Qiuyan Ren,Hua Jiang,Wenyang Xiang,Yang Nie,Song Xue,Haijian Zhi,Kai Li*,Junyi Gai*
MARA National Center for Soybean Improvement/National Key Laboratory for Crop Genetics and Germplasm Enhancement/MARA Key Laboratory of Biology and Genetic Improvement of Soybean,Nanjing Agricultural University,Nanjing 210095,Jiangsu,China
Keywords:MADS-box gene Soybean mosaic virus (SMV)Expression analysis Virus-induced gene silencing (VIGS)Overexpression
ABSTRACT Soybean mosaic virus (SMV) is a member of the genus Potyvirus that extensively impairs global soybean production.The full-length coding sequence of the MADS-box transcription factor GmCAL was cloned from the SMV-resistant soybean cultivar Kefeng 1.SMV-induced expression analysis indicated that GmCAL responded quickly to SMV-SC8 infection in Kefeng 1 but not in NN1138-2.GmCAL was expressed at high levels in flowers and pods but at lower levels in leaves.The gene was localized to the nucleus by subcellular localization assay.Virus-induced gene silencing did not increase the accumulation of SMV in GmCAL-silenced Kefeng 1 plants(with silencing efficiency~80%)after SC8 inoculation.GmCAL-silencing plants still conferred resistance to SC8 that might be owing to incomplete silencing of genes with lower expression.SMV content decreased significantly in GmCAL-overexpressing NN1138-2 plants after SMVSC3,SMV-SC7,and SMV-SC8 inoculation in comparison with a vector control,showing that overexpression of GmCAL conferred broad-spectrum resistance to multiple SMV strains.These results confirm that GmCAL,a key regulator but not a specific SC8 resistance gene(Rsc8),is a positive regulatory transcription factor involved in soybean resistance to SMV.
Soybean[Glycine max(L.)Merr.]is a major world source of edible vegetable oil and protein.More than 30 diseases pose severe threats to soybean production [1].Among them,Soybean mosaic virus (SMV) disease can severely damage soybean seed yield and quality [2].Developing SMV-resistant cultivars is the most economical,effective,and eco-friendly approach to managing the disease.To achieve this,knowledge of the resistance mechanism to SMV in soybean is essential.
Owing to the long coevolution between the pathogen SMV and its host,as well as to differentiation in SMV pathogenicity,numerous SMV strains with diverse virulences have formed.Based on the pathogenicity difference of SMV isolates in differential hosts,five SMV strains have been reported in Japan [3],seven in the USA[4]and 21 in China,and six new strains in the USA and the Republic of Korea[5–7].Among them,SC3,SC7,and SC8 are the predominant strains [8].In the U.S.,four dominant disease resistance genes:Rsv1,Rsv3,Rsv4,and Rsv5,have been identified [9–10].Rsv1 on chromosome 13 in accession PI96983,confers resistance to SMV strains G1–G6 but not G7 [11],Rsv3 on chromosome 14 was susceptible to strains G1–G4 but resistant to strains G5–G7[12],and Rsv4 on chromosome 2 was resistant to strains G1–G7.A new gene symbol Rsv5 on chromosome 13 was assigned to the resistance gene in ‘York’ to replace the old allele name Rsv1-y[13].Ishibashi et al.[14] reported the positional cloning of a broad-spectrum SMV-resistance locus Rsv4,which encodes an RNase H-family protein with dsRNA degradation activity.An active-site mutant of Rsv4 did not inhibit virus reproduction and was associated with an active viral RNA polymerase complex in infected cells.These results indicate that Rsv4 enters the viral replication compartment and degrades viral dsRNA.Tran et al.[15]reported that transient silencing of the candidate gene Glyma.14 g204700 caused high accumulation of an avirulent strain in the SMV-resistant variety L29 carrying Rsv3.Overexpression of Glyma.14 g204700 in leaves inoculated with SMV inhibited viral infection in a soybean genotype lacking Rsv3,suggesting that Glyma.14 g204700 might be the Rsv3 gene itself,conferring specific resistance to SMV strains.
SC8 is one of the predominant SMV strains in Chinese soybean production areas.We studied the inheritance of soybean resistance to SC8 and fine-mapped a resistance gene in a~200-kb interval on chromosome 2 [16].Zhao et al.[17] used a larger F2population(2122 plants) to further fine-map the resistance gene to a 30.8-kb interval.Two candidate genes encoding MADS-box proteins,Glyma.02G121500 and Gm.02G121600,are likely involved in resistance to SC8.Based on these results,we performed functional characterization of the two genes during soybean and SMV interaction and found that Gm.02G121600(here GmCAL)acts as a positive regulator of soybean resistance to SMV.Here,we cloned GmCAL from Kefeng 1 and NN1138-2 and then performed expression profile analysis in soybean organs and in SMV-inoculated leaves,and subcellular localization.We then used a bean pod mottle virus-mediated virus-induced gene silencing (VIGS) system and overexpression system to verify the function of GmCAL in soybean.
The SMV-resistant soybean variety Kefeng 1,the SMVsusceptible variety NN1138-2,Nicotiana benthamiana,and SMV strains SC3,SC7,and SC8 were provided by the National Center for Soybean Improvement,Nanjing Agricultural University,Nanjing,Jiangsu,China.
Soybean seeds were sown in pots (20 cm in diameter) filled with sand and grown in a growth chamber under a temperature cycle of day 25 °C/night 21 °C,75% relative humidity,and a photoperiod cycle of 16 h light/8 h dark.
The SMV strains were propagated on the susceptible variety NN1138-2.At the seedling stage,1 g of NN1138-2 leaves with typical SMV symptoms were removed and ground into homogenate in a mortar with 10 mL of phosphate-buffered saline (PBS,0.01 mol L1,pH 7.4) and a small amount of carborundum (600 mesh).The inoculum was used to inoculate the first fully developed leaves with a paintbrush.The inoculated leaves were rinsed with tap water.
The sequence of GmCAL,obtained from the soybean reference genome annotation database (https://soybase.org/gb2/gbrowse/gmax2.0/),was used to design gene-specific primers with Primer Premier 5.0 (PREMIER Biosoft International,Palo Alto,CA,USA)[18](Table S1).cDNA from SMV-infected soybean leaves was used as templates for gene amplification.The physicochemical properties of the protein encoded by GmCAL were characterized with ProtParam (http://expasy.org/tools/protparam.html) [19].Amino acid sequence alignment was performed with DNAMANTM 8.0(Lynnon Biosoft,Pointe-Claire,Quebec,Canada) [20].A phylogenetic tree of GmCAL and MADS box homologs from other species was constructed with MEGA 7.0 [21].
The resistant and susceptible varieties were separately planted in an insect-free greenhouse,and SMV was inoculated after the primary leaves were fully expanded.One gram of leaves was collected at multiple time points (0,1,2,4,8,12,24,48,and 72 h,and 7 d)post-inoculation.Three biological replicates were taken.Roots,stems,leaves,flowers and bean pods of NN1138-2 were also sampled in three biological replicates.Samples were chilled in liquid nitrogen and stored at 80 °C until further use.
Total RNA was extracted with an RNA Simple Total RNA Kit(Tiangen,Beijing,China).First-strand synthesis of cDNA was performed using oligo (dT) primers and a PrimeScript II 1st Strand cDNA Synthesis Kit (TaKaRa,Kusatsu,Shiga,Japan).Specific primers for GmCAL and the internal reference gene Tubulin α were designed(Table S1).The expression was measured by quantitative real-time polymerase chain reaction (qRT-PCR).Three technical replicates were examined.
The Gateway system(Invitrogen,San Diego,CA,USA)was used to construct the subcellular localization fusion expression vector pGWB5::GmCAL::GFP.Gene-specific primers used for vector construction are shown in Table S1.The construct was transferred into the Agrobacterium tumefaciens strain EHA105,and positive clones were sequenced.The strain with pGWB5::GFP was cultured to log phase(OD600between 0.2 and 0.6)in yeast extract broth liquid medium containing kanamycin (50 mg L1) and rifampicin antibiotics (100 mg L1).The supernatant was removed by centrifugation,and 5 mL of permeate was added to suspend the cells at normal temperature for 2–3 h.Healthy tobacco plants were placed under an incandescent lamp for 2 h to open leaf stomata and then gently injected with suspended bacterial solution using a 1-mL syringe.The injection area was marked,and the plants were placed into a growth chamber for 48–72 h.Transient fluorescence was observed using an LSM780 laser confocal microscope (Zeiss,Oberkochen,Germany) and photographed.
The full sequence (with promoter region and coding sequence)of GmCAL was analyzed using the NCBI online primer-design tool(https://www.ncbi.nlm.nih.gov/tools/primer-blast/) to select a 241-nt sequence region,which was further evaluated for potential off-target silencing by alignment against the soybean reference genome annotation database (https://soybase.org/gb2/gbrowse/gmax2.0/).The 241-nt fragment of GmCAL was amplified from Kefeng 1 with the primer pair VIGS-GmCAL-F/R (Table S1).The target fragment contained partial 5′-UTR and partial 5′CDS sequences of GmCAL (Fig.S1).The fragment of interest was inserted into the pBPMV-IA-V2-R2 vector to construct the recombinant viral vector pBPMV-IA-V2-R2-GmCAL.pBPMV-IA-V2-R2-GmCAL and pBPMVIA-R1M were mixed at a molar ratio of 1:1 and then inoculated immediately on the primary leaves of NN1138-2 to generate GmCAL-silenced plants,designated VIGS-GmCAL.Plants inoculated with empty vector (pBPMV-V2) served as controls.
A fragment of GmCAL was inserted into the pBPMV-IA-V4-R2 vector to construct the recombinant viral vector pBPMV-GmCAL,with the negative control pBPMV-V4 (the empty vector).Plants inoculated with pBPMV-GmCAL vector were designated OE-GmCAL.A mixture of pBPMV-RNA1 and pBPMV-RNA2 (1:1 in molar) was inoculated into the primary leaves of NN1138-2 treated for 24 h in the dark.Typical diseased BPMV leaves were sampled,grounded into inoculum and then was inoculated into NN1138-2 for testing.Expression was measured after inoculation with BPMV.Then,SMV was inoculated,and SMV content was detected by qRT-PCR and double antibody sandwich-enzyme-linked immunosorbent assay(DAS-ELISA) 14–21 d post inoculation.The expression level of VIGS-0 and OE-0 was defined as 1.0.
qRT-PCR was performed on a Light Cycler 480 Real-Time PCR System (Roche Diagnostics,Mannheim,Germany) using the reaction system of SYBR Premix Ex Taq (TAKARA,Japan),and the relative expression levels of GmCAL and SMV CP were quantified by a relative quantitative (2ΔΔCT) method [22].The ratios of the relative expression levels of the samples inoculated with SMV to those of plants mock-inoculated with PBS were taken as the relative expression levels induced by SMV at the different time points.The relative expression level of each treatment was the mean ± SE of three biological replicate samples.
An ELISA kit with the SMV antibodies (ACD Inc.,Cat #V094-R1,Beijing,China) was used to perform DAS-ELISA.To ensure the reproducibility of DAS-ELISA,samples were collected from all of the plants in each treatment.The optical density at 405 nm wavelength was measured with an INFINITE M200 PRO microplate reader (Tecan Austria GmbH,Gr?dig,Austria).The raw DAS-ELISA readings of each sample were averaged and converted to multiples of the negative controls (mock-inoculated plants).Samples with a relative ratio greater than 2.0 were considered to be positive for SMV.
Statistical analysis was performed using Statistical Analysis System software (SAS 9.4,SAS Institute,Inc.,Cary,NC,USA),and means were compared by Student’s t-test at P <0.01.
GmCAL,encoding a MADS-box transcription factor,was successfully amplified from the SMV-resistant Kefeng 1 and the SMV-susceptible NN1138-2.The CDS (732 bp in length) encodes a protein consisting of 243 amino acids with an isoelectric point of 8.65 and a relative molecular mass of 28.2 kDa.There was no difference in the GmCAL CDSs between Kefeng 1 and NN1138-2(Fig.S2).The amino acid sequence of GmCAL (XP_006574961.1)was highly similar to those of MADSs from other plants.For example,its similarities with GmMADS1 (soybean,XP_003516454.1),CcMADS (Cajanus cajan,XP_020212024.1),SsMADS (Spatholobus suberectus,TKY69526.1),and VrMADS (Vigna radiata,XP_014522264.1) were 93.83%,92.18%,92.59%,and 91.77%,respectively.Comparison of the GmCAL sequence with those of other members of the MADS protein family revealed that it contains the characteristic MADS-box domain as well as the I,K,and C domains (Fig.1).The phylogenetic tree showed that GmCAL belongs to the AP1 subgroup of the MADS-box transcription factor family (Fig.2).
After inoculation with SMV,the expression level of GmCAL in Kefeng 1 reached the highest level at 2 h,with a value significantly higher than those at the other time points.There was no significant change in NN1138-2 at each time point.The expression level of GmCAL was significantly different between Kefeng 1 and NN1138-2 at 2 and 24 h (Fig.3A).Its expression in Kefeng 1 was significantly higher than that in NN1138-2.
The expression patterns of GmCAL in roots,stems,leaves,flowers,and pods were further assessed.GmCAL was highly expressed in flowers,followed by bean pods,stems,leaves,and roots(Fig.3B).
pGWB5::GmCAL::GFP was injected into N.benthamiana epidermal cells to determine the subcellular location of GmCAL.The fusion protein pGWB5::GmCAL::GFP revealed a fluorescent signal in the nucleus,while the empty vector pGWB5::GFP was expressed throughout the entire cell (cell membrane,cytoplasm,and nucleus) (Fig.3C).We concluded that GmCAL is localized in the nucleus.
To verify the function of GmCAL during the interaction between soybean and SMV,we performed VIGS on the resistant cv.Kefeng 1 using the bean pod mottle virus (BPMV) VIGS system.First,we silenced the soybean phytoene desaturase (GmPDS) gene,which causes leaf bleaching in soybean and was used as a positive control to determine the effectiveness of this system.The expression level of GmPDS decreased by approximately 80% when it was silenced(Fig.S3A).The photobleaching phenotype was clearly observed 7–21 d post-GmPDS gene silencing(Fig.S3B).These results showed that the VIGS system was effective.The constructed plasmids were then inoculated onto the fully expanded primary leaves,and the silencing level was measured.The expression level of GmCAL in the VIGS-GmCAL plants was reduced by approximately 80% compared to the empty vector (Fig.4A).
SC8 was inoculated on leaves containing BPMV,and two methods were used to detect SMV accumulation at 14 and 21 d.qRTPCR results showed that the SMV content increased slightly at 21 d relative to that at 14 d.However,SMV-SC8 accumulation did not increase significantly relative to the empty vector (Fig.4B).DAS-ELISA results showed that SMV accumulation in Kefeng 1-silenced plants was negative (Fig.4C).This result indicates that the reduced transcript levels had no effect on Kefeng 1 resistance to SC8 (Fig.4).It was hypothesized that the low expression of GmCAL in Kefeng 1 leaves (Fig.3B) and the incomplete silencing of GmCAL resulted in its proper function in response to SMV attack.Accordingly,whether GmCAL was involved in disease resistance to SMV was further investigated.
To further confirm the function of GmCAL in response to SMV infection,GmCAL was overexpressed in the susceptible cultivar NN1138-2 using the BPMV overexpression system.The constructed vectors were inoculated onto the fully expanded primary leaves.Compared with the empty vector,the expression level of GmCAL in the GmCAL-overexpressing plants increased by~200-fold (Fig.5A).The symptoms of BPMV after inoculation of vectors are shown in Fig.5B.
We inoculated the SMV strains SC3,SC7,and SC8 separately into NN1138-2 leaves overexpressing GmCAL and assessed SMV accumulation at 14 and 21 d after inoculation.The qRT-PCR results showed almost no accumulation of SC3 and SC7 14 and 21 d after inoculation compared to that of the empty vector(Fig.6A).A small amount of SC8 was accumulated from 14 to 21 d after inoculation.However,the increased level was nearly negligible in comparison with the mock treatment.DAS-ELISA assays were negative for SC3 and SC7 but weakly positive for SC8,while the content was significantly lower than that of the positive control (Fig.6B).The phenotypic symptoms after SMV inoculation are shown in Fig.6C.Thus,the accumulation level of SMV measured by DASELISA was consistent with that by qRT-PCR.These results demonstrate that GmCAL is involved in SMV resistance.As it is not for a specific strain,it is likely that the gene is involved in broadspectrum resistance to SMV.
Fig.1.Alignment of predicted amino acid sequences of GmCAL with its homologs in soybeans and other species.GmCAL(XP_006574961.1),GmMADS1(XP_003516454.1),Abrus precatorius ApMADS (XP_027359868.1), Cajanus cajan CcMADS (XP_020212024.1), Lupinus angustifolius LaMADS (XP_019443748.1), Phaseolus vulgaris PvMADS(XP_007153481.1), Spatholobus suberectus SsMADS (TKY69526.1), Vigna radiata var.radiata VrMADS (XP_014522264.1), Vigna unguiculata VuMADS (XP_027913264.1).
Fig.2.Phylogenetic analysis of GmCAL and its homologs in soybeans and other species.A rooted gene tree (using majority-rule consensus from 1000 bootstrap replicates)was constructed using the neighbor-joining option in MEGA 7.0.Bootstrap values are indicated at each branch node.
In the present work,we cloned the CDS sequence of soybean Rsc8 candidate gene GmCAL in SMV-resistant soybean cv.Kefeng 1 and SMV-susceptible cv.NN1138-2.Amino acid sequence of GmCAL has the typical domains that are highly conserved in MADS-box genes[23].Interestingly,CDS sequences of GmCAL were identical between Kefeng 1 and NN1138-2,whereas there was a marked difference between their resistance phenotypes.Similar results had been reported previously.Li et al.[24] reported that the target gene LOC_Os03 g32230 was found in the resistant variety Digu and the susceptible variety LTH when during the study of a transcription factor involved in broad-spectrum resistance to rice blast.The coding region in the sequence was the same,and there were 21 SNP and 2 InDel differences between Digu and LTH in the promoter region of the LOC_Os03 g32230 gene [24].Whether the differences in resistance were due to variation in the promoters and other regulatory sequences,whether there were different protein level modifications between resistant and susceptible varieties,and whether homodimers and heterodimers were formed with other homologous MADS in soybean to affect the expression of other disease resistance genes,awaits further study.
Fig.3.Expression analysis of GmCAL.(A) SC8-induced expression analysis of GmCAL in Kefeng 1 and NN1138-2.(B) Tissue-specific expression of GmCAL.(C) Subcellular localization of GmCAL (Scale bar,50 μm).**, P <0.01.
Fig.4.VIGS of GmCAL in soybeans.(A) Silencing level of GmCAL measured by qRT-PCR.(B) Detection of SMV CP transcripts at 14 and 21 d post inoculation by qRT-PCR.(C)SMV accumulation in GmCAL-silenced plants.SMV content was measured by DAS-ELISA at 21 d post inoculation.(D) Symptoms of BPMV and SMV on soybeans after SMV inoculation.Mock,soybean plants inoculated with PBS solution.VIGS-0 means empty-vector control.**,significant difference at P <0.01 by Student’s t-test.
VIGS is a special reverse genetic tool that can be used to generate mutant phenotypes for delivering functions to unknown genes.It has many advantages over other methods;it is fast and does not require plant transformation [25].In the VIGS system,viruses are designed to carry partial sequences of known or candidate genes to link their function to the mutant phenotype.Recombinant viral replication and the production of dsRNA intermediates trigger an RNA-mediated host defense system that results in the degradation of RNA,which shows sequence identity with recombinant viruses including mRNAs of the interested gene [26–28].However,VIGS also has some shortcomings.The symptoms caused by viral infection may mask the phenotype associated with the silencing of the target gene.Observation of phenotypes is difficult because BPMV and SMV symptoms are similar,and plants may show mottling,mosaic,and shrinkage.In addition,a VIGS-silenced target gene cannot be completely silenced which makes it difficult to characterize the function of genes with lower transcript levels.In this work,although expression level of GmCAL was~58.86-fold higher at 2 h in Kefeng 1 inoculated with SMV when compared with that in non-inoculated leaves,it was lower throughout the test period(Fig.3A).When GmCAL was silenced via VIGS system,a probably 80% silencing was measured in Kefeng 1 (Fig.4).The incomplete silencing might be one of the explanations why GmCAL-silenced Kefeng 1 plants conferred SMV resistance.Based on this observation,it is speculated that the VIGS system may not be suitable for functional characterization of genes with low expression level.
Fig.5.Overexpression of GmCAL in NN1138-2.(A) Relative expression level of GmCAL.OE-0 means empty-vector control.OE-GmCAL represents GmCAL-overexpressing plants.(B) Symptoms of soybeans infected by vectors after plasmid inoculation:mottled,mosaic leaves.**,significant difference at P <0.01 by Student’s t-test.
Fig.6.Detection of SMV accumulation after overexpression.(A) Relative expression of SMV CP genes of SC3,SC7,and SC8 strains in NN1138-2 by qRT-PCR.OE-0 means empty-vector control.(B)Accumulation of SC3,SC7,and SC8 in GmCAL-overexpressing plants.SMV content was measured by DAS-ELISA 21 d post-inoculation.(C)Symptoms of BPMV and SMV on soybeans after SMV inoculation:healthy,mottled,mosaic leaves.**,significant difference at P <0.01 by Student’s t-test.
Transcription factors (TFs) regulate the expression of related genes in response to adverse conditions such as drought,salt,extreme temperatures,oxidative stress and pathogen attack [29].Previous studies [30–31] have shown that the MADS-box transcription factor family acts mainly in plant growth and development and in plant defense responses,especially to abiotic stresses [32].MADS-box genes were also found associated with disease resistance.Zhang et al.[33] showed that the MADS1-H2O2-NO pathway in tobacco mediates a variety of bacterial harpinXoo-triggered responses,including stomatal closure,allergic cell death,and defense-related gene expression,suggesting that MADS1 has a function in regulating the defense response of tobacco to harpinXoo.OsMADS26 was downregulated in transgenic rice while conferred increased resistance to rice blast and bacterial wilt[24],indicating that OsMADS26 acts as a negative regulator of pathogen resistance[34].The antiviral signal caused by viral infection was transmitted to OsRDR1 (RDR1,a key component of the antiviral RNA silencing pathway),and the overexpression of miR444,a key factor in the expression of OsRDR1,increased resistance to rice stripe virus(RSV),in which the three targets of miR444,OsMADS23,OsMADS27a,and OsMADS57,form homodimers and heterodimers,inhibiting OsRDR1 expression by directly binding the CArG motif of its promoter,thereby activating the OsRDR1-dependent antivirus RNA-silencing pathway [35].Similar results were observed in this work,expression of GmCAL was significantly induced post SMV inoculation (Fig.3) and its overexpression in NN1138-2 remarkably reduced SMV accumulation and thus improved plants resistance to SMV (Fig.5).These results suggest that the MADS-box gene functions in defense response to plant pathogens.
The GmCAL-silenced Kefeng 1 plants obtained via the VIGS system were observed still confer resistance to the SC8 strain,probably because of the incomplete silencing of GmCAL.The expression level of GmCAL was low before SMV inoculation while significantly increased after SMV inoculation,suggesting that GmCAL is an inducible gene.When the gene was cloned,we extracted RNA not from normal leaves but from leaves 1–3 d post-SMV inoculation.After overexpression of GmCAL in the susceptible cv.NN1138-2,the accumulation of SC3,SC7 and SC8 was reduced,especially that of SC3 and SC7.We speculate that GmCAL does not confer resistance specifically to a single strain,SC8,but is involved in broadspectrum resistance to SMV.MADS-box proteins are combinatorial TFs that interact with different cofactors and function in diverse and important biological activities [31].The detailed regulatory mechanisms governing resistance to SMV for GmCAL await further investigation.We conclude that GmCAL is a positive regulator of interaction between soybean and SMV.
Declaration of competing interest
Authors declare that there are no conflicts of interest.
CRediT authorship contribution statement
Qiuyan Ren:Investigation,Methodology,Formal analysis,Visualization,Writing-Original draft.Hua Jiang:Methodology,Software,Writing-Original draft.Wenyang Xiang:Methodology,Formal analysis.Yang Nie:Methodology,Data Curation.Song Xue:Methodology,Data Curation.Haijian Zhi:Resources,Conceptualization.Kai Li:Conceptualization,Resources,Project administration,Writing-Review &Editing,Funding acquisition.Junyi Gai:Conceptualization,Writing-Review & Editing,Project administration.
Acknowledgments
This work was supported by the National Key Research and Development Program of China (2017YFD0101500),the National Natural Science Foundation of China (31671718),and China Agriculture Research System of MOF and MARA(CARS-04),the Jiangsu Collaborative Innovation Center for Modern Crop Production(JCICMCP),Collaborative Innovation Center for Modern Crop Production co-sponsored by Province and Ministry (CIC-MCP),and the Program for Changjiang Scholars and Innovative Research Team in University (PCSIRT_17R55).
Appendix A.Supplementary data
Supplementary data for this article can be found online at https://doi.org/10.1016/j.cj.2021.10.003.