**3. Results**

### *3.1. Total Phenols in Cones*

In the factorial experiment of plant vigour × year, a significant interaction was found for total phenols (Figure 3), meaning that the response of this variable to the field of plants of different vigour depended on the year and/or vice versa. Observing the effect of each factor separately, significant differences were found between years but not between fields. In 2017 total phenols were particularly higher than in 2016 and 2018. The average values were 19.0, 11.9 and 15.1 mg g<sup>−</sup><sup>1</sup> in 2017, 2018 and 2016, respectively.

**Figure 3.** Total phenols as a function of year and hop plant vigour (weak—W, fair—F, good—G and very good—VG) and year. Error bars are the standard errors (α = 0.05).

In the factorial experiment of plant vigour × foliar treatment × year, a significant interaction was found for total phenols for the combination of the three factors and for plant vigour × year and foliar treatment × year (Figure 4). Thus, the year seems to be the factor that adds more variability to the results, influencing the accumulation of total phenols in plants of different vigor and subject to different foliar treatment. By analysing the factors separately, differences in total phenols between plots were found, but without any relation to the vigour of the plants. Foliar sprays did not cause a significant effect on total phenols, but in 2017 the values were significantly higher than those of 2018. The average values were 21.2 and 14.7 mg g<sup>−</sup>1, respectively, in 2017 and 2018.

**Figure 4.** Total phenols as a function of year, plant vigour (weak—W, fair—F, good—G and very good—VG) and foliar treatment (Fnut, nutrient-rich foliar spray; Algae, algae-rich foliar spray; and Control). Error bars are the standard errors (α = 0.05).

In the factorial experiment of different plots × liming × year, significant interaction for total phenols only occurred between plot × year (Figure 5), meaning that total phenol accumulation in plants from different plots was dependent on the year effect. In this experiment, the effect of the plot and year was not statistically significant, and lime's application significantly reduced the content of total phenols. The average values of total phenols were 17.8 and 16.5 mg g<sup>−</sup>1, respectively, in control and limed plots.

**Figure 5.** Total phenols as a function of year, plot (P5 and P6) and liming (L, limed; and C, not limed). Error bars are the standard errors (α = 0.05).

In the factorial experiment cultivars × year, significant interaction was found for total phenols, which means that the response of the cultivars depended on the year (Figure 6). A separate observation of the effect of each of the factors indicated that Nugget showed significantly lower values than Columbus and Cascade, and the values of 2017 were significantly higher than those of 2018. The average values of Nugget, Columbus and Cascade were 16.7, 19.9 and 19.6 mg g<sup>−</sup>1, respectively, and the average values of 2017 and 2018 were 19.9 and 17.6 mg g<sup>−</sup>1, respectively.

**Figure 6.** Total phenols as a function of year and cultivar. Error bars are the standard errors (*α* = 0.05).

### *3.2. Principal Component Analysis*

The PCA applied to data collected from 2016 to 2018 concerning total phenols and nutrient concentration in hop cones resulted in four principal components (PC1 to PC4), which accounted for 70.02% of the variance explained. The main differences in the variance explained were between PC1 (23.35%) and PC4 (11.77%). All variables presented high scores for at least one, or more than one, PC (Table 2). The positive association with N, P, Mg and negative association with K seems to explain greater variance. The higher loading of total phenols was negative and registered in PC3 (−1.606), but scores were also high in PC4 (0.753) and in PC1 (−0.730). These results seem to indicate a negative association of total phenols with Zn and B.


**Table 2.** PCA results for total phenols and nutrient concentrations on hop cones from 2016 to 2018.

PC—principal component; values in bold correspond to the higher loadings of each variable in the respective PC.

Correlation analysis (Table 3) indicates total phenols significantly and negatively correlated with Zn followed by Cu, N and Fe in decreasing order. Positive and significant correlations of total phenols with other nutrients were not recorded. On the other hand, cone N concentration presented positive correlations with other nutrients and most significantly with Mg and P, whereas K was significant and positively correlated with Mn.


**Table 3.** Correlation matrix of total phenols (TPH) and nutrient in hop cones, with Spearman correlation coefficients.

\*, \*\* Significant correlations according to selected significance levels, 0.05 and 0.01, respectively.

### *3.3. Phenolic Compounds Identification and Quantification*

Data on the chromatographic characteristics (retention time, UV in the maximum absorption, molecular ion, and main MS<sup>2</sup> fragments) and tentative identification of the phenolic compounds found in the extracts of hop cones are described in Table 4. A total of 13 phenolic compounds were tentatively identified in the samples, namely, 5 phenolic acids (*p*-coumaroyl- and caffeoylquinic acid derivatives) and 8 *O*-glycosylated flavonoids (quercetin and kaempferol derivatives).

**Table 4.** Retention time (Rt), wavelengths of maximum absorption (λmax), mass spectral data, and identification of the phenolic compounds present in hop cones extract: 3-CQA (3- *O*-Caffeoylquinic acid), *cis* 3-*p*-CoQA (*cis* 3-*p*-Coumaroylquinic acid), *trans* 3-*p*-CoAD (*trans* 3-*p*-Coumaroylquinic acid), 4-CQA (4- *O*-Caffeoylquinic acid), 5-CQA (5- *O*-Caffeoylquinic acid), Q-3-2Rh-Ru (Quercetin-3- *O*-(2-rhamnosyl)-rutinoside), K-3-2Rh-Ru (Kaempferol-3- *O*-(2-rhamnosyl)- rutinoside), Q-3-Ru (Quercetin-3- *O*-rutinoside), Q-3-H (Quercetin-3- *O*-hexoside), Q-3-6M-G (Quercetin-3- *O*-(6-O-malonyl)-glucoside), K-3-Ru (Kaempferol-3- *O*-rutinoside), K-3-G (Kaempferol-3- *O*-glucoside), K-3-6M-G (Kaempferol-3- *O*-(6-O-malonyl)-glucoside).


Data on the quantification of phenolic compounds in the three different cultivars of hop cones are described in Tables 5 and 6. An example phenolic profile chromatogram of the Cascade cultivar is presented in Figure 7. The quantification of the individual phenolic compounds from the first trial (Plant Vigour × Year) revealed that some of the compounds were not detected in plants with good and very good vigour, in particular, *O*-glycosylated kaempferol derivatives and caffeoylquinic acid derivatives (data not shown). Plants of weak vigour were generally higher in quercetin and kaempferol derivatives. The concentration of phenolic compounds in hop cones was very similar among foliar fertilizer treatments, although for most of the compounds the values were slightly higher in the control treatment and slightly lower in the algae treatment (Table 5). In comparison with the control, plants on limed soil presented a significantly higher concentration of kaempferol-3-*O*-(2-rhamnosyl)- rutinoside and 4-*O*-caffeoylquinic acid though not significantly.

**Table 5.** Phenolic compound quantification (mean ± standard deviation) in hop cone samples from 2017 as a function of foliar treatments (Fnut, nutrient-rich foliar spray; Algae, algae-rich foliar spray; and Control) and liming: 3-CQA (3-*O*-Caffeoylquinic acid), *cis* 3-*p*-CoQA (*cis* 3-*p*-Coumaroylquinic acid), *trans* 3-*p*-CoAD (*trans* 3-*p*-Coumaroylquinic acid), 4-CQA (4-*O*-Caffeoylquinic acid), 5-CQA (5-*O*-Caffeoylquinic acid), Q-3-2Rh-Ru (Quercetin-3-*O*-(2-rhamnosyl)-rutinoside), K-3-2Rh-Ru (Kaempferol-3-*O*-(2-rhamnosyl)-rutinoside), Q-3-Ru (Quercetin-3-*O*-rutinoside), Q-3-H (Quercetin-3-*O*-hexoside), Q-3-6M-G (Quercetin-3-*O*-(6-O-malonyl)-glucoside), K-3-Ru (Kaempferol-3-*O*rutinoside), K-3-G (Kaempferol-3-*O*-glucoside), K-3-6M-G (Kaempferol-3-*O*-(6-O-malonyl) glucoside).


Means followed by the same letter are not statistically different by Tukey HSD (Foliar Treatment) or *t*-Student (Limestone treatment) tests (α = 0.05).

**Table 6.** Phenolic compound quantification (mean ± standard deviation) in hop cone samples from 2017 as a function of the cultivar: 3-CQA (3-*O*-Caffeoylquinic acid), *cis* 3-*p*-CoQA (*cis* 3-*p*-Coumaroylquinic acid), *trans* 3-*p*-CoAD (*trans* 3-*p*-Coumaroylquinic acid), 4-CQA (4-*O*-Caffeoylquinic acid), 5-CQA (5-*O*-Caffeoylquinic acid), Q-3-2Rh-Ru (Quercetin-3-*O*-(2-rhamnosyl)-rutinoside), K-3-2Rh-Ru (Kaempferol-3-*O*-(2-rhamnosyl)-rutinoside), Q-3-Ru (Quercetin-3-*O*-rutinoside), Q-3-H (Quercetin-3-*O*-hexoside), Q-3-6M-G (Quercetin-3-*O*-(6-O-malonyl)-glucoside), K-3-Ru (Kaempferol-3-*O*rutinoside), K-3-G (Kaempferol-3-*O*-glucoside), K-3-6M g (Kaempferol-3-*O*-(6-O-malonyl)-glucoside).


Means followed by the same letter are not statistically different by Tukey HSD tests (α = 0.05).

Between cultivars, the differences in phenolic compound quantification were significant for most of the compounds, though not for the total sum of phenolic compounds. Cascade presented lower concentrations of *p*-coumaroylquinic acid (*p*-CoQA) and 4-*O*caffeoylquimic acid (4-CQA), but was generally higher in quercetin and kaempferol derivatives (Table 6). Nugget and Columbus were overall very similar in their phenolic profile.
