatile organic compounds (VOCs) were dis- tinguished in this study and they comprised around 80% of all the detected VOCs (Tab. 2). Most of them were not influenced signifi- cantly by K fertilization, while the main variations were related to a Tab. 1: Mean values and standard deviation of taste-related attributes calculated for each K level (n=4) within the three cultivars Lyterno F1, Primavera, and Yellow Submarine. TSS: total soluble solids, TA: titratable acids. Color-a: estimated for Lyterno F1 and Primavera. Color-b: determined for Yellow Submarine. n.d. not determined. Letters indicate significant differences at p<0.05 between the K treatments. K1 low 0.5; K2 medium 2.19; and K3 optimal 3.66 g/plant. Instrumental Lyterno F1 Primavera Yellow Submarine analyzed Attributes K1 K2 K3 K1 K2 K3 K1 K2 K3 K-content (%) 1.42b ± 0.11 2.48a ± 0.12 2.44a ± 0.17 1.21c ± 0.07 2.34b ± 0.07 2.66a ± 0.15 1.60c ± 0.11 2.29b ± 0.07 2.54a ± 0.15 Color-a 21.09a ± 0.92 20.32a ± 0.7 20.77a ± 0.71 12.16b ± 0.45 16.81a ± 1.72 17.67a ± 1.25 n.d. n.d. n.d. Color-b n.d. n.d. n.d. n.d. n.d. n.d. 44.82b ± 2.71 49.48a ± 1.33 50.24a ± 1.94 Firmness 1.21b ± 0.34 1.59ab ± 0.08 1.79a ± 0.32 0.69a ± 0.03 0.82a ± 0.09 0.70a ± 0.11 0.75b ± 0.01 0.95a ± 0.10 1.03a ± 0.14 (kg/cm) TSS (%) 5.80b ± 0.43 6.45ab ± 0.44 7.30a ± 0.62 6.75b ± 0.3 8.45a ± 0.44 8.52a ± 0.19 8.25b ± 0.01 9.05b ± 0.01 10.45a ± 0.00 TA (%) 0.26c ± 0.03 0.48b ± 0.03 0.53a ± 0.01 0.25b ± 0.02 0.48a ± 0.04 0.51a ± 0.02 0.34c ± 0.02 0.51b ± 0.08 0.65a ± 0.08 DM (%) 7.76a ± 1.18 7.94a ± 0.73 8.99a ± 0.64 8.35b ± 0.42 9.92a ± 0.3 9.95a ± 0.61 10.18b ± 0.66 10.93b ± 0.76 12.44a ± 0.26 Tab. 2: Mean values and standard deviation of identified and unknown VOCs calculated for each K level (n=4) within the three cultivars Lyterno F1, Primavera, and Yellow Submarine. Values below the limit of detection (LOD) were indicated. Letters indicate significant differences at p<0.05 between the K treatments. VOCs Lyterno F1 Primavera Yellow Submarine Identified (%) K1 K2 K3 K1 K2 K3 K1 K2 K3 hexanal 32.82a ± 4.93 19.86b ± 6.19 19.46b ± 2.47 40.29a ± 7.47 40.47a ± 1.55 39.24a ± 1.72 27.27a ± 2.64 26.72a ± 2.45 27.38a ± 1.28 (E)-2-hexenal 4.22a ± 0.81 3.65ab ± 0.32 3.01b ± 0.39 8.08a ± 3.73 6.67a ± 1.85 7.01a ± 1.47 9.12a ± 2.73 10.76a ± 3.01 9.89a ± 3.08 octanal 5.26a ± 0.82 4.04ab ± 0.68 3.96b ± 0.35 4.48a ± 1.11 3.64a ± 0.31 4.14a ± 1.89 1.35a ± 0.96 1.73a ± 0.21 1.59a ± 0.12 β-ionone 1.06a ± 0.23 0.36b ± 0.42 0.16b ± 0.33 1.98a ± 0.68 1.73a ± 0.23 1.83a ± 0.73