Decrypting tubby-like protein gene family of multiple functions in starch root crop cassava

Abstract

Tubby-like proteins (TLPs) are ubiquitous in eukaryotes and function in abiotic stress tolerance of some plants. Cassava (Manihot esculenta Crantz) is a high-yield starch root crop and has a high tolerance to poor soil conditions and abiotic stress. However, little is known about TLP gene characteristics and their expression in cassava. We identified cassava TLP genes, MeTLPs, and further analysed structure, duplication, chromosome localization and collinearity, cis-acting elements in the promoter regions and expression patterns of MeTLPs, and three-dimensional structure of the encoded proteins MeTLPs. In conclusion, there is a MeTLP family containing 13 members, which are grouped into A and C subfamilies. There are 11 pairs of MeTLPs that show the duplication which took place between 10.11 and 126.69 million years ago. Two MeTLPs 6 and 9 likely originate from one gene in an ancestral species, may be common ancestors for other MeTLPs and would most likely not be eligible for ubiquitin-related protein degradation because their corresponding proteins (MeTLPs 6 and 9) have no the F-box domain in the N-terminus. MeTLPs feature differences in the number from TLPs in wheat, apple, Arabidopsis, poplar and maize, and are highlighted by segmental duplication but more importantly by the chromosomal collinearity with potato StTLPs. MeTLPs are at least related to abiotic stress tolerance in cassava. However, the subtle differences in function among MeTLPs are predictable partly because of their differential expression profiles, which are coupled with various cis‑acting elements existing in the promoter regions depending on genes.

Cassava (Manihot esculenta Crantz) is a high-yield starch root crop of high tolerance to poor soil conditions and abiotic stress. Tubby-like proteins (TLPs) are ubiquitous in eukaryotes and function in the abiotic stress tolerance of some plants. A cassava TLP gene, MeTLP, family containing 13 members was identified and characterized at the genomic level, showing some differences from TLP genes of other plants. The expression of MeTLPs indicates that they are related to abiotic stress tolerance. This is the first report about MeTLPs of cassava, providing clues for further study on the molecular mechanism of cassava stress resistance.

Introduction

Tubby-like proteins (TLPs) are ubiquitous in eukaryotes (Liu 2008) which were first found to function in obese mouse (Kleyn et al. 1996). Thereafter, the TLPs were discovered in other species, such as human (North et al. 1997), chicken (Heikenwälder et al. 2001), rice (Kou et al. 2009), Arabidopsis (Lai et al. 2004), wheat (Hong et al. 2015) and poplar (Yang et al. 2008). A typical TLP protein contains a highly conserved C-terminal tubby domain that is composed of a β-barrel enclosing a central α-helix (Santagata et al. 2001). The tubby domain is associated with folding, solubility and subcellular localization of the TLPs (Kim et al. 2017). Unlike animal TLPs, the N-terminus of the TLPs from plants contains a highly conserved F-box domain in addition to the C-terminal tubby domain (Gagne et al. 2002; Lai et al. 2004). The F-box domain mediates ubiquitin proteolysis by interacting with other target proteins (Patton et al. 1998). The tubby and F-box domains have a co-evolutionary relationship (Kile et al. 2002; Yang et al. 2008).

More extensive functional studies have been performed on the TLPs of animals. Mutations in some TLP genes can cause obesity (Coleman et al. 1990; Kleyn et al. 1996); and vision and hearing loss, and infertility and insulin resistance (Noben-Trauth et al. 1996). Functional identification of the TLP genes, named TLPs, has been conducted only in the limited plant species. In rice, the expression of the OsTLP is pathogen induced (Kou et al. 2009). In Arabidopsis, AtTLP9 and AtTLP3 play important roles in the abscisic acid signalling pathway associated with seed germination (Bao et al. 2014). In chickpea, expression of CaTLP1 is significantly upregulated under dehydration stress (Bhushan et al. 2007). Overexpression of CaTLP1 in tobacco enhances the tolerance of transgenic tobacco to salt, dehydration and oxidative stresses (Wardhan et al. 2012). Expression of MdTLP7 from Malus domestica enhances abiotic stress tolerance in Arabidopsis (Xu et al. 2019).

The number of the TLP genes in plants is much larger than in animals mainly because of segmental duplication, random translocation and insertion (Yang et al. 2008). It has been found that there are 11 AtTLPs in Arabidopsis (Lai et al. 2004), 14 OsTLPs in rice (Liu 2008), 11 PtTLPs in poplar (Yang et al. 2008), 4 TaTLPs in wheat (Hong et al. 2015), 15 ZmTLPs in maize (Chen et al. 2016) and 9 MdTLPs in apple (Xu et al. 2016).

Cassava (Manihot esculenta Crantz) is a high-yield starch root crop that is economically and socially significant (Oliveira et al. 2014; Perera et al. 2014). It has high tolerance to poor soil conditions and drought (Hu et al. 2016b; Wei et al. 2016). However, little is known about TLPs in cassava, named MeTLPs, although the genome of this crop has been sequenced and a wealth of omics data has been generated. Our hypothesis is that there is also such a MeTLP family in cassava, which may have some characteristics different from TLPs of other crops. To confirm this assumption and also to lay a foundation and provide clues for future functional identification of MeTLPs, in this study, we identified a cassava MeTLP family containing 13 genes, and further analysed chromosomal location, gene duplication, three-dimensional structure, gene structure and expression patterns of MeTLPs.

Materials and Methods

Cassava materials and public databases

The target plant species included cassava, three dicots (Arabidopsis, poplar and potato) and three monocots (rice, maize and sorghum). The genomic DNA sequences, coding sequences of the TLPs and/or amino acid sequences of these plants were obtained from the Phytozome database (https://phytozome.jgi.doe.gov/pz/portal.html).

Genome-scale identification of MeTLPs

Unless otherwise specified, the software running parameters mentioned in the text were the default values.

The amino acid sequences of AtTLPs from the TAIR database (http://www.arabidopsis.org/) and OsTLPs from the RGAP database (http:// rice.plantbiology.msu.edu/) were downloaded. With amino acid sequences of the downloaded TLPs, the local HMMER model was built following the previous methods (Finn et al. 2011; http://www.ebi.ac.uk/Tools/hmmer/).

With the HMMER model, the data of all proteins of cassava in the Phytozome database (https://phytozome.jgi.doe.gov/pz/portal.html) were scanned to search for the candidate cassava TLPs, MeTLPs, under an E-value of ≤0.01. Additionally, the amino acid sequences of both AtTLPs and OsTLPs were aligned through BLASTp tool with amino acid sequences of all cassava proteins in the Phytozome database under an E-value of ≤0.01. Subsequently, the amino acid sequences of redundant proteins resulting from the abovementioned two approaches were removed, generating the candidate MeTLPs. All candidate MeTLPs were further verified by using two databases, CDD (Marchler-Bauer et al. 2017) and Pfam (Finn et al. 2016), under an E-value of ≤0.01.

The conserved domains, including the F-box domain and tubby domain, of MeTLPs were analysed by using the ClustalW (Larkin et al. 2007) and the BoxShade software (https://embnet.vital-it.ch/software/BOX_form.html). The phylogenetic tree was constructed using the neighbour-joining method (Tamura et al. 2013) by using MEGA6.0 software, with 1000 bootstrap replicates.

Analyses of MeTLP properties and MeTLP structures

The molecular weight and isoelectric point of proteins were predicted by using the ExPaSy software (Artimo et al. 2012; http://expasy.org/). Predications for the subcellular localization of proteins were conducted through the CELLO approaches (Yu et al. 2006; http://cello.life.nctu.edu.tw/). The conserved amino acid sequence motifs were analysed by using the MEME software (Bailey et al. 2015) under the maximum 16 motifs, and then annotated by using InterProScan software (Mitchell et al. 2019; http://www.ebi.ac.uk/interpro/).

Intron–exon structure of the genes was analysed based on the Mesculenta_305_v6.1.gene.gff3 file of cassava from the Phytozome database by using the tool of Amazing Optional Gene Viewer function in the TBtools 0.655 (Chen et al. 2018; https://github.com/CJ-Chen/TBtools).

Homologous modelling of three-dimensional structures of MeTLPs

The three-dimensional structure of MeTLPs was constructed following the homologous modelling by using the SWISS-MODEL online analysis tool (https://swissmodel.expasy.org/interactive) followed by PyMOL software-based visualization analyses.

Analyses of gene duplication, chromosome localization and collinearity of the TLPs

The genome DNA file and amino acid sequences of all proteins of the target plant species in this study were downloaded from the Phytozome database and then aligned with cassava’s total proteins through the BLASTp program under an E-value of <10⁻⁵ (Kou et al. 2009), resulting in an m8 format output file. Both the m8 format output file and genome gff3 file_6.1 used as input files were analysed by using MCScanX software (Wang et al. 2012) to identify the duplicate genes. The duplicate genes were classified as five types, singleton duplication, dispersed duplication, proximal duplication, tandem duplication and segmental duplication, by using the duplicate gene classifier in MCScanX software.

Within cassava, the collinearity relationships and chromosome localization of the MeTLPs were drawn by using the circlize package (Gu et al. 2014; https://github.com/jokergoo/circlize).

Collinear maps between the MeTLPs and TLPs of other plant species were established by using the Dual Synteny Plotter tools in TBtools software (Yu et al. 2006; https://github.com/CJ-Chen/TBtools). Non-synonymous (K_a) and synonymous (K_s) rates for duplicate gene pairs were analysed by using ParaAT (Zhang et al. 2012) and KaKs_Calculator 2.0 software (Wang et al. 2010). The approximate duplicate time (T) of the genes was estimated following a formula (T = K_s/(2λ), λ = 1.5 × 10⁻⁸) (Koch et al. 2000; Yang et al. 2008).

Prediction of cis-acting elements in the promoter regions of MeTLPs

Genomic DNA fragments, 1.5 kb upstream of the gene ATG start codon, were selected as candidate promoter regions and then analysed following the approaches in the PlantCARE database (Lescot et al. 2002; http://bioinformatics.psb.ugent.be/webtools/plantcare/html/). The cis-acting elements were categorized according to a method in the literature (Hou et al. 2019).

Transcriptome analysis of MeTLPs

As for the expression profiles of the genes in different tissues, the fragments-per-kilobase-per-million fragments mapped (FPKM) values related to cassava were download from the Gene Expression Omnibus database according to the accession number of GSE82279, which were submitted by Wilson et al. (2017). These data resulted from 11 tissues of 3-month-old TME 204 cassava plants that were grown in a greenhouse, including storage roots, fibrous roots, stems, petioles, leaves, lateral buds, midveins, friable embryogenic calli, somatic organized embryogenic structures, root apical meristems and shoot apical meristems.

To explore the expression profiles of the genes among different cassava varieties, the data related to two cassava cultivars (KU50 and Arg7) and one wild species W14 were downloaded from the RNA-seq read archives (SRA) (Hu et al. 2016b). These data resulted from 11 cassava tissues, including early storage roots (ESRs; 75 days after planting), medium tuber roots (120 days after planting) and late storage roots (150 days after planting), and stems (70 days after planting) and leaves (70 days after planting). All SRA accession numbers of the data used in this study are listed in Supporting Information—Table S1, and the SRA accession number-based data files were converted into FASTQ files by using the SRA Toolkit 2.9.2. The quality of the FASTQ files was verified by the FastQC 0.11.8 followed by filtering the low-quality sequences by using the Trimmomatic 0.38, resulting in the trimmed paired reads. The trimmed paired reads were aligned to the Bowtie2-indexed cassava reference genome 6.1 deposited in the Phytozome database by using the TopHat version 2.1.1 (Trapnell et al. 2012), resulting in the bam files of the aligned sequence reads. The gene expression level was represented by the FPKM values (Li and Dewey 2011), which were calculated based on the bam files by using the Cufflinks 2.2.1 software. The data of the FPKM values were converted into log2 values. The log2-based FPKM values were used to create the heat maps of the gene expression that were drawn through the Amazing Simple HeatMap tool in TBtools software 0.655 (Chen et al. 2018).

Abiotic stress treatments of pot-grown cassava

Cassava was planted with stem cuttings strictly following the potting methods indicated in the literature (Chen et al. 2018). The abiotic stress treatment was conducted on 15-day old cassava plantlets after stem cutting planting. For the water deficit stress treatment, the plantlets were then transferred into a 20 % polyethylene glycol (PEG) 8000 solution and incubated for 1, 3, 6 and 9 h, respectively. For salt stress, the plantlets were then transferred into a solution containing 200 mM NaCl and incubated for 1, 3, 6 and 9 h, respectively. For low-temperature stresses, the plantlets were transferred into a growth chamber at 4 °C and stayed for 1, 3, 6 and 9 h, respectively. For phytohormone treatment, the rootlets of the plantlets were soaked for 1, 3, 6 and 9 h in 100 μM abscisic acid and in 100 μM salicylic acid solution, respectively. The controls were set up in parallel under the conditions without special treatments.

Quantitative PCR (qPCR)

Total RNA was extracted by using a Plant RNA Kit (CWBIO, Beijing, China). One microgram of total RNA was used for the synthesis of the first-strand cDNA in a 20-µL reaction volume by using a PrimeScript™ II First-Strand cDNA Synthesis Kit (TaKaRa, Dalian, China). The reaction solution containing the first-strand cDNA was diluted 10 times with RNAase-free water and stored at −80 °C for qPCR. The qPCR was conducted with sequence-specific primers (Table 1) and the SYBR Green Mix (Vazyme, Nanjing, China) on a StepOnePlus™ Real-Time PCR System (Thermo Fisher Scientific, USA). The qPCR was performed in a 20-µL reaction volume under the thermal cycles of 95 °C for 3 s, 95 °C for 5 s and 60 °C for 30 s. The cassava gene, numbered cassava4.1_006776, was used as an internal control gene (Hu et al. 2016a). The relative expression level of the genes was calculated by the 2− ^ΔΔCT method (Livak and Schmittgen 2001), where ΔΔCt = [(Ct_{target gene} – Ct_{actin gene}) under stress] – [(Ct_{target gene} – Ct_{actin gene}) under control conditions]. All expression analysis of the genes had three biological replicates. Differential expression of the genes was defined at a significance level of P < 0.05 through Duncan’s multiple range test.

Table 1.

qPCR primers used for analysis of expression of cassava MeTLPs

Gene	Forward primer	Reverse primer	Length (bp)
MeTLP1	GAGCCAGGCGGTTTCGTC	GGCACTGCTAAACTCGGTTG	117
MeTLP2	AGGAAGAACGAACACGGCAA	CAGAGACTGCATACCCTCCG	122
MeTLP3	TTCTGGAAGCCCCTCAGAGT	AACACGCCCTCGGAAATTCA	127
MeTLP4	GCACTTTAGGAGTTGTCTTCCT	TCGGTGGGCGTAGATTTCTG	118
MeTLP5	ACCCGAAATTATTGGCTCTCA	CCAACTTGCATCTACAACCCA	140
MeTLP6	TCCCAGTTGTGATGCGATTGA	CAACCGACTGACCGACTGAT	148
MeTLP7	CGGGCAACGAGTACCAATTT	GGATTGCATTGCTGGATGTGG	150
MeTLP8	CCTAGTTGAAGGTTGGGTGGG	TGTCAAGCACACAAGATTAGCA	119
MeTLP9	GTGGGCAAAGGTTGCTTATGG	TGCGAATGACAAAGTTCCAGTG	150
MeTLP10	GCTGTGGCTGACAAACATCA	AGCAAGTAGCAACAGTAACCTT	123
MeTLP11	ACGAGTTGAAGAGAGCGAGG	TCCGATATTGCGGGTAGACCT	121
MeTLP12	GGTAACACGCTGATGGGTTG	GAAGCACAGGAACCAACCTT	131
MeTLP13	TCCTGCCAGTGCTTCTTACA	GCTCAGGTATCACAGTCAGCA	108
Cassava4.1_006776 (internal control gene)	TGGTCAGCACATTTGTTCGT	AGCAGACCCCGTCATTGTAG	106

MeTLP, cassava tubby-like protein gene; qPCR, quantitative PCR.

Results

MeTLPs in cassava

A total of 13 cassava MeTLPs with a tubby domain were identified from the cassava genome. The MeTLPs ranged from 380 (MeTLP4) to 424 (MeTLP13 and MeTLP3) amino acids in length, 41.9 (MeTLP4) to 47.8 (MeTLP3) kDa in relative molecular mass and 8.95 (MeTLP2) to 9.62 (MeTLP13) in isoelectric point (Table 2). Detailed information on the corresponding genes (MeTLPs) could be fetched according to the gene ID shown in Table 2.

Table 2.

Characteristics of MeTLPs and MeTLPs of cassava

MeTLP		MeTLP
Name	Locus ID in Phytozome database (https://phytozome.jgi.doe.gov/pz/portal.html)	Length (amino acid residues）	Subcellular location	Amino acid residue localization of F-box domain in primary structure	Amino acid residue localization of tubby domain in the primary structure	Theoretical molecular weight (Da)	Theoretical isoelectric points
MeTLP1	Manes.01G214100	406	Nuclear	52–97	118–399	45 494.2	9.41
MeTLP2	Manes.02G095300	414	Nuclear	55–97	119–407	46 229.8	8.95
MeTLP3	Manes.02G179400	424	Nuclear	53–95	117–417	47 774.6	9.49
MeTLP4	Manes.03G100600	380	Nuclear	52–106	117–373	42 235.2	9.47
MeTLP5	Manes.03G190900	391	Nuclear	36–81	102–384	43 655.1	9.38
MeTLP6	Manes.04G100600	400	Nuclear	-	151–392	44 584.14	9.5
MeTLP7	Manes.05G067900	406	Nuclear	52–97	118–399	45 573	9.26
MeTLP8	Manes.05G157900	415	Nuclear	56–98	120–408	46 235.4	9.51
MeTLP9	Manes.11G069900	399	Nuclear	-	157–391	42 094.9	9.36
MeTLP10	Manes.15G017200	388	Nuclear	34–81	100–381	43 378.9	9.55
MeTLP11	Manes.15G095300	381	Nuclear	52–107	118–374	41 902	9.44
MeTLP12	Manes.18G023100	414	Nuclear	54–100	120–407	46 294.2	9.56
MeTLP13	Manes.18G091900	424	Nuclear	53–95	117–417	47 550.3	9.62

MeTLP, cassava tubby-like protein gene; MeTLP, cassava tubby-like protein.

Phylogenetic evolution of TLPs

A contiguous junction tree of 64 TLPs was constructed, including 13 MeTLPs, 11 AtTLPs, 15 ZmTLPs, 14 OsTLPs and 11 PtTLPs (Fig. 1a). Detailed information on these TLPs, except for cassava MeTLPs, can be found by searching with the corresponding gene ID number [seeSupporting Information—Table S2]. These TLPs could be divided into three subfamilies A, B and C (Fig. 1a). The A subfamily was further divided into subgroups A1, A2 and A3, of which the A1 subgroup was largest, including 24 TLPs (5 AtTLPs, 7 MeTLPs, 3 OsTLPs, 6 PtTLPs and 3 ZmTLPs), the A2 subgroup contained 19 TLPs and the A3 subgroup was smallest, containing only 12 members. The C subfamily contained 6 TLPs. The B subfamily was smallest and contained only three TLPs (ZmTLPs 10 and 15 and AtTLP4) but did not contain any MeTLPs (Fig. 1a). Two MeTLPs 6 and 9 that were categorized into the subfamilies C were found to be on an evolutionary branch different from that of the other 11 MeTLPs (Fig. 1a).

Figure 1.

Phylogenetic trees of TLPs (a) of different plants and MeTLPs (b), conservative motifs of MeTLPs (c), and exon–intron structure of MeTLPs (d) in cassava.CDS, coding sequence; MeTLP, cassava tubby-like protein gene; MeTLP, cassava tubby-like protein; UTR, untranslated region.

Putative conserved sequence motifs of MeTLPs and genomic structure of MeTLPs

A total of 16 conserved amino acid sequence motifs were found in MeTLPs, including 1 F-box domain (motif 3) in the N-terminus, six tubby domains (motifs 1, 2, 4, 5, 7 and 12) in the C-terminus, and nine unknown motifs (motifs 6, 8–10, 11 and 13–16) (Fig. 1b; Supporting Information—Table S3). In general, the same subfamily of MeTLPs (Fig. 1b) shared similar putative conserved sequence motifs (Fig. 1c). Eleven MeTLPs 1–5, 7, 8 and 10–13 had both F-box and tubby domains, and two MeTLPs 6 and 9 had no F-box domains. All the MeTLPs had three tubby domains of motifs 1, 5 and 7. The tubby domains of motif 12 were present in only MeTLPs 6 and 9 (Fig. 1c; Supporting Information—Table S3).

Eleven MeTLPs in the A subfamily (Fig. 1a and b) contained three to six introns (Fig. 1d). Two MeTLPs 6 and 9 in the C subfamily (Fig. 1a and b) had eight introns (Fig. 1d).

The conserved amino acid residues in F-Box and tubby domains of MeTLPs were shown in Fig. 2.

Figure 2.

Conserved amino acid residues in the tubby domain and F-box domain of MeTLPs. MeTLP, cassava tubby-like protein.

Chromosomal distribution and collinearity of MeTLPs

A total of 13 MeTLPs were mapped to seven of the 18 cassava chromosomes (Fig. 3). Eleven pairs of MeTLP duplications, which belonged to whole-genome duplications/segmental duplications, were found [seeSupporting Information—Table S4]. The K_a/K_s values of 11 pairs of MeTLP duplications were all <1, therefore, their duplication likely took place between 10.11 million years and 126.69 million years ago [seeSupporting Information—Table S4].

Figure 3.

Chromosomal distribution and duplication of MeTLPs in the cassava genome. The coloured curves indicate segmental duplication genes. MeTLP, cassava tubby-like protein gene.

There were 11, 11 and 13 MeTLPs that showed collinearity with the TLPs of three dicots, Arabidopsis, poplar and potato (Fig. 4; Supporting Information—Table S5), respectively. There were 9, 6 and 9 MeTLPs that had collinearity with the TLPs of three monocots, rice, maize and sorghum (Fig. 4; Supporting Information—Table S5), respectively. Interestingly, five MeTLPs 2, 3, 5, 10 and 13 had collinearity with the TLPs from both dicots and monocots (Fig. 4; Supporting Information—Table S5). Three MeTLPs 4, 8 and 11 showed collinearity with the TLPs of only three dicots, Arabidopsis, poplar and potatoes (Fig. 4; Supporting Information—Table S5), respectively.

Figure 4.

Collinearity between cassava MeTLPs and TLPs from other plants. The grey lines in the background indicate the collinear blocks within the genomes of cassava and other plants, whereas the red lines highlight the syntenic TLP pairs. The Arabic numerals after Chr, such as 01, indicated chromosome number. A., Arabidopsis; Chr., chromosome; M., Manihot; MeTLP, cassava tubby-like protein gene; O., Oryza; P., Populus; S., Solanum; Sor., Sorghum; Z., Zea.

Three-dimensional structure of MeTLPs

As shown in Fig. 5, nine MeTLPs 1, 2, 4, 5–8, 11 and 13 presented a typical tubby architecture formed by a closed β-barrel that was composed of one central α-helix and 12 anti-parallel strands. Three MeTLPs 3, 9 and 10 contained an incomplete β-barrel. MeTLP3 had one central α-helix and 11 antiparallel strands. Two MeTLPs 9 and 10 showed one central α-helix and 10 anti-parallel strands. MeTLP12 just had 12 antiparallel strands but no central α-helixes.

Figure 5.

Homology modelling of the three-dimensional structure of MeTLPs. The α helixes are shown in green, and β-folds are shown in purple. MeTLP, cassava tubby-like protein.

Cis-acting elements in potential promoter regions of MeTLPs

A total of 68 cis-acting elements were found in potential promoter regions of 13 MeTLPs, including 28 functionally unknown elements. These cis-acting elements could be divided into the following seven types (Table 3; Supporting Information—Table S6): development, environmental stress, hormone-responsive, light-responsive, promoter-related, site-binding and other aspects. Of the cis-acting elements, the number of light-responsive elements was up to 16, accounting for 23.9 % of the total cis-acting elements.

Table 3.

The number and types of cis-acting elements in the promoter regions of MeTLPs

MeTLP	Development	Environmental stress	Hormone responsive	Light responsive	Promoter related	Site binding	Others
MeTLP1	0	2	5	9	2	0	10
MeTLP2	0	3	7	4	2	0	12
MeTLP3	1	3	2	6	2	0	9
MeTLP4	0	4	2	3	2	0	9
MeTLP5	0	2	3	7	2	0	10
MeTLP6	1	1	5	4	2	0	8
MeTLP7	0	2	3	6	2	0	8
MeTLP8	0	3	3	4	2	0	15
MeTLP9	1	3	2	6	2	0	8
MeTLP10	1	2	6	7	2	1	11
MeTLP11	1	5	5	4	2	2	11
MeTLP12	0	0	0	1	2	0	0
MeTLP13	1	3	2	5	2	1	8

MeTLP, cassava tubby-like protein gene.

For the development-related elements (Table 3; Supporting Information—Table S6), they were found in the promoters of six MeTLPs not including MeTLPs 1, 2, 4, 5, 7, 8 and 12, which could be subdivided into the following groups: CAT-box and CCGTCC-box related to the development of meristems; GCN4 motif relevant to the development of endosperm; and RY-element associated with seed-specific regulation.

Of six environmental stress-related elements (Table 3; Supporting Information—Table S6), both anaerobic induction (ARE) and GC motifs were essential for anaerobic induction, of which AREs existed in the promoters of 11 MeTLPs. Low-temperature responsiveness and MYB-binding sites (MBSs) were related to the plant response to low temperature and drought, respectively. LTRs responsive to low temperature were found in the promoters of MeTLPs 1, 2, 4, 9, 11 and 13. MBSs were found in MeTLPs 3, 4, 11 and 13. No abiotic stress-related elements were found in the promoters of MeTLP12.

Of 10 hormone-responsive elements (Table 3; Supporting Information—Table S6), both AuxRR-core and TGA elements were responsive to auxin, and abscisic acid-responsive element (ABRE) responded to abscisic acid. Both CGTCA motifs and TGACG motifs were responsible to methyl jasmonate. Both gibberellin responsive element motifs and P-boxes responded to gibberellin. TCA motifs and ethylene-responsive elements (EREs) were associated with salicylic acid and ethylene, respectively. One ERE was present in the promoters of 9 MeTLPs 2,4–7, 9–11 and 13. One ABRE was present in the promoters of eight MeTLPs 1, 3, 4–6, 8, 10 and 11. No hormone-responsive elements were found in the promoters of MeTLP12.

Expression patterns of MeTLPs based on public transcriptome databases

There were 10 MeTLPs, except MeTLPs 6, 9 and 10, which had a high expression level in 11 tissues of 3-month-old TME 204 cassava (Wilson et al. 2017). MeTLP6 showed a high expression level in eight cassava tissues except for fibrous and storage roots, and leaves. MeTLP9 had a high expression level in only friable embryogenic calli and shoot apical meristems. MeTLP10 showed a high expression level in 10 tissues except for root apical meristems (Fig. 6a; Supporting Information—Table S7).

Figure 6.

Heat maps of expression of MeTLPs in 11 tissues of 3-month-old TME 204 cassava (a) and in the tissues among two cassava cultivars of Arg7 and KU50, and cassava wild species W14 (b). ESR, early storage root; Fec, friable embryogenic callus; LSR, last storage root; MeTLP, cassava tubby-like protein gene; MSR, middle storage root; Oes, somatic organized embryogenic structure; Ram, root apical meristem; Sam, shoot apical meristem.

Using data in the high-throughput SRA (Hu et al. 2016b), expression of MeTLPs in wild species W14 and two cultivars (KU50 and Arg7) was analysed (Fig. 6b). Five MeTLPs 1, 3, 7, 11 and 13 had a high expression level in all tissues of these cassava materials. Three MeTLPs 5, 6 and 9 showed a low expression level in all cassava tissues. MeTLP2 had a low expression level only in Arg7 leaves, MeTLP4 had a low expression level only in ESRs of Arg7 and middle storage roots (MSRs) of W14. MeTLP8 had a high expression level only in stems and leaves of Arg7 and stems of W14. MeTLP10 had a high expression level only in MSRs, stems and leaves of W14. MeTLP12 had a low expression level only in ESRs and leaves of Arg7 (Fig. 6b; Supporting Information—Table S8).

Expression of MeTLPs in rootlets pot-grown cassava Arg7 in response to abiotic stresses, abscisic acid and salicylic acid

Under NaCl treatment (Fig. 7a), expression of MeTLPs could be grouped into the following patterns: significantly upregulated at or after 3 h of the stress, as was the case for MeTLPs 3, 4, 8, 11 and 12; significantly downregulated throughout the stress, as was the case for only MeTLP6; and unchanged throughout the stress, as was the case for MeTLPs 2, 5 and 7. Under PEG treatment (Fig. 7b), expression of MeTLPs could be grouped into the following patterns: significantly upregulated at only 1 h of the stress, as was the case for MeTLPs 5, 6 and 10; significantly upregulated throughout the stress, as was the case for only MeTLP12; significantly upregulated at only 9 h of the stress, as was the case for MeTLPs 3, 8, 11 and 13; and unchanged throughout the stress, as was the case for only MeTLP2. Under low-temperature treatment (Fig. 8), expression of MeTLPs could be grouped into the following patterns: significantly downregulated throughout the stress, as was the case for MeTLPs 3, 5, 6 and 10; and unchanged throughout the stress, as was the case for only MeTLP9. Under abscisic acid treatment (Fig. 9a), expression of MeTLPs could be grouped into the following patterns: significantly upregulated at or after 3 h of treatment, as was the case for MeTLPs 4, 11 and 12; significantly upregulated at 9 h of treatment, as was the case for MeTLPs 3, 6 and 9; and unchanged throughout abscisic acid treatment, as was the case for MeTLPs 7, 8 and 13. Under salicylic acid treatment (Fig. 9b), expression of MeTLPs could be grouped into the following patterns: significantly upregulated at or after 3 h of treatment, as was the case for MeTLPs 4, 8, 11 and 12; significantly downregulated throughout the treatment, as was the case for only MeTLP6; and unchanged throughout salicylic acid treatment, as was the case for only MeTLP2.

Figure 7.

Expression of MeTLPs in roots of plantlets of pot-grown cassava Arg7 under 200 mM NaCl (a) and 20 % PEG (b) stress. The expression of MeTLPs in roots of 15-day-old cassava plantlets of cassava Arg7 after stem cutting planting was analysed by qPCR. Three biological repeats were conducted. Different letters on the columns indicate a statistically significant difference at a significance level of P < 0.05 by Duncan’s multiple range test. MeTLP, cassava tubby-like protein gene; qPCR, quantitative PCR. PEG, polyethylene glycol.

Figure 8.

Expression of MeTLPs in roots of plantlets of pot-grown cassava Arg7 under a low temperature at 4 °C. The expression of MeTLPs in roots of 15-day-old cassava plantlets of cassava Arg7 after stem cutting planting was analysed by qPCR. Three biological repeats were conducted. Different letters on the columns indicate a statistically significant difference at a significance level of P < 0.05 by Duncan’s multiple range test. MeTLP, cassava tubby-like protein gene; qPCR, quantitative PCR.

Figure 9.

Expression of MeTLPs in roots of plantlets of pot-grown cassava Arg7 in response to 100 μM abscisic acid (a) and 100 μM salicylic acid (b). The expression of MeTLPs in roots of 15-day-old cassava plantlets of cassava Arg7 after stem cutting planting was analysed by qPCR. Three biological repeats were conducted. Different letters on the columns indicate a statistically significant difference at a significance level of P < 0.05 by Duncan’s multiple range test. MeTLP, cassava tubby-like protein gene; qPCR, quantitative PCR.

Discussion

It has been known that cassava is highly tolerant to drought and soil infertility (Hu et al. 2016b; Wei et al. 2016). Reportedly, TLPs play an important role in plant growth, development and abiotic stress responses (Bao et al. 2014; Du et al. 2014; Xu et al. 2016). In this study, a total of 13 MeTLPs were identified in cassava, different in number from that of TLPs of other plants (Lai et al. 2004; Yang et al. 2008; Hong et al. 2015; Chen et al. 2016; Xu et al. 2016).

The TLPs from maize, poplar, Arabidopsis and rice were evolutionally classified into three subfamilies of A, B and C (Fig. 1), being in accordance with previous reports in Arabidopsis (Yang et al. 2008), apple (Xu et al. 2016) and maize (Chen et al. 2016). However, no MeTLPs in the B subfamily suggests that the functional evolution of MeTLPs is not exactly the same as that of Arabidopsis, apple and maize.

It should be pointed out that the β-barrel of MeTLPs 3, 9 and 10 was incomplete (Fig. 5), showing findings analogous to those in apples (Xu et al. 2016). According to classification of the TLP family (Wang et al. 2018), 14 MeTLPs 1–5, 7, 8 and 10–16 should belong to class III because they had both the tubby domain and F-box domain (Fig. 1b and c),and MeTLPs 6 and 9 from the C subfamily may be assigned to class II because they do not have the F-box domain in the N-terminus, different from the finding in Arabidopsis that only AtTLP8 out of the AtTLPs does not contain the conserved F-box domain (Lai et al. 2004), also not fully agreeing with previous reports that the tubby and F-box domains have a co-evolutionary relationship (Kile et al. 2002; Yang et al. 2008). However, the results could suggest that the two corresponding genes, MeTLPs 6 and 9, likely originate from one gene in an ancestral species and may be common ancestors for other MeTLPs.

Similar to the findings in other species (Yang et al. 2008; Du et al. 2014), the MeTLPs belonging to the same subfamily had similar exon–intron structures (Fig. 1b and d) and the corresponding MeTLPs possessed similar conserved amino acid sequence motifs [seeSupporting Information—Table S3], suggesting that there were similar but redundant functions among these genes. Of 16 conserved sequence motifs, three tubby domains of motifs 1, 5 and 7 were found in all MeTLPs (Fig. 1b and c), indicating the signatures of the tubby domains (Lai et al. 2004; Kou et al. 2009; Xu et al. 2016).

Gene duplication is also known as the source of differentiation in novel genes during evolution (Moore and Purugganan 2003; Kong et al. 2007). A total of 11 pairs of segmental duplications of MeTLPs [seeSupporting Information—Table S4] were different from five pairs of segmental duplications of OsTLPs in rice and six pairs of segmental duplications of PtTLPs in poplar (Yang et al. 2008), respectively, indicating that segmental duplication, rather than other duplications, is more important for MeTLP duplications in cassava than for TLP duplication in rice and poplar.

The collinearity of MeTLPs with the TLPs of six other representative plant species (Fig. 4; Supporting Information—Table S5) provides a reference for cloning and functional studies of cassava MeTLPs. More MeTLPs showed collinearity with potato StTLPs (Fig. 4; Supporting Information—Table S5), suggesting that these two types of TLPs have the conserved gene order within the corresponding chromosomal segment in these two crops (Keller and Feuillet 2000). MeTLPs 2, 3, 5, 10 and 13 showed collinearity with TLPs of analysed plant species (Fig. 4; Supporting Information—Table S5), indicating that these five orthologous pairs had already existed before the divergence of dicots and monocots, which is a phenomenon similar to the duplication of the WRKY gene family in pineapples (Xie et al. 2018). MeTLPs 4, 8 and 11 showed collinearity with the TLPs of three dicots, Arabidopsis, poplar, and potato, indicating that they occurred after the divergence of dicots.

The TLPs from different classes are very different in function (Lai et al. 2012). The F-box family proteins are characterized by a signature of the F-box domain (Jia et al. 2013). So, MeTLPs 1–5, 7, 8 and 10–16 can also be simultaneously categorized as the F-box family.

The F-box proteins are involved in ubiquitin-dependent proteolysis in cell cycle regulation and signal transduction, which function in many aspects such as cell elongation and division, injury response, floral differentiation, circadian clock and response to the plant growth regulators auxin and jasmonic acid in plants (Craig and Tyers 1999; del Pozo and Estelle 2000; Ho et al. 2008; Baute et al. 2017). Relatively to the F-box proteins, most plant TLPs remain a mystery in function although they have been found across plant species (Lai et al. 2012; Wang et al. 2018). The limited reports indicate that plant TLPs play roles in responses to multifarious stresses including biotic and abiotic stress (Bhushan et al. 2007; Kou et al. 2009; Wardhan et al. 2012; Bao et al. 2014; Wang et al. 2018; Xu et al. 2019). Two MeTLPs 6 and 9 would most likely not be eligible for ubiquitin-related protein degradation in the early stages of evolution because their corresponding MeTLPs lack the F-box domain (Fig. 1d; Table 2).

The increases in abiotic stressors are important constraints for food production and farming worldwide, including drought/water deficit, salt, and cold/low temperature (Roychoudhury et al. 2013; Razzaq et al. 2019). The tolerance to abiotic stressors in plants depends on some abiotic stress-responsive genes and phytohormone signalling pathways such as abscisic acid (Roychoudhury et al. 2013) and salicylic acid (Eremina et al. 2016). The changes in expression profiles with the tissues (Fig. 6) as well as expression responses to salt and PEG (Fig. 7), low temperature (Fig. 8), and abscisic acid and salicylic acid (Fig. 9) suggest that, like the TLPs in other plants, MeTLPs may make contributions to abiotic stress tolerance in cassava. Although further functional validation is required, subtle differences in function among MeTLPs are predictable partly because of their differential expression profiles depending on cassava cultivars and tissues (Fig. 6) and/or under different treatments (Fig. 7). The part of the reason for this is likely related to various cis‑acting elements existing in the promoter regions depending on genes (Table 3; Supporting Information—Table S6).

CONCLUSION

There is a MeTLP family containing 13 members, which can be divided into A and C subfamilies. There are 11 pairs of MeTLPs that show the duplication which likely took place between 10.11 and 126.69 million years ago. Two MeTLPs 6 and 9 likely originate from one gene in an ancestral species, may be common ancestors for other MeTLPs, and their corresponding proteins (MeTLPs 6 and 9) would most likely not be eligible for ubiquitin-related protein degradation in the early stages of evolution because they just have the tubby domain in the C-terminus but no the F-box domain in the N-terminus. MeTLPs feature differences in the number from TLPs in wheat, apple, Arabidopsis, poplar and maize, and are highlighted by segmental duplication but more importantly by the chromosomal collinearity with potato StTLPs. Like TLPs in other plants, MeTLPs are at least related to abiotic stress tolerance in cassava. However, the subtle differences in function among MeTLPs are predictable partly because of their differential expression profiles depending on cassava cultivars and tissues and/or under different treatments, which are coupled with various cis‑acting elements existing in the promoter regions depending on genes.

SUPPORTING INFORMATION

The following additional information is available in the online version of this article —

Table S1. The accession number of the public high-throughput RNA-seq read archives databases submitted by Hu et al. (2016b).

Table S2. The ID number for TLPs in plant species analysed in this study.

Table S3. The amino acid sequences and annotation of conserved motifs of MeTLPs.

Table S4. The segmental duplication events of cassava MeTLP family.

Table S5. The pair relationships of orthologous between cassava MeTLPs and TLPs of other plants.

Table S6. The potential cis-acting elements in the promoter region of cassava MeTLPs.

Table S7. The expression profiles of cassava MeTLPs in 11 tissues of 3-month-old TME 204 cassava plants based on the GEO database submitted by Wilson et al. (2017).

Table S8. The expression of cassava MeTLPs in different tissues of cultivars Ku50 and Arg7, and wild species W14 based on the high-throughput SRA databases submitted by Hu et al. (2016b).

Sources of Funding

This work was supported by the Innovation Project of Guangxi Graduate Education in 2018 (YCBZ2018020) from the Department of Education of Guangxi Zhuang Autonomous Region, Guangxi, China; and the funding (SKLCUSA-a201804) from the State Key Laboratory for Conservation and Utilization of Subtropical Agro-bioresources, Guangxi Univiersity, Guangxi, China.

Contributions by the Authors

M.-Y.D. conducted data analysis and experiments and wrote the first draft. X.-W.F. assisted in the design and management of experiments. X.-Y.P. helped to conduct the gene expression experiments. Y.-Z.L. conceived the project, designed the experiments and wrote the manuscript as a supervisor.

Conflicts of Interest

None declared.