Phenolic acid-degrading Paraburkholderia prime decomposition in forest soil

Plant-derived phenolic acids are catabolized by soil microorganisms whose activity may enhance the decomposition of soil organic carbon (SOC). We characterized whether phenolic acid-degrading bacteria enhance SOC mineralization in forest soils when primed with 13C-labeled p-hydroxybenzoic acid (pHB). We further tested whether pHB-induced priming could explain differences in SOC content among mono-specific tree plantations in a 70-year-old common garden experiment. pHB addition primed significant losses of SOC (3–13 µmols C g−1 dry wt soil over 7 days) compared to glucose, which reduced mineralization (-3 to -8 µmols C g−1 dry wt soil over 7 days). The principal degraders of pHB were Paraburkholderia and Caballeronia in all plantations regardless of tree species or soil type, with one predominant phylotype (RP11ASV) enriched 23-fold following peak pHB respiration. We isolated and confirmed the phenolic degrading activity of a strain matching this phylotype (RP11T), which encoded numerous oxidative enzymes, including secretion signal-bearing laccase, Dyp-type peroxidase and aryl-alcohol oxidase. Increased relative abundance of RP11ASV corresponded with higher pHB respiration and expression of pHB monooxygenase (pobA), which was inversely proportional to SOC content among plantations. pobA expression proved a responsive measure of priming activity. We found that stimulating phenolic-acid degrading bacteria can prime decomposition and that this activity, corresponding with differences in tree species, is a potential mechanism in SOC cycling in forests. Overall, this study highlights the ecology and function of Paraburkholderia whose associations with plant roots and capacity to degrade phenolics suggest a role for specialized bacteria in the priming effect.

(DEPC) for inactivation of RNase activity. Briefly, ~0.5g soil was transferred to 2-mL Lysing Matrix E tubes (MP Biomedicals) and 0.5 mL CTAB (hexadecyltrimethylammonium bromide) buffer (10% CTAB in 0.7 M NaCl, 240 mM potassium phosphate buffer, pH 8) and 0.5 mL phenol (pH>7.8) : chloroform:isoamyl alcohol (25:24:1) were added. Tubes were homogenized for 45 sec in a Mini-Bead Beater-8 instrument (Biospec Products), and then centrifuged at 16,000 g for 5 minutes at 4°C. The aqueous layer was transferred to a new 1.5-mL microcentrifuge tube containing an equal volume of chloroform:isoamylalcohol (24:1), vortexed, then centrifuged again. The aqueous layer was transferred to a new tube containing two volumes of 30% (wt/vol) polyethylene glycol 6000/1.6 M NaCl. RNA/DNA was precipitated at 4°C overnight. RNA/DNA pellets were collected by centrifugation (10 min, 16,000 g, 4°C), washed with ice-cold 70% ethanol, air-dried, and resuspended in 30 µl nuclease-free water. Total nucleic acid concentrations were determined using a Qubit® 3.0 fluorometer (ThermoFisher Scientific) and extracts were stored at -80°C. 16S rRNA gene and 16S rRNA libraries were constructed from extracts from the soil priming experiment, while only 16S rRNA gene libraries were constructed for the in situ SIP experiment.

Density gradient ultracentrifugation and recovery of 13 C-labeled DNA
Briefly, 5 µg of total nucleic acid were diluted to 1.20 mL with sterile GB buffer (0.1 M Tris, 0.1 M KCl, 1 mM EDTA), mixed with 4.80 mL of 7.163M cesium chloride, and loaded into ultracentrifuge tubes which were then heat-sealed. Tubes were centrifuged at 140,000 g (41,900 rpm; Vti80 rotor) for 66 hours at 20°C. CsCl density was determined based on the refractive index of solution measured with an AR200 digital refractometer (Leica Microsystem). Reagent blanks (no added DNA) and a positive controls were processed identically, the latter consisted of 5 µg each of 12 C-and 13 C-DNA, prepared from cultures of Pseudomonas putida G7 grown in with either 12 C-or 13 C6-glucose as the sole carbon source (Sigma-Aldrich, St. Louis, MO).

Characterizing pHB-degrading activity of isolates
Nine isolates from the THFP (7 from RPL and 2 from SML soils) were tested for their ability to degrade pHB. Briefly, one colony of each isolate was transferred to test tubes containing 5 ml of either: 1) MSM or 2) MSM + 3 mM pHB. Uninoculated tubes were included as controls. Tubes were incubated for 3 days at 20 °C with gentle shaking (150 rpm) and growth was measured periodically by UV absorbance at 600 nm (Spectronic 21 spectrophotometer, Bausch & Lomb).
After the 3-day incubation, cultures were analyzed by GC/MS for the production of protocatechuate (aerobic metabolite of pHB degradation) and for the presence of residual pHB. Briefly, 5 mL of each culture was extracted twice with 5 mL of ethyl acetate. Extracts were dried over sodium sulfate and concentrated to 300 µL using a TurboVap LV evaporator (Zymark). Next, 100 µL of extract was derivitized with 25 µL of N,O-Bis(trimethylsilyl)trifluoroacetamide (BSTFA) and 1 µL of sample was injected into a Hewlett-Packard Model 6890 gas chromatograph equipped with an HP-5 fused silica capillary column (5% phenylmethyl silicone, 30 m x 0.25mm x 0.25 μm film thickness; Hewlett-Packard) connected to a Hewlett-Packard Model 5973 quadrupole mass selective detector (operated at an electron energy of 70 eV and a detector voltage of 2000-3000). A splitless injection was used with a 1-min delay before septum purge. Helium was the carrier gas (linear gas velocity of 30 cm/s). The injector and detector temperatures were 250 and 300°C, respectively. The ion source pressure was maintained at 1.0 x 10-5 Torr. The oven temperature program began at 50°C and increased to 250°C at 10 degree/min.

Quantifying expression of pobA
The two paralogs of p-hydroxybenzoate 3-monooxygenase genes (pobA1 and pobA2) from the genome of P. madseniana RP11 T were used to design four primer sets (two per gene) using the NCBI Primer-BLAST tool [16]. The efficacy of each primer set was tested in qPCR assays using mRNA extracted from soil microcosms (RPLnC) dosed with pHB. We found that only the pobA1 transcripts increased in abundance following the addition of pHB, which did not occur in wateronly controls. Total 16S rRNA gene expression and pobA1was consistent between the pHB-dosed microcosms and water-only controls. The chosen pair targeted a 75-bp fragment of the pobA1 gene identified: pobA1_f (5'-AGA TCG AAT CCA CCA TCC GC-3') and pobA1_r (5'-TTC AAA GCC GTG ATG CAA CG -3').
Prior to reverse transcription, DNA was eliminated from RNA extracts by treatment with DNAse I (Invitrogen). Resulting cDNA was PCR reactions contained 1X Power SYBR qPCR Master Mix (Applied Biosystems), 0.5 µM of forward and reverse primer, and 1 µL template. Thermal cycling conditions were as follows: 95°C for 10 minutes, 45 cycles of 95°C for 15 seconds and 58°C for 1 minute, and a 7-min extension at 72°C. A standard curve was generated by diluting PCR amplified pobA from RP11 T genomic DNA, spanning 2.76 x 10 2 to 2.76 to 10 7 copies per uL. Our pobA1 primer set covered approximately 40% of the genes present in clade 6 with a maximum of a single base mismatch, while our custom HMMs models (available in Supplementary Data) were 100% accurate and identified six additional homologs missed using BLAST homology searches.

Discrepancies in ASV Naming
Due to differences in trimming parameters used in the processing of 16S rRNA gene libraries from the soil priming experiment versus in situ DNA-SIP experiment, the ASV identifiers differ between the two datasets. For example, the ASV for P. madseniana RP11 T is '5b18237cf385326c2cda6c3441ba8d40' and '6b64ff37fff389a42eccf7387ffac8b9', respectively. The first 5 nt from 515F region of V4 were truncated during the processing of the soil priming amplicon data. The corresponding ASV names for all pHB-degraders is available in Table S3, based on 100% identity across the full amplicon sequences.

Evolutionary reconstruction of pobA paralogs
The phylogeny of all Burkholderiaceae was determined based on a multi-locus sequence alignment (MLSA). All Burkholderiaceae genomes were downloaded from NCBI refseq_genomic on June 20 th , 2019 (NCBI taxid 119060; n = 3,991). Open reading frames were predicted with Prodigal (v. 2.6.4) [9] and the translated peptide sequences for a common set of 24 single-copy genes were identified using HMMs from BUSCO [17]. Peptide sequences were independently aligned with MAFFT (algorithm: FFT-NS-2 & progressive method) [18] and then concatenated into the MLSA containing 13,366 positions and a total of 3,559 genomes. A maximum likelihood tree was built from the MLSA using RAxML (nbootstraps = 100) with GTR substitution and gamma heterogeneity models [19]. The 'isolation source' for each genome was recovered from the NCBI BioSample database.
The phylogeny of all pobA encoded by Burkholderiaceae was determined using similar methods. pobA were recovered from genome according to a DIAMOND BLASTp [20] homology of >= 40% identity across 95% of full-length of the pobA identified in Paraburkholderia madseniana RP11 T (pobA1 and pobA2). An alignment and maximum likelihood tree were built using the previously described software and methods. The tree was divided into seven clades for which HMMs were built using hmmbuild [11] from HMMER with the input MAFFT alignments for each clade. The hmm profiles were concatenated and are provided in the Supplementary Data package. The accuracy and sensitivity of each hmm was validated against all Burkholderiaceae genomes. The phylogenetic conservation of pobA paralogs was assessed using consenTRAIT based on mapping genomes encoding paralogs onto the MLSA tree and contrasting against the distribution of other dioxygenase genes involved in degrading aromatics: gentisate 1,2 dioxygenase, benzoate 1, 2-dioxygenase, catechol 1, 2-dioxygenase and protocatechuate 4,5dioxygenase [21]. The phylogenetic dispersion for the same genomic traits was assessed using Purvis and Fritz's D [22] based on the same MLSA using the 'phylo.d' function in the R-package caper [23]. These analysis of phylogenetic conservation and dispersion results are presented in the Supplementary Extended Results section.

Assessing function of pobA paralogs
A structural analysis of the pobA paralogs in the RP11 T genome was performed based on a TCoffee 'Expresso' alignment [24] using homologs whose structure and function have been elucidated [25][26][27]. The analysis was performed to determine whether both contained the correct active and substrate-binding site residues required for p-hydroxybenzoate 3-monooxygenase activity.

Comparing pobA genes to SIP-lignin study
A targeted assembly of pobA in metagenomes from a related SIP-lignin [10] was performed using megaGTA [28] and our custom pobA HMMs (accessions in Table S6). The taxonomy of assembled pobA was determined using the Lowest Common Ancestor algorithm implemented by MEGAN [29] based on DIAMOND BLASTp searches against the NCBI 'nr' database (downloaded Feb. 3rd, 2017).

Soil Priming Experiment
A soil microcosm experiment was performed to link the activity of phenolic-acid degrading populations with soil priming. Soil treatments were aimed at identifying differences according to substrate: (i) water-only (baseline), (ii) 13 C-glucose and (iii) 13 C-pHB, or the role of pobA and Paraburkholderia sp. RP11 T activity: (iv) 13 C-pHB + pobA inhibitor, (v) pobA inhibitor-only control, (vi) 13 C-labeled RP11 T cells + dilute 12 C-pHB and (vii) unlabeled RP11 T cells + dilute 13 C-pHB (overview in Figure S2).
For each treatment, the sum total of mg C per gram dry weight soil required for all replicates was weighed in bulk on an analytical scale (0.1 mg accuracy) and dissolved in the correct volume of water needed to aliquot four replicates. p-hydroxybenzoic acid (pHB) and methyl 4-hydroxy-3iodobenzoate ('m-pHB', the pobA inhibitor) were dissolved with the assistance of sonication at 60 °C for approximately 30 minutes. A saturated solution of m-pHB (0.486 mM) was added to soils, resulting in slight differences in the total mg C of m-pHB amended. This was done to maximize inhibitory strength, which was accounted for by supplying the same amount to controls. PobA activity was inhibited with methyl 4-hydroxy-3-iodobenzoate (Fisher Scientific) at a saturating concentration (0.48 mM), which fully inhibited growth of RP11 T for 72 hrs in mineral salts media containing 25 mM pHB as the sole carbon source (see Supplementary Results).
The individual amendments of glucose (cat #: G-8270; Sigma-Aldrich) and pHB (cat #: H-3766; Sigma-Aldrich) were diluted to 17.5 atom % 13 C using corresponding compounds with natural abundance in order to provision enough C for an amendment rate of 0.5 mg C per g dry wt soil. The final isotopic concentration of amended solutions was verified using EA-IRMS at the Cornell Stable Isotope Lab. Prior to EA-IRMA, solutions were diluted with 100 mg/mL of 12 Cglucose to prevent saturation of the detector. Final atom % 13 C values for glucose (17.46 % 13 C) and pHB (17.23 % 13 C), matching desired labeling concentrations, were obtained.
Soils were amended with Paraburkholderia madseniana RP11 T cells to isolate its role in priming. Cells were added in solution with dilute 13 C-labeled pHB (25 µg C per g dry wt. soil) to stimulate pHB-degrading activity prior to amendment. Two treatments were necessary to isolate the mineralization of native SOC (i.e., priming). To account for CO2 derived from RP11 T biomass, cells were grown in MSM with 13 C-glucose (99% atom 13 C; Sigma-Aldrich) as the sole carbon source (5% w/v) to label cells. The 13 C-labeled cells were amended to soil with dilute, unlabeled pHB. Conversely, to account for the respiration of the dilute pHB amendment, unlabeled RP11 T cells were grown in MSM with unlabeled glucose and amended with the same 13 C-pHB amended in the other pHB treated soils, except diluted 1:20. RP11 T cultures were grown for 30 hrs on 0.2% glucose ( 13 C and 12 C). Cells were washed by two cycles of pelleting (3,000 rcf for 20 min) and resuspension in saline solution (0.85%). Plate counts were performed on the washed cell prior to equally dividing cell biomass into three portions to treat soil from each ecoplot. Cells were pelleted and resuspended in dilute pHB in a volume corresponding to the desired water holding capacity of each soil. The resuspended cell slurry was incubated for 20 minutes to stimulate activity prior to amendment. A total of 1.8 x 10 7 cells per g dry wt soil were added for each treatment.