Impact of agricultural management on bacterial laccase-encoding genes with possible implications for soil carbon storage in semi-arid Mediterranean olive farming

Experimental design

The main characteristics of the experiment and soils are described in Castro et al. (2008) and Moreno et al. (2009). The study was conducted in Jaen in south-east Spain (Supplemental Information 1), with a random block experimental plot design consisting of four treatments and four replicates. Each plot consisted of 16 olive trees, with the central four monitored trees surrounded by guard rows. The experiment, involving bare vs. covered soils, was begun in 1976 and was carried out over a period of 30 years. Samples were collected in May 2011. The treatments tested were:

• Tillage (T): Bare soils where weeds were eliminated by 3–4 annual passes with a disk harrow (at 30 cm deep) and/or a cultivator in spring, followed by a tine harrow in summer.

• Non-tillage and no-cover (NC): Bare soils where weeds were eliminated by applying the pre-emergence herbicide oxyfluorfen in autumn. In the spring, glyphosate was applied locally.

• Cover crop + herbicides (CH): Covered soils where weeds were left to grow each year to be eliminated in spring with herbicide glyphosate.

• Cover crops + mower (CM): Covered soils where weeds were left to grow each year to be eliminated in spring with various passes of a chain mower.

Two samples were collected from the centre of each plot, using a modified soil-sample ring kit (Eijkelkamp), which consists of a 20 cm-deep sampling cylinder specifically manufactured for this purpose, and then bulked. To isolate the nucleic acids, subsamples of fresh soil were immediately frozen in liquid nitrogen. The samples were stored at −80 °C until molecular analyses were carried out.

Humic substances and isoelectric focusing

In this study, humic C was measured in sodium pyrophosphate extract (pH 7.1), a good extractant of humic compounds (Nannipieri et al., 1974; Piccolo et al., 1998). Humic substances were extracted at 37 °C for 24 h under shaking, using Na₂P₄O₇ (0.1 M, pH 7.1) in a 1:10 w/v dry soil:extract ratio (Ceccanti et al., 1978). The suspension was then centrifuged at 12,400 g and filtered through a 0.22 μm Millipore membrane. The > 10⁴ Da humic-derived fraction was obtained through ultrafiltration on the AMICON PM10 membrane of the extract. The C content of humic C (> 10⁴ Da) was determined by acid digestion with 0.17 M K₂Cr₂O₇ and concentrated (97%) H₂SO₄ at 140 °C for 2 h. A spectrophotometric method was used to quantify the Cr³⁺ produced by the reduction of Cr⁶⁺ (λ = 590 nm; Yeomans & Bremner, 1988).

Isoelectric focusing (IEF) of humic C (> 10⁴ Da) was carried out in cylindrical gel rods (0.5 × 8 cm) containing polyacrylamide gel (5% w/v) and carrier ampholines in the pH 4–6 range (Bio-Rad Laboratories, Richmond, CA, USA) at a final concentration of 2% (Ceccanti et al., 1986; Ceccanti, Bonmati-Pont & Nannipieri, 1989). N,N,N′,N′-Tetramethyl-1,2-diaminomethane and ammoniumperoxy-disulfate were also added in a gel solution at 0.03%. At the top of the gel rod (cathode), 100 μl of humic C (> 10⁴ Da) at 4.4% of glycerine were applied. A small amount (5 μl) of glycerine at 2.2% was added to the top of the sample to avoid mixing with the cathodic solution (NaOH 0.02 N); 0.01 M H₃PO₄ was used for the anodic cell. A pre-run of 1 h at the same current intensity and voltage used for the samples run was carried out for each gel tube (1.5 mA for each tube, 100–800 V); the samples run was then carried out for 2 h or longer until stable IEF banding was reached. The electrophoretic bands were scanned using a Bio-Rad GS 80 densitometer, giving a typical IEF profile for each soil investigated. The IEF peak area was determined for each soil IEF profile, assuming as 100% the area under the entire IEF profiles (representative of the total loaded C). Gel pH was measured at 0.5 cm intervals with an Orion microprocessor (model 901, Orion research) connected to a microelectrode gel-pHiler (Bio-Rad Laboratories, Richmond, CA, USA).

Aromatic hydrocarbons

Aromatic hydrocarbons were extracted ultrasonically from 5 g of freeze-dried soils using 10 ml of hexane/acetone (1:1 v/v). The 50 μl of C3-naphthalene in dichloromethane was added as a GC internal standard. After centrifugation and concentration, the solvent was dried by passing it over anhydrous sodium sulfate. The aromatic hydrocarbons were analyzed using a large volume injection-gas chromatography mass spectrometry (LVI–GC–MS) method. GC analyses were conducted on a Varian 450-GC fitted with a 1079 injector in split/splitless mode, a CTC Analytics CombiPal refrigerated autosampler and a Varian 240 Ion Trap as a mass spectrometer detector. The samples were injected in large volume mode using the 1079 PTV injector. The 10 μl were injected with the split open (split ratio of 1/100) for 0.5 min to purge the solvent, the split was then closed for 3 min, with the injector temperature ramped up, and the split was finally opened again (split ratio of 1/50). The injector temperature started at 60 °C for 0.5 min and then ramped up by 100 °C/min to 280 °C for 8 min. A Thermo TG-5SILMS fused silica capillary column (30 m × 0.25 mm × 0.25 μm) was used. Temperature ramp: 70 °C (3.5 min) > 25 °C/min to 180 °C (10 min) > 4 °C/min to 300 °C (13 min). The 10 μl of the extracts were injected according to the LVI method. The carrier gas was He at 0.9 ml min⁻¹. Electron-impact ionization and detection in full scan (50–500 m/z) mode were used. Structural and quantitative information were obtained by comparing the mass spectra and retention times of the samples and standards with the aid of MS Workstation 6.9.1 software.

Nucleic acid isolation from soils and cDNA synthesis

For each soil-sample replicate, the total community-DNA was separately extracted from four 1 g subsamples using the bead-beating method by following the manufacturer’s instructions for the MoBioUltraClean Soil DNA Isolation kit (MoBio Laboratories, Solana Beach, CA, USA). The extracts were pooled and further concentrated at 35 °C to a final volume of 20 μl using a Savant Speedvac® concentrator.

Total RNA was extracted from four 2 g subsamples of each replicate following the manufacturer’s instructions for the MoBio RNA PowerSoil Total RNA Isolation kit (MoBio Laboratories, Solana Beach, CA, USA). To remove residual DNA, DNase I enzyme was added using the RNase-Free DNase Set (Roche Applied Science, Penzberg, Germany) following the manufacturer’s instructions. The extracts were pooled and further concentrated at 35 °C to a final volume of 80 μl using a Savant Speedvac® concentrator. The cDNA was synthesized from 1–2 μg of total DNase-treated RNA using the Transcriptor High Fidelity cDNA Synthesis Kit according to the manufacturer’s instructions (Roche Applied Science, Penzberg, Germany). The synthesis reaction was carried out at 50 °C for 30 min. The concentration and quality of the final DNA/RNA/cDNA samples were checked with the aid of a Nanodrop® ND-100 spectrometer (Nanodrop Technologies, Wilmington, DE, USA).

Quantitative real-time PCR assays

Real-time PCR quantification of DNA/cDNA for bacterial functional LMCO-encoding genes and transcripts was based on amplification with primers Cu1AF (5′-ACM WCB GTY CAY TGG CAY GG-3′) and Cu2R (5′-G RCT GTG GTA CCA GAA NGT NCC-3′), designed by Kellner et al. (2008). Agarose-gel electrophoresis was carried out to confirm the specific PCR product, resulting in a single band of the expected size (approximately 144 bp), with primer sequences excluded. To determine whether the primers were specific for amplification and detection of bacterial LMCO genes and transcripts of the soils tested, an LMCO gene-clone library was established after PCR of soil DNA samples. A total of 46 sequences were analysed: 38 showed homology to LMCO-genes, 6 showed homology to the cloning vector, 2 were badly sequenced and 1 did not show homology to the LMCO-genes (Supplemental Information 2).

Each 21-μl PCR reaction contained between 4 and 5.5 ng of DNA or between 0.2 and 0.3 μg of cDNA, 10.5 μl 2× iQ SYBR Green Supermix (Bio-Rad, Munich, Germany) and 400 nM of each primer. For each DNA/cDNA extracted, real-time PCR experiments were carried out three times, with the threshold cycle (Ct) determined in triplicate. The real-time PCR program consisted of 2 min at 50 °C for carryover prevention, 5 min at 94 °C for enzyme activation, followed by 5 cycles consisting of 94 °C for 15 s, 57–54 °C for 15 s, while decreasing the temperature by 0.5 °C in every cycle, 72 °C for 30 s, and 30 cycles consisting of 94 °C for 15 s, 54 °C for 15 s and 72 °C for 30 s; the fluorescence signal was measured at the 72 °C step for both DNA and cDNA. The final step lasted 7 min at 72 °C. The PCR amplification procedure was checked using a heat dissociation protocol (from 70 to 100 °C) following the final PCR cycle. The LMCO-encoding gene copy number was quantified on an iQ5 thermocycler using iQ5-Cycler software (Bio-Rad, Munich, Germany).

A standard curve was generated by using a recombinant plasmid containing one copy of the LMCO gene from soil bacteria. The curves were drawn according to the method described by Moreno et al. (2013). The relationship between the threshold cycle (Ct) and target-gene copy number and the copy numbers of the real-time standard was calculated as described by Qian et al. (2007) using the formula Ct = −3.089 × log₁₀ (LMCO) + 30.254, R² = 0.984. Target molecules were linear from 10⁴ to 10⁹ copies.

The potential presence of qPCR inhibitors was tested by mixing 1 μl (4–8 ng) of soils DNA extracts or 1 μl (50–350 ng) of soils cDNA extracts with a known amount of recombinant plasmid DNA (pCR®2.1, Invitrogen, Carlsbad, CA, USA) with the appropriate primers. Controls, where DNA and cDNA templates were replaced by filter-sterilized milliQ water, were carried out simultaneously. Ct values did not significantly differ between the DNA/cDNA extracts and the controls.

PCR-DGGE analyses

The community structure of bacterial LMCO-encoding genes and transcripts was studied by DGGE analysis, for which Cu1AF-DGGE and Cu2R primers were used. Primer Cu1AF-DGGE contained the same sequence as Cu1AF but with an additional 40-nucleotide GC-rich sequence (GC clamp) at the 5′ end of the primer.

DGGE analyses were conducted using 550 ng of the final PCR product loaded onto a 45–60% urea-formamide–polyacrylamide gel. An INGENYphorU System (Ingeny International BV, The Netherlands) was run at 200 V for 10 min followed by 75 V for 18 h at 58 °C in order to separate the fragments. Gels were silver stained with the Bio-Rad Silver Stain according to the standard DNA-staining protocol and photographed under UV light (λ = 254 nm) using a UVItec Gel Documentation system (UVItec Limited, Cambridge, UK).

Data analyses

Results are the means of 12 replicates (three per plot) for chemical analyses. With regard to nucleic acids, four extractions of each sample were carried out and then grouped together to obtain a single sample per plot. The results were subjected to a factorial analysis of variance (ANOVA) using the STATISTICA software program (StatSoft Inc., Tulsa, OK, USA). The post hoc Tukey HSD test in a one-way ANOVA was used. P-values of under 0.05 were considered to be evidence of statistical significance.

Band patterns in different DGGE lanes were compared with the aid of UVImap Analysis software (UVItec Limited, Cambridge, UK). DGGE banding data were used to estimate the Raup & Crick (1979) probability-based index of similarity (SRC). Raup and Crick’s (1979) probability method was used to determine whether similarities within and between samples were stronger or weaker than the randomized prediction. For this purpose, band-matching data were stored as a binary matrix and analyzed using the aforementioned probability-based index. The SRC index measures the probability of the randomized similarity being greater than or equal to the similarity actually observed; SRC values of over 0.95 or under 0.05 indicate similarity or dissimilarity, respectively, which are not random assortments of the same species (bands or OTUs) (Rowan et al., 2003). SRC and cluster analyses were carried out using the PAST (Paleontological Statistics) software program version 1.82b (Hammer, Harper & Ryan, 2001).

Organic carbon fractions

We found that the high-molecular-weight (C > 10⁴ Da) humic acid (HA) fractions behaved in a significantly different way from total SOC (Table 1). HA levels were much higher in CM, comparable in CH and NC and considerably lower in the T treatment. Figure 1 shows the IEF patterns of HAs. All the treatments showed the highest optical-density values (IEF peaks) in the pH range 4.5–4.2 (bands 1 and 2).

Aromatic hydrocarbons

The results of the analysis of major aromatic hydrocarbons are shown in Table 2. Three main polycyclic aromatic hydrocarbons were detected in soil samples. 2,6-diisopropylnaphthalene (m/z = 239 + 199) was found to be present in all soils, although values were close to the detection limit in the CM treatment. The levels of pp-DDA, a polar metabolite of DDT also found in all soils, were significantly higher in the NC treatment. The third compound, present only in the NC and CH soils, was a mixture of four aromatic hydrocarbons (m/z = 265 + 266 + 331), whose mass spectra did not match any compounds in the library. To gain a better understanding of this mixture, its mass spectra are shown as supplementary material (Supplemental Information 3). The data appear to suggest that the C and D compounds, containing an odd number of N atoms, are dimers of A and B, respectively.

Real-time PCR assays and LMCO gene expression

Table 3 shows LMCO-encoding gene and transcript-copy numbers in the four treatments, both of which were much higher in CM soil and comparable in the other treatments.

LMCO-encoding gene and transcript diversity

Figure 2 shows the DNA- and RNA-based DGGE dendrograms generated by Raup & Crick (1979) cluster analyses (see Supplemental Information 4). The SRC values, obtained by comparing the four treatments, are also summarized in Fig. 2. With respect to LMCO-encoding genes (Fig. 2A), T and NC soils differed significantly (SRC < 0.05). For the other pairs of DNA-DGGE profiles compared, the similarity was no greater than that the random prediction (0.95 > SRC > 0.05). However, SRC values comparing CM and the other treatments were very close to the dissimilarity boundary (SRC < 0.05), as evidenced by the solitary cluster where these soils were located.

When the LMCO transcripts were evaluated (Fig. 2B), T and NC soils clustered together, showing a high degree of similarity (SRC ≈ 1). CH and CM soils showed low similarity (< 10%). After comparing CM with the other soils, the similarity was found to be no greater than that randomly predicted (0.95 > SRC > 0.05).