**4. Materials and Methods**

## *4.1. Plant Materials and Sinorhizobium Fredii Inoculation*

Nodules were harvested from a cultivated soybean C08 (*Glycine max*), which is closely related to Williams 82 from Illinois, USA [48]. Seeds of C08 were surface-sterilized with chlorine gas for 16 h and germinated in the dark in sterilized vermiculite with de-ionized water in the greenhouse. At three days post-germination, seedlings were transferred to individual containers with sterilized vermiculite with 1X low nitrogen nutrient solution (6.35 μM Ca(NO3)2, 133.59 μM CaSO4, 50.30 μM KCl, 12.17 μM MgSO4•7H2O, 39.04 μM K2HPO4, 15.31 μM ferric citrate, 2.31 μM H3BO3, 0.6 μM MnSO4, 0.07 μM ZnSO4, 0.16 μM CuSO4•5H2O, 0.01 μM H2MoO4) in a 16h/8h light/dark cycle at 25–30 ◦C [49] and inoculated with *S. fredii* strain *CCBAU45436* [50]. A set of uninoculated soybean control was also prepared alongside this. The *Rhizobium* strain was cultured on TY medium [51] at 28 ◦C with shaking at 180 rpm for 40 h. The cells were then pelleted and diluted in saline (0.9% *w*/*v* NaCl) to a final concentration of 10<sup>20</sup> cells mL−<sup>1</sup> (OD600 = 0.2) for inoculation. Ten-day-old seedlings were inoculated with 1 mL inoculum per plant. On the 28th day after inoculation, UR, N, and SR samples were collected separately and frozen immediately in liquid nitrogen and stored at −80 ◦C for RNA extraction. For mitochondria isolation, uninoculated roots, nodules, and the stripped roots were harvested and kept on ice for immediate isolation [5].

## *4.2. RNA Extraction and Sequencing*

RNA was extracted using TRIzol (Life Technologies, Carlsbad, CA, USA) following the manufacturer's protocol. Three biological replicates were prepared for each of the three conditions to produce a total of nine samples. Nine strand-specific RNA-seq libraries were generated using TruSeq RNA Sample Preparation Kit (Illumina, San Diego, CA, USA). Messenger RNA (mRNA) was enriched by depleting ribosomal protein RNA using Ribo-Zero Plant kit, rather than poly-A enrichment. These libraries were sequenced on Illumina Hiseq series platforms (sequencing service provided by Groken Bioscience, Hong Kong, China).

#### *4.3. Bioinformatics Analysis*

*G. max* Williams 82 v275 reference genome was downloaded from the Phytozome database (v9.0: https://phytozome.jgi.doe.gov/). The complete soybean mitochondrion genome was downloaded

from the NCBI database (ID: NC\_020455.1) [8]. The sequences were combined into one reference genome for subsequent analysis. Genomic variations (SNPs and Indels) between C08 and the reference genome were identified using the GATK pipeline (version 4.0.5.2) [52] according to GATK best practice workflow for germline short variant discovery (https://software.broadinstitute.org/gatk/best-practices/ workflow?id=11145). Briefly, C08 DNA sequencing reads were mapped to the reference genome using BWA-MEM (version 0.7.15) implementation with default parameters. Duplicated reads were marked with GATK MarkDuplicates implementation. Then, base quality score recalibration was performed with known variation sites downloaded from the Phytozome database (v12). Then, SNPs and Indels between C08 and the reference genome were called based on mapped and quality score recalibrated reads using GATK HaplotypeCaller implementation.

RNA-seq reads were mapped to reference genome using Tophat2 (version 2.1.1) [53]. Properly mapped read pairs were assigned to each annotated gene by featureCounts from the subread package [54]. Read pair count of each mitochondrial gene was normalized by transcript length and total read count to calculate FPKM value.

To identify RNA editing sites, RNA-seq reads that could be mapped to the mitochondrial genome were extracted. SNPs were identified, and allele frequency was calculated for each polymorphic site using samtools mpileup (version 1.7) [55] and varScan2 (version 2.4.3) [56]. SNPs that were also identified by GATK with DNA reads were assumed to be germline variations and filtered. The remaining SNPs were considered as candidate RNA-editing sites.

#### *4.4. RT-PCR and Quantitative RT-PCR*

cDNAs were synthesized with M-MLV RT (200 U/μL) (Invitrogen, Hong Kong, China) and random hexamers (Invitrogen, Hong Kong, China), according to the manufacturer's instructions. To verify the differential editing of mitochondrial *matR* transcript observed in RNA-seq data, specific primers were designed to amplify regions that contain each *matR* editing site by RT-PCR (Supplementary Table S5). PCR products were sent out to BGI-Shenzhen for Sanger sequencing.

Quantitative reverse transcription PCR (qRT-PCR) analysis was carried out using the same batch of RNA samples. Primers used in qRT-PCR were derived from a previous *A. thaliana* study in which specific oligonucleotides were designed to target intron-exon and exon-exon regions [23]. New primers were designed based on the homology between *A. thaliana* and soybean mitochondrial genomes (Supplementary Table S5). SYBR Green Master Mix (ABIsystems, Hong Kong) was used in a 10 μL volume PCR reactions. Tubulin gene (Gene ID: Glyma20g27280) was used as the internal house-keeping control. The assessment of relative expression levels was calculated using the Ct comparative threshold method [15,57]. The expression levels of spliced mRNA and unspliced mRNA were first calculated and the ratio was defined as the splicing efficiency. To compare the splicing efficiencies between samples, the splicing efficiencies of SR and N were divided by that of UR (Figure 3a).

#### *4.5. Isolation of Soybean Mitochondria*

Soybean mitochondria were isolated as previously described with modifications [58,59]. All procedures were done at 4 ◦C including sample harvest and centrifugation. A total of 20 g of the UR, 20 g of SR, and 10 g of the N were sampled and ground in 50 mL grinding buffer (pH 7.5) containing 0.3 M sucrose, 25 mM Tetrasodiumpyrophosphate, 2 mM EDTA, 10 mM KH2PO4, 1.0% (*w*/*v*) PVP-40, 1% (*w*/*v*) BSA, 20 mM ascorbate and L-cysteine. After 2 min of grinding, the homogenates were filtered through a double layer of Miracloth and rinsed again with 50 mL grinding buffer and centrifuged at 4000 *g* for 5 min. The supernatant was transferred to a tube and centrifuged at 10,000 *g* for 15 min. The pellet was resuspended in a wash buffer (0.3 M sucrose, 10 mM TES, 0.1% (*w*/*v*) BSA, pH 7.5) and layered on 30 mL of wash buffer containing 45% (*v*/*v*) Percoll in a tube and centrifuged at 40,000 *g* for 30 min. The crude mitochondria located in a tight brown band near the top of the tube were transferred and diluted at least 5-fold with the wash buffer and concentrated by centrifuging at 15,000 *g* for 10 min. The pellet was resuspended in around 5 mL of wash buffer before loading to a

continuous gradient solution containing 0 to 4.4% (*v*/*v*) PVP-40 and 28% (*v*/*v*) Percoll in the wash buffer. After centrifugation at 40,000 *g* for 30 min, the mitochondria were concentrated in a pale-yellow band located near the bottom of the tube. This layer was then transferred to a new polycarbonate centrifuge tube with the BSA-free wash buffer and centrifuged at 2450 *g* for 15 min. After 3-4 wash steps, the mitochondria pellet was resuspended in the wash buffer without BSA. After the protein concentration was determined by the Bradford protein assay (BIO-RAD, Hercules, CA, USA), the mitochondria were stored in aliquots at −80 ◦C.

## *4.6. Blue Native-Polyacrylamide Gel Electrophoresis (BN-PAGE) and In-Gel Enzyme Activity Staining*

The mitochondrial protein complex extraction and BN-PAGE were carried out as previously published with modifications [60]. An equal amount of mitochondria was collected by centrifugation at 14,300 g for 10 min at 4 ◦C and resuspended in 5% (*w*/*v*) digitonin extraction buffer to a final ratio of 10:1 (*w*/*v*) of protein to detergent and incubated on ice for 20 min. The solubilized proteins were then centrifuged at 18,300 g for 20 min at 4 ◦C. The supernatant (100 μg per sample) was transferred to a new tube supplemented with 5% (*v*/*v*) Serva blue G250 solution by a final ratio of 100:1 (*w*/*v*) of protein to dye and was loaded to a standard 1.0 mm <sup>×</sup> 10 well NativePAGETM 3–12% Bis-Tris Gel (Invitrogen, Hong Kong, China). The cathode buffer (50 mM Tricine, 15 mM bis-Tris, 0.02% (*w*/*v*) Serva Blue G250, pH 7.0) and the anode buffer (50 mM Bis-Tris, pH 7.0 with HCl) were freshly prepared. The electrophoresis was carried out at 4 ◦C at 75 V for 30 min, followed by 100 V for 30 min, 125 V for 30 min, 150 V for 1 h, 175 V for 30 min, and then set to a constant voltage 200 V until the sample reached the bottom of the gel.

After electrophoresis, the gels were washed twice with MiliQ water for 10 min. Then, the gels were equilibrated in the appropriate reaction buffer without reagents for 10 min. The gel was then incubated in a fresh reaction buffer of complex I (0.1 M Tris, 0.2 mM NADH, 0.2% (*w*/*v*) nitro-blue tetrazolium, pH 7.4) for 30 min. The other gel was incubated in a fresh reaction buffer of complex II (50 mM KH2PO4, 0.1 mM ATP, 0.2 mM Phenazine methosulphate, 10 mM succinate, 0.2% (*w*/*v*) nitro-blue tetrazolium) for 2 h. The reactions were terminated by fixing the gels in 40% (*v*/*v*) methanol and 10% (*v*/*v*) acetic acid for at least 1 h. The gels were destained overnight in 20% (*v*/*v*) methanol to remove residual Serva Blue G. For Western blot analysis, 10 μg mitochondrial proteins were run into SDS-PAGE and the proteins were transferred to Hybond-P nitrocellulose membranes (GE Healthcare, Hong Kong, China). The following antibodies were used: anti-beta subunit of ATP synthase (ATP4, PhytoAB PHY0587S, 1:1000; ATPβ, Agrisera AS05 085, 1:4000); anti-cytochrome oxidase subunit II (COXII, Agrisera AS04 053A, 1:2000); anti-cytochrome oxidase subunit III (COXIII, PhytoAB PHY0580S, 1:1000); anti-NAD4 (PhytoAB PHY0511S, 1:1000); anti-NAD9 (from Dr. G. Bonnard, 1:50,000); anti-51kDa (PhytoAB PHY0525S, 1:1000). The signals were developed by the Enhanced Chemiluminescence method (ECL; GE Healthcare, Hong Kong, China).

**Supplementary Materials:** The following are available online at http://www.mdpi.com/1422-0067/21/24/9378/s1. Figure S1: Sanger sequencing of three biological replicates of each sample. Table S1: Normalized read count for each mitochondrial transcript in the three tissues. Table S2. RNA edited sites in the mitochondrial genome identified in this study. Table S3. Differentially edited RNA editing sites. Table S4: Expression of nuclear intron maturases in the three tissues. Table S5. Primers used in this study.

**Author Contributions:** B.L.L. and H.-M.L. designed and coordinated the study; Y.S. led the experimental works related to Sanger sequencing intron splicing, isolation of mitochondria, in-gel activity assays, and Western blotting; M.X. performed bioinformatic analysis of the RNA and DNA sequencing data; Z.X. participated in the in-gel activity assays; K.C.C. and J.Y.Z. participated in the isolation of mitochondria and Western blotting; K.F. and J.W.-B. grew the soybean, performed inoculation, and prepared the RNA samples; B.L.L., H.-M.L. and Y.S. wrote the manuscript. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by the Hong Kong Research Grants Council Area of Excellence Scheme (AoE/M-403/16), and the Innovation and Technology Fund (Funding Support to State Key Laboratory of Agrobiotechnology) of the HKSAR, China. Any opinions, findings, conclusions, or recommendations expressed in this publication do not reflect the views of the Government of the Hong Kong Special Administrative Region or the Innovation and Technology Commission.

**Acknowledgments:** We thank Zeta Mui for her help in sample preparation. Man-Wah Li and Qianwen Wang assisted with the analysis of genomic and transcriptome data.

**Conflicts of Interest:** The authors declare no conflict of interest.
