**Identification and Expression Analysis of Hormone Biosynthetic and Metabolism Genes in the 2OGD Family for Identifying Genes That May Be Involved in Tomato Fruit Ripening**

**Qiangqiang Ding 1,**† **, Feng Wang 1,**† **, Juan Xue <sup>1</sup> , Xinxin Yang <sup>1</sup> , Junmiao Fan <sup>1</sup> , Hong Chen <sup>2</sup> , Yi Li <sup>3</sup> and Han Wu 1,\***


Received: 30 June 2020; Accepted: 23 July 2020; Published: 28 July 2020

**Abstract:** Phytohormones play important roles in modulating tomato fruit development and ripening. The 2-oxoglutarate-dependent dioxygenase (2OGD) superfamily containing several subfamilies involved in hormone biosynthesis and metabolism. In this study, we aimed to identify hormone biosynthesis and metabolism-related to 2OGD proteins in tomato and explored their roles in fruit development and ripening. We identified nine 2OGD protein subfamilies involved in hormone biosynthesis and metabolism, including the gibberellin (GA) biosynthetic protein families GA20ox and GA3ox, GA degradation protein families C19-GA2ox and C20-GA2ox, ethylene biosynthetic protein family ACO, auxin degradation protein family DAO, jasmonate hydroxylation protein family JOX, salicylic acid degradation protein family DMR6, and strigolactone biosynthetic protein family LBO. These genes were differentially expressed in different tomato organs. The GA degradation gene *SlGA2ox2*, and the auxin degradation gene *SlDAO1*, showed significantly increased expression from the mature-green to the breaker stage during tomato fruit ripening, accompanied by decreased endogenous GA and auxin, indicating that *SlGA2ox2* and *SlDAO1* were responsible for the reduced GA and auxin concentrations. Additionally, exogenous gibberellin 3 (GA3) and indole-3-acetic acid (IAA) treatment of mature-green fruits delayed fruit ripening and increased the expression of *SlGA2ox2* and *SlDAO1*, respectively. Therefore, *SlGA2ox2* and *SlDAO1* are implicated in the degradation of GAs and auxin during tomato fruit ripening.

**Keywords:** genome-wide identification; expression analysis; 2OGD family; hormone biosynthetic and metabolism genes; tomato fruit ripening

#### **1. Introduction**

Tomato is used as a model to study climacteric fruit ripening, which is mediated by the hormone ethylene. Other hormones are also involved in tomato ripening. For example, exogenous auxin treatment, or increasing the endogenous auxin level by silencing the expression of the auxin-degradation gene *SlGH3.2*, delayed tomato fruit ripening [1–3]. Indeed, exogenous application of gibberellins (GAs) delayed fruit ripening, while decreasing the endogenous levels of GAs via overexpression of the GA-catabolic gene *SlGA2ox1* accelerated fruit ripening [4]. Therefore, changes in hormone concentrations play important roles in tomato fruit ripening, and identification and functional analysis of hormone biosynthetic and metabolism genes are prerequisites for understanding their roles in tomato fruit ripening.

The 2-oxoglutarate-dependent dioxygenase (2OGD) superfamily is the largest enzyme family and facilitates numerous oxidative reactions, including hydroxylation, halogenation, desaturation, epimerization, etc. [5]. The 2OGD superfamily contains many proteins involved in hormone biosynthesis and metabolism. To date, nine hormone biosynthesis and metabolism-related protein families have been identified in the 2OGD family, including GA biosynthetic protein families GA20-oxidase (GA20ox) and GA3-oxidase (GA3ox), GA degradation protein families GA2-oxidases (C19-GA2ox and C20-GA2ox), auxin degradation protein family Dioxygenase for Auxin Oxidation (DAO), ethylene biosynthetic protein family 1-aminocyclopropane-1-carboxylic acid oxidase (ACO), jasmonate (JA) hydroxylation protein family JASMONATE-INDUCED OXYGENASE (JOX), salicylic acid (SA) degradation protein family Downy Mildew Resistant6 (DMR6) and DMR6-LIKE OXYGENASE (DLO), and strigolactone (SL) biosynthetic protein family LATERAL BRANCHING OXIDOREDUCTASE (LBO). In detail, GA20oxs and GA3oxs catalyze the final two steps of GA biosynthesis: GA20oxs catalyze the conversion of GA<sup>12</sup> and GA<sup>53</sup> to GA<sup>9</sup> and GA20, which are converted by GA3oxs to bioactive GA<sup>1</sup> and GA<sup>4</sup> [6]. GA2oxs are GA-oxidation enzymes that convert bioactive GAs or their precursors into inactive forms [6]. DAOs catalyze the irreversible conversion of active auxin into inactive 2-oxindole-3-acetic acid (oxIAA) [7]. ACO proteins function in the last step of ethylene biosynthesis by converting ACC into ethylene [8]. JOX proteins hydroxylate jasmonate (JA) into inactive 12-OH-JA [9]. DMR6s, as SA 5-hydroxylases, hydroxylate active salicylate (SA) at the C5 position of the phenyl ring to produce inactive 2,5-DHBA [10]. In *Arabidopsis*, LBO converts methyl carlactonoate into an unidentified strigolactone (SL)-like compound that may be the final product of SL biosynthesis [11]. All of these 2OGD-family hormone biosynthetic and metabolism genes play key roles in maintaining endogenous hormone homeostasis, thereby regulating plant growth and development, and the response to stresses.

2OGDs are non-heme iron-containing proteins. Their catalytic core contains a double-stranded β-helix fold (DSBH) with a highly conserved His-X-Asp-(X)n-His (HxD . . . H) motif, which is responsible for binding Fe (II) to form a catalytic triad [12]. Using Fe (II) as a cofactor and 2-oxoglutarate (2OG) as a co-substrate, 2OGD proteins catalyze oxidation of the substrate and concomitant decarboxylation of 2OG to produce succinate and CO2. In addition, a conserved Arg-X-Ser/Thr (RxS/T) motif at the subfamily-conserved position within the secondary structure of the DSBH fold likely binds the C5-carboxy group of 2OG, which is the co-substrate for all known members of the subfamily except isopenicillin N synthase (IPNS), 1-aminocyclopropane-1-carboxylic acid oxidase (ACO) and (S)-2-hydroxypropylphosphonic acid epoxidase (HPPE) [13]. 2OGD-family proteins have been identified in several species [14]. The 2OGD superfamily can be divided into DOXA, DOXB, and DOXC subfamilies based on the amino acid sequence [14]. DOXA proteins contain a 2OG-FeII\_Oxy\_2 conserved domain, and the DOXA protein AlkB of *Escherichia coli*, which has homologs in *Arabidopsis* and rice, participates in the oxidative demethylation of alkylated nucleic acids and histones [15]. DOXB proteins typically have a conserved 2OG-FeII\_Oxy\_1 domain; most studies have focused on prolyl-4-hydroxylase, which is involved in the synthesis of cell-wall proteins in plants and algae [16]. DOXC proteins, including those involved in hormone biosynthesis and metabolism, have a conserved 2OG-FeII\_Oxy domain [14].

In this study, we identified hormone biosynthesis- and metabolism-related proteins from DOXC family in tomato. Based on analysis of their structures, we predicted their motifs with the aim of determining their molecular mechanisms of action. We also analyzed the transcript levels of these genes in tomato to assess their roles in tomato growth and development, and focused on the correlations between their expression levels and tomato fruit ripening to identify proteins that degrade GAs and auxin during tomato fruit ripening.

#### **2. Results**

#### *2.1. Identification and Phylogenetic Analysis of Hormone Biosynthetic and Metabolism Proteins in 2OGD Superfamily*

Currently known hormone biosynthetic and metabolism proteins in the 2OGD superfamily are exclusively present in the DOXC subfamily. To identify all hormone biosynthetic and metabolism proteins of DOXC family in tomato, we used DOXC-specific 2OGD domain 2OG-FeII\_Oxy (PF03171) as a key query in hmmersearch to identify all DOXC proteins in *Arabidopsis*, rice, and tomato. The result showed that 99, 90, and 159 proteins were identified in *Arabidopsis*, rice, and tomato, respectively. A phylogenic tree was constructed using the best-fit model in MEGA6.0, based on the complete sequences of the 348 identified proteins (Figure S1). Nine hormone biosynthetic and metabolism protein families in DOXC family were identified: the GA biosynthetic protein families GA20ox and GA3ox, GA degradation protein families C19-GA2ox and C20-GA2ox, auxin degradation protein family DAO, ethylene biosynthetic protein family ACO, JA hydroxylation protein family JOX, SA degradation protein family DMR, and SL biosynthetic protein family LBO. The bootstrap values were >80%, suggesting high reliability of the results. The numbers of these subfamilies in *Arabidopsis*, rice, and tomato were as follows: 20 GA20oxs, 10 GA3oxs, 20 C19-GA2oxs, 8 C20-GA2oxs, 20 ACOs, 6 DAOs, 11 JOXs, 9 DMR6s, and 3 LBOs. A phylogenetic tree constructed using the above proteins showed that there were 11 GA20oxs, 4 GA3oxs, 9 C19-GA2oxs, 3 C20-GA2oxs, 7 ACOs, 3 DAOs, 3 JOXs, 2 DMR6s, and 1 LBO in tomato, of which 10 GA20oxs (SlGA20ox1-SlGA20ox10), 6 GA2oxs (SlGA2ox2, SlGA2ox4, SlGA2ox5, SlGA2ox7, SlGA2ox8, and SlGA2ox9), 3 DAOs (SlDAO1-SlDAO3), and 5 ACOs (SlACO1-SlACO3, SlACO4, and SlACO6) clustered together to form a monophyletic group (Figure 1). Therefore, these genes emerged via lineage-specific expansion events in tomato. In addition, the identified hormone biosynthetic and metabolism proteins in tomato comprised 104–380 amino acids, and most of them also containing a DIOX\_N domain (Table S1).

#### *2.2. Synteny and Duplication Analysis of Hormone Biosynthetic and Metabolism Proteins in 2OGD Superfamily*

Synteny was performed to assess the relationships of the hormone biosynthetic and metabolism 2OGD genes among *Arabidopsis*, rice, and tomato. The result showed that there were 27 collinear gene pairs, of which 25 were between tomato and *Arabidopsis*: 5 pairs in the *ACO* family, 2 in the *GA3ox* family, 4 in the *C19-GA2ox* family, 2 in the *C20-GA2ox* family, 6 in the *JOX* family, 4 in the *GA20ox* family, and 2 in the *DMR6* family. There was only one collinear gene pair in the *JOX* family between tomato and rice, as and one between rice and *Arabidopsis* (Figure 2, Table S2). This result is consistent with the evolutionary relationship between monocotyledons and dicotyledons.

**2. Results** 

*Superfamily* 

*2.1. Identification and Phylogenetic Analysis of Hormone Biosynthetic and Metabolism Proteins in 2OGD* 

Currently known hormone biosynthetic and metabolism proteins in the 2OGD superfamily are

exclusively present in the DOXC subfamily. To identify all hormone biosynthetic and metabolism proteins of DOXC family in tomato, we used DOXC-specific 2OGD domain 2OG-FeII\_Oxy (PF03171) as a key query in hmmersearch to identify all DOXC proteins in *Arabidopsis*, rice, and tomato. The result showed that 99, 90, and 159 proteins were identified in *Arabidopsis*, rice, and tomato, respectively. A phylogenic tree was constructed using the best-fit model in MEGA6.0, based on the complete sequences of the 348 identified proteins (Figure S1). Nine hormone biosynthetic and metabolism protein families in DOXC family were identified: the GA biosynthetic protein families GA20ox and GA3ox, GA degradation protein families C19-GA2ox and C20-GA2ox, auxin degradation protein family DAO, ethylene biosynthetic protein family ACO, JA hydroxylation protein family JOX, SA degradation protein family DMR, and SL biosynthetic protein family LBO. The bootstrap values were >80%, suggesting high reliability of the results. The numbers of these subfamilies in *Arabidopsis*, rice, and tomato were as follows: 20 GA20oxs, 10 GA3oxs, 20 C19-GA2oxs, 8 C20-GA2oxs, 20 ACOs, 6 DAOs, 11 JOXs, 9 DMR6s, and 3 LBOs. A phylogenetic tree constructed using the above proteins showed that there were 11 GA20oxs, 4 GA3oxs, 9 C19-GA2oxs, 3 C20- GA2oxs, 7 ACOs, 3 DAOs, 3 JOXs, 2 DMR6s, and 1 LBO in tomato, of which 10 GA20oxs (SlGA20ox1- SlGA20ox10), 6 GA2oxs (SlGA2ox2, SlGA2ox4, SlGA2ox5, SlGA2ox7, SlGA2ox8, and SlGA2ox9), 3 DAOs (SlDAO1-SlDAO3), and 5 ACOs (SlACO1-SlACO3, SlACO4, and SlACO6) clustered together to form a monophyletic group (Figure 1). Therefore, these genes emerged via lineage-specific expansion events in tomato. In addition, the identified hormone biosynthetic and metabolism proteins in tomato comprised 104–380 amino acids, and most of them also containing a DIOX\_N

**Figure 1.** Phylogenetic tree of hormone biosynthetic and metabolism proteins of DOXC family in *Arabidopsis*, rice, and tomato. The phylogenetic tree was constructed by MEGA6 with Maximum likelihood.

The chromosomal location of the hormone biosynthetic and metabolism 2OGD genes in tomato was analyzed based on genome annotation data. The result showed that the identified hormone biosynthetic and metabolism 2OGD genes were unevenly distributed on tomato 12 chromosomes (Figure S2). There was one gene on chromosomes 4, 8, and 12, seven on chromosome 2, and six on chromosome 7. Further, the genes exhibited the following duplication events: nine dispersed gene pairs in *SlGA20ox* (Figure S2, Table S3); two segmental duplication genes (one WGD or segmental duplication events) and two dispersed gene pairs in *SlGA3ox*; seven dispersed gene pairs in *C19-SlGA2ox*; two segmental duplication genes (one WGD or segmental duplication events) in *C20-SlGA2ox*; two tandem duplication events in *SlDAO*; four segmental duplication genes (two WGD or segmental duplication events) in *SlACO*; two segmental duplication genes (one WGD or segmental duplication events) in *SlJOX*; two segmental duplication genes in *SlDMR6* (one WGD or segmental duplication events); and one dispersed gene pair in *SlLBO*.

likelihood.

*Superfamily* 

**Figure 1.** Phylogenetic tree of hormone biosynthetic and metabolism proteins of DOXC family in *Arabidopsis*, rice, and tomato. The phylogenetic tree was constructed by MEGA6 with Maximum

Synteny was performed to assess the relationships of the hormone biosynthetic and metabolism 2OGD genes among *Arabidopsis*, rice, and tomato. The result showed that there were 27 collinear gene pairs, of which 25 were between tomato and *Arabidopsis*: 5 pairs in the *ACO* family, 2 in the *GA3ox*  family, 4 in the *C19-GA2ox* family, 2 in the *C20-GA2ox* family, 6 in the *JOX* family, 4 in the *GA20ox*  family, and 2 in the *DMR6* family. There was only one collinear gene pair in the *JOX* family between

*2.2. Synteny and Duplication Analysis of Hormone Biosynthetic and Metabolism Proteins in 2OGD* 

with the evolutionary relationship between monocotyledons and dicotyledons.

**Figure 2.** Synteny analysis of hormone biosynthetic and metabolism 2-oxoglutarate-dependent dioxygenase (2OGD) genes among *Arabidopsis*, rice, and tomato. Chromosome numbers of *Arabidopsis* (At), rice (Os), and tomato (Sl) are indicated on the inner side. Red, green, and blue colors represent *Arabidopsis*, rice, and tomato chromosomes. Gene pairs with a collinear relationship are joined by lines. Red lines represent collinear pairs between *Arabidopsis* and tomato, blue lines represent collinear pairs between *Arabidopsis* and rice, green lines represent collinear pairs between rice and tomato. **Figure 2.** Synteny analysis of hormone biosynthetic and metabolism 2-oxoglutarate-dependentdioxygenase (2OGD) genes among *Arabidopsis*, rice, and tomato. Chromosome numbers of *Arabidopsis* (At), rice (Os), and tomato (Sl) are indicated on the inner side. Red, green, and blue colors represent*Arabidopsis*, rice, and tomato chromosomes. Gene pairs with a collinear relationship are joined by lines. Red lines represent collinear pairs between *Arabidopsis* and tomato, blue lines represent collinear pairsbetween *Arabidopsis* and rice, green lines represent collinear pairs between rice and tomato.

#### The chromosomal location of the hormone biosynthetic and metabolism 2OGD genes in tomato was analyzed based on genome annotation data. The result showed that the identified hormone *2.3. Multiple Sequence Alignment and Motif Composition Analysis of Hormone Biosynthetic and Metabolism 2OGD Proteins*

biosynthetic and metabolism 2OGD genes were unevenly distributed on tomato 12 chromosomes (Figure S2). There was one gene on chromosomes 4, 8, and 12, seven on chromosome 2, and six on To determine the functional similarity of hormone biosynthetic and metabolism 2OGD proteins of tomato with those of *Arabidopsis* and rice, we performed multiple sequence alignments and motif composition analysis. Two 2OGD-family proteins of known three-dimensional structure—OsGA2ox3 and OsDAO [17], and seven hormone biosynthetic and metabolism 2OGD-family proteins—AtGA20ox1, AtGA3ox1, AtGA2ox7, SlACO1, AtJOX1, AtDMR6, and AtLBO1—which have been functionally characterized were aligned to identify conserved domains or motifs in 2OGD family. The result showed that the above 2OGD proteins had the HxD . . . H and RxS/T conserved motifs in OsGA2ox3 and OsDAO (Figure 3a), which recruit Fe(II) as a cofactor and co-substrate. Further, among the hormone biosynthetic and metabolism 2OGD proteins in *Arabidopsis*, rice, and tomato, SlGA20ox8, SlGA20ox9, SlGA20ox10, SlGA2ox12, and OsACO6 did not have an HxD . . . H motif, while SlGA20ox7, SlGA20ox10, OsGA2ox10, SlGA2ox12, and OsACO6 lacked an RxS/T motif (Figures S3–S11), suggesting that these proteins do not have 2OGD biological activity.

*Metabolism 2OGD Proteins* 


chromosome 7. Further, the genes exhibited the following duplication events: nine dispersed gene pairs in *SlGA20ox* (Figure S2, Table S3); two segmental duplication genes (one WGD or segmental duplication events) and two dispersed gene pairs in *SlGA3ox*; seven dispersed gene pairs in *C19- SlGA2ox*; two segmental duplication genes (one WGD or segmental duplication events) in *C20- SlGA2ox*; two tandem duplication events in *SlDAO*; four segmental duplication genes (two WGD or segmental duplication events) in *SlACO*; two segmental duplication genes (one WGD or segmental duplication events) in *SlJOX*; two segmental duplication genes in *SlDMR6* (one WGD or segmental

To determine the functional similarity of hormone biosynthetic and metabolism 2OGD proteins of tomato with those of *Arabidopsis* and rice, we performed multiple sequence alignments and motif composition analysis. Two 2OGD-family proteins of known three-dimensional structure— OsGA2ox3 and OsDAO [17], and seven hormone biosynthetic and metabolism 2OGD-family proteins—AtGA20ox1, AtGA3ox1, AtGA2ox7, SlACO1, AtJOX1, AtDMR6, and AtLBO1—which have been functionally characterized were aligned to identify conserved domains or motifs in 2OGD family. The result showed that the above 2OGD proteins had the HxD…H and RxS/T conserved motifs in OsGA2ox3 and OsDAO (Figure 3a), which recruit Fe(II) as a cofactor and co-substrate. Further, among the hormone biosynthetic and metabolism 2OGD proteins in *Arabidopsis*, rice, and tomato, SlGA20ox8, SlGA20ox9, SlGA20ox10, SlGA2ox12, and OsACO6 did not have an HxD…H

*2.3. Multiple Sequence Alignment and Motif Composition Analysis of Hormone Biosynthetic and* 

(Figure S3–S11), suggesting that these proteins do not have 2OGD biological activity.

duplication events); and one dispersed gene pair in *SlLBO*.

**Figure 3.** Sequence alignment and conserved motif analysis of functionally characterized hormone biosynthetic and metabolism 2OGD proteins. (**a**) Sequence alignment of functionally characterized hormone biosynthetic and metabolism 2OGD proteins in *Arabidopsis*, rice, tomato. The putative His-X-Asp-(X)n-His (HxD . . . H) and Arg-X-Ser/Thr (RxS/T) motif locations are highlighted in red and black dotted boxes, respectively. (**b**) The motif composition of functionally characterized hormone biosynthetic and metabolism 2OGD proteins. The motif enclosed by red boxes is specific motifs in each group.

However, what is the difference of protein structure among different hormone biosynthetic and metabolism 2OGD protein families? Next, we used MEME to identify conserved motifs in DOXC-family proteins of *Arabidopsis*, rice, and tomato (Tables S4 and S5). The result showed that seven hormones biosynthetic and metabolism 2OGD protein families had uniquely conserved motifs—motifs 29, 40, 35, 45, 25, 44, and 38 were unique to the GA20ox, GA3ox, C19-GA2ox, C20-GA2ox, DAO, ACO, and JOX families, respectively (Figure 3b). However, no specific conserved motif was identified in the DMR6 or LBO families (Table S4). Further, sequence alignments showed that SlGA20ox7, SlGA20ox8, SlGA20ox9, and SlGA20ox10 did not have motif 29 (Figure S3), OsGA2ox10 did not have motif 35 (Figure S5), and OsACO4 did not have motif 25 (Figure S7), suggesting that these six proteins are not related to hormone biosynthesis or metabolism. In addition, SlGA2ox6 and SlGA2ox9 were truncated proteins with several missing amino acids in the N-terminal region (Figure S5). In conclusion, from the result of multiple sequence alignment and motif composition, the results suggesting that SlDAO1-SlDAO3, SlGA20ox1-SlGA20ox6, SlGA3ox1-SlGA3ox4, SlGA2ox1-SlGA2ox5, SlGA2ox7-SlGA2ox8, SlGA2ox10-SlGA2ox11, SlACO1-SlACO7, SlJOX1-SlJOX3, SlDLO1-SlDLO2, and SlLBO1 may have the ability of hormone biosynthesis and metabolism in tomato.

#### *2.4. Expression of Hormone Biosynthetic and Metabolism 2OGD Genes in Tomato*

To assess the function of identified hormone biosynthetic and metabolism 2OGD genes in tomato, we analyzed online transcriptome data of tomato roots, leaves, flowers, and developing fruits. Most genes exhibited distinct spatial and temporal expression patterns (Figure 4). Three *GA3ox* genes exhibited the highest expression in flowers, *SlGA3ox1* had moderate expression in roots and early developing fruits, and *SlGA3ox2* had moderate expression in leaves. No *GA3ox* gene was expressed during fruit ripening (Figure 4a). Regarding the *GA20ox* family, *SlGA20ox1*, *SlGA20ox2*, and *SlGA20ox3* were highly expressed in flowers and early developing fruits; *SlGA20ox1* and *SlGA20ox3* were also expressed in roots, and *SlGA20ox1* and *SlGA20ox2* were expressed in leaves. *SlGA20ox4* was specifically expressed in unopened flowers. Only *SlGA20ox3* was expressed during fruit ripening, during which its expression increased continuously (Figure 4b). Five GA2ox-family genes (*SlGA2ox3*, *4*, *5*, *7*, and *10*) showed high expression in roots, three (*SlGA2ox2*, *3*, and *10*) in leaves, and six (*SlGA2ox1*, *2*, *4*, *5*, *7*, and *10*) in flowers. In addition, four genes (*SlGA2ox2*, *4*, *5*, and *7*) had high expression in early developing fruits, which increased during fruit ripening (from the mature-green stage to the breaker stage) (Figure 4c). Among the *DAO* family, the expression of *SlDAO1* was high in ripening fruits, moderate in early fruits, and low in roots, leaves, and flowers. *SlDAO2* was expressed mainly in flowers and early fruits, while the expression of *SlDAO3* was negligible in all organs. Notably, *SlDAO1* expression increased significantly from the mature-green to the breaker stage, suggesting a role in fruit ripening (Figure 4d). The expression of the three *JOX*-family genes was highest in flowers, while that of *SlJOX1* and *SlJOX2* was moderate in roots, leaves, and early developing fruits, and *SlJOX2* was expressed in breaker fruits (Figure 4e). Regarding the *ACO* family, three genes (*SlACO2*, *3*, and *4*) were expressed in roots, two (*SlACO4* and *5*) in leaves, and five (*SlACO1*, *2*, *3*, *4*, and *6*) in flowers. Further, four genes (*SlACO1*, *3*, *4*, and *6*) had high expression in early developing fruits, and the expression of four other genes (*SlACO1*, *3*, *5*, and *6*) increased from mature-green to breaker fruit (Figure 4f). The *DLO*-family gene *SlDLO1* showed high expression in roots, leaves, flowers, and early fruits, and decreased expression in ripening fruits, while *SlDLO2* was expressed only in flowers and early fruits (Figure 4g). The only *LBO* gene in tomato, *SlLBO1,* was expressed mainly in roots and flowers, suggesting roles in root and flower development (Figure 4h). In conclusion, a variety of 2OGD hormone biosynthetic and metabolism genes play roles in organ development and fruit ripening in tomato.

#### *2.5. Expression of SlGA2ox and SlDAO Genes during Tomato Fruit Ripening*

Ethylene is the major hormone regulating tomato fruit ripening, while auxin and GAs regulate fruit ripening via the ethylene pathway [2–4]. The endogenous auxin and GA concentration was decreased during tomato fruit ripening (Figure S12) [3,4], so we investigated the roles of auxin- and GA-degradation genes on tomato fruit ripening. Tomato pericarps at four stages (mature-green, breaker, yellow-ripening, and red-ripening) were collected from the tomato cultivars 'Ai Ji Qiao Li' and 'Micro-Tom' for qPCR analysis (Figure 5a). The *SlDAO1* expression level was higher than that of *SlDAO2* in Ai Ji Qiao Li and Micro-Tom during fruit ripening (Figure 5b,d). Notably, the expression of *SlDAO1* significantly increased, about two-fold, in Ai Ji Qiao Li, and tenfold in Micro-Tom from the mature-green to the breaker stage; its expression level remained elevated in the yellow- and red-ripening stages. However, *SlDAO2* expression did not significantly change from the mature-green to the breaker stage, and remained very low in the yellow- and red-ripening stages (Figure 5b,d). Thus, *SlDAO1*, rather than *SlDAO2*, likely plays a role in the transition from the mature-green to the breaker stage and subsequent fruit ripening. In addition, the expression of *SlGA2ox2* was 100-fold higher than that of *SlGA2ox4* and *SlGA2ox5*, while *SlGA2ox4* and *SlGA2ox5* expression was negligible in Ai Ji Qiao Li and Micro-Tom (Figure 5c,e). *SlGA2ox2* expression was increased threefold in Ai Ji Qiao Li and thirty-fold in Micro-Tom from the mature-green stage to the breaker stage, and decreased slightly in the yellow- and red-ripening stages (Figure 5c,e); this suggested that *SlGA2ox2* participates in tomato fruit ripening.

development (Figure 4h). In conclusion, a variety of 2OGD hormone biosynthetic and metabolism

**Figure 4.** Expression pattern of hormone biosynthetic and metabolism 2OGD genes in tomato. (**a**–**h**) Expression pattern of *SlGA3ox*, *SlGA20ox*, *SlGA2ox*, *SlDAO*, *SlJOX*, *SlACO*, *SlDLO*, and *SlLBO* group genes. Gray boxes represent the expression of genes was undetectable. Unopened flowers (UF); Opened flowers (F); 1cm fruits (1cm F); 2cm fruits (2cm F); 3cm fruits (3cm F); mature-green fruits (Mg F); breaker fruits (Br F); breaker+10 days' fruits (Br+10 F); roots (R); leaves (L). The detailed descriptions of the stages and tissues were on the website (http://ted.bti.cornell.edu/cgibin/TFGD/digital/home.cgi). **Figure 4.** Expression pattern of hormone biosynthetic and metabolism 2OGD genes in tomato. (**a**–**h**) Expression pattern of *SlGA3ox*, *SlGA20ox*, *SlGA2ox*, *SlDAO*, *SlJOX*, *SlACO*, *SlDLO*, and *SlLBO* group genes. Gray boxes represent the expression of genes was undetectable. Unopened flowers (UF); Opened flowers (F); 1 cm fruits (1 cm F); 2 cm fruits (2 cm F); 3 cm fruits (3 cm F); mature-green fruits (Mg F); breaker fruits (Br F); breaker+10 days' fruits (Br+10 F); roots (R); leaves (L). The detailed descriptions of the stages and tissues were on the website (http://ted.bti.cornell.edu/cgi-bin/TFGD/ digital/home.cgi).

Ethylene is the major hormone regulating tomato fruit ripening, while auxin and GAs regulate fruit ripening via the ethylene pathway [2–4]. The endogenous auxin and GA concentration was decreased during tomato fruit ripening (Figure S12) [3,4], so we investigated the roles of auxin- and GA-degradation genes on tomato fruit ripening. Tomato pericarps at four stages (mature-green, breaker, yellow-ripening, and red-ripening) were collected from the tomato cultivars 'Ai Ji Qiao Li' and 'Micro-Tom' for qPCR analysis (Figure 5a). The *SlDAO1* expression level was higher than that of *SlDAO2* in Ai Ji Qiao Li and Micro-Tom during fruit ripening (Figure 5b,d). Notably, the expression of *SlDAO1* significantly increased, about two-fold, in Ai Ji Qiao Li, and tenfold in Micro-Tom from the mature-green to the breaker stage; its expression level remained elevated in the yellow- and redripening stages. However, *SlDAO2* expression did not significantly change from the mature-green to the breaker stage, and remained very low in the yellow- and red-ripening stages (Figure 5b,d). Thus,

*2.5. Expression of SlGA2ox and SlDAO Genes during Tomato Fruit Ripening* 

*SlDAO1*, rather than *SlDAO2*, likely plays a role in the transition from the mature-green to the breaker stage and subsequent fruit ripening. In addition, the expression of *SlGA2ox2* was 100-fold higher than that of *SlGA2ox4* and *SlGA2ox5*, while *SlGA2ox4* and *SlGA2ox5* expression was negligible in Ai Ji Qiao Li and Micro-Tom (Figure 5c,e). *SlGA2ox2* expression was increased threefold in Ai Ji Qiao Li and thirty-fold in Micro-Tom from the mature-green stage to the breaker stage, and decreased slightly in the yellow- and red-ripening stages (Figure 5c,e); this suggested that *SlGA2ox2* participates in tomato

**Figure 5.** Expression analysis of *SlDAOs* and *SlGA2oxs* genes during tomato fruit ripening in the pericarp. (**a**) Different ripening stages of Ai Ji Qiao Li and Micro-Tom. (**b**) Expression levels of *SlDAOs*  in Ai Ji Qiao Li. (**c**) Expression levels of *SlGA2ox* genes in Ai Ji Qi Li. (**d**) Expression levels of *SlDAOs*  in Micro-Tom. (**e**) Expression levels of *SlGA2ox* genes in Micro-Tom. Mg: mature-green; Br: breaker; Yr: yellow-ripening; Rr: red-ripening. \* The asterisk at the top of each column indicates a significant **Figure 5.** Expression analysis of *SlDAOs* and *SlGA2oxs* genes during tomato fruit ripening in the pericarp. (**a**) Different ripening stages of Ai Ji Qiao Li and Micro-Tom. (**b**) Expression levels of *SlDAOs* in Ai Ji Qiao Li. (**c**) Expression levels of *SlGA2ox* genes in Ai Ji Qi Li. (**d**) Expression levels of *SlDAOs* in Micro-Tom. (**e**) Expression levels of *SlGA2ox* genes in Micro-Tom. Mg: mature-green; Br: breaker; Yr: yellow-ripening; Rr: red-ripening. \* The asterisk at the top of each column indicates a significant difference compared to Mg fruits at *p* < 0.05 (*n* = 3) by students t-test.

#### *2.6. E*ff*ects of Auxin, GA3, and Ethylene on the Expression of SlDAO1, SlDAO2, and SlGA2ox2*

difference compared to Mg fruits at *p* < 0.05 (*n* = 3) by students t-test.

*2.6. Effects of Auxin, GA3, and Ethylene on the Expression of SlDAO1, SlDAO2, and SlGA2ox2*  To study the response of *SlDAO1*, *SlDAO2*, and *SlGA2ox2* to auxin, GAs, and ethylene, we treated Micro-Tom mature-green fruits with IAA, GA3, and ethylene, and analyzed their expression after 2 and 4 days. Consistent with previous reports, IAA and GA3 delayed tomato fruit ripening (Figure 6a). Further, the expression of *SlDAO1* was significantly induced by IAA, but was unaffected by GA3 and ethylene at 2 and 4 days, while *SlDAO2* expression was not significantly affected in auxin- , GA-, or ethylene-treated mature-green fruits (Figure 6b). In addition, *SlGA2ox2* showed higher expression in GA3-treated fruits, but similar expression in IAA- and ethylene-treated fruits, compared To study the response of *SlDAO1*, *SlDAO2*, and *SlGA2ox2* to auxin, GAs, and ethylene, we treated Micro-Tom mature-green fruits with IAA, GA3, and ethylene, and analyzed their expression after 2 and 4 days. Consistent with previous reports, IAA and GA<sup>3</sup> delayed tomato fruit ripening (Figure 6a). Further, the expression of *SlDAO1* was significantly induced by IAA, but was unaffected by GA<sup>3</sup> and ethylene at 2 and 4 days, while *SlDAO2* expression was not significantly affected in auxin-, GA-,or ethylene-treated mature-green fruits (Figure 6b). In addition, *SlGA2ox2* showed higher expression in GA3-treated fruits, but similar expression in IAA- and ethylene-treated fruits, compared to the control (Figure 6b). In conclusion, the expression of *SlDAO1* and *SlGA2ox2* was induced by auxin and GAs, respectively, suggesting that *SlDAO1* and *SlGA2ox2* are responsible for regulating auxin and GA catabolism during tomato fruit ripening.

auxin and GA catabolism during tomato fruit ripening.

to the control (Figure 6b). In conclusion, the expression of *SlDAO1* and *SlGA2ox2* was induced by

**Figure 6.** Expression analysis of *SlDAO1*, *SlDAO2*, and *SlGA2ox2* after auxin, GA3, and ethylene treatments. (**a**) Photos of mature-green fruits after indole-3-acetic acid (IAA) and gibberellin 3 (GA3) treatment, respectively. (**b**) Expression analysis of *SlDAO1, SlDAO2*, and *SlGA2ox2* after auxin, GA3, and ethylene treatments. \* The asterisk at the top of each column indicates a significant difference at *p* < 0.05 (*n* = 3) by students t-test. **Figure 6.** Expression analysis of *SlDAO1*, *SlDAO2*, and *SlGA2ox2* after auxin, GA<sup>3</sup> , and ethylene treatments. (**a**) Photos of mature-green fruits after indole-3-acetic acid (IAA) and gibberellin 3 (GA<sup>3</sup> ) treatment, respectively. (**b**) Expression analysis of *SlDAO1, SlDAO2*, and *SlGA2ox2* after auxin, GA<sup>3</sup> , and ethylene treatments. \* The asterisk at the top of each column indicates a significant difference at *p* < 0.05 (*n* = 3) by students t-test.

#### **3. Discussion 3. Discussion**

#### *3.1. Identification of Hormone Biosynthetic and Metabolism Genes from 2OGD Family*

*3.1. Identification of Hormone Biosynthetic and Metabolism Genes from 2OGD Family*  The 2OGD superfamily is widespread in microorganisms, fungi, mammals, and plants. In plants, 2OGD proteins are classified as DOXA, DOXB, and DOXC [14]. DOXA proteins are involved in the oxidative demethylation of alkylated nucleic acids and histones, while DOXB proteins are involved in proline 4-hydroxylation in cell-wall protein synthesis, and DOXC proteins in the metabolism of various phytochemicals, such as phytohormones and flavonoids. The number of 2OGDs of the DOXA and DOXB classes is constant across plant species, whereas that of the DOXC class is extremely variable, suggesting that the latter has diversified during the evolution of land plants. The vast majority of 2OGDs from land plants are of the DOXC class, including all hormone biosynthesis- and metabolism-related proteins of the 2OGD family. In this study, the number and classifications of DOXC hormone biosynthesis- and metabolism-related proteins were consistent with the report by Kawal et al. [14]. DOXC proteins are involved in the biosynthesis and metabolism of the phytohormones auxin, GAs, ethylene, JA, SA, and SLs, which play important roles in plant growth and development. Furthermore, the number of DOXC hormone biosynthetic and metabolism genes increases from ancient lower land plants to higher plants, consistent with the high complexity The 2OGD superfamily is widespread in microorganisms, fungi, mammals, and plants. In plants, 2OGD proteins are classified as DOXA, DOXB, and DOXC [14]. DOXA proteins are involved in the oxidative demethylation of alkylated nucleic acids and histones, while DOXB proteins are involved in proline 4-hydroxylation in cell-wall protein synthesis, and DOXC proteins in the metabolism of various phytochemicals, such as phytohormones and flavonoids. The number of 2OGDs of the DOXA and DOXB classes is constant across plant species, whereas that of the DOXC class is extremely variable, suggesting that the latter has diversified during the evolution of land plants. The vast majority of 2OGDs from land plants are of the DOXC class, including all hormone biosynthesis- and metabolism-related proteins of the 2OGD family. In this study, the number and classifications of DOXC hormone biosynthesis- and metabolism-related proteins were consistent with the report by Kawal et al. [14]. DOXC proteins are involved in the biosynthesis and metabolism of the phytohormones auxin, GAs, ethylene, JA, SA, and SLs, which play important roles in plant growth and development. Furthermore, the number of DOXC hormone biosynthetic and metabolism genes increases from ancient lower land plants to higher plants, consistent with the high complexity and diversity—and specialized metabolism—of higher plants.

and diversity—and specialized metabolism—of higher plants. Although the 2OGD superfamily is highly diverse, structural studies suggest that its members have a highly conserved Fe(II) binding HxD/E…H triad motif and a less conserved 2OG C5 carboxy group binding motif (RxS/T) [13]. In this study, forty-three hormone biosynthetic and metabolism proteins of the DOXC family were identified in tomato, but five SlGA20ox7, SlGA20ox8, SlGA20ox9, SlGA20ox10 and SlGA2ox12) lacked the HxD/E…H or RxS/T motif (Figure S3,6), suggesting a lack of 2OGD activity. In addition, we identified family-specific conserved motifs in DAOs, GA20oxs, GA3oxs, C19-GA2oxs, C20-GA2oxs, ACOs, and JOXs (Figure 3b); however, their function was Although the 2OGD superfamily is highly diverse, structural studies suggest that its members have a highly conserved Fe(II) binding HxD/E . . . H triad motif and a less conserved 2OG C5 carboxy group binding motif (RxS/T) [13]. In this study, forty-three hormone biosynthetic and metabolism proteins of the DOXC family were identified in tomato, but five SlGA20ox7, SlGA20ox8, SlGA20ox9, SlGA20ox10 and SlGA2ox12) lacked the HxD/E . . . H or RxS/T motif (Figures S3 and S6), suggesting a lack of 2OGD activity. In addition, we identified family-specific conserved motifs in DAOs, GA20oxs, GA3oxs, C19-GA2oxs, C20-GA2oxs, ACOs, and JOXs (Figure 3b); however, their function was unclear. A MdACO1 protein with mutated conserved Lys296 and Arg299 residues in the C-terminal helix retained only 15–30% of the activity of the wild-type, possibly because these two residues are important for ACO activity and may be involved in binding bicarbonate, the unique activator of ACOs [18].

Notably, these two amino acids are located in the ACO-specific conserved motif identified in this study (Figure S7). Therefore, the subfamily-specific conserved motifs may play important roles in the functional differentiation of 2OGD subfamilies.

#### *3.2. Functional Analysis of Hormone Biosynthetic and Metabolism Genes in 2OGD Family*

GAs, ethylene, auxin, JA, SA, and SLs regulate many aspects of plant growth and development, and the response to stresses. Several 2OGD genes involved in hormone biosynthesis and metabolism have been functionally analyzed in *Arabidopsis* and rice, and these genes participate in the development of roots, stems, flowers, fruits, and seeds. In tomato, the *SlGA20oxs* GA-biosynthetic genes, particularly *SlGA3oxs*, which function in the final step of GA biosynthesis, were mainly expressed in tomato roots, leaves, flowers, and early developing fruits, suggesting that GAs play a role in the development of these tissues/organs (Figure 4a,b). Consistently, RNAi-mediated silencing of *SlGA20ox1*, *SlGA20ox2*, or *SlGA20ox3* affected the development of tomato stems, leaves, fruit, and seeds [19], and inhibitors of GA biosynthesis decreased tomato fruit growth and fruit set; also, exogenous GA<sup>3</sup> induced parthenocarpic fruits [20,21]. The *SlGA2oxs* GA-metabolism proteins also play key roles in regulating endogenous GA levels. The silencing of *SlGA2ox1-SlGA2ox5* increased the active GA<sup>4</sup> content, induced parthenocarpic fruits, and inhibited lateral branching in tomato plants [22]. In this study, the newly identified genes *SlGA2ox7* and *SlGA2ox10*, mainly expressed in roots, leaves, flowers, and early developing fruits (Figure 4c), had the same conserved motif as *SlGA2ox1* to *SlGA2ox5* (Figure S5), suggesting a role for *SlGA2ox7* and *SlGA2ox10* in the metabolism of GAs during the development of these tissues/organs.

Although auxin regulates the growth and development of various plant tissues and organs, studies of auxin in tomato have focused on fruit set and development. Exogenous auxin treatment could induce parthenocarpic fruits, and altering the expression of auxin response genes also affected tomato fruit set and development [21,23]. *DAO*-family proteins irreversibly degrade auxin, and a *dao* mutant in rice displayed defective pollen fertility and seed development [7]; meanwhile, a *dao1* mutant in *Arabidopsis* displayed larger cotyledons, increased lateral root density, and elongated pistils [24]. *DAO* has three homologs in tomato; the expression of *SlDAO2* was higher in flowers and early developing fruits compared to *SlDAO1* and *SlDAO3*, suggesting a role in regulating the auxin level for fruit set and development (Figure 4d). Ethylene plays important roles in fruit set and development [25], especially fruit ripening, likely due to high expression of the ethylene-biosynthetic genes *SlACO1*, *SlACO3*, and *SlACO6* in flower, early developing fruits, and ripening fruits (Figure 4f). Other *ACO* genes (*SlACO2* and *SlACO4*) may contribute to ethylene production for root and flower development. In addition, three JA-metabolism *SlJOX* genes showed high expression in tomato flowers (Figure 4e), indicating roles in regulating JA homeostasis for flowering [26]. *AtDMR6,* the product of which degrades salicylic acid, was involved in plant growth and resistance to pathogens, and the *dmr6* mutant displayed smaller size, early senescence, and a loss of susceptibility to *Pseudomonas syringae* pv tomato DC3000 [10]. In tomato, the homolog *SlDLO1* was highly expressed in roots, leaves, flowers, and fruits (Figure 4g), and CRISPR-Cas9 mediated the mutagenesis of *SlDLO1* in tomato conferred broad-spectrum disease resistance; however, vegetative growth and development were not significantly affected, and its role in reproductive organs was not investigated [27]. *SlDLO2* is highly expressed only in flowers and fruits, suggesting roles in regulating the SA level in reproductive organs. SLs are plant hormones that regulate plant root and branch development, as well as stress tolerance [28,29]. High expression of SL biosynthetic and signaling genes in tomato or strawberry fruit indicated roles in fruit development [30]. *LBO* acts in the final stages of SL biosynthesis to produce active SLs in *Arabidopsis*, and its homolog *SlLBO1* is only expressed in roots and flowers (Figure 4h). This suggests that SLs are synthesized in tomato roots and flowers, but does not mean that SLs have no effect on fruit development; they could be transported to fruit from other organs or tissues.

#### *3.3. SlGA2ox2 and SlDAO1 May Play a Role in GA and Auxin Metabolism for Normal Ripening of Tomato Fruits*

Tomato is a model plant for studying the ripening of climacteric fruits, and ethylene regulates tomato fruit ripening. In this study, exogenous GA<sup>3</sup> treatment of tomato fruits at the mature-green stage delayed fruit ripening, while overexpression of the GA catabolism gene *SlGA2ox1* specifically in tomato fruits led to early ripening [4]. We have previously shown that GAs play negative roles in the ethylene pathway by inhibiting the expression of ethylene biosynthetic genes (*SlACS2*, *SlACS4,* and *SlACO1*) and signaling genes (*SlETRs* and *SlEINs*) [4]. Therefore, the concentration of GAs in fruits influences fruit ripening in tomato. In plants, the GA level is regulated by the balance between biosynthesis and metabolism. GA20oxs and GA3oxs catalyze the rate-limiting step of active GA biosynthesis, and GA2oxs converts bioactive GAs or their immediate precursors into inactive forms. In this study, although the expression of one *GA20ox* gene (*SlGA20ox3*) increased from the mature-green to the breaker stage (Figure 4b), no *GA3ox* genes, which encode enzymes that catalyze the last step of GA biosynthesis, were expressed (Figure 4a), suggesting the absence of GA biosynthesis in mature-green and breaker fruits. Further, the expression of three GA-metabolism genes (*SlGA2ox2*, *SlGA2ox4*, and *SlGA2ox5*) was increased, and that of *SlGA2ox2* was highest, and dramatically increased, from the mature-green to the breaker stage (Figure 4c). It has been reported that the concentrations of endogenous active GAs (GA<sup>1</sup> and GA4) in the fruit pericarp of tomato decrease significantly from the mature -green to the breaker stage (Figure S12) [4]. Therefore, we speculate that *SlGA2ox2* may be vital for GA metabolism from the mature-green to the breaker stage, and the reduced GA level caused by the increase in *SlGA2ox2* expression promotes tomato fruit ripening.

Auxin also negatively regulates tomato fruit ripening. Exogenous applications of IAA reduced expression of ethylene biosynthetic and consequently reduced ethylene production, and also the ethylene signaling genes, resulting in delayed tomato fruit ripening [1,2]. The concentration of endogenous auxin in tomato fruit pericarps is reduced from the mature-green to the breaker stage (Figure S12) [3]. In plants, auxin is synthesized by tryptophan (Trp)-dependent and -independent pathways [31]. Our knowledge of the genes and intermediates of the Trp-independent pathway is limited, but the complete Trp-dependent pathway has been established. YUCCA (YUC) family proteins function in the final step of Trp-dependent auxin biosynthesis, and play a crucial role in auxin biosynthesis in various plant species. In tomato, six *YUC* genes were identified, the transcript levels of five of which were negligible, whereas one *YUC* gene (*ToFZY4*) displayed high expression during ripening of tomato fruit [32]. It is not clear why the auxin concentration was decreased, but the expression of a key gene in auxin biosynthesis was increased in ripening tomato fruit. One explanation for this is that there is a change from the Trp-dependent to the Trp-independent pathway for auxin biosynthesis between the mature and red-ripe stages of tomato fruits [33], and *ToFZY4* may have a novel function related to tomato fruit ripening rather than auxin biosynthesis. Auxin can be deactivated by conjugation to amino acids, or by chemical oxidation. Conjugation of IAA to amino acids is catalyzed by *GH3*-family proteins and yields, for instance, indole-3-acetic acid aspartic acid (IAA-Asp) and indole-3-acetic acid glutamic acid (IAA-Glu). The chemical oxidation of auxin is catalyzed by DAO-family proteins to produce oxIAA. In tomato, 24 *GH3* genes were identified, only 4 (*SlGH3-1*, *SlGH3-2*, *SlGH3-5*, and *SlGH3-24*) of which showed high expression during fruit ripening [3]. Silencing of *SlGH3-2* in tomato increased the auxin level and reduced lycopene accumulation in ripening fruit, suggesting that *SlGH3-2* plays a role in deactivating free auxin to maintaining normal ripening of tomato fruit [3]. However, oxIAA is a major IAA catabolite, where up to 10–100 folds more oxIAA than the major IAA conjugates IAA-Glu and IAA-Asp was detected in *Arabidopsis* [34,35]. More importantly, oxIAA oxidized by DAO is biologically inactive, and is formed rapidly and irreversibly in plant tissues [34–36]. *DAO* is likely involved in maintaining the basal level of active auxin under normal growth conditions, while *GH3* functions in the response to various environmental factors [37]. In this study, we identified three DAO genes in tomato. *SlDAO3* had lost some sequences in the N-terminal (Figure S8), suggesting that it may be not involved in IAA degradation. *SlDAO2* expression was

negligible, but that of *SlDAO1* was high and increased from mature-green to breaker fruits (Figure 4d); moreover, it was significantly induced by auxin in mature-green fruits (Figure 6b). These results implicate *SlDAO1*, rather than *SlDAO2* and *SlDAO3*, in auxin metabolism from the mature-green to the breaker stage during tomato ripening. In addition, the reduction in auxin level caused by the increase in *SlDAO1* expression may play an important role in maintaining normal ripening of tomato fruit.

#### **4. Materials and Methods**

#### *4.1. Identification and Phylogenetic Analysis of Hormone Biosynthesis and Metabolism Related DOXC Proteins*

To find proteins belonging to DOXC family, we used 2OG-FeII\_Oxy (PF03171) domain as query in hmmsearch BLAST of *Arabidopsis*, rice, and tomato protein databases downloaded from JGI [38]. All sequences (length <sup>≥</sup> 100 aa) with an E-value cutoff <sup>1</sup> <sup>×</sup> <sup>10</sup>−<sup>4</sup> were retrieved. The obtained sequences were submitted to Pfam [39] and SMART [40] to verify the existence of 2OG-FeII\_Oxy domain. In order to better understand the relationship among all members of the DOXC and identify proteins involved in hormone biosynthesis and metabolism, we then used all verified protein sequences to construct a phylogenetic tree by MEGA6 with Maximum likelihood. The best model JTT + F was selected by Model Generator software. According to hormone biosynthesis and metabolism related genes with known function in *Arabidopsis* and rice, all proteins which clustered into hormone biosynthesis and metabolism related protein subfamilies were selected to construct a new phylogenetic tree.

#### *4.2. Chromosomal Location and Synteny Analysis*

Genome annotation files were downloaded from the *Arabidopsis*, rice, and tomato databases to obtain chromosomal location information of these hormone biosynthetic and metabolism genes, then the Circos software was used to draw location pictures. A method similar to that developed for the Plant Genome Duplication Database (PGDD) [41] was used to identify syntenic blocks in *Arabidopsis*, rice, and tomato. Potential homologous sequences were initially identified by BLASTP (E-value <sup>&</sup>lt; <sup>1</sup> <sup>×</sup> <sup>10</sup>−<sup>5</sup> , top 5 matches). MCScanX was used for synteny analysis [42]. Additionally, MCScanX was further used to detect duplicate types of these biosynthetic and metabolism genes in tomato.

#### *4.3. Multiple Sequence Alignment and Motif Composition Analysis*

To detect the HxD/E . . . H and RxS/T motifs, multiple sequence alignments were performed by submitting protein sequences to ClustalW with the default parameters in BioEdit software. Motif composition analysis was performed by submitting protein sequences to MEME [43] with the following parameters: the maximum number of motifs was 50 and the maximum motif length was 15 amino acids.

#### *4.4. Expression Analysis*

Transcriptome datasets of different tomato organs were downloaded from Tomato Functional Genomics Database [44]. RPKM values of related genes were transformed in log<sup>2</sup> level, and a heatmap was shown using MeV4.8 software (Dana-Farber Cancer Institute, Boston, MA, USA).

#### *4.5. Plant Materials and Hormone Treatments*

Two tomato cultivars Ai Ji Qiao Li grown in greenhouse and Micro-Tom grown in climate chamber were chosen as plant materials. The fruit was collected at four different ripening stages: mature-green (Mg), breaker (Br), yellow-ripening (Yr), and red-ripening (Rr). The fruit pericarp sample without placenta and seeds was collected and then immediately frozen in liquid nitrogen prior to storage at −80 ◦C until RNA extraction.

Tomato cultivars Micro-Tom grown in climate chamber was used for hormone treatments of fruits. Flowers were tagged at the date of pollination. After 36 days, mature-green fruits on the plants were

injected with 0.1 mM IAA, 0.1 mM GA3, and 0.1 mM ethephon, respectively, distilled water was used as the control. The amount of injection was about 50 µL per fruit. Twelve fruits for each treatment were performed. The fruit pericarp without placenta and seeds was collected at two days and four days after treatments, and were immediately frozen in liquid nitrogen, and then stored at −80 ◦C. Plant growth conditions was: 16-h light (25 ◦C)/8-h dark (18 ◦C) photoperiod cycle and 65% relative humidity. In addition, detached mature-green fruits were injected with 0.1 mM IAA and 0.1 mM GA3, respectively, distilled water was used as the control. Then the fruit was placed under dark at 25 ◦C and 90% relative humidity, photos were taken after eight days.

#### *4.6. RNA Extraction and qPCR Analysis of Selected Genes*

Total RNA was extracted with a modified CTAB method [4]. cDNA library was generated by Primerscript RT reagent Kit with gDNA Erase (Takara, Beijing, China) according to the manufacturer's protocol. qPCR was carried out using SYBR Premix Ex Taq II (Takara, Beijing, China). Primer sequences were listed in Table S6. Three biologicals with triplicates were performed and results were analyzed using the 2−∆CT method. *Actin* gene (gene ID: Solyc11g005330) was used as the reference.

#### **5. Conclusions**

We have identified 43 hormone biosynthetic and metabolism genes of nine subfamilies of the 2OGD family, which were related to GAs, ethylene, auxin, JA, SA, and SLs in tomato. The subfamily-specific conserved motifs identified in this study might play roles in the functional differentiation of 2OGD subfamilies, and the different expression profiles suggest that these genes play diverse roles in tomato organ growth and development. Especially, the expression levels of the auxin-degradation gene *SlDAO1* and the GA-degradation gene *SlGA2ox2* were significantly increased from the mature-green to the breaker stage during tomato fruit ripening, accompanied by decreased endogenous IAA and GAs levels. In addition, the expression of *SlDAO1* and *SlGA2ox2* was increased by IAA and GA3, respectively, indicating that *SlDAO1* and *SlGA2ox2* may be responsible for reducing IAA and GA concentrations to maintain normal ripening of tomato fruit.

**Supplementary Materials:** Supplementary materials can be found at http://www.mdpi.com/1422-0067/21/15/ 5344/s1, Figure S1: Phylogenetic tree of DOXC family proteins identified in *Arabidopsis*, rice, and tomato, Figure S2: Chromosomal location and duplication analysis of hormone biosynthetic and metabolism genes in tomato, Figure S3: Sequence alignment of GA20ox group proteins, Figure S4: Sequence alignment of GA3ox group proteins, Figure S5: Sequence alignment of GA2ox (C19) group proteins, Figure S6: Sequence alignment of GA2ox (C20) group proteins, Figure S7: Sequence alignment of ACO group proteins, Figure S8: Sequence alignment of DAO group proteins, Figure S9: Sequence alignment of Jox group proteins, Figure S10: Sequence alignment of DMR6 group proteins, Figure S11: Sequence alignment of LBO group proteins, Figure S12: The concentrations of GA<sup>1</sup> , GA<sup>4</sup> , and IAA decrease from the mature-green (Mg) to the breaker (Br) stage, Table S1: Identification and characterization of hormone biosynthetic and metabolism proteins in *Arabidopsis*, rice, and tomato, Table S2: Microsynteny relationships of hormone biosynthetic and metabolism genes in *Arabidopsis*, rice, and tomato, Table S3: Duplication modes of hormone biosynthetic and metabolism genes in tomato, Table S4: Motif composition in DOXC class of 2OGD family in *Arabidopsis*, rice, and tomato, Table S5: Motif sequence identified by MEME tools in *Arabidopsis*, rice, and tomato DOXC class of 2OGDs, Table S6: Primers used in this study.

**Author Contributions:** Data curation, X.Y. and H.W.; formal analysis, Q.D.; funding acquisition, H.W.; investigation, Q.D. and F.W.; methodology, Q.D. and J.X.; software, F.W. and J.F.; supervision, H.W.; validation, H.W.; visualization, H.C.; writing—original draft, Q.D. and H.W.; writing—review and editing, Y.L. and H.W. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by the National Key R&D Program of China (2016YFD0100506); The Special Financial Grant Program from the China Postdoctoral Science Foundation [2016T90471]; The General Financial Grant Program from the China Postdoctoral Science Foundation [2015M581812]; The Priority Academic Program Development of Jiangsu Higher Education Institutions (PAPD).

**Acknowledgments:** We thank Helin Tan for his support for the experiment in this study.

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

## **Abbreviations**


## **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## *Article* **Silencing of** *TaCKX1* **Mediates Expression of Other** *TaCKX* **Genes to Increase Yield Parameters in Wheat**

**Bartosz Jabło ´nski <sup>1</sup> , Hanna Ogonowska <sup>1</sup> , Karolina Szala <sup>1</sup> , Andrzej Bajguz <sup>2</sup> , Wacław Orczyk <sup>3</sup> and Anna Nadolska-Orczyk 1,\***


Received: 19 June 2020; Accepted: 3 July 2020; Published: 7 July 2020

**Abstract:** *TaCKX*, *Triticum aestivum* (cytokinin oxidase/dehydrogenase) family genes influence the development of wheat plants by the specific regulation of cytokinin content in different organs. However, their detailed role is not known. The *TaCKX1*, highly and specifically expressed in developing spikes and in seedling roots, was silenced by RNAi-mediated gene silencing via *Agrobacterium tumefaciens* and the effect of silencing was investigated in 7 DAP (days after pollination) spikes of T<sup>1</sup> and T<sup>2</sup> generations. Various levels of *TaCKX1* silencing in both generations influence different models of co-expression with other *TaCKX* genes and parameters of yield-related traits. Only a high level of silencing in T<sup>2</sup> resulted in strong down-regulation of *TaCKX11 (3)*, up-regulation of *TaCKX2.1*, *2.2*, *5,* and *9* (*10*), and a high yielding phenotype. This phenotype is characterized by a higher spike number, grain number, and grain yield, but lower thousand grain weight (TGW). The content of most of cytokinin forms in 7 DAP spikes of silenced T<sup>2</sup> lines increased from 23% to 76% compared to the non-silenced control. The CKs cross talk with other phytohormones. Each of the tested yield-related traits is regulated by various up- or down-regulated *TaCKX* genes and phytohormones. The coordinated effect of *TaCKX1* silencing on the expression of other *TaCKX* genes, phytohormone levels in 7 DAP spikes, and yield-related traits in silenced T<sup>2</sup> lines is presented.

**Keywords:** wheat; cereals; *TaCKX1*; *TaCKX* expression; grain yield; cytokinins; phytohormones; gene silencing; RNAi; wheat spikes

### **1. Introduction**

Wheat (*Triticum aestivum* L.) is the third most economically important crop in the world after corn and rice, and probably the most important in moderate climates. It provides approximately 20% of human calories and protein [1]. The large genome of this high-yielding species, composed of three (AABBDD) genomes, has been very challenging for improving traits [2]. However, it might be a great reservoir to sustain a further increase of grain productivity [3]. The continuous increase of wheat production is necessary to feed the rapidly growing world population [4]. Biotechnological tools implemented in the process of increasing wheat productivity are expected to be beneficial.

Cytokinins (CKs) are important regulators of plant growth and development, influencing many agriculturally important processes [5]. This regulation might occur at the posttranscriptional and/or posttranslational level [6,7], or by the modulation of context-dependent chromatin accessibility [8]. CKs modulate the expression of other genes involved in the control of various processes including

meristem activity, hormonal cross talk, nutrient acquisition, and various stress responses [9]. There is growing evidence on their key role in seed yield regulation [10]. In cereals and grasses, an increased content of CKs has been reported to positively affect sink potential in developing grains [11] and maintain leaf chlorophyll status during plant senescence [12] and grain filling [13].

The majority of naturally occurring CKs in plants belong to isoprenoid cytokinins grouping *N*<sup>6</sup> -(12-isopentenyl) adenine (iP), *trans*-zeatin (tZ), *cis*-zeatin (cZ), and dihydrozeatin (DZ) derived from tRNA degradation or from isopentenylation of free adenine nucleosides catalysed by isopentenyltransferase (IPT) or tRNA-IPT. The second, smaller group comprise N6-aromatic CKs, represented by benzyladenine (BA) [14]. To better characterize their physiological role, CKs are classified into such -base active forms as tZ, cZ, and iP, translocation forms (nucleosides) as tZ-ribosides (tZR), which exhibit a low level of activity, and sugar conjugates (*O*-glucosides), which are storage and inactivated forms [14,15].

CKs function as local or long-distance regulatory signals, but the mechanisms of their precise spatial and temporal control are still largely unknown [16]. They are produced in roots as well as in various sites of the aerial part of plants [17]. The level of CKs in respective cells and tissues is dependent on many processes, including biosynthesis, metabolism, activation, transport, and signal transduction. Active CKs can be metabolized via oxidation by cytokinin oxidase/dehydrogenase (CKX) or by activity of glycosyltransferases. Many reports have demonstrated that the irreversible degradation step by the CKX enzyme plays an important role in the regulation of cytokinin level in some cereals, namely maize [18], rice [19], barley [20,21], and wheat [22].

The *CKX* gene families in plants show different numbers of genes and various expression patterns, which are tissue- and organ-specific, suggesting gene-specific functions. The specificity of expression of 11 *TaCKX* in developing wheat plants were assigned to four groups: highly specific to leaves, specific to developing spikes and inflorescences, highly specific to roots and expressed through all the organs tested [23]. The *TaCKX* genes co-operated inside and among organs. Their role in plant productivity has been described in many plants including model plants and some cereals. Knock-out mutation or silencing by RNAi of *OsCKX2* in rice significantly increased grain number [19]. The same effect of elevated grain number, spike number, and yield was reported for RNAi-silenced *HvCKX1* in barley [20,21,24] and repeated for the same gene under field conditions [25]. Moreover, significantly increased grain number per spike was found as the effect of the *TaCKX2.4* gene silencing by RNAi [26]. Knock-out mutation of *HvCKX1* by CRISPR/Cas9 editing had a limited effect on yield productivity, however significantly decreased CKX enzyme activity in young spikes and 10-day old roots corresponded to greater root length, numbers of root hairs and increased surface area [27]. In contrast, roots of knock-out mutants of ckx3 were smaller.

The role of other *TaCKX* genes in wheat was analysed based on natural *TaCKX* variation. Haplotype variants of *TaCKX6a02* and *TaCKX6-D1* were related to higher filling rate and grain size [28,29]. Quantitative trait locus (QTL) found in recombinant inbred lines containing a higher copy number of *TaCKX4* was associated with higher chlorophyll content and grain size [30].

To arrange the numbering of *TaCKX* family genes, a new annotation for the first two was suggested by Ogonowska et al. (2019) based on the Ensembl Plants database [31] and phylogenetic analysis. *TaCKX6a02* was annotated as *TaCKX2.1*, *TaCKX6-D1* (JQ797673) was annotated as *TaCKX2.2* and *TaCKX2.4* was annotated as *TaCKX2.2*. Annotations for these genes were maintained in the recently published review on the *TaCKX* [22], however tested in this research *TaCKX10* was renamed as *TaCKX9* and *TaCKX3* was renamed as *TaCKX11*. Newly revised by Chen et al. [22], naming is applied and former names are given in brackets.

Due to the size and complexity of the wheat genomes, the knowledge about the role of *TaCKX* genes, containing homologues from three genomes, is more difficult to obtain, because of the limited number of natural mutants. Most homoeologous genes are expected to have overlapping functions [32], therefore the effect of gene mutations might be masked by the other genomes. One solution to silence all of them is to apply RNAi-mediated gene silencing, which allows silencing of all the homologues. Moreover, this tool made it possible to obtain a number of lines with different levels of silencing, which in the case of genes coding proteins of key importance for life gave a possibility to regenerate plants for analysis [33]. The introduction of a silencing cassette by stable transformation results in a stable, and inherited to T4, effect of silencing [21,34]. The applicability of *Agrobacterium*-mediated transformation compared to a biolistic one for gene silencing of the developmentally regulated gene *HvCKX10* (*2*) was proved to be reliable [24].

We present the first report on the role of *TaCKX1* in the co-regulation of expression of other *TaCKX* genes, phytohormone content, and their joint participation in the regulation of yield-related traits in wheat. Various levels of gene silencing in T<sup>1</sup> and T<sup>2</sup> have been related to different patterns of other *TaCKX* expression, strongly influencing yield-related traits. Models of regulation of phytohormone levels and phenotypic traits in non-silenced and highly silenced T<sup>2</sup> plants by the coordinated expression of *TaCKX* genes are proposed.

#### **2. Results**

### *2.1. Expression Levels of Silenced TaCKX1 in Segregating T<sup>1</sup> and T<sup>2</sup> Plants*

Expression levels of *TaCKX1* were measured in 44 segregating T<sup>1</sup> plants from 8 T<sup>0</sup> PCR+ lines. In 14 T<sup>1</sup> plants relative expression (related to the control = 1.00) ranged from 0.39 to 0.88 with the mean of 0.67 (±0.14). In 30 T<sup>1</sup> plants, relative expression ranged from 0.90 to 1.52 with the mean of 1.16 (±0.18) (Figure 1). The proportion of silenced to non-silenced plants changed in the T<sup>2</sup> generation. There were 42 silenced from 0.24 to 0.88 plants with the mean of 0.54 (±0.14) and 20 non-silenced plants. Eight of them, with low relative expression ranging from 0.24 to 0.40 (mean 0.33 ±0.14) and representing different T<sup>1</sup> lines, were selected for further analysis.

**Figure 1.** Relative expression level of silenced *TaCKX1* in segregating T<sup>1</sup> (**a**) and T<sup>2</sup> (**b**) plants. The level of expression is related to the control set as 1.00.

## *2.2. Co-Expression of Silenced TaCKX1 with Other TaCKX Genes in T<sup>1</sup> and T<sup>2</sup> and CKX Enzyme Activity*

Mean relative expression of *TaCKX1* in the selected 8 lines was 0.67 in T<sup>1</sup> and was decreased to 0.33 in T<sup>2</sup> (Figure 2). Similarly, in the case of *TaCKX11* (*3*) related gene expression was 0.81 in T<sup>1</sup> and was decreased to 0.34 in T2. Relative expression levels of *TaCKX2.2* and *TaCKX9* (*10*) were decreased in T<sup>1</sup> to 0.51 and 0.39 and increased in T<sup>2</sup> slightly above the control level, to 1.08 and to 1.10 respectively. Mean relative values for *TaCKX2.1* were similar to control in T<sup>1</sup> (1.05) and slightly increased in T<sup>2</sup> (1.17). Relative expression of *TaCKX5*, which was in T<sup>1</sup> below the control level (0.84), was significantly increased to 1.82 in T2. The relative values of CKX enzyme activity in both generations were around the control, 1.00.

**Figure 2.** Comparison of means of relative CKX enzyme activity and selected gene expression levels in T1 (bars) and T<sup>2</sup> (line) generation of silenced lines. \*—significant at *p* < 0.05; \*\*—significant at *p* < 0.01.

The effect of *TaCKX1* silencing on the levels of expression of selected *TaCKX* genes is presented by the expression ratio indicator (Table 1), which is a quotient of the mean relative value in silent per mean relative value in non-silent, control plants. In the case of *TaCKX1* and *TaCKX11* (*3*), the ratio indicator, significantly decreased in T1, was strongly decreased in T2. The value of the ratio indicator for *TaCKX2.2* was not changed in T<sup>1</sup> compared to the control and was only slightly decreased in T2. The expression ratio indicator of *TaCKX9 (10)*, strongly decreased to 0.59 in T1, rose above the control level (1.15) in T2. Already high in T1, the expression ratio indicator for *TaCKX2.1* (1.22) increased to 1.32 in T2. The phenotype ratio indicator for CKX enzyme activity was 1.01 in T<sup>1</sup> and 0.99 in T2.


**Table 1.** Effect of *TaCKX1* silencing on expression levels of selected *TaCKX* genes presented by expression ratio indicator (mean value in silent/mean value in non-silent, control plants) in T<sup>1</sup> and T<sup>2</sup> generations.

\*—significant at *p* < 0.05.

In T<sup>1</sup> segregating plants, CKX enzyme activity significantly correlated with spike length (0.51; n = 16) and grain weight (0.50; n = 16), but in T<sup>2</sup> these correlations were not significant.

#### *2.3. Influence of TaCKX1 Silencing on Phenotypic Traits and Chlorophyll Content in Flag Leaves of T<sup>1</sup> and T<sup>2</sup> Plants*

The values of phenotypic traits in T<sup>1</sup> plants with slightly decreased relative expression of *TaCKX1* (0.67 ± 0.14) compared to control plants (1.00) were on the same level in the case of plant height and lower for number of spikes, spike length, grain number, and grain yield (Supplementary Table S2). Higher values were obtained for TGW. Data for chlorophyll content measured by SPAD in the flag leaves of first spikes and the next spikes were similar. All these differences were not significant. Opposite results were obtained for some traits in T<sup>2</sup> plants with highly silent *TaCKX1* (0.33 ± 0.06) compared to the control (1.00) (Supplementary Table S3). Silent T<sup>2</sup> plants were substantially smaller, had a higher number of spikes, number of grains, grain yield, seedling root weight, and lower SPAD values for the flag leaves of first spikes. TGW and spike length were significantly lower than in control plants.

These differences between the slightly silenced T<sup>1</sup> and highly silent T<sup>2</sup> generation are expressed by comparison of ratio indicators of phenotypic traits in both generations (Figure 3). There were no changes in plant height, TGW or spike length in T<sup>1</sup> plants compared to the control. However, these values were respectively 7%, 10%, and 25% lower in T<sup>2</sup> plants. Opposite phenotype ratio indicators for number of spikes per plant and number of grains per plant were about 21% and 30% lower in T<sup>1</sup> and 57% and 29% higher in T2. These differences for spike number, grain number, and TGW were significant.

**Figure 3.** Comparison of phenotypic effect of silencing of *TaCKX1* in T<sup>1</sup> and T<sup>2</sup> generations based on ratio indicators. \*—significant at *p* < 0.05; \*\*—significant at *p* < 0.01.

The levels of expression of *TaCKX1* in 7 DAP spikes of all T<sup>1</sup> significantly correlated with number of grains, grain weight, spike length and spike number (0.47, 0.39, 0.42 and 0.33 respectively; n = 42) and grain weight correlated with enzyme activity (0.33; n = 42). The *TaCKX9 (10)* expression level significantly correlated with grain number (0.51; n = 16).

Correlation coefficients among the expression of all tested *TaCKX* genes and enzyme activity, and phenotypic traits in non-silent and highly silent T<sup>2</sup> are included in Supplementary Table S4A,B. All these correlations are graphically presented in Figures and described in Section 2.6.

#### *2.4. Phytohormone Content in 7 DAP Spikes of T<sup>2</sup>*

tZGs, which were mainly composed of tZ9G, tZ7G, tZOG and tZ9GOG, were the most abundant cytokinin group in 7 DAP spikes (Figure 4a). Their mean content in control plants was 6.97 ng/g biomass and in silent T<sup>2</sup> was 6.24 ng/g biomass respectively. The second most abundant was tZ with the level of 3.74 ng/g biomass in the control and 4.59 ng/g biomass in silent T2. The content of cZ was slightly lower to tZ (2.90 ng/g biomass) in control but higher (5.10 ng/g biomass) in silent plants. cZOG was more abundant in the control than the groups of silent plants, and the content was 1.27 and 0.57 ng/g biomass respectively. The concentration of DZGs (sum of DZ7G, DZOG, DZ9G and DZOGR) was higher in silent (1.61 ng/g biomass) than in control plants (1.11 ng/g biomass). Low concentrations (below 0.5 ng/g biomass) were measured for iP and BA. The concentration of IAA was also low and on a comparable level in control and in silent plants (0.23 and 0.24 ng/g biomass respectively). In the case of ABA, the concentration in the control was slightly decreased in silent plants (2.61 and 2.29 ng/g biomass respectively). The concentration of GA was increased from 0.28 ng/g biomass in the control to 2.93 ng/g biomass in silent plants, which was more than a 10-fold increase.

**Figure 4.** Phytohormone content (ng/g biomass) measured in the group of control and silent T<sup>2</sup> plants (**a**). Phytohormone ratio indicators (mean value in silent per mean value in not silent, control plants) in silent T<sup>2</sup> plants (**b**). \*—significant at *p* < 0.05. Small amounts (≤1.00 ng/g biomass): tZR, tZOGR, cZOGR, DZOG, DZ7G, DZ9G, DZOGR, iP, iP7G, BA, IAA. Trace amounts (≤0.05 ng/g biomass) or not detected: cZ9G, cZR, DZ, DZR, iPR, IBA, IPA, NAA, PAA.

Most of the phytohormone ratio indicators in the group of six silent T<sup>2</sup> plants (Figure 4b) were much higher than in control plants. There were the following cytokinins: tZ (1.23), tZ7G (3.53), tZ9GOG (2.15), tZOG (1.11), cZ (1.76), sum of DZGs (1.45) and iP (1.32). The ratio indicators for some of them were significantly lower, as in the case of BA (0.27), cZOG (0.45) and tZ9G (0.53). Similar values were observed for IAA (1.04), and slightly lower for ABA (0.88), but much higher for GA (10.42).

## *2.5. Coordinated E*ff*ect of TaCKX1 Silencing on Expression of Other TaCKX Genes and Phytohormone Level in 7 DAP Spikes as Well as Phenotype in T<sup>2</sup>*

A graphic presentation of the coordinated effect of *TaCKX1* silencing on expression of other *TaCKX* genes and phytohormone levels in 7 DAP spikes as well as the phenotype of T<sup>2</sup> plants is presented in Figure 5. The significant decrease of expression of *TaCKX1* was coordinated with the significant decrease of *TaCKX11* (*3*), which presumably resulted in a significant increase of most CKs: tZ, tZGs, cZ, DZGs, iP, as well as GA. The increased phytohormone level in the first 7 DAP spikes positively influenced traits such as spike number and grain number, reaching the ratio indicators 1.57 and 1.29, respectively, and negatively influenced TGW (0.78), spike length (0.86), plant height (0.93), and flag leaf senescence (0.95). Opposing data were obtained for *TaCKX2.1* and *TaCKX9* (*10*), which showed increased expression in silenced 7 DAP spikes (1.32 and 1.15 respectively). This might have influenced the decreased ratio indicators for phytohormones—cZOG (0.45), BA (0.27), and ABA content (0.88), and slightly increased ratio indicators for yield-related traits: root weight and grain yield (1.07 and 1.03 respectively). Expression ratio indicators for *TaCKX5* and *TaCKX2.2* were both close to 1.00, but their expression significantly increased compared to T<sup>1</sup> and positively correlated with the expression of *TaCKX2.1* and *TaCKX9* (*10*) respectively.

**Figure 5.** Graphic presentation of coordinated effect of *TaCKX1* silencing on expression of other *TaCKX* genes, phytohormone levels as well as phenotype in 7 DAP spikes of T<sup>2</sup> plants based on ratio indicators. \*—significantly increased comparing to T<sup>1</sup> ; ?—expected changes.

*2.6. Models of Co-Regulation of Phytohormone Levels and Phenotype Traits by Coordinated Expression of TaCKX Genes in Non-Silenced and Silenced T<sup>2</sup> Plants*

Two different models of co-regulation of *TaCKX* expression, phytohormone levels and phenotypic traits in non- silenced and silenced plants of the T<sup>2</sup> generation are proposed (Figure 6a–h) based on correlation coefficients (Table S4A,B).


**Figure 6.** Models of regulation of phytohormone levels and phenotypic traits by coordinated expression of *TaCKX* genes based on correlation coefficients (cc) in non-silenced and silenced wheat plants (**a**–**h**). (cc)—correlation coefficient between expression and trait; (?) – lack of correlation with expression of any gene; bold—strong, significant correlations at *p* ≤ 0.05 (cc above 0.82); grey—cc from 0.5 to 0.6.

Plant height (Figure 6a). There was no correlation between plant height and expression values of any *TaCKX* expressed in 7 DAP spikes of non-silent as well as silent plants. In the first group of plants this trait negatively correlated with BA and positively with IAA and GA content. By contrast, in silent plants the values of plant height were negatively correlated with growing concentration of tZ and tZGs, which resulted in a smaller plant phenotype.

Spike length (Figure 6b) in non-silent plants was positively correlated with BA, and negatively with cZ and ABA content. These correlations determined longer spikes and the trait negatively correlated with spike number and grain number. A strong positive correlation between CKX activity and spike length was noted in silent plants. The values of enzyme activity correlated positively with

slightly increased *TaCKX5* expression, which negatively correlated with increasing content of cZ and tZGs. Spike length in silent plants was positively correlated with grain yield.

TGW (Figure 6c). There was no correlation of TGW with expression of any *TaCKX* expressed in 7 DAP spikes of non-silent plants. However, the trait was strongly negatively correlated with cZ content and positively with GA. The grains in this group of plants were larger and TGW higher. By contrast, in silent plants there was a strong negative correlation of the trait with growing expression of *TaCKX2.1*, which positively regulated tZ, cZ, iP, and GA content. Moreover, the values of expression of down-regulated *TaCKX11* (*3*) positively correlated with decreasing content of cZOG, negatively with highly growing GA and positively with the trait. Altogether it resulted in lower TGW compared to non-silenced plants. The trait in silent plants was strongly and positively correlated with grain yield (0.82) and root weight (0.77).

Grain yield (Figure 6d). Expression levels of *TaCKX1*, *TaCKX2.2* and *TaCKX5* in non-silent plants positively correlated with tZ and iP and negatively with BA content. However, expression of *TaCKX11* (*3*) and *TaCKX2.1* regulates the same CKs in opposite way. Altogether, it resulted in lower grain yield comparing to silenced plants, and the trait was strongly positively correlated with spike number (0.93) and grain number (0.99). The increasing expression of *TaCKX2.1* positively correlated with a growing content of tZGs and cZ and negatively with the trait in silent plants. Decreasing expression of *TaCKX11* (*3*), which was positively correlated with decreased cZOG content and negatively with GA content, positively correlated with the trait. A positive correlation was observed between CKX activity and grain yield in this group of plants, which was higher than in non-silent plants. Moreover, CKX activity negatively correlated with tZGs. The trait was strongly correlated with TGW (0.82) and root weight (0.66).

Spike number (Figure 6e) and grain number (Figure 6f) in non-silenced plants were positively regulated by *TaCKX1*, *TaCKX2.2* and *TaCKX5*, and their expression was positively correlated with tZ, iP and negatively with BA. On the other hand, expression levels of *TaCKX2.1* plus *TaCKX11* (*3*) were negatively correlated with the traits as well as with tZ, iP and positively with BA. Both groups of genes finally affected lower spike and grain numbers in non-silent plants in comparison to silent plants and were strongly and positively correlated with each other (0.91) and grain yield (0.93 and 0.99 respectively). In silent plants decreasing expression of *TaCKX1* is negatively correlated with both spike and grain number and the gene negatively regulates decreasing BA content. In the case of grain number, the main player positively correlated with the trait is *TaCKX5*, increased expression of which was correlated with slightly higher IAA content, which resulted in higher grain number. Spike number is also positively regulated by *TaCKX5* co-expressed with *TaCKX2.1*, and both genes were positively correlated with growing CKs, DZGs and iP as well as GA, determining higher spike number. Both traits are highly correlated (0.88) with each other.

Seedling root weight (Figure 6g). There was strong, positive correlation between *TaCKX9* (*10*) expression in 7 DAP spikes and seedling root weight in non-silenced plants. Moreover, CKX activity negatively correlated with tZ (in spikes) and the trait, which finally resulted in lower root weight. The decreasing expression of *TaCKX11* (*3*) in the case of silent plants was positively correlated with decreasing content of cZOG and strongly positively correlated with the trait. Increasing expression levels of *TaCKX9* (*10*) plus *TaCKX2.2* negatively correlated with decreasing content of cZOG and root weight.

Chlorophyll content measured by SPAD in flag leaves of first spikes (Figure 6h). There was no correlation between expression level of any *TaCKX* measured in 7 DAP spikes of non-silent plants and the trait. The only correlations were between phytohormone content and the trait, positive for cZ and negative for GA, which resulted in higher SPAD values (chlorophyll content). Increasing expression of *TaCKX2.1* was strongly positively correlated with growing values of tZ, tZGs, cZ, and DZGs as well as GA in silent plants. A strong negative correlation was observed between the gene expression and chlorophyll content, which means that increasing expression of *TaCKX2.1* in 7 DAP spikes results in lower chlorophyll content in silent plants.

#### **3. Discussion**

First, 7 DAP spike was chosen as a research objective in wheat since decreased *HvCKX1* expression at this stage in barley resulted in higher yield due to the higher spike and grain number [20,21]. The *TaCKX1* gene is an orthologue of *HvCKX1* and both genes are specifically expressed in developing spikes [23], indicating their possibly important role in the regulation of yield-related traits. The samples were taken from the middle part of the spikes, when anthesis starts, in order to ensure a similar developmental stage of spikelets for research. The 7 DAP spikes of wheat represent the middle of cell division/cell expansion stage [35,36].

#### *3.1. Various Levels of TaCKX1 Silencing Influence Di*ff*erent Models of Co-Expression with Other TaCKX Genes and Parameters of Yield-Related Traits*

Various levels of silencing of *TaCKX1* in T<sup>1</sup> and T<sup>2</sup> generate different results of co-expression with other *TaCKX* genes and plant phenotype. For example, the expression of *TaCKX9* (*10*) was highly and significantly correlated with *TaCKX1* only in T1. However, a new and strong positive correlation between *TaCKX9* (*10*) and *TaCKX2.2* in highly silenced T<sup>2</sup> was observed. Slightly decreased co-expression of silenced *TaCKX1* together with *TaCKX11* (*3*) in T<sup>1</sup> was much stronger in T2, indicating their positive co-regulation. It should be underlined that there is no homology between the sequence of *TaCKX1* used for silencing and sequences of other *TaCKX* genes tested. Therefore, the process of RNAi silencing was specifically addressed to *TaCKX1* silencing. It indicates that the level of silencing of the modified gene affected variable levels of expression of the other *TaCKX* genes in a co-operative process maintaining homeostasis of CKX enzyme in the research object. The models of co-regulation of other *CKX* by highly silenced *TaCKX1* and knock-out *HvCKX1* [27] differ between these species.

The differences in the levels of expression of *TaCKX1* and various co-expression of other *TaCKX* genes in T<sup>1</sup> and T<sup>2</sup> resulted in opposite phenotypic effects. Since spike number, grain number, and grain yield were reduced in T1, the same yield-related traits were significantly higher in highly silenced T<sup>2</sup> plants. High-yielding phenotype occurred when highly silenced *TaCKX1* co-operated with down-regulated *TaCKX11* (*3*) but up-regulated *TaCKX5*, *TaCKX2.2*, *TaCKX2.1*, and *TaCKX9* (*10*). These differences showed that both levels of silencing might be helpful to better understand the function of developmentally regulated genes. Unexpectedly, changes in the expression levels of co-working *TaCKX* did not result in different enzyme activity, even in highly silenced T<sup>2</sup> plants. This might be explained by the fact that down-regulation of *TaCKX1* and *TaCKX11* (*3*) is compensated for by the up-regulation of *TaCKX2.2*, *TaCKX5*, and *TaCKX9* (*10*), and therefore the contribution of isozymes encoded by the genes in the general pool of CKX enzyme activity is the same. Since CKX enzymes indicate different specificities for the particular cytokinin hormone [37], the cytokinin contribution and phenotypic traits of modified plants were changed accordingly, with consequent differences in the active pool of CKs influencing phenotype.

#### *3.2. Co-Operating E*ff*ect of TaCKX on the Level of Active CKs in Silenced Plants*

Since CKX isozymes specifically degrade CKs, the highly decreased expression of *TaCKX1* and *TaCKX11* (*3*) in 7 DAP spikes is expected to result in the observed increase of most major forms of CKs: tZGs, tZ, and cZ in silenced plants. We documented that both tZ and cZ, which are isomers of zeatin, together with their derivatives are a major group of isoprenoid CKs in 7 DAP spikes. It has already been shown that trans-zeatin is the predominant form after anthesis [36,38], but comprehensive analysis of cytokinins during spike, spikelet, ovule and grain development has not yet been reported for wheat using LC-MS/MS [22]. The content of DZGs increased by 40% in silent compared to non-silent wheat plants, suggesting that this less known isoprenoid form of CKs might also play an important role in plant productivity. Interestingly, isoprenoid iP was represented in 7 DAP spikes of non-silent plants at very low quantities, but its content in 7 DAP spikes of silent plants was increased by 32%. A similar relationship between the reduced expression of selected *CKX* family genes and cytokinin

accumulation in reproductive organs has been observed in other species including *A*. *thaliana* [39], rice [19], and barley [25], but detailed data are not comparable to our research in wheat.

The physiological significance of these isoprenoid forms is still not very well known. tZ and iP, which are susceptible to CKX, were found the most abundant and bioactive CKs in maize, whereas cZ, which shows low affinity to CKX was reported to have a weak biological impact and unknown biological role [40,41]. However, the cZ concentrations changed significantly during development in maize grain, as well as in shoot and root tissues [42,43]. High levels of cZ at the first developmental stage of barley spike observed by Powell et al. [44] might indicate an important role of this form in early barley embryo development, what is also documented in our results (discussed further below).

The BA is represented in 7-DAP spikes of wheat at trace amounts but their content was significantly decreased in silent plants. However, their correlations with the *TaCKX* genes as well as yield-related traits of non-silenced plants indicate their importance (discussed in more detail below). Interestingly, BA was found to participate in posttranscriptional and/or posttranslational regulation of protein abundance in *Arabidopsis*, showing high specificity to shoots and roots, and affected differential regulation of hormonal homeostasis [45].

#### *3.3. Cross Talk of CKs with Other Phytohormones*

Negative correlations between ABA content and *TaCKX2.2* and *TaCKX9* (*10*) expression, and positive with *TaCKX11* (*3*), were associated with a slight decrease of ABA content in 7 DAP spikes of silenced plants. Moreover, ABA was strongly positively correlated with BA. The main auxin, IAA, remained at the same level. A ten-fold increase of GA content in silenced comparing to non-silenced plants was observed. Such cross regulation of CKs and other plant hormones is documented in other species. In maize kernels the *CKX1* gene is up-regulated by cytokinin and ABA, and abiotic stress [18]. In tobacco altered cytokinin metabolism affected cytokinin, auxin, and ABA contents in leaves and chloroplasts [46], which host the highest proportion of CK-regulated proteins [47]. Moreover, auxin, ABA and cytokinin are involved in the hormonal control of nitrogen acquisition and signalling [48], which often limits plant growth and development. All four phytohormones, CKs, GA, IAA, and ABA, were found to be involved in the regulation of grain development in drought conditions [49]. Moreover, in shoots, BA up-regulated the abundance of proteins involved in ABA biosynthesis and the ABA response, whereas in the roots, BA strongly up-regulated the majority of proteins in the ethylene biosynthetic pathway [45]. We proved that IAA, GA, and ABA contents are also co-regulated by CKs in non-silenced and silenced 7 DAP spikes. Up-regulation of major CKs and down-regulation of some minor ones in silent plants influence GA, ABA, and IAA content in a similar manner as in abiotic stress conditions.

#### *3.4. Coordinated E*ff*ect of TaCKX Gene Expression on the Content of CKs, Other Phytohormones and Yield-Related Traits*

Plant height in non-silenced plants is down-regulated by BA and up-regulated by IAA and GA content in the first 7 DAP spikes, resulting in taller plants. Oppositely, increased content of tZ and tZGs negatively correlated with the trait in silent plants, stimulated plant height. As it was already showed [50,51] and similarly to our results, plant height and root weight are regulated by CKs and IAA in opposite ways. This may be dependent on basipetal auxin flow in the stem, which suppresses axillary bud outgrowth, and similarly as in pea, auxin derived from a shoot apex suppresses the local level of CKs in the nodal stem through the regulation of *CKX* or *IPT* genes [52].

The main role in spike length seemed to be played by cZ and its glucoside. Increased content of cZOG in non-silenced plants negatively correlated with ABA, resulting in longer spikes. In silent plants the trait is positively regulated by *TaCKX2.2* together with *TaCKX5*, and the latter is a positive regulator of enzyme activity and negative of cZ content. Consequently, a higher content of cZ in 7 DAP spikes led to shorter spikes. cZOG, found as a positive regulator of longer spikes, is a sugar conjugate of cZ-0-glucoside, which is the inactivated form of cZ, showing metabolic stability against CKX activity [53]. Moreover, 0-glucosylation of cZ is catalysed by a specific 0-glucosyltransferase, cisZOG1, discovered in maize [54], and this form mainly functions in the early stages of seed development. Knowledge of function of cZ degradation pathways via the CKX enzyme is limited. Interestingly, two *Arabidopsis* genes, *CKX1* and *CKX7*, expressed in stages of active growth, were shown to have high preference for cZ [37]. In our case the *TaCKX5* positively regulated CKX activity and negatively cZ content.

None of the tested individual *TaCKX* genes was involved in high TGW in non-silenced plants, but a negative correlation with cZ and positive with GA was found. Otherwise a significant negative correlation of *TaCKX2.1* and a positive correlation of *TaCKX11* (*3*) in determining low TGW were observed in silenced plants. Unexpectedly increased expression of the first one positively influenced tZ, cZ, and iP content and negatively GA content, and the opposite was true for the second gene, resulting in lower TGW. Therefore both *TaCKX2.1* and *TaCKX11* (*3*), acting in an opposite manner, maintain homeostasis of CKX enzyme activity and co-regulate TGW in silenced plants. A greater concentration of CKs, especially tZ, was observed during the grain filling stage of high-yielding cultivars [44]. We might suppose that the observed higher concentrations of tZ and other CKs at the 7 DAP stage, which originally was a consequence of *TaCKX1* silencing, might accelerate germination of the grains, which resulted in smaller grains/lower TGW than in non-silenced plants. The silenced *TaCKX1* co-work with down-regulated *TaCKX11* (*3*) in increasing CK content as well as up-regulating *TaCKX2.1*, with seems to play a regulatory role. The involvement of GA in TGW and other traits demonstrated by us might be the effect of co-regulation of *CKX* and other gibberellin-responsive genes regulating yield-related traits as well [55,56]. Fahy et al. [57] suggested that final grain weight might be largely determined by developmental processes prior to grain filling. This is in agreement with our observations, in which yield-related traits are differently regulated in two groups of plants, non-silent and silent. Therefore, we might suppose that the coordinated co-regulation of expression of *TaCKX* genes and related CKs takes place during whole plant and spike development and small seeds in silenced plants are determined at earlier stages.

Grain yield, which is very strongly correlated with grain and spike number in non-silent plants but with TGW in silent plants, is a more complex feature. Two groups of genes up-regulating or down-regulating grain yield in non-silent plants have been found. The first one includes *TaCKX1*, *2.2*, and *5* positively regulating iP content but negatively BA. The second comprises *TaCKX11* (*3*) acting in down-regulation of tZGs. Both groups might determine lower grain yield. It is worth to mention that *TaCKX5*, which is highly expressed in inflorescences and leaves might be a main player of this trait. Higher grain yield was positively regulated by enzyme activity and both, down-regulated *TaCKX11* (*3*) as well as up-regulated *TaCKX2.1* in silenced plants. Again, the *TaCKX2.1* positively regulated tZGs and cZ content just like for TGW, which is rather untypical for a gene encoding a CKX enzyme degrading CKs. Therefore, the positive regulation of the main CK content by *TaCKX2.1* observed by us supports its role in regulation of expression of other genes rather than encoding the CKX isozyme.

As observed in barley cultivars, changes in cytokinin form and concentration in developing kernels correspond with variation in yield [44]. Interestingly, the authors observed no peaks and no differences in CKX activity at the particular stages of spike development. This is in agreement with the homeostasis of the pool of isozymes in 7 DAP spikes of wheat, as suggested by us, which is independent of the level of silencing of *TaCKX1* but is rather a consequence of co-regulation of expression of other *TaCKX* genes. A similar effect of increased grain yield, which was a consequence of higher spike and grain number, was obtained in barley with silenced by RNAi *HvCKX1*, an orthologue of *TaCKX1* [20,21,25]. In this research, CKX activity was decreased, however according to Zalewski et al. [20], it was measured not in 7 DAP spikes, but in 0 DAP spikes and seedling roots. Therefore this inconsistency might be result of measurements in various organs/developmental stages. Another explanation is that these two cereal species varied three times in ploidy level, what might influence differences in action of both orthologues. The *TaCKX* homologues located on A, B and D chromosomes might significantly affect homeostasis of pooled CKX isozymes in wheat. Incomparable to the results obtained for RNAi silenced *TaCKX1* and

*HvCKX1*, no changes in yield parameters were observed in mutant lines with knock-out of *HvCKX1* (Gasparis et al., 2019). These essential phenotypic differences between RNAi-silenced *TaCKX1* and *HvCKX1* or knocked out by CRISPR-Cas9 *HvCKX1* might be the result of different processes involved in inactivation of the gene. The first one is regulated at the posttranscriptional and the second at the transcriptional level. Since CKs might regulate various developmental and physiological processes at the posttranscriptional level [6,7] or by modulation of context-dependent chromatin accessibility [8], the way of deactivating *TaCKX* function seemed to be important.

Spike number and grain number are highly correlated in both non-silent and silent plants and are regulated by the same groups of *TaCKX* genes as well as phytohormones. The first group includes *TaCKX1*, *2.2* and *5* positively regulating iP but negatively BA. The second comprises *TaCKX11* (*3*) and *2.1* acting in the opposite way, and homeostasis of these hormones in non-silenced plants maintains a lower spike number. The main role in controlling higher spike and grain number in silent plants seemed to be played by *TaCKX5*, highly expressed in seedling roots, leaves, inflorescences and 0 DAP spikes. These correlations are not significant because they were measured in a stage of plant development in which the number of spikes and seed number have already been set. As reported, the higher spike number was the consequence of a higher tiller number, which was positively correlated with the content of endogenous zeatin in the field-grown wheat after exogenous hormonal application [58]. Shoot branching might also be dependent on the acropetal transport of cytokinin [52].

Root weight was positively correlated with lower expression of *TaCKX9* (*10*) in 7 DAP spikes of non-silent plants and, negatively with increased expression of this gene in silenced plants. Therefore the gene might determine lower root weight in the first group of plants, but higher in the second. Increased expression of *TaCKX9* (*10*) down-regulated cZOG. The same cZOG was up-regulated by *TaCKX11* (*3*), but expression of this gene in 7 DAP spikes of silent plants is strongly decreased. Both cZ and cZOG are involved in spike length regulation as well as TGW and grain yield in the group of silenced plants. Although both tested organs are in different developmental stages, correlations between *TaCKX9* (*10*) and *TaCKX11* (*3*) expression in 7 DAP spikes and weight of seedling roots are reasonable. The *TaCKX9* (*10*) is mainly expressed in younger organs from seedling roots to 0 DAP spikes and highly expressed in leaves. The *TaCKX11* (*3*) is expressed in all organs tested [23] and both seemed to regulate seedling roots as well, although in the opposite manner. Therefore, we should take into consideration the possible action of cytokinin transport and signalling genes as well as other phytohormones which take part in hormonal crosstalk to control the regulation of root growth [59]. Accordingly, cZ type CKs found as the major forms in phloem are translocated from shoots to roots [60,61]. Some *CKX* genes might be induced by transcription factors [62,63], what is also observed in our unpublished yet data.

The lower plant height and higher root weight observed in the group of silenced plants of wheat is in agreement with opposed regulation of these traits by CKs and IAA mentioned above [64,65]. Up-regulated content of active cZ in 7 DAP spikes, might influence down-regulation of this CK in roots. It has been documented that such suppressing cZ levels mediated by overexpression of *AtCKX7* affected root development in *Arabidopsis* [66]. A higher weight of seedling root was also obtained by silencing via RNAi or knock-out via CRISPR/Cas9 of *HvCKX1* in barley plants, as in wheat, and the trait corresponded with decreased activity of CKX enzyme measured in roots (Zalewski et al., 2010; Gasparis et al., 2019).

Leaf senescence was determined in the flag leaf of the first spike by measuring chlorophyll content. Increased expression of *TaCKX2.1* in silent plants up-regulated tZ, tZGs and cZ content in 7 DAP spikes and down-regulated the trait. The gene functions in a similar way, by up-regulating these CKs in determining lower TGW and higher grain yield in silent plants. A higher content of active CKs as well as GA in 7 DAP spikes of silent plants is expected to down-regulate CKs in the flag leaves, accelerating their senescence, what is documented by the results.

It was previously demonstrated that level of chlorophyll content in flag leaves is associated with the senescence process, in which CKs suppress inhibition of senescence [67]. During this processes, proteins are degraded and nutrients are re-mobilised from senescing leaves especially to the developing

grains [68]. We might suppose that slower spike ripening in non-silent plants, which is dependent on lower CK content in the 7 DAP spike, causes a slower flow of micronutrients as well as CKs from flag leaf to spike. Therefore, prolonged chlorophyll content in the flag leaf of the first spike negatively correlated with TGW but positively with plant height. Opposite data were obtained for flag leaves of silent plants, in which higher content of CKs in 7 DAP spikes might be the result of faster flow accelerating leaf senescence. The reduced chlorophyll content in flag leaves of the first spike of silent plants positively correlated with grain yield. The important role of tZ and less active cZ in the suppression of senescence was proven in maize leaves [69] and in an oat-leaf assay [37]. It was also documented that delayed senescence of wheat stay-green mutant, tasg1, at the late filling stage was related to high cytokinin and nitrogen contents [70].

#### **4. Materials and Methods**

#### *4.1. Vector Construction*

The hpRNA type of silencing cassette was constructed in pBract207 (https://www.jic.ac. uk/technologies/crop-transformation-bract/). It contains the Hpt selection gene under the 35S promoter and cloning sites for the cloning silencing cassette under the Ubi promoter. The vector is compatible with the gateway cloning system. For cloning purposes a coding sequence of *TaCKX1* (NCBI JN128583) 378 codons long was used. In the first step, the cassette was amplified using: EAC11-F: 50 -TTGAATTCGACTTCGACCGCGGCGTTTT-30 and EAC12-R: 50 -TTGAATTC ATGTCTTGGCCAGGGGAGAG-30 and cloned into the entry vector pCR8/GW/TOPO (Invitrogen). In the next step, the cassette was cloned to the destination Bract7 vector in the gateway reaction. The presence of the silencing cassette in the vector was verified by restriction analysis and sequencing. The vector was electroporated into the AGL1 strain of *Agrobacterium tumefaciens* and used for transformation.

#### *4.2. Plant Material, Agrobacterium-Mediated Transformation and In-Vitro Culture*

The spring cultivar of common wheat (*Triticum aestivum* L.) Kontesa was used as a donor plant for transformation experiments as well as transgenic plants. Seeds were germinated into Petri dishes for one day at 4 ◦C and then five days at room temperature in the dark. Six out of ten seedlings from each Petri dish were replanted into pots with soil. The plants were grown in a growth chamber under controlled environmental conditions with 20 ◦C/18 ◦C day/night temperatures and a 16 h light/8 h dark photoperiod. The light intensity was 350 µmol·s −1 ·m−<sup>2</sup> .

*Agrobacterium*-mediated transformation experiments were performed according to our previously described protocols for wheat [71,72]. Putative transgenic plants were regenerated and selected on modified MS media containing 25 mg·L <sup>−</sup><sup>1</sup> of hygromycin as a selectable agent.

First, 7 days after pollination, (DAP) spikes from T1, T2, and control plants were collected for RT-qPCR and phytohormone quantification. Only 1 in 3 of the middle part of each spike was used for experiments (upper and lower parts were removed).

#### *4.3. PCR Analysis*

Genomic DNA was isolated from well-developed leaves of 14-day plants according to the modified CTAB procedure [73] or by using the KAPA3G Plant PCR Kit (Roche Sequencing and Life Science, Kapa Biosystems, Wilmington, MA, USA). The PCR for genomic DNA isolated by CTAB was carried out in a 25 mL reaction mixture using Platinum Taq DNA Polymerase (Invitrogen by Thermo Fisher Scientific, Waltham, MA, USA) and 120 ng of template DNA. The reaction was run using the following program: initial denaturation step at 94 ◦C for 2 min, 35 cycles of amplification at 94 ◦C for 30 s, 65 ◦C for 30 s, 72 ◦C for 30 s with a final extension step at 72 ◦C for 5 min. The PCR for genomic DNA isolated by KAPA3G was carried out in a 50 µL reaction mixture using 1 U of KAPA3G Plant DNA Polymerase and a 0.5 × 0.5 mm leaf fragment. The reaction was run using the following program:

initial denaturation step at 95 ◦C for 3 min, 40 cycles of amplification at 95 ◦C for 20 s, 68 ◦C for 30 s, 72 ◦C for 30s with a final extension step at 72 ◦C for 2 min.

Putative transgenic T<sup>0</sup> and T<sup>1</sup> plants were tested with two pairs of specific primers amplifying a fragment of the *hpt* selection gene. The sequences of the primers for the first pair were: hygF1 5 0 -ATGACGCACAATCCCACTATCCT-30 and hygR1 50 -AGTTCGGTTTCAGGCAGGTCTT-30 , and the amplified fragment was 405 bp. The sequences of the primers for the second pair were: hygF2 5 0 -GACGGCAATTTCGATGATG-30 and hygR2 50 -CCGGTCGGCATCTACTCTAT-30 , and the amplified fragment was 205 bp.

Non-transgenic null segregants were used as a control.

#### *4.4. RNA Extraction and cDNA Synthesis*

Total RNA from 7 DAP spikes was extracted using TRI Reagent (Sigma-Aldrich, Hamburg, Germany) and 1-bromo-3-chloropropane (BCP) (AppliChem GmbH, Darmstadt, Germany) according to the manufacturer's protocol. The purity and concentration of the isolated RNA were determined using a NanoDrop spectrophotometer (NanoDrop ND-1000) and the integrity was checked by electrophoresis on 1.5% (w/v) agarose gels. To remove the residual DNA the RNA samples were treated with DNase I, RNase-free (Thermo Fisher Scientific, Waltham, MA, USA). Each time 1 µg of good quality RNA was used for cDNA synthesis using the RevertAid First Strand cDNA Synthesis Kit (Thermo Fisher Scientific) following the manufacturer's instructions. The obtained cDNA was diluted 20 times before use in RT-qPCR assays.

#### *4.5. Quantitative RT-qPCR*

RT-qPCR assays were performed for 6 target genes: *TaCKX1* (JN128583), *TaCKX2.1* (JF293079)/*2.2* (FJ648070), *TaCKX11* (*3*) (JN128585), *TaCKX5* (Lei et al., 2008), *TaCKX9* (*10*) (JN128591). Primer sequences designed for each gene as well as for the reference gene are shown in Table S1. All real-time reactions were performed in a Rotor-Gene Q (Qiagen) thermal cycler using 1× HOT FIREPol EvaGreen qPCR Mix Plus (Solis BioDyne), 0.2 µM of each primer, and 4 µL of 20 times diluted cDNA in a total volume of 10 µL. Each reaction was carried out in 3 technical replicates at the following temperature profile: 95 ◦C—15 min initial denaturation and polymerase activation (95 ◦C—25 s, 62 ◦C—25 s, 72 ◦C—25 s) × 45 cycles, 72 ◦C—5 min, with the melting curve at 72–99 ◦C, 5 s per step. The expression of *TaCKX* genes was calculated according to the two standard curves method using ADP-ribosylation factor (*Ref 2*) as a normalizer.

Relative expression/silencing of *TaCKX1* was related to mean expression of the gene in non-silenced control plants set as 1.00. Relative expression of other *TaCKX* genes was related to each tested gene set as 1.00 in non-silenced plants.

Statistical analysis was performed using Statistica v13.3 software (StatSoft, Kraków, Poland). The normality of data distribution was tested using the Shapiro–Wilk test. To determine whether the means of two sets of data of expression levels, phytohormone concentrations, and yield-related traits between non-silenced and silenced lines are significantly different from each other (for *p* value less than *p* < 0.05), either the Student's *t*-test or the Mann–Whitney test was applied. Correlation coefficients were determined using parametric correlation matrices (Pearson's test) or a nonparametric correlation (Spearman's test).

#### *4.6. Quantification of ABA, Auxins, Cytokinins and GA<sup>3</sup>*

Chemicals used for quantification were: the standard of ABA, five standards of auxins: IAA, indole-3-butyric acid (IBA), indole-3-propionic acid (IPA), 1-naphthaleneacetic acid (NAA), and 2-phenylacetic acid (PAA); twenty-seven standards of CKs: tZ, *trans*-zeatin riboside (tZR), *trans*-zeatin-9-glucoside (tZ9G), *trans*-zeatin-7-glucoside (tZ7G), *trans*-zeatin-*O*-glucoside (tZOG), *trans*-zeatin riboside-*O*-glucoside (tZROG), *trans*-zeatin-*9*-glucoside-*O*-glucoside (tZ9GOG), *trans*-zeatin-9-glucoside riboside (tZ9GR), *c*Z, *cis*-zeatin-riboside (cZR), *cis*-zeatin *O*-glucoside

(cZOG), *cis*-zeatin 9-glucoside (cZ9G), *cis*-zeatin-*O*-glucoside-riboside (cZROG), dihydrozeatin (DZ), dihydrozeatin-riboside (DZR), dihydrozeatin-9-glucoside (DZ9G), dihydrozeatin-7-glucoside (DZ7G), dihydrozeatin-*O*-glucoside (DZOG), dihydrozeatin riboside-*O*-glucoside (DZROG), *N*<sup>6</sup> -isopentenyladenine (iP), *N*<sup>6</sup> -isopentenyladenosine (iPR), *N*<sup>6</sup> -isopentenyladenosine-7-glucoside (iP7G), *para*-topolin (*p*T), *meta*-topolin (*m*T), *ortho*-topolin (*o*T), 6-benzylaminopurine (6-BAP), and standard of GA3.

For the measurement of phytohormones, 200 mg of plant powders were placed into the 2-mL Eppendorf tubes, suspended in 1 mL of (*v*/*v*) 50% ACN, and homogenized in a bead mill (50 Hz, 5 min) using two 5-mm tungsten balls. Then, samples were homogenized using the ultrasound processor VCX 130 (max. power 130 W, max. frequency 20 kHz, 5 min) equipped with titanium probe and mixed in laboratory shaker (90 rpm, dark, 5 ◦C, 30 min). Samples were centrifuged (9000× *g*, 5 min) and collected in a glass tube. For the quantification of ABA, AXs, CKs, and GA3, [2H6](+)-*cis*,*trans*-ABA (50 ng), [2H5] IAA (15 ng), [2H6] iP (50 ng), [2H5] *t*Z (30 ng), [2H5]-*t*ZOG (30 ng), [2H3]-DZR (30 ng), and [2H2] GA<sup>3</sup> (30 ng) were added to samples as internal standards.

Prepared extracts were purged using a Waters SPE Oasis HLB cartridge (Waters Corporation, Milford, MA, USA), previously activated and equilibrated using 1 mL of 100% MeOH, 1 mL water, and 1 mL of (*v*/*v*) 50% ACN [74]. Then, extracts were loaded and collected to the Eppendorf tubes and eluted with 1 mL of 30% ACN (*v*/*v*). Samples were evaporated to dryness by centrifugal vacuum concentrator, dissolved in 50 µL of (*v*/*v*) 30% ACN and transferred into the insert vials. Detection of analyzed phytohormones was performed using an Agilent 1260 Infinity series HPLC system (Agilent Technologies, Santa Clara, CA, USA) including a Q-ToF LC/MS mass spectrometer with Dual AJS ESI source; 10 µL of each sample was injected on the Waters XSelect C<sup>18</sup> column (250 mm × 3.0 mm, 5 µm), heated up to 50 ◦C. Mobile phase A was 0.01% (*v*/*v*) FA in ACN and phase B 0.01% (*v*/*v*) FA in water; flow was 0.5 mL min−<sup>1</sup> . Separation of above hormones was done in ESI-positive mode with the following gradient: 0–8 min flowing increased linearly from 5 to 30% A, 8–25 min 80% A, 25–28 min 100% A, 28–30 min 5% A.

For the optimization of MS/MS conditions, the chemical standards of analyzed phytohormones were directly injected to the MS in positive ([M + H]+) ion scan modes, then areas of detected standard peaks were calculated. [M + H]<sup>+</sup> was chosen because of its significantly better signal-to-noise ratios compared to the negative ion scan modes.

Chlorophyll content was measured using an SPAD chlorophyll meter.

#### **5. Conclusions**

Based on the 7 DAP spike as a research object, we have documented that silencing of *TaCKX1* by RNAi strongly influenced up- or down-regulation of other *TaCKX* genes, as well as phytohormone levels and consequently phenotype. This co-regulation is dependent on the level of silencing of the gene and is independent of cross-silencing of other *TaCKX* genes. Detailed analysis revealed that each tested yield-related trait is regulated by various up- or down-regulated *TaCKX* genes and phytohormones. Key genes involved in the regulation of grain yield, TGW, or root weight in highly silenced plants are *TaCKX2.1* and *TaCKX11* (*3*) acting antagonistically, and increased expression of the first one determines growth of tZ, tZ derivatives, and cZ, whereas decreased expression of the second down-regulates content of cZOG. A key role in determination of the high-yielding phenotype seemed to be played by the growing content of tZ in 7 DAP spikes, which might accelerate maturation of immature grains by speeding up nutrient flow from flag leaves. This finally led to reduction of TGW but enhancement of grain number and yield. The latter traits are the result of a higher spike number, which is determined in the early stages of plant development.

**Supplementary Materials:** Supplementary materials can be found at http://www.mdpi.com/1422-0067/21/13/4809/ s1. Table S1: Primer sequences designed for reference gene and each of 6 tested *TaCKX* genes and amplicon length. Table S2: Phenotypic traits and ratio indicator in silent T<sup>1</sup> and not silent, control plants. Table S3: Phenotypic traits and ratio indicator in silent T<sup>2</sup> and not silent, control plants. Table S4: A. B. Correlation coefficients among

expression of all tested *TaCKX* genes and enzyme activity, and phenotypic traits in not-silent (A) and highly silent T<sup>2</sup> plants (B). \* non-parametric analysis; in bold: significant at *p* < 0.01.

**Author Contributions:** Conceptualization, A.N.-O. and W.O.; methodology, B.J., H.O., K.S., and A.B.; software, H.O. and K.S.; validation, B.J.; formal analysis, B.J.; investigation, B.J., H.O., and K.S.; data curation, B.J. and A.B.; writing—original draft preparation, A.N.-O.; writing—review and editing, A.N.-O.; visualization, A.N.-O. and H.O.; supervision, A.N.-O. and W.O.; project administration, A.N.-O.; funding acquisition, A.N.-O. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by the National Science Centre, grant UMO-2014/13/B/NZ9/02376 and a statutory grant of PBAI-NRI.

**Acknowledgments:** We thank Malgorzata Wojciechowska, Izabela Skuza, Agnieszka Glowacka, Maja Boczkowska, and Agnieszka Onysk for excellent technical assistance.

**Conflicts of Interest:** The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
