Soluble Biobased Substances in soil or Salicylic Acid on leaves affect the foliar pH and soil biovariability of Grapes-as explained by the NIR Spectroscopy of Litterbags and Teabags

As highlighted in the introduction, the studies on treatments with HA on the soil are not numerous and there are even fewer that concern soil biodiversity, while none are available about leaf treatments with SA. In this experiment, we have studied an alternative to HA, namely a non-commercial alkaline product derived from municipal biowastes, which is extracted from non-renewable sources. The alternative simple and non-joint effects of a soil treatment with SBS and an SA leaf treatment have been compared in this completely randomized experiment. The phenotypic evaluation on the plants took into consideration the leaf composition, as disclosed by the NIR spectra, and especially the leaf pH, the grape production, and the quality of the grapes pertaining to their technological and phenolic maturity. Indirect methods, based on the detection of the bioactivity of hay or of red and green Tea litters, coupled with NIR spectroscopy, were applied to evaluate the soil microorganisms. An attempt was also made to predict the bacterial and fungal biota from the NIR spectra of green Tea to evaluate the soil biodiversity.

The trial was conducted in the western Po valley (45.08180636477463, 7.958740094755527) on an adult vineyard of Malvasia red wine, located on a steep South oriented slope. A conventional management of the vineyard, following an integrated pest management protocol, is adopted, and it had not been fertilized in the three years before the research. The vines are pruned using the mono lateral guyot system with espalier training and at a density of 3.460 plants per hectare. During the vegetative period, the soil management involves three grass mulchings between the rows and three mechanical processings under the rows. The topsoil composition is reported in Table 1.

An experimental alkaline (pH 8.40) complex of Soluble Biobased Substances (SBS) was used for the soil treatment to provide some humic substances, derived from biogas residues (Table 2). A total of 30 ml of the complex, diluted in 750 ml of water, was distributed over each six plants in a single solution on July 6, 2022.

The humic acid extraction process was performed using compost obtained from the anaerobic digestion of municipal green biowastes. The process ²⁴ involves hydrolyzing the compost in an k-alkaline solution and producing value added (SBS). The suspended solids were separated and the liquid solution, containing around 7% of dry matter, was used for the above-mentioned treatments. A typical solution of SBS has 4% of humic extract: humic acids + fulvic acids in a 3:2 ratio (which means 2.4% w/w humic and 1.6% w/w fulvic acids).

The other treatment considered in the experiment was based on Salicylic Acid (SA), a slightly soluble substance (2000 ppm) that was used at a concentration of 150 ppm. The solution reached pH 7.12 (-4% vs. water at pH 7.44) and was sprayed onto the leaves of six plants every two weeks at a dose of 50 ml plant.

Two leaves per plant were sampled seven times from July 6 till September 6. The leaves were analyzed for the petiole pH using a Hamilton Peek Double-Pore F, / Knick combined 35 x 6 (LxØ) glass-plastic electrode, at two decimals. The NIR spectroscopy of the leaf blade was conducted in single scan mode using an NIR-SCiO^TM smart device, as described in Masoero et al. ²⁵. The foliar pH data were analyzed, by means of the XLSTAT 2019.4.1 ²⁶ software, according to a bifactorial ANOVA, with three treatments and seven times replicates. The chemical components of the leaves were predicted, by means of NIRS, through the application of previously published PLS models ^27,28, calculated using WinISI-III (FOSS NIRSystem/Tecator, Infrasoft International, LLC) software.

A total of 30 berries were collected randomly from each plant at two different times near the harvest and subsequently taken to the laboratory. Three subsets of 10 berries were composed and weighed, then the pulp, skins and seeds were divided and each berry was examined for pH using the Hamilton Double-Pore F electrode and by a bench refractometer for the degree brix, while a triplicate NIRS scan was applied to the pooled skins and seeds. The technological index was calculated as pH² * brix, whereas the phenolic index was expressed as the ratio of the non-extractable to the extractable phenols and was predicted from the NIR spectra according Cugnetto et al. ²⁹.

The litterbag-NIRS method was developed by Masoero et al. ³⁰ and has been applied in numerous trials with biofertilizers ^31,32,33,34. It consists of examining the NIR spectrum of ground alpine hay placed in a package of permeable sachets that have remained buried for 60 days. The chemometric elaboration of the spectra allows a discrimination of the various groups under study to be carried out. The method is that of the Partial Least Squares Discriminant (PLSD), which consists of first preparing the N*G incidence matrix, composed of N observations and G groups, in whose cells a value of 1 is assigned if an observation is present and 0 if an observation is absent. The PLS-Regression method is then used, and G equations fitted to the values 1 and 0 are calculated. The predictions obtained from the equations generate a matrix of values whose G columns are submitted to a linear discriminant analysis to ascertain the percentage of correct classifications (%Correct), which is the diagonal of the G² confusion matrix. In our case, we used the WinISI-III software (Infrasoft) to obtain the PLS equations from the NIR spectra and the XLSTAT software ²⁶ to perform the discriminant analysis with cross-validation of the predictions.

Based on previous equations ³⁵, the NIR spectra allowed us to predict 23 Litterbag-NIRS variables. These predictions were the final bromatological composition parameters of the hay residuals in the litter in relation to a feeding value for ruminants. Furthermore, some soil properties previously related to spectra (NO₃^-N, NH₄⁺N, Soil Induced Respiration) and the mycorrhizal index were estimated. Finally, in agreement with the proposal of Fontaine et al.³⁶ and Blagodatskaya et al.³⁷, it was possible to obtain an indirect evaluation of the microbial activity by comparing the initial composition of the hay and the final one of the litters. This scheme considers two conventional types of microbes: the so-called “r-” (r-labile-strategist) microbe, which refers to a negative allometric variation of substances that decrease during the decay, as opposed to a category of slow “k+” (k+ resistant-strategist) microbes, which refer to a positive allometry of the substances that increase in percentage in the litter. In the works of Masoero et al.³³and Hett et al.³⁴, it was highlighted that the components that formed last in the herbs (acid detergent fiber, acid detergent lignin, crude fiber, cellulose) are the first to degrade, while the substances that formed first(ash, hemicellulose, crude protein) are of type "k", whose relative presence in litters increase during the decay process.

The Teabag Index (TBI) method of Keuskamp et al. ³⁸ involves burying two bags of Lipton green tea (EAN 87 22700 05552 5) and rooibos tea (EAN 87 22700 18843 8) for 90 days. At the end of the test, the two sachets are removed and weighed. In the TBI, two parameters are determined by means of two formulas: the stabilization factor TBI_S and the decomposition rate TBI_K. These parameters were elaborated as univariates by means of the ANCOVA model. In this work, for the first time, we have improved the TBI method by applying NIR spectroscopy to green and red tea residues. The spectra were elaborated by means of a PLS-D type procedure, exactly as was done for the hay Litterbags. At present, we cannot consider the allometry of the decomposition of the two types of tea with respect to the original ones, because there is a lack of chemical data to calibrate the procedure. Instead, starting from the NIR spectra of green tea, which were taken from the study of Staffolani et al ³⁹in vineyards, it was possible to obtain a set that was highly correlated with the bacterial and fungal biomass. In that study, which was carried out on nine clusters of Nebbiolo vines, the triplicate samples of soil, kept in a refrigerated environment between 1 and 6°C, reached the Bioma Makers company laboratory within 72 hours of sampling. The analysis performed by the laboratory was based on the sequencing and characterization of the 16SrRNA gene and ITS markers of bacteria, fungi and Archaea using Illumina MiSeq technology. The results reported the presence of taxa expressed as a percentage of the total identified taxa. In the present work, the PLS equations, recalculated after the cited study (unpublished), resulted an R² coefficient of 0.68 ± 0.13 for both Bacteria and Fungi, and these models were applied to the green tea spectra of the present study.

The SBS lowered the leaf pH by 5% (Table 3).

This did not happen at the beginning of the test but after 10 days, and it continued until the end (Figure 1). It is worth noting that the pH trend during the summer reached its maximum around August 12, and then decreased .

Figure 1.Trend of the leaf pH at seven moments of the trial in the three treatments. (a>b…>i, P<0.05)

Based on the analyses of the NIR spectra (Table 3), it was observed that the treatments had a strong effect on the leaves, as demonstrated by the 1-VR coefficient, which reached values of 0.47 in the comparison Control vs. Salicylic and of 0.66 in the comparison Control vs. SBS-Humic. The treatments affected the confusion matrix and the correct classifications. In fact, when the Control was set at a 75% level, the classification dropped to 44% for Salicylic and 38% for SBS-Humic, although the latter was not significant (P = 0.34).

%Class = %Classification; 1-VR ≡ R-squares in cross-validation; a>b P<0.05.

This result originated from modifications in the components of the leaf blade (Table 4), with several decreases in hemicellulose (-17% and -15% for SA and SBS, respectively), crude protein (-5% and -1%) and dry matter (-1% and -2%). An increase was instead observed for the cellulose, ADF and not digestile NDF, so that the overall net effect was a rise in the Crop Maturity Index (+3% and +1%).

The average classification rate (Table 5) was 80% from the NIRS of the Litterbags, 96% from the NIRS of the rooibos and 74% from the NIRS of the green tea. The calculations based on the weight loss were explanatory of the experiment, with an average classification of 67% (statistically significant different from the random value of 33%), but inferior to the level obtained from the NIR spectra. Therefore, it could be convenient to use rooibos as a marker of soil bioactivity as it can enhance the resolution power of the experiment.

The two parameters of the TBI showed different features in relation to the acidic treatments. In fact, the regression of TBI_K on TBI_S, which was negative in the Control, became positive in SA and strongly positive in SBS (Figure 2), while at same time, TBI_S increased by 80% in SA, and the TBI_K factor increased by 116% in the SBS treatment (Table 6).

Figure 2.Regression of the TBI_K (Decomposition) on the TBI_S (Stabilization) parameters from the Teabag-Index in the three treatments (SBS-Humic, SA-Salicylic, C-Control).

The Litterbag-NIRS appeared preferable as it allowed the mechanics of the results to be interpreted on the basis of well-established equations. The results of the litterbag decomposition (Table 7) have been very instructive. The microbial k+ resistant-strategists decreased to a great extent (-13 and -28% in SA and SBS, respectively), whereas the r-labile-strategists decreased less (-8 and -4%) and their r-k difference therefore rose, especially in SBS (+19 and +115%). Another sign of a lower microbial activity in the SBS-Humic treatment was the crop maturity index, which decreased by 11%. However, the soil respiration capacity increased (SIR +9 and +5%), albeit not significantly. The most penalized variables of the litter were the type "k+ resistant" ones (protein, ash, hemicellulose), which are linked to the k-strategists, while the "r-labile" variables, which represent the main components of the wall (crude fiber, ADL, NDF not-digestible, ADF, cellulose), were tendentially increased by the treatments. Interestingly, the Mycorrhizal index increased in S (+8%) but clearly decreased in SBS (-7%). A further result was a strong increase in the predicted N content of the soil, that is, of about 17% in S and 46% in SBS, with a prevalence in the NH₄ form.

The high-degree relationships between the NIR spectra and TBI parameters were assessed through a calibration-validation process (Table 8). In fact, it was possible to predict the two TBI-parameters from all the NIR spectra, but especially from the rooibos.

The abundance of Bacteria and Fungi predicted from the NIR spectra of the green tea (Table 9) was not different, but the Taxa profiles showed marked deviations from the Control. In particular, the SBS treatment altered the Fungi composition and showed a drop of 35% in Glomeromycota, which was correlated with the reduction in the mycorrhizal index. Among the Bacteria, the most favored were Proteobateria and Actinobacteria, while Mortierellomycotaappeared to be the most penalized.

However, the regression line of the SA effect paired with the SBS effect (Figure 3) was positive for the Bacteria, whereas it was negative for the Fungi, due to the deviations of the Glomeromycota

Figure 3.Regression of the Salicylic effect (SA, X axis) on the alkaline SBS-Humic effect (SBS, Y axis) for the Taxa of Phyla Bacteria (blue) and Taxa of Phyla Fungi (red).

As far as the mineral soil N contents (NO₃^—N as well as NH₄⁺-N) regressed on the Taxa are concerned, the results (not shown in the table) enhanced the negative coefficients more than the positive ones for Fungi, with a minimum for Ascomycota (-0.045). Similarly, the minima negative taxa from Bacteria were Planctomycetota(-0.06) and Chloroflexi (-0.04), as opposed to the most positive one (Mucoromycota, +0.05).

The treatments did not modify the productions (Table 10).

The berries sampled from the treated groups (Table 11) resulted to be smaller than the Control (SA -9% and SBS -4%). In addition to the reduction in the mass of the berries, it has been observed that the treatments delayed the technological maturity by around 10%. This was due to a state of greater acidity, with a -4% reduction in pH for both treatments, albeit for the same amount of sugar. As far as the NIR spectra of the skin is concerned, a higher variability was observed in the SBS group, which was classified at 35%, a non-significant value, and was significantly lower than the other more homogeneous Control (67%) and SA (61%) groups.

The seeds of the three groups were highly and equally characterized by the NIR spectra (about 87%) (Table 12). As for the components of the seeds with similar total phenols, the non-extractables were 6% lower for SBS. This corresponds to a lower phenolic maturity of around 18%, with a prevalence of proanthocyanidines (+9%).

Therefore, the treatment with humic acids generated a maturation delay of 18%.

How is it possible to imagine that a minimal treatment with acids could affect the bioactivity of the soil so much? All soil investigation systems show differences between treatments. In our case, the couple of TBI parameters derived from weighing the residues indicated that the TBI_S (Stabilization) factor, which depended on the green tea degradation, increased by 80% in SA, while the TBI_K (Decomposition) factor, linked to the rooibos degradation, increased by 116% in SBS. According to Mori ⁴⁰,the two parameters are related in the positive mode, and greater emphasis should be placed on the TBI_S parameter, linked to green tea. Therefore, in the framework of the regression features exposed in Figure 2, the Control group with the strong negative trend should be considered an outlier, thus making the comparison results invalid. As far as the chemical composition and quality degradation of the two teas are concerned, the study by Duddigan et al.⁴¹ was based on Nuclear Magnetic Resonance Spectroscopy, where the variability depended on the relative abundance of the components that are traditionally considered labile (O-alkyl C and carbohydrate C) against more recalcitrant components (alkyl C and methoxyl C of lignin), and the tea was disregarded as a feed decomposed in a rumen, which is the approach adopted in the LBN method. The mass decay of the green tea was biphasic, with a rapid decline to 14 d, and the loss of soluble tannins, followed by a slower decline rate to 91 d, whereas the mass of rooibos tea declined at a slow rate over the whole incubation period. The percentages of decomposition at 91 d were 80% and 52% for green and rooibos tea, respectively.

Since the NIR spectra and residual weight were well correlated (Table 8), the conclusions obtained from the weight tend to converge, but the spectra, as they also reflected finer quality variations, showed higher power of discrimination. TBI only has agroecological purposes, and is only rarely used for agronomic purposes

In a framework of a multifaceted experiment, Tóth et al. ⁴² compared set-aside areas and cultivated areas in Hungary, which were cultivated using only rooibos buried for one month. The decomposition rate in the cultivated fields was 18.7%, whereas it was 25.83 (+32%) in the set-aside fields. However, in that study, the type of habitat did not affect the bacterial alpha diversity or community composition to any great extent (P = 0.122).

The Litterbag-NIRS variables predicted from the hay NIR spectra clearly indicate that SBS decreased the components of the k+ type (protein, ash, crop maturity index, NDF digestibility, hemicellulose) by about 28% and the r-type traits by about 4%. The SA treatment showed a similar but lower reduction of the k-variables (-13%), which was stronger in the r-variables (-8%). In short, the bioactivity detected from the quality of the hay litters was limited by both the SBS and SA. Our results are in agreement with those of Allison⁹ about the strong HA inhibition of enzyme activity, but are not in agreement with the results of a study on HA in pots of chicory ⁸, in which the increase in bacterial biomass was related to the availability of carbonaceous substrates for heterotrophic growth: in fact their dose threshold resulted to be 2 g kg^-1 of soil, a value that is not comparable with the amendment used for our vines, which was instead around 0.125 g per kg of vineyard soil, that is, 16 times lower.

Contrary to expectations, no significant differences were found between the cultivated and set-aside soils in the Hungarian experiment Tóth et al ⁴², regarding bacterial alpha diversity and community composition. In the present trial, we have been able to profile the reduction in bioactivity in the litters through a biodiversity pattern in the bacterial and fungal taxa, a result that was beyond any expectation.

Our data confirm the rise in mineral N levels observed by Jin et al.⁷in Panax ginseng, which amounted to 90-125% after dosing humic and fulvic acid at 0.8 g⋅Kg^{− 1} soil, an 8 times higher level than that applied in the present work. Moreover, their finding indicated that the microbial diversity of the rhizosphere soil and bulk soil was reduced after the humic acidic treatments. The results of the present work confirm the reduction in Chloroflexi, Actinobacteria and Acidobacteria.

The treatments affected the foliar composition, regardless of whether they were administered per leaf or in the soil. The most surprising result was the lowering of the leaf pH following the treatment with SBS. This physiological measure is not usual in agronomy. A first comparative analysis on forty-nine Mediterranean species in different environmental conditions ⁴³ was followed by observations on tree species located along 3300 km of a water gradient ⁴⁴ and on desert species in China ⁴⁵. From a physiological point of view, a lowering of pH can be determined by a stress condition, be it water or heat ⁴³, such as an excess of UV-A, B ²⁵ or by a reduction of the solar cycle over the years ⁴⁶. A negative relationship between artificial mycorrhization and leaf pH has been observed on corn ⁴⁷, sorghum ²⁷, and on seven crops ⁴³, in the sense that a biofertilization with mycorrhizae corresponded to a lowering of pH.

The injection of SA into grapevine berries is an elicitor of phenolics and hydrolytic enzymes that delay, or even inhibit the ripening of the berries ⁴⁸. Ranjbaran et al.⁴⁹, who sprayed SA on harvested berries,enhanced the increase of the total phenolic content of the berry skins after the shelf-life period. Similarly, in our trial, SA delayed the commercial maturity of the grape by 9%, but preserved the phenolic maturity of the seeds, and both these results could be interesting for winemakers to counteract climatic warming.

However, we cannot comment on the benefic effects obtained from massive SBS spraying on mango pointed out by Patel et al ²³ (SBS @ 1.5% or SA @ 1500 ppm) or on the advantages in grape quality pointed out by Popescu and Popescu ³.