GT Transcription Factors of Rosa rugosa Thunb. Involved in Salt Stress Response

Simple Summary Salt stress produced ion toxicity on plant cells and limited the of culture of cultivated Rosa rugosa. GT genes in salt stresses responses have been emerging. From the GT gene family of the salt-tolerant wild Rosa rugosa, four NaCl stress responsive genes (RrGT-1, RrSIP1, RrSIP2, RrGTγ-4) were identified. RrSIP1 and RrGTγ-4, RrGT-1 and RrSIP2 located on chloroplasts and cell nucleus, respectively. RrSIP1, RrSIP2 and RrGTγ-4 could play roles in regulation of sodion and potassium transport. And RrGT-1 expressed higher specifically in wild Rosa rugosa than in the salt-sensitive cultivated Rosa rugosa. These four genes would be candidates for further study of regulation mechanism of salt-tolerance of wild Rosa rugosa and would supply gene resources for tolerance improvement of cultivated Rosa rugosa. Abstract Rosa rugosa was a famous aromatic plant while poor salt tolerance of commercial cultivars has hindered its culture in saline-alkali soil. In many plants, the roles of GT (or trihelix) genes in salt stresses responses have been emerging. In the wild R. rugosa, a total of 37 GTs (RrGTs) were grouped into GT-1, GT-2, GTγ, SH4, and SIP1 lineages. SIP1 lineage expanded by transposition. The motifs involved in the binding of GT cis-elements were conserved. Four RrGTs (RrGT11/14/16/18) significantly differentially expressed in roots or leaves under salt stress. The responsive patterns within 8 h NaCl treatment indicated that RrGTγ-4 (RrGT18) and RrGT-1 (RrGT16) were significantly induced by salt in roots of R. rugosa. Subcellular localizations of RrSIP1 (RrGT11) and RrGTγ-4 were on chloroplasts while RrGT-1 and RrSIP2 (RrGT14) located on cell nucleus. Regulation of ion transport could be the most important role of RrSIPs and RrGTγ-4. And RrGT-1 could be a halophytic gene with higher transcription abundance than glycophytic GT-1. These results provide key clue for further investigations of roles of RrGTs in salt stress response and would be helpful in the understanding the salt tolerance regulation mechanism of R. rugosa.

Salinity stress is one of the most severe abiotic stresses and poses a continuing threat to economic crops [27], and for this reason the research are starting to consider and valorize wild species suitable for saline environments [28,29]. Rosa rugosa cultivars are widely used for spicery of food industry or essential oil of cosmetics industry [30,31]. These R. rugosa cultivars lost salt tolerance along with the breeding processes of floral traits, resulting in their limiting planting areas although there were vast saline-alkali soils in China [32][33][34]. e.g., the commercial cultivar R. rugosa 'Zizhi' (Zizhi) which planted in the narrow hilly lands of Shandong Province (China) was a typical glycophyte. While the wild R. rugosa which distributed naturally in the coastal area of northeast China belong to halophytes. The wild R. rugosa kept strong salt tolerance to adapt to the high salinity beach, as observed in other plant species in the coastal areas in the world [35]. In genetic engineering of salt-tolerance for glycophytic crop, homologous genes from halophytes should be more efficient since halophytes are more salt-tolerant than glycophytes [28,36]. The mining of salt responsive TF of wild R. rugosa would supply a base for salt tolerance improvement of R. rugosa cultivars [34].
With the genome of wild R. rugosa, this study aimed to screen GTs of R. rugosa (RrGTs) involved in the salt response. The phylogeny, synteny and sequence analyses would give a systematic understanding of lineage, gene duplication events, conserved motifs and gene structures of RrGT family. Expression profiles of salt treated R. rugosa were built to detect significantly induced/reduced RrGTs. These candidates were pointcuts of further study of salt stress responses regulation. Our study preliminary studied the roles of RrGTs under slat stress and would be helpful in understanding the regulatory mechanism of salt tolerance of R. rugosa.

Identification of RrGT Family
The R. rugosa and Rosa chinensis Jacq. genomes were obtained from GDR ( 1 May 2022), a neighbor-join (NJ) phylogenetic tree was constructed by MEGA 7 with 1000 bootstrapping replications [38]. Besides, the p-distance model of the substitution type, pairwise deletion of Gaps/Missing data and uniform rates among sites were selected for the phylogenetic analysis.

Synteny Analysis of RrGTs
The homologous gene pairs (E < 10 −5 , top five matches) within the R. rugosa genome or among R. rugosa, R. chinensis and Fragaria vesca L. genomes (obtained from GDR) were identified by BLASTP (BLAST+ 2.13.0) search. Based on the location of homologous pairs, MCScanX (mcscan2) [39] identified the syntenic regions and predicted the gene duplication events. RrGTs and corresponding homologous GTs on syntenic regions were highlighted by Synteny plot tool of TBtools [40].

Gene Structure, Motif Analysis and Cis-Acting Elements of RrGT Family
The top 10 conserved motifs were predicted by MEME web tools (https://memesuite.org/meme/, accessed on 2 May 2022) under default parameters. Cis-elements on the 2000 bp sequence upstream to the initiation codon were predicted by PLANTCARE (http: //bioinformatics.psb.ugent.be/webtools/plantcare/html/, accessed on 2 May 2022). The gene structures and motifs were illustrated by the Gene structure view tool of TBtools [40].

Expression Analysis under Salt Stress
Previous transcriptome data [32,41] of wild R. rugosa and R. rugosa 'Zizhi' provided the per kilobase of exon model per million mapped fragments (FPKM) of RrGTs and the fold-changes of differentially expressed RrGTs.
One-month-old wild R. rugosa seedlings were treated with 340 mM NaCl solution for 0.5 h, 1 h, 2 h, 4 h and 8 h and roots of these samples were collected with three biological repetitions. Total RNAs were extracted and reverse-transcribed as cDNA templets by RNAprep Pure plant kit (Tiagen, Beijing, China) and HiScript ® III RT SuperMix (Vazyme, Nanjing, China). Quantitative real-time polymerase chain reaction (qRT-PCR) was conducted using ChamQ SYBR Color qPCR Master Mix (Vazyme, Nanjing, China) on the CFX96 platform (Bio-Rad, China). All the steps of RNA extraction, reverse-transcription and qRT-PCR were conducted following the manufacturers' recommended instructions. Table S5 listed the primers of RrGTs and reference genes (phD and 5.8s).
regions which including less than 20 gene pairs ( Figure S1), only one redundant pair was credible ( Figure 2B, red line). The consistent collinearity indicated that the lineage evolution of GT families among R. rugosa, R. chinensis and F. vesca was conserved ( Figure 2B).
The expression patterns of the four candidate genes were checked in wild R. rugosa seedings treated by water (CK) and 340 mM NaCl solution ( Over half RrGTs (20 RrGTs, 54.05%) included 2 exons and 10 RrGTs (27.03%) were intron-free. One RrGT contained 3, 7, 8 and 9 exons and two RrGT genes contained 5 exons, respectively. Exons of RrGT33 were up to 17 but most (exon 1-exon 12) coded the long nonconservative N-terminal with no conserved motifs.
Four differentially expressed genes were predicted as salt responsive RrGTs ( Figure  4,

Subcellular Localization of RrGT Candidates
The predicted subcellular localizations (Table S2) of RrGT family were diverse (Nuclei, chloroplasts, mitochondria, cytoplasm and so on). And different prediction tools were inconsistent in prediction of several RrGTs. e.g., RrSIP1 and RrSIP2 were predicted as proteins located on nucleus by Plant-mPLoc but on chloroplasts by WoLF PSORT. We detected the subcellular localizations of above 4 candidates (RrGTγ-4, RrSIP1, RrSIP2 and RrGT-1) to exclude the inconsistent prediction.

Subcellular Localization of RrGT Candidates
The predicted subcellular localizations (Table S2) of RrGT family were diverse (Nuclei, chloroplasts, mitochondria, cytoplasm and so on). And different prediction tools were inconsistent in prediction of several RrGTs. e.g., RrSIP1 and RrSIP2 were predicted as proteins located on nucleus by Plant-mPLoc but on chloroplasts by WoLF PSORT. We detected the subcellular localizations of above 4 candidates (RrGTγ-4, RrSIP1, RrSIP2 and RrGT-1) to exclude the inconsistent prediction.

The Species Specific Expansion of SIP1 Lineage
Expansion of SIP1 lineage has been observed in the Brassica plants. The majority of SIP1 genes of Brassica rapa L. retained two or three copies of corresponding AtGTs and it was more than other four lineages (one copy) [42]. The Brassica specific whole genome

The Species Specific Expansion of SIP1 Lineage
Expansion of SIP1 lineage has been observed in the Brassica plants. The majority of SIP1 genes of Brassica rapa L. retained two or three copies of corresponding AtGTs and it was more than other four lineages (one copy) [42]. The Brassica specific whole genome triplication contributed to the SIP1 lineage expansion since its divergence from A. thaliana [42,43]. But the conserved number of SIP1 genes of R. rugosa relative plants (8 GTs of F. vesca and 10 GTs of R. chinensis) indicated genome triplication or WGD did not contribute to R. rugosa SIP1 lineage expansion, and the specific expansion happened after the divergence from R. chinensis (Table S1). The 7 RrGTs (RrTH1/12/17/21/31/37) without corresponding collinear RcGTs and FvGTs were predicted as dispersed duplication types, which indicated proximal duplication and tandem duplication did not contribute to the SIP1 lineage expansion. And 5 of above genes (RrTH1/12/21/31/37) were clustered into the Rosa-specific lineages of the dendrogram with RrTH34/35 and transcript abundances of the 5 genes were extremely low in roots and leaves (Figure 4). These observations indicated part of RrGTs in SIP1 lineage could arise from transposition, which contributed to SIP1 lineage expansion of R. rugosa.
In GTγ lineage, HRA1 (HYPOXIA RESPONSE ATTENUATOR 1, AT3G10040) attenuates the anaerobic response induced by ERF-VIIs by protein interaction with RAP2.12 [11]. In rice, three OsGTγ genes were induced by salt, abscisic acid (ABA) or other abiotic stresses [24,44]. OsGTγ-1 was significantly induced by salt and its mutant increased salt stress sensitivity while OsGTγ-1 overexpression enhanced salt tolerance [24]. Similar, knockout and overexpression of OsGTγ-2 increased salt stress sensitivity and tolerance, respectively. In our study, RrGTγ-4 clustered with OsGTγ-1 and HRA1 in same cluster of GTγ lineage. RrGTγ-4 only highly expressed in roots under salt stress (FPKM = 271.54) and its abundance was much higher than other RrGTs (FPKM < 83). qRT-PCR proved the rapid and dramatic salt-inducing of RrGTγ-4 in roots within 1 h and peaked at 2 h, which was similar to OsGTγ-1. RrGTγ-4 should be a key candidate involved in salt tolerance regulation of R. rugosa roots.
In SIP1 lineage, NtSIP1 was firstly identified from Agrobacterium 6b-interacting proteins of Nicotiana tabacum L. [45]. AST1 (At3g24860) binds to a Novel AGAG-Box of stress tolerance genes to regulate Arabidopsis salt tolerance positively [12]. BnSIP1-1 overexpression improved both osmotic and salt stresses tolerance during seed germination, but only improved osmotic stress tolerance of transgenic B. napus plants [20]. In our study, RrSIP1, RrSIP2 were highly expressed in roots and leaves (FPKM > 20) and downregulated under salt stress only in leaves. Though most studies focus on stress induced GT genes, the two salt-stress-reduced RrGTs could be candidates involved in the negative regulation of salt tolerance.
Though no candidate RrGTs of GT-2 lineage significantly induced or reduced by salt, some studies have indicated GT-2 lineage also played roles in stress responses. e.g., Arabidopsis gtl1 (GTL1, AT1G33240) mutant improved water use efficiency by reducing leaf transpiration [46], and GT2L (AT5G28300) induced significantly by salt, drought, cold and ABA and it responded to cold and salt stresses by interaction with calcium/calmodulin [13].

Potential Target Genes and Regulation Roles of RrGTs
Previous study indicated that R. rugosa responds to ion stress by gene expression regulation but rather gene dosage since its ion transporter gene number was conserved [34]. In rice, OsGTγ-2 directly interacted with the GT-1 element 'GAAAAA' of three ion transporter genes (OsHKT2; 1, OsNHX1 and OsHKT1; 3) [44]. In the ion transporter genes of the R. rugosa, GT-1 elements were found in the promoters (1000 bp upstream) of RrHKT (evm.model.Chr5.2560, only one HKT in the genome) and 6  Most transporter genes contained 1-4 GT-1 elements (-300 to -100 mostly) while only evm.model.Chr5.7017 (AtNHX5/6 homolog) identified 7 GT-1 elements all across the promoter (−700 to −100). It indicated that RrGTγ-4 could be regulator of these ion transporter genes and coordinated ion transport under salt stress.
SIPs mostly take part in ABA signaling to resist growth inhibiting effect of salt stress. In apple, MdSIP1-2 promoted lateral root development promotion which was associated with ABA sensitivity, drought and salt stress tolerance [47]. AST1 and BnSIP1-1 reduced water loss rat in the overexpressed seedings [12,20]. The osmotic stress marker genes were highly expressed both in the BnSIP1-1 transgenic plants and seedings whereas ion transporter genes (BnSOS1, BnNHX1, and BnHKT) were only significantly higher expressed in transgenic seedings (not in plants). SIP genes seem to exert different regulatory mechanisms along with different development process. Interestingly, the AGAG-Box 'GGTAAA' was found in two RrNHXs (evm.model.Chr5.2640, evm.model.Chr4.534) which lacked GT-1 elements. It indicated that excess ion response and plant developing regulation were crossed in R. rugosa by RrSIPs.
The genes of halophytes might be more superior for plant growth in saline-rich areas than their orthologs of glycophytes [28,29]. The 4 RrGT candidates were homologous to corresponding glycophytic GTs of Zizhi (over 90% protein identity). In roots, only RrGT-1 (FPKM = 60.15, this study) expressed significantly higher than GT-1 of Zizhi (FPKM = 2.34 [41]). RrGT-1 could be a halophytic gene with higher transcription levels in wild R. rugosa. While the post translational regulation not transcription contributed mostly of salt-tolerance difference between halophytes and glycophytes. Whether the GTs were involved in the post translational regulation was not studied yet and it could be an important way to study the roles of other 3 RrGTs specific to halophytes [28,29].

Conclusions
In this study, we identified the RrGT family with 37 members from the R. × rugosa genome. RrGTs belonged to 5 lineages and SIP1 lineage expanded significantly. The conserved motif groups corresponding to trihelix domains were most conserved. RrGTγ-4 (RrGT18), RrGT-1 (RrGT16) located on chloroplasts and RrSIP1 (RrGT11), RrSIP2 (RrGT14) located on nucleus. The four genes significantly differentially expressed under salt stress in roots or leaves. Regulation of ion transport could be the most important role of RrSIP genes and RrGTγ-4 in response to salt stress of wild R. rugosa. And RrGT-1 could be a halophytic gene with higher transcription abundance than glycophytic GT-1. Together, ion transport regulation roles of GT needed to be illuminated and the regulation role of GTs specific to halophytes would be a breakthrough point for further researches.