Identification of BBX gene family and its function in the regulation of microtuber formation in yam

BBX proteins play important roles in all of the major light-regulated developmental processes. However, no systematic analysis of BBX gene family regarding the regulation of photoperiodic microtuber formation has been previously performed in yam. In this study, a systematic analysis on the BBX gene family was conducted in three yam species, with the results, indicating that this gene plays a role in regulating photoperiodic microtuber formation. These analyses included identification the BBX gene family in three yam species, their evolutionary relationships, conserved domains, motifs, gene structure, cis-acting elements, and expressional patterns. Based on these analyses, DoBBX2/DoCOL5 and DoBBX8/DoCOL8 showing the most opposite pattern of expression during microtuber formation were selected as candidate genes for further investigation. Gene expression analysis showed DoBBX2/DoCOL5 and DoBBX8/DoCOL8 were highest expressed in leaves and exhibited photoperiod responsive expression patterns. Besides, the overexpression of DoBBX2/DoCOL5 and DoBBX8/DoCOL8 in potato accelerated tuber formation under short-day (SD) conditions, whereas only the overexpression of DoBBX8/DoCOL8 enhanced the accelerating effect of dark conditions on tuber induction. Tuber number was increased in DoBBX8/DoCOL8 overexpressing plants under dark, as well as in DoBBX2/DoCOL5 overexpressing plants under SD. Overall, the data generated in this study may form the basis of future functional characterizations of BBX genes in yam, especially regarding their regulation of microtuber formation via the photoperiodic response pathway. Supplementary Information The online version contains supplementary material available at 10.1186/s12864-023-09406-1.


Introduction
Yam (Dioscorea spp.) is an important tuberous crop belonging to the Dioscoreaceae family, grown in various regions globally [1,2]. D. opposita 'Tiegun' is a popular Chinese yam cultivar that has been commercially grown in Jiaozuo city of Henan province, China for a long time due to its medicinal and nutritional benefits [3]. However, vegetative propagation of yam through its tubers year after year leads to the spread of viral infections and declining quality [4][5][6]. Researchers have successfully developed microtuber culture techniques, providing virus-free tubers suitable for transportation and storage [7]. † Yingying Chang and Haoyuan Sun contributed equally to this work. Tuberization in yam [8] and other tuberous crops [9,10] may involve hormomes (e.g., gibberellins (GAs) and abscisic acid (ABA)) and genes in GA and ABA synthesis, catabolism, and signalling [11][12][13]. Tuberization in model plant species, potato (Solanum tuberosum) [14] may also be regulated by photoperiodic genes such as circadian-regulated gene CONSTANS (CO) and the CO/ FT module [15,16]. In potato, StCO inhibited the transcription of the tuberigen StSP6A in leaves and repressed tuberization in a photoperiod-dependent manner. However, it is unknown whether tuberization in yam involves photoperiod-dependent genes, such as the BBX family-ZFP transcription factors.
The BBX family, also known as CONSTANS-LIKE (COL) proteins, are important in photomorphogenesis, seed germination, shade avoidance, photoperiodic regulation of flowering and tuberization [14,17]. In Arabidopsis, the BBX gene family has 32 members divided into five structural groups based on their domains [18]. AtBBXs are mainly involved in the regulation of flowering. It has been reported that some BBX genes function to regulate flowering and are also involved in the regulation of tuberization. For example, AtBBX1/AtCO promotes flowering under LD conditions, but inhibites tuber development under SD conditions [19,20]. AtBBX6/COL5 accelerates flowering [21], while over-expression of COL5 from lotus (Nelumbo nucifera) in potato increased tuber weight, but without changing the number of tubers under SD conditions [22]. Besides, AtBBX21 promoting seedling photomorphogenesis, also increased tuber yield in potato [23].
We obtained genomes and transcriptomes of the white Guinea yam (D. rotundata), the greater yam (D. alata), and D. opposita 'Tiegun' , and analyzed the BBX gene family in the species with a focus on physical and chemical properties, conserved domains, gene structure, chromosome distribution, collinearity, and cis-acting elements. We identified two candidate DoBBX genes in 'Tiegun' , and investigated their expression patterns in different tissues, under diurnal cycle and continuous light, and at five tuberization stages. The function of DoBBX2/DoCOL5 and DoBBX8/DoCOL8 in regulation of tuberization in potato was identified through heterologous overexpression. This study provides insights into the BBX gene family and the function of DoBBXs in microtuber formation.

Identification of the BBX gene family in yam
To identify the yam BBX genes, the B-box domain (Pfam00643/cl00034) was downloaded from the PFAM database (https:// pfam. xfam. org/) [28] and was used as the query sequence to find the predicted proteins in yam using the HMMER 3.0 program with the threshold of E-value < 10 -20 [29]. The predicted proteins containing the B-box domain were used to construct a yamspecific HMM file via hmmbuild from HMMER 3.0 [29]. The yam-specific BBX HMM was used as query against the predicted proteins of yam. The peptide sequences with the threshold of E-value < 10 -10 and containing the B-box domain identified by the PFAM database [28] and NCBI-CDD tools (https:// www. ncbi. nlm. nih. gov/ Struc ture/ bwrpsb/ bwrpsb. cgi) [30] were selected as candidate proteins. The ClustalW program in MEGA 7.0 and chromosomal locations of candidate genes learned from GFF3 files were used to remove the repetitive sequences and the redundant alternatively spliced sequences [31]. The IDs, characteristics and sequences of BBX proteins from three yam species were listed in Supplementary Tables 1-3, and were named according to their relationships with homologous genes in A. thaliana.

Multiple sequence alignment and phylogenetic analyses of BBX gene family in yam
All the amino acid residues of DoBBXs, DrBBXs, DaBBXs, AtBBXs and OsBBXs were aligned by MEGA 7.0 [31], and the phylogenetic tree was constructed via the neighborjoining (NJ) method with 1000 bootstrap replicates.

Chromosomal distribution and gene duplications
The lengths of the chromosomes and the physical locations of DrBBXs and DaBBXs were obtained from the genome annotation information (gff3) of D. rotundata and D. alata, respectively. MapGene2Chromosome v2 software (http:// mg2c. iask. in/ mg2c_ v2.0/) was used to map the distribution of DrBBX and DaBBX genes [32]. MCScanX software was used to detect the genome replication gene pairs in D. alata and between different species [33]. Nucleotide sequences with alignment ratios and similarity ratios greater than 75% and with distances between genes on the same chromosome of less than 100 kb were selected as tandem duplications. Moreover, the genes located in the duplicated regions and nucleotide sequences with alignment ratios greater than 75% were selected as resulting from segmental duplications [34]. Tbtools was used to analyze collinearity of the BBX genes and to visualize the duplicated gene pairs [35,36]. The KaKs_Calculator1.2 software was used to calculate the nonsynonymous (Ka), synonymous (Ks) substitution rates, and Ka/Ks values of the different gene duplication pairs [37]. The Ks values were used to estimate the approximate date of every duplicated event occurred in yam, using the formula: T = Ks/2λ × 10 -6 Mya (λ = 6.5 × 10 −9 ) (Supplementary Table 4).

Cis-acting elements in the promoter regions of DrBBXs and DaBBXs
We considered the sequence 2000 bp upstream of the initiation codons as the proximal promoter region sequences. Promoters were predicted in the PlantCARE database (http:// bioin forma tics. psb. ugent. be/ webto ols/ plant care/ html/) [44] and were categorized for functional groups and visualized using EvolView [40,41].

Excavation of the differentially expressed BBX genes during microtuber formation in D. opposita 'Tiegun'
The transcriptome sequencing datasets during tuberization of a traditional Chinese medicinal plant D. opposita 'Tiegun' were retrieved form the NCBI Sequence Read Archive (SRA) repository (http:// www. ncbi. nlm. nih. gov/ sra? term= SRP06 1414). The transcriptome data was processed following Kim et al. (2018) [45]. In brief, the SRA files were converted into Fastqa format using the SRAToolkit software [46]. The raw reads were then subjected to quality control using Trimmomatic and FastQC softwares [47,48]. Subsequently, high-quality reads were aligned to the reference genome from D. alata using the HISAT2 (version: 2.0.5). The mapped reads of were assembled and quantified using the StringTie program [49]. The FPKM was obtained for further analysis. The expression patterns of DoBBX family members during tuberization were examined using the FPKM and the expression heatmap was generated using the pheatmap R package [50].

Plant materials and growth condition
D. opposita 'Tiegun' were cultured on the MS medium containing 3 g·L −1 agar and 30 g·L −1 sucrose in Engineering Technology Research Center of Nursing and Utilization of Genuine Chinese Crude Drugs in Henan Province (Henan Normal University in Xinxiang, China). The growth conditions were kept at 23 ± 2 °C with 16 h light/8 h dark photoperiod under 38 μm·sec −1 ·m −2 light intensity, with medium change every four weeks. For microbuer induction, plants from four-week-old D. opposita 'Tiegun' were grown on MS medium containing and 60 g·L −1 sucrose with shaking (120 r·m −1 ) at 23 ± 2 °C under dark.
A Chinese potato variety 'E-potato-3' (S. tuberosum 'E3') kindly provided by Huazhong Agricultural University was used for genetic transformation. The potato seedlings were cultured on the potato medium (P0: MS + 30 g·L −1 sucrose + 3 g·L −1 agar) under the following: the growth conditions 23 ± 2 °C with 16 h light/8 h dark photoperiod (long day condition, LD) under 38 μm·sec −1 ·m −2 light intensity, with medium change every four weeks. For tuberization, buds at the 2 th ~ 4 th nodes from the shoot apex from two-week-old 'E3' were grown on the P0 medium for two weeks, and then transferred to the potato tuber induction medium (PTI: MS + 6 g·L −1 sucrose + 3 g·L −1 agar). The culture was maintained for 60 days at 18 ~ 20℃ in 38 μm·sec −1 ·m −2 on a 8 h/16 h day/night cycle (short day condition, SD) for tuberization.

Isolation and qRT-qPCR analyses of DoBBX2 and DoBBX8
All fresh samples were immediately frozen in liquid nitrogen after collection and stored at -80℃ for the expression analysis of DoBBX2 and DoBBX8.
The total RNA was extracted from the leaves at the 3 th ~ 4 th nodes from the four-week-old D. opposita 'Tiegun' using the RNA purification kit (TaKaRa, China). DoBBX2 and DoBBX8 were amplified using primers designed in Primer Premier 5.0 (Premier, Canada). The full CDS sequences of DoBBX2 and DoBBX8 were cloned into PBI121 with the homologous recombination technology (Vazyme, Nanjing, China).
The tissue-specific expression patterns of DoBBX2 and DoBBX8 were examined from root (R), stem (S), leaf (L), main bud (MB) and accessory bud (AB) of four-week-old D. opposita 'Tiegun' as well as the MBs 0, 7, 14, 28, 35, and 42 d after microtuber induction. The 3 rd leaf from the shoot apex of the 4-week old D. opposita 'Tiegun' plants under 16 h light/8 h dark was collected in the dark period every 3 h for the daily expression patterns. Then, the plants were transferred to continuous light for circadian expression analysis. The 3 rd leaf was collected every 3 h for 60 h after the initial adaptational period (8 h) for continuous light. The experiments were repeated three times, each time with 10 planets.
cDNAs were synthesized using a HiScriptII 1st Strand cDNA Synthesis Kit (Vazyme, Nanjing, China) and transcript levels were quantified with qPCR run in the Roche Real-Time PCR Detection System with the AceQ qPCR SYBR Green Master Mix (Vazyme, Nanjing, China). All reactions were done in triplicate. The primers for gene isolation and qRT-PCR analysis were listed in Supplementary Table 5.
The SMART (http:// smart. embl-heide lberg. de/) was employed to determine conserved domains of DoBBX2 and DoBBX8 proteins. The NJ tree was constructed using MEGA 7.0 software with 1,000 bootstrap replicates among full-length amino acid residues of BBX2/ COL5 and BBX8/COL8 homologous genes in D. opposita  Table 6).

Transformation in potato and tuberization
Slices of tuber from 'E3' potato were used for Agrobacterium tumefaciens-mediated transformation. The vector containing 35S::DoBBX2 and 35S::DoBBX8 were transformed into A. tumefaciens strain LBA4404 using electroporation, individually. The overexpression (OE) lines were selected on kanamycin (100 mg·L −1 ) MS medium, verified by semi-RT-PCR and qRT-PCR (specific primers were listed in Supplementary Table 5), and propagated by growing single nodes on P0 medium under LD. Two-week-old seedlings of OE lines and 'E3' were transferred to SD or dark for the observation of tuberization, and tuberization rate-related parameters were documented including tuberization time, number of tubers per plants, and tuber weight per tuber. The experiments were repeated three times and each time with 10 plants for each treatment, i.e., SD, or dark.

Statistical analyses
Statistical analysis of collected data was carried out by Excel and SPSS in this chapter. The Fisher LSD test (P < 0.05) were performed to test whether there were significant differences at P = 0.

Identification and characteristics of BBX genes in yam Identification and phylogenetic analysis of BBX genes in yam
Sequence identity analyses using HMM identified 20, 20, and 19 BBX genes in D. rotundata, D. alata, and D. opposita, respectively (Supplementary Tables 1-3).
The phylogenetic analysis suggested that there are five groups of the BBX families in D. rotundata, D. alata, and D. opposita (Fig. 1, Supplementary Tables 1-3). Group IV had the most members (7 Dr, 7 Da and 7 Do), while Group V had the fewest members (1 Dr, 2 Da and 2 Do) ( Fig. 1). In addition, the orthologous genes identified among three Dioscorea species are highly conserved.

Chromosomal location and gene duplication events of BBX in yam
In D. rotunda, 17 BBX genes are located on 11 of the 20 chromosomes, whereas the other three DrBBXs are found on scaffolds. All of the 20 BBX genes are all located on 12 of the 20 chromosomes in D. alata (Fig. 2a and Supplementary Figure 1). Intraspecific collinearity analysis in D. alata suggests that there are both segmental and tandem duplications in this species, producing four and one BBX gene pairs, respectively (Fig. 2a).
Homologous gene analysis between D. alata and D. rotundata showed that 23 Da-Dr orthologous gene pairs were located in collinear regions (Fig. 2b, Supplementary  Table 4). While 23 homologous genes pairs are colinear, genes are slightly different in their locations on the chromosomes between D. alata and D. rotundata.
D. alata has more colinear BBX genes with rice (26 pairs) than with A.thaliana (7 pairs). The DaBBX genes showed one-to-one (6), two-to-one (17), and three-toone (3) corresponding relationships with BBX genes in rice, while showed one-to-one (1), two-to-one (2), and one-to-two (4) Table 4), indicating that they were under strong purifying selection during their evolution and a conserved evolutionary pattern was shared among BBX gene family in yam.

Conservative domain, motif and gene structure analyses of DrBBXs and DaBBXs
In the predicted protein domains, there is conserved B-box1 domain in all of DrBBX and DaBBX proteins, in the BBX gene groups I, II, III, IV and V, however, the conserved domains included the B-box1 + B-box2 + CCT, B-box1 + B-box2 + CCT, B-box1 + CCT, B-box1 + B-box2, and B-box1 domain combinations, respectively. Besides, all members in group I and 9 individual members in Group IV contained a valine-proline (VP) motif in their C-terminal region (Fig. 3a, Supplementary Figure 2). Protein sequence alignment and logo analysis show that B-box1, B-box2, CCT domains and VP motifs are highly conserved ( Supplementary Figures 2 and 3).
DrBBXs and DaBBXs from same groups are similar in motif number, exon/intron number, length and arrangement (Fig. 3b,c). Motif 1 + Motif 10 are present in all of the DrBBX and DaBBX groups, Motif 2 is present in group I, II and III, Motif 3 + Motif 4 are present in group I, II and IV, whereas Motif 8 was specific in group I and Motif 6 + Motif 9 was specific in group II. Motif 5 was widely distributed in all of the DrBBX and DaBBX members (except DaBBX20), while Motif 7 was only   (Fig. 3b,c).

Cis-elements in the promoter regions of DrBBXs and DaBBXs
We identified 28 cis-elements. Besides the conventional cis-acting elements (TATA-box, CAAT-box) in the promoter, the other 26 cis-acting elements include 8 light responsive, 8 hormone-responsive, 8 stress responsive, and 2 growth and development groups (Fig. 4). The most frequent elements were Box 4 (the total number was 176) and G-Box (95) (involved in light responsiveness), ABRE (ABA responsive element, 98), and ARE (essential for the anaerobic induction, 92).

Excavation and functional analysis of BBXs in microtuber formation in D. opposita 'Tiegun' Excavation of differentially expressed BBXs during the microtuber formation
Based on the data obtained from RNA-seq during the microtuber formation in D. opposita 'Tiegun' , the accumulation of DoBBXs during microtuber formation were depicted by heat map (Fig. 5). On the basis of the expression patterns of the BBX genes during the microtuber formation, the DoBBXs were clustered into three main classes: class A, class B and class C (Fig. 5). Six DoBBX members in class B were nearly absent during microtuber formation. Whereas 13 DoBBX genes showed differentially expression patterns during microtuber formation, which included Class A (DoBBX8, DoBBX14 and DoBBX4) were upregulated, and Class C (DoBBX2, DoBBX3, DoBBX7, DoBBX10, DoBBX12, DoBBX15, DoBBX16, and DoBBX18) were down regulated. Among them, DoBBX2 transcripts were the highest at EXP stage, and a significant reduction (~ 10.922-fold) was found from EXP to MTV stages. Besides, we observed the highest increase of DoBBX8 expression (~ 3.880-fold) from EXP to MTV stages. Further study would be focused on DoBBX2 and DoBBX8.

Spatial expression patterns of DoBBX2 and DoBBX8
DoBBX2 and DoBBX8 were isolated from D. opposita 'Tiegun' and submitted to NCBI's GenBank (DoBBX2/ DoCOL5, NCBI Accession No: CP897490; DoBBX8/ DoCOL8, NCBI Accession No: CP897491). The fulllength sequence of DoBBX2 cDNA was 1332 base pair (bp), ORF was 1041 bp length and encoded a deduced  Figure 4). Similar expression patterns were observed for the two DoBBX genes in 5 tissues with the highest expression in leaf (L), moderate in root (R), stem (S) and accessory buds (AB), and the lowest in main buds (MB) (Supplementary Figure 5, Fig. 6a). Specifically, the expression levels of DoBBX2 and DoBBX8 in L was ~ 3.90 times and ~ 7.61 times that in MB, respectively. The transcript levels of DoBBX2 were significantly higher than that of DoBBX8. This finding implied DoBBX2 and DoBBX8 may participate in the regulation of microtuber formation by receiving signals in leaf and then travel to the tissues forming microtubers.
In order to verify the involvement of DoBBX2 and DoBBX8 genes in microtuber formation, the expression levels of these two DoBBXs were detected at five stages during microtuber formation (Fig. 6b). The DoBBX2 expression gradually decreased. However, the expression level of the DoBBX8 increased from EXP stage (0 d) to MTV stage (28 d), then decreased from Given that numerous plant BBX proteins are implicated in multiple light-regulated growth and developmental processes, we then sought to study the daily oscillation and circadian expression pattern of DoBBX2 and DoBBX8 (Fig. 6c). The changes in the expression profiles of DoBBX2 and DoBBX8 were observed in the day/night cycle oscillate following about 24-h rhythm. The transcript level of DoBBX2 increased during the day and reached a maximum in the middle of the light phase, whereas that of DoBBX8 had two main peaks throughout the 24-h diurnal cycle, the first with maximal expression in the light phase and the second with maximal expression in the end of light phase (the point of 15 h under light) or in the beginning of dark phase (before -6 h under dark, not shown). In the 8 h light adaptations period, the transcript level of DoBBX2 decreased, while that of DoBBX8 decreased first and then increased. During the continuous 44 h light period, the expression profiles of DoBBX2 were in synchrony with those observed during the diurnal cycle, but the maximum values were decreased. The expression of DoBBX8 was altered by prolonging the total duration at the phase of diurnal cyclic. It suggested that the diurnal expression of DoBBX2 was not directly regulated by light, but by the circadian clock; while diurnal oscillation of DoBBX8 expression required a dark period.

Overexpression of DoBBX2 and DoBBX8 regulates tuberization through photoperiodic pathway in potato
We produced more than 10 transgenic potato lines with 35S::DoBBX2 and 35S:: DoBBX8 constructs as detected using semi-PCR and qRT-PCR (Supplementary Figure 6).
The tuberization in transgenic and control potato lines were investigated under both SD and dark conditions (Fig. 7). Almost all of the transgenic lines (except for DoBBX8-2) formed tuber earlier than 'E3' under SD (Fig. 7a,c,f ). The OE lines of DoBBX2-1, DoBBX2-2, DoBBX2-3, DoBBX8-1 and DoBBX8-3 formed tubers after 21, 34, 34, 34, 34 days of being transferred to SD, respectively, while a minimum of 36 days was required for 'E3' . Under SD for 60 days, we observed a higher percentage of tuberization in the OE lines of DoBBX2 and DoBBX8 than in 'E3' (Fig. 7e). In addition, tuber yield was greater in the OE lines of DoBBX2-1, DoBBX2-2 and DoBBX2-3 than in the 'E3' , with higher average of tubers produced per plant but without changing weight per tuber (Supplementary Figure 7, and Fig. 7g,h). However, the OE of DoBBX8 produced fewer tubers per plant (Fig. 7g,h). It indicated that both DoBBX2 and DoBBX8 advanced tuberization under SD, while number of tubers per plant was higher in OE DoBBX2 lines compared to 'E3' , but it was slightly reduced in OE DoBBX8 lines.
Under dark, the tuberization time of 'E3' and OE DoBBX8 lines was significantly earlier than that under SD (Supplementary Figure 7, Fig. 7b,d,f ). Under dark for 60 days, the tuberization rate, mass per tuber and yield were all reduced in 'E3' and OE lines (Fig. 7e,g,h). The number of tubers per plant in OE DoBBX8-2 and DoBBX8-3 lines had no significant change, while it demonstrated an obvious reduction in the 'E3' and OE DoBBX2 lines than that under SD (Fig. 7g). This indicates that although both over-expression of DoBBX2 and DoBBX8 can accelerate tuber formation under dark, only the overexpression of DoBBX8 enhanced the promoting effect under dark on tuber induction.

Discussion
The BBXs are zinc-finger transcription factors and play vital roles in plant growth, development and response to biotic and abiotic stresses [52][53][54]. BBX gene family has been identified from several plant species, such as Arabidopsis [18], rice [55], orchid [56], bamboo [57], tomato [58], potato [59], pepper [60], grapes [61], cottons [62], pear [63], and apple [64]. However, little information is known about the BBX gene family in yam, which is one of the important tuberous crops. In this study, we performed systematic genome-wide identification and analyses of the BBX gene family in three yam species. Based on the transcriptome data during microtuber formation in D. opposita 'Tiegun' , we selected two candidate genes for further investigation. Furthermore, we focus on the expression patterns and the potential function of the two candidate genes in potato tuberization.
In the present species, segmental duplication counts for 20% (4 Dr-Dr paralogs) of DrBBX and 25% (5 Da-Da paralogs) of DaBBX (Supplementary Table 4 and Fig. 2). The percentages are smaller than those in rice (60%, 18 OsBBX) [55], Phyllostachys edulis (88.89%, 24 PeBBX) [57], Solanum lycopersicum (40%, 12 SlBBXs) [58], but greater than those in D. officinale (10.53%, 2 DofBBXs) and P. equestris (12.50%, 2 PeqBBXs) [56]. It may be the reason for the difference in BBXs number among these species. We found five groups of the BBX genes and they varied in conserved domains, motifs and gene structures. Group IV with the largest number of the BBX genes in D. rotundata, D. alata, Arabidopsis and rice, it may have undergone a much greater gene expansion [17]. However, the reasons are unclear. The differences between groups might be linked with a wide functional diversity in BBX gene family [65].
The different cis-regulatory elements in the promoter regions may also be important for functional diversity [66]. The G-box and its variants in the promoter region of BBX gene are the binding sites of central regulators like HY5 and PIFs of photomorphogenesis [67]. The homologous genes of DoBBX2 and DoBBX8 in D. rotundata and D. alata possess the G-box as well as the ABA/GAresponsive cis-elements. 85% of DaBBXs (17) and 90% DrBBXs (18), including homologous genes of DoBBX2 and DoBBX8, have ABRE motifs, whereas 85% of DaB-BXs (17) and 85% of DrBBXs (17) contain CGTCA-motif (Fig. 4). The BBX1/CO gene was initially identified in Arabidopsis as an important regulator of flowering in the photoperiodic pathway [19,68,69] and also play a role in regulating the tuber formation in potato [16,53,55]. Therefore, DoBBX2 and DoBBX8 may have the potential functions of photoperiodic and hormone-regulated tuberization.
Our spatio-temporal expression analyses indicate that DoBBX2 and DoBBX8 proteins may function as components of circadian clock signals during microtuber formation. This is consistent with studies in Arabidopsis and rice [59,70]. However, specific functions regulated by the circadian clock in DoBBX2 and DoBBX8 have not yet been clarified.
The length of the dark period is critical for tuberization [15]. D. opposita 'Tiegun' MBs were previously induced to form microtuber under dark [8]. Tubers of potato plants overexpressing DoBBX8 (the OE DoBBX8 lines) form earlier under dark than under the SD (Fig. 7). Surprisingly, the early tuber formation in the OE DoBBX8 lines occurs in both SD and dark conditions. The overexpression of DoBBX8 also increases the number of tubers per potato plant under dark conditions. However, DoBBX2 is down regulated during the microtuber formation, accelerating tuberization in potato under SD, but delaying tuberization under dark. Besides, DoBBX2 increases the tuber yield per plant both under SD and dark (Supplementary Figure 7). Our results are similar to those of the overexpression of NnCOL5 in potato, which affects the expression levels of NnCOL8, and the related genes in CO-FT and GA signal pathways [22]. Therefore, one of the future studies may focus on elucidating the regulatory relationships between DoBBX2, DoBBX8, hormoneregulated and photoperiod related genes.

Conclusion
In this study, we conducted a systematic genome-wide analysis of the BBX gene family in three yam species and identified 20 DrBBX, 20 DaBBX and 19 DoBBX genes in D. rotundata, D. alata and D. opposita 'Tiegun' transcriptome, respectively. The BBX genes form five major groups that are characterized by duplications, conserved domains, motifs, gene structure and cis-elements. Our RNA-seq data suggest that DoBBX2 and DoBBX8 are functional BBX genes during D. opposita 'Tiegun' microtuber formation under SD and dark condition. The overexpression of DoBBX2 and DoBBX8 in potato accelerates tuberization under SD and increase the number of tubers per plant under both SD and dark. Overall, this study forms the basis for future functional characterizations of yam BBX gene family, especially regarding regulation of tuberization via the photoperiodic pathway.