Transcriptome analysis reveals long noncoding RNAs involved in the tuber anthocyanin synthesis of potato (Solanum tuberosum L.)

Abstract

Anthocyanin plays an important role in improving the overall appearance and nutritional value of plants, and its has physiological functions of anti-cancer, aging prevention, cardiovascular disease on humans. Although the molecular mechanism has been described by genome-wide association studies (GWAS), metabolomics analysis, proteome and transcriptomic analysis, the mechanism by which the long non-coding RNA manipulates the anthocyanin biosynthesis in potato remains unknown. In present study, we using white potato (Atlantic) and purple potato (Huasong66) as materials, and detection of anthocyanin content showed that there were significant differences between the two materials. Three development stage tubers, include formation, bulking and the maturity, of Atlantic (D1, D2, D3) and Huasong66 (H1, H2, H3) were collected and used for deep transcriptome sequencing. A total of 15,191 long non-coding RNAs (lncRNAs) were identified, including 8,124 lincRNAs, 1,959 antisense-lncRNA, 2,302 intronic-lncRNA and 2,806 sense overlapping lncRNAs. Among them, 990 lncRNAs were preferentially expressed in the Huasong 66 and 679 lncRNAs were preferentially expressed in the Atlantic. A total of 1,831 lncRNAs were common differential expression across three tuber development periods between Atlantic and Huasong 66. And partly of their target genes mainly participate in phenylalanine metabolism, anthocyanin biosynthesis, flavone and flavonol biosynthesis, phenylalanine, tyrosine and tryptophan biosynthesis. We found the lncRNA TCONS00021404 and its target genes, C88_C11H1G019590 and C88_C10H3G081620, which encode flavonoid 3’,5’-hydroxylase and anthocyanidin 3-O-glucosyltransferase, were all showed higher expression patterns in Huasong 66 rather than in Atlantic, not only during the tuber formation and bulking stages, but also during the maturity stage. In addition, we found there were 243 lncRNAs act as precursors of 53 known miRNAs, 917 lncRNAs were targeted by 205 miRNAs, and part of lncRNAs could acted as eTMs of miR172b and miR482b respectively. This study established the transcriptomic repertoires in tuber development and provided evidence for the potential functions of lncRNAs in anthocyanin synthesis in potato.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12870-026-08701-9.

Introduction

With the development of next generation sequencing (NGS), more and more transcriptome data have been received. Although most of researches have revealed that ~ 90% of eukaryotic genomes are transcribed [1], a massive amount of experiments showed that DNA sequences encode proteins occupy only a small amount (2–25%). This result suggest that there are a large proportion of non-coding RNAs (ncRNAs) [2], which can be divided into housekeeping ncRNAs (rRNAs, tRNAs, snRNAs and snoRNAs) and regulatory ncRNAs [3]. Long non-coding RNAs (lncRNAs) are transcripts of longer than 200 nucleotides (nt) in length, it possess no coding sequence (CDS) or open reading frame which coding for more than 100 amino acids [4]. They are mainly transcribed by RNA polymerase II in animals [5, 6] and also can be transcribed by RNA polymerase IV and V in plants [7]. LncRNAs are divided into long intergenic non-coding RNAs (lincRNAs), long non-coding natural antisense transcripts (antisense), intronic lncRNAs (sense_intronic) and sense overlapping non-coding RNAs depending on their relative positions with/in protein-coding genes [8]. LncRNAs shows low expression, poor sequence conservation and complex functions, these characteristics lead difficult to identify lncRNAs from genome sequences through computational methods [9]. Although studies on plant lncRNAs are far behind those in animals [10], there are a number of lncRNAs have been identified in plant species with the rapid development of sequencing technology [11–16].

Anthocyanins belong to the class of flavonoids, which are water-soluble pigments widely present in plants. They usually make plants appear red, blue or purple, and play an important role in improving the overall appearance and nutritional value of plants [17]. Anthocyanins are currently recognized as the best natural antioxidants, which have physiological functions, such as anti-cancer, aging prevention, cardiovascular disease and beauty [18]. Anthocyanin synthesis is an important branch of the flavonoid metabolic pathway [19], and the related enzymes are mainly encoded by structural genes, which can be divided into early biosynthetic genes (EBGs), including: chalcone synthase gene (CHS), chalcone isomerase gene (CHI), flavanone 3-hydroxylase gene (F3H) and flavonoid 3-hydroxylase gene (F3’H); And late biosynthetic genes (LBGs), including flavonoid 3’, 5’ -hydroxylase (F3’,5’H), dihydroflavonol reductase (DFR), flavonoid glycoside transferase (UFGT) and anthocyanin synthase (ANS) [20]. The biosynthesis of anthocyanins in plants starts from phenylalanine, which is catalyzed by phenylalanine ammonia-lyase (PAL), cinnamate hydroxylase (C4H) and 4-coumaroyl-CoA ligase (4CL) to produce 4-coumaroyl-CoA, the precursor of flavonoids [20]. Under the catalysis of CHS, coumaroyl-CoA and malonyl-CoA generate the first colored substance: Chalcone, in the flavonoid synthesis pathway [21]. Subsequently, CHI catalyzes chalcone to generate naringenin, so CHI is one of the key genes in the biosynthesis of flavonoid compounds, and its expression pattern has an important impact on the flavonoid content in plants [21]. F3H is another key structural gene in the upstream of anthocyanin metabolism pathway [22]. It encodes an enzyme that catalyzes the addition of a hydroxyl group to the C3 position of flavanone to form dihydroflavonols [21]. Dihydroflavonols can form colorless anthocyanins under the action of DFR, and can also form flavonols under the action of flavonol synthetase. Finally, the colorless anthocyanins were transformed into corresponding colored anthocyanins under the catalysis of ANS, and the stable structure of anthocyanins was formed under the action of UFGT [23].

Except for structural genes, the synthesis, transport and accumulation of anthocyanins are specifically and synergistically regulated by multiple transcription factors [23]. At present, the three most studied and most important transcription factors in the anthocyanin synthesis pathway are MYB, bHLH and WDR [20]. These transcription factors can promote or inhibit anthocyanin biosynthesis by recognizing specific regions of structural gene promoters [24]. In addition to individually regulating structural genes, the three types of transcription factors can also combine with each other to form MBW complex to jointly regulate structural genes [25].

Although protein-coding genes related to anthocyanin biosynthesis have been identified, few studies about lncRNAs involving anthocyanin have been reported up to now. Apple MdWRKY1 regulates the cis-interacting element W-box in the promoter region of lncRNA MdLNC499, and activation of MdWRKY1 under light induction can promote the expression of MdLNC499. MdLNC499 further induced the upregulation of anthocyanin synthesis structural genes by regulating the expression abundance of its adjacent gene MdERF109, and finally promoted anthocyanin synthesis [26]. In radish, overexpression of LINC15957 was significantly increased anthocyanin accumulation in leaves, and the expression levels of structural genes and transcription factors related to anthocyanin biosynthesis were also significantly increased [27]. Through the multiple high-throughput sequencings analysis of sea buckthorn fruit at three development stages, Zhang et al. assembled and annotated 9,008 lncRNAs in sea buckthorn fruits, and identified 118 differentially expressed lncRNAs [28]. Silenced the two lncRNAs (LNC1 and LNC2) in vivo lead to increased and decreased anthocyanin content in transgenic plants [28].

Potato (Solanum tuberosum L.) is the fourth largest grain crop in the world. In addition to starch, protein and fatty acid, potato tubers are also rich in mineral elements such as calcium, phosphorus, iron and vitamins [29]. Potato can be divided into two categories: ordinary potato and colored potato. Colored potato not only contains all the nutrients of ordinary potato, but also is rich in anthocyanins [30], which has good market application value. In recent years, the analysis of anthocyanin synthesis mechanism of colored potato tubers has become a research hotspot. Du et al. used 135 diploid potato germplasms to perform GWAS (Genome-wide association studies) analysis of potato tuber anthocyanin traits, and identified a major locus Pf on chromosome 10. Two candidate genes, StMYB200 and StMYB210, were obtained by combining genome and transcriptome analysis. Further verification revealed that StMYB200 and StMYB210 formed a complex with StbHLH1 to activate the transcription of downstream anthocyanin synthesis genes and promote the accumulation of anthocyanin [31]. The research results from Zhang et al. [32] suggested that StWRKY13 could promote anthocyanin biosynthesis by activating transcription of StCHS, StF3H, StDFR, and StANS in potato tubers. The transcriptomic and metabolomic profile from three colored potato clones at three developmental stages were analyzed in tubers, through weighted gene co-expression network analysis (WGCNA), the results showed that both StMYB3 and StMYBATV could interact with the co-factor StbHLH1 and repress anthocyanin biosynthesis [33]. By comparing the transcriptome expression profiles of the tubers of two potato varieties, a total of 36 differentially expressed genes (DEGs) which involved in flavonoid biosynthesis were obtained. Among them, the StWRKY44 was proved acted as a possible regulator of anthocyanin accumulation [34]. Also using transcriptome and metabolome analysis, Zhang et al. proved that the expression level of F3’5’H is significantly correlated with the content of anthocyanins in purple flesh, and F3’5’H was responsible for metabolic divergence between red and purple samples [35]. In our previous research, through using the high-throughput sequencing and systematic bioinformatics analysis, a total of 7,176 significantly DEGs were discovered from the purple potato Huasong 66 tubers at different developmental stages. GO and KEGG enrichment analysis showed that 43 DEGs were mainly enriched in the biological processes which are closely related to anthocyanin biosynthesis, these DEGs include the protein-coding genes PAL, CHS, CHI, 4CL, F3H, UFGT, LAR, and the TFs MYB, bHLH, and HY5 [20].

The regulatory factors of anthocyanin synthesis are complex and diverse. So far, there has been no report on lncRNA regulating anthocyanin synthesis in potato tubers. To identify functional lncRNAs related to anthocyanin synthesis in potato, we performed whole transcriptome sequencing to detect lncRNA changes during the three development stages of colored potato Huasong 66 and white potato Atlantic. Together with 18 available potato RNA-seq datasets, we systematically identified lncRNAs and performed functional analysis on some of these anthocyanin related lncRNAs. Our results indicated that a lncRNA TCONS_00021404 is highly expressed in colored tubers, and its target gene C88_C10H3G081620 (UFGT: Anthocyanidin 3-O-glucosyltransferase) and C88_C11H1G019590 (F3’5’H: Flavonoid 3’,5’-hydroxylase) also showed higher expression abundance in colored potato. Taken together, these studies helped us to assess the functions of lncRNAs during anthocyanin synthesis, and deepened our understanding of the putative regulatory roles of lncRNAs in potato tuber development processes.

Materials and methods

Potato materials collection

Two tetraploid potato genotypes, Huasong 66 and Atlantic were used in this study. Huasong 66 with purple tuber skin and flesh, Atlantic with yellow-skinned and white fleshed, all of them were kindly provided by HuaSong Seed Industry Co., Ltd. (Beijing, China). Seeds were sown on the farm of Agricultural College of Inner Mongolia Agricultural University (Hhhot, China) (40°46`N, 110°45`E) in early May 2023. After conventional weeded, fertilized, and watered during their developmental phases, the potato tubers of three different developmental phases (H1, H2, and H3 for Huasong 66; D1, D2 and D3 for Atlantic) were collected between July 15 and September 30, 2023 at 42-day intervals. And the last time samples (H3, D3) were sampling when the stems and leaves had withered. According to the classification standards of potato growth stages [36], the sample H1 and D1 were during the tuber formation stage, and H2 and D2 were during the tuber bulking stage, and H3 and D3 were maturation stage. Three tubers from the same plant were used as three biological replicates. All samples were used for anthocyanin extracted and expression profiles analysis.

Total anthocyanins content determination

Potato anthocyanins were extracted according to the method described by Huan et al. [37]. Take 1 g fresh potato tuber flesh and ground completely, put it into a test tube with 15 mL extracting solution (1:1 mixture of 0.1 mol· L^{− 1} HCl : 95% ethanol [v/v]), then water bath at 60 ℃ for 1.5 h. After filtered the supernatant, the extraction was repeated once by adding another 15 mL of the extracting solution to the filter residue. The supernatant of the two extractions was mixed and the pH was adjusted to 1.0 and 4.5. The samples were kept in the dark for 20 to 30 min, then the absorption (A) was measured at λ = 520 nm and λ = 700 nm using a visible spectrophotometer (TU-1810SPC, Beijing). The diluted test portions were read versus a blank cell filled with distilled water. Final absorbance values were calculated through the pH differential method as follows:

Library preparation and sequencing

Total RNA samples used for lncRNA-seq were extracted from the three different development stages (H1, H2, H3, D1, D2 and D3) according to the method of Nie et al. [38]. A total amount of 1.5 µg RNA per sample was used as input material for rRNA removal using the Ribo-Zero rRNA Removal Kit (Epicentre, Madison, WI, USA). Sequencing libraries were generated using NEBNext^R Ultra™ Directional RNA Library Prep Kit for Illumina^R (NEB, USA) following manufacturer’s recommendations and index codes were added to attribute sequences to each sample. Briefly, fragmentation was carried out using divalent cations under elevated temperature in NEBNext First Strand Synthesis Reaction Buffer (5X). First strand cDNA was synthesized using random hexamer primer and reverse transcriptase. Second-strand cDNA synthesis was subsequently performed using DNA Polymerase I and RNase H. Remaining overhangs were converted into blunt ends via exonuclease/polymerase activities. After adenylation of 3’ ends of DNA fragments, NEBNext adaptor with hairpin loop structure were ligated to prepare for hybridization. In order to select insert fragments of preferentially 150–200 bp in length, the library fragments were purified with AMPure XP Beads (Beckman Coulter, Beverly, USA). Then 3 µl USER enzyme (NEB, USA) was used with size-selected, adaptor-ligated cDNA at 37 ℃ for 15 min before PCR. Then PCR was performed with Phusion High-Fidelity DNA polymerase, Universal PCR primers and Index (X) Primer. At last, PCR products were purified (AMPure XP system) and library quality was assessed on the Agilent Bioanalyzer 2100 and qRT-PCR. The clustering of the index-coded samples was performed on acBot Cluster Generation System using TruSeq PE Cluster Kitv3-cBot-HS (Illumina) according to the manufacturer’s instructions. After cluster generation, the library preparations were sequenced on an Illumina Hiseq platform and paired-end reads were generated.

Quality control and lncRNA analysis

Raw data of fastq format were firstly processed and the clean data were obtained through removing the impure reads, including adapter, ploy-N and low quality reads, from raw data. The transcriptome was assembled using the Cufflinks and Scripture based on the reads mapped to the reference genome (spuddb_solanum_tuberosum_C88_v1). The assembled transcripts were annotated using the Cuffcompare program from the Cufflinks package. The unknown transcripts were used to screen for putative lncRNAs. Three computational approaches include CPC/CNCI/CPAT [39–41] were combined to sort non-protein coding RNA candidates from putative protein-coding RNAs in the unknown transcripts. Putative protein-coding RNAs were filtered out using a minimum length and exon number threshold. Transcripts with lengths more than 200 nt and have more than one exons were selected as lncRNA candidates. LncRNA candidates transcripts which with FPKM ≤ 0.1 (fragments per kilobase of exon per million mapped fragments) were also eliminated. As well as the different types of lncRNAs include lincRNA, intronic-lncRNA, sense overlapping and anti-sense lncRNA were selected using cuffcompare.

Differential expression analysis and validation

Differential expression analysis of two samples was performed using the Cuffdiff software. P-values were adjusted using the Benjamini and Hochberg’s approach. Genes with an adjusted P-value < 0.05 and |log₂(FC)| ≥ 1 found by Cuffdiff were assigned as differentially expressed.

To validate the expression patterns of lncRNAs, qRT-PCR was carried out according to the manufacturer instructions. The specific primer pairs were designed using Primer Premier 5 software [42] and the products length was 80–500 bp. The constitutively expressed gene StActin was used as internal controls for normalization of lncRNA expression patterns. All reactions were run in triplicates, and the result of qRT-PCR was analyzed using the 2^−△Ct method [43].

Interactions between lncRNAs and miRNAs

All the identified lncRNAs were analyzed for miRNA precursors using local Blastn, the potato miRNAs sequences were taken from our previous study (Unpublished). Based on similar relationships, lncRNA sequences that had less than four mismatches and possess similar secondary structure with miRNA precursors were categorized into the precursors of miRNA. The lncRNAs act as miRNA targets were predicted using psRNATarget webserver (http://plantgrn.noble.org/psRNATarget/) with an expectation ≤ 3. All lncRNAs candidates were used to predict endogenous target mimics using the online software psRobot (http://omicslab.genetics.ac.cn/psRobot) with loose parameters (penalty score threshold = 5, five prime boundary of essential sequence = 2, three prime boundary of essential sequence = 31, maximal number of permitted gaps = 3, and position after which with gaps permitted = 1).

Target gene prediction and functional enrichment analysis

Protein-coding genes located on upstream or downstream within 100 kb were cis regulated by lncRNA. The trans-acting lncRNAs and their target protein-coding genes were predicted using LncTar software [44] based on nucleotide complementary pairing. For all the target protein-coding genes, GO analysis was carried out using the Web Gene Ontology Annotation Plot (WEGO, http://wego.genomics.org.cn/). The GO annotations of all the Unigenes of S. tuberosum were taken as the background. The KOBAS 2.0 web server was used to test the statistical enrichment of genes in Kyoto Encyclopedia of Genes and Genomes (KEGG) biological pathways. The resulting P-values were adjusted using the Benjamini and Hochberg’s approach, statistically significant were assigned as adjusted P-value < 0.01.

Results

Anthocyanin content varies between the three different stages in purple‑fleshed cultivars

Total anthocyanin content of purple fleshed potato were significant higher than that of white potato. It is worth noting that the anthocyanin content in the white potato tubers is extremely low, and no specific values were detected in the tubers of the three developmental stages of Atlantic (Fig. 1). In colored potato Huasong 66, the content of anthocyanins shows a trend of increasing first and then decreasing during the tuber formation period, the bulking period and the maturity period. And in the mature tubers, the anthocyanin content is 47.85 mg/kg FW (Fig. 1). The above results indicate that comparative transcriptome analysis of these two potato varieties maybe helpful to elucidate the mechanism of anthocyanidin synthesis, and the genes specifically expressed during the formation and bulking periods of the colored potato tuber maybe play a crucial regulatory role in the synthesis of anthocyanins.

Fig. 1.

Potato cultivars used in this study and total anthocyanin content of tuber flesh tissues. Notes: H: purple tuber skin and flesh potato Huasong 66, D: yellow tuber skin and white flesh potato Atlantic. H1, H2, and H3 represent the tuber during the formation stage, bulking stage, and maturation stage for Huasong 66 respectively; D1, D2 and D3 represent the tuber during the formation stage, bulking stage, and maturation stage for Atlantic respectively. Potato anthocyanins were extracted according to the method described by Huan et al. [37], the specific figures were obtained through calculation using a formula as follow: Anthocyanin quantity (mg kg^{− 1} FW)= ((A520 - A700) _pH1.0 - (A520-A700) _pH4.5) /31,600*30*449.2*1000

A computational approach for the genome-wide identification of lncRNAs in potato

To systematically identify lncRNAs related to potato anthocyanins synthesis, we performed whole transcriptome ssRNAseq of potato tubers during the formation period (H1, D1), the bulking period (H2, D2) and the maturity period (H3, D3). A total of six samples from colored potato Husong 66 (H) and white potato Atlantic (D) were utilized in three biological replicates, including H1, D1, H2, D2, H3 and D3. After discarding low-quality reads, removing filtering 5′ contaminants, and trimming 3′ adaptor reads, more than 220 gigabases (Gb) clean reads with Q30 > 96% were obtained and used for further analysis (Table S1). The sequences for each sample were mapped to S. tuberosum genome (spuddb_solanum_tuberosum_C88_v1) via TopHat2 and Cufflinks programs. A set of unique and non-overlapping transcripts were obtained after Cuffmerge assembly. All of transcripts from 6 samples were used for distinguish potential lncRNAs through several sequential stringent filters (Fig. 2A). After using those stringent criteria, 15,191 transcripts were considered as high-confidence lncRNAs (Fig. 2B).

Fig. 2.

Properties of potato lncRNAs. Notes: (A) An integrative computational pipeline for the systematic identification of lncRNAs in potato; (B) Screen the non-protein coding RNA candidates using CPC/CNCI/CPAT, CPC: Coding Potential Calculator, CNCI: Coding-Non-Coding Index, CPAT: Coding-Potential Assessment Tool; (C) The proportion of different types lncRNAs; (D) ORF size distributions for lncRNAs, known protein-coding transcripts and novel protein-coding transcripts; (E) Transcript length of lncRNAs, known protein-coding transcripts and novel protein-coding transcripts; (F) The number of exon per transcript for lncRNAs, annotated mRNA and novel mRNAs

Characteristic features of potato lncRNAs

We characterized the basic genomic features of the obtained lncRNAs and compared these features with the available features of potato protein-coding genes. According to the locations relative to the nearest protein coding genes, lncRNAs were further classified into the different types, include lincRNA, antisense, sense_intronic, and sense overlapping lncRNAs. In this study, we found that only a small fraction of the sequence was either antisense overlapped by protein-coding mRNA (antisense-lncRNA) or transcribed from inside genes (most from introns, intronic-lncRNA or sense overlapping), accounting for 12.9%, 15.2% and 18.5% of all potato lncRNAs, respectively. Otherwise, the majority of lncRNAs (53.5%) were located in intergenic regions (lincRNA) (Fig. 2C). The lncRNAs in both plants and animals have been showed to be shorter and to consist of fewer exons compared with mRNAs, and potato lncRNAs also share the similar features. The mean length of ORF of lncRNA is 101 bp, significantly less than the 441 bp of mRNA ORF mean length ((Fig. 2D). Similarly, the results of Fig. 2E indicated that the distribution of the transcript length of these lncRNA ranged from 201 bp to 18,836 bp, which is approximately 98.2% of the lncRNAs ranging in size from 205 to 4,000 bp (Fig. 2E). In contrast, for the protein-coding transcripts, ~ 6% of them were > 4,000 bp, and the mean length of mRNAs was 1,817 bp, which was quite higher as compared to mean length of lncRNAs (883 bp). Most (~ 85%) of the potato lncRNAs contain one or two exons, and the exon number larger than 3 only accounting for 15%, while the majority of mRNAs consist of more than 3 exons (~ 72%) (Fig. 2F). Due to the more number of exons in mRNAs, it might have resulted in lesser exon length (median length of 297 nucleotides) than lncRNAs (median length of 524 nucleotides).

Lower expression abundance of potato lncRNAs

We mapped the lncRNA, miRNAs, known mRNAs and novel mRNAs onto the S. tuberosum (C88) reference genome. A circos plot clearly showed that most of potato lncRNAs were have lower densities in the pericentromeric heterochromatin regions than in the euchromatin (Fig. 3A). Interestingly, most protein-coding genes and miRNAs also showed similar distribution in potato chromosome. These results suggest that lncRNAs might have concordant transcriptional features with the mRNAs and miRNAs in potato (Fig. 3A).

Fig. 3.

Localization and expression analysis of potato lncRNAs. Notes: (A) Chromosome-wise distribution of known genes (II), novel genes (III), lncRNAs (IV) and miRNAs (V) in potato, (I) represents the chromosome of the potato; (B) Correlation analysis of the expression levels of various sample for genes; (C) Correlation analysis of the expression levels of various sample for lncRNAs; (D) PCA analysis of various sample for lncRNAs; (E) PCA analysis of various sample for genes; (F) The expression patterns of protein coding genes in different samples of potato; (G) The expression patterns of lncRNAs in different samples of potato

Then, we systematically estimated the expression level of all transcripts identified including different types of lncRNAs and mRNAs using FPKM value. The FPKM was determined using Cufflinks, and the correlation analysis results of FPKM values showed that the correlation coefficients between the three biological replicates are all greater than 0.8 (Fig. 3BC), indicating that the biological replicates are good and the data can be used for the subsequent expression pattern analysis. Principal component analysis revealed that the three biological replicates of the same material group together and different materials separate in the principal component space. Especially, the samples of purple potato and white potato showed a very distinct dispersion (Fig. 3DE). Compare the expression abundances of lncRNAs and mRNAs in potato, the results showed that the expression abundance of lncRNAs was lower than those of mRNAs in RNA-seq samples (Fig. 3F), indicating lncRNAs were less transcribed (Fig. 3G) in potato.

Prediction and identification of interactions between lncRNAs and miRNAs

More and more miRNAs which participated in anthocyanin synthesized have been identified in Capsicum annuum, litchi and blackberry [45–47]. Long non-coding RNAs can be involved in diverse biological functions depending on a variety of modes of action. A large number of studies have demonstrated that lncRNAs can serve as miRNA precursors, potential targets or endogenous target mimics to interfere with the miRNA-mediated regulation of their intrinsic target genes in plants. In our previous research, we found there were 264 miRNAs involved in the tuber development process and maybe participate in anthocyanin metabolism of Atlantic and Huasong 66 (Unpublished). In this investigation, based on the alignment of 264 miRNA precursors with 15,191 lncRNAs, we found there were 243 lncRNAs act as precursors of 53 known miRNAs (include stu-MIR156d, stu-MIR482c, stu-MIR172a, stu-MIR399j, and et al.) and 35 novel miRNAs in potato (Table S2; Fig. 4CD). This result indicated that one lncRNA may act as precursor of multiple miRNAs. In addition, to predict whether lncRNA transcripts could be targeted by miRNAs, psRNATarget (http://plantgrn.noble.org/psRNATarget/) with an expectation ≤ 3 was used based on the structural complementary between miRNAs and lncRNAs. Among 264 potato miRNAs and 15,191 lncRNAs, 1,496 miRNA-lncRNA interactions were found, involved in 205 miRNAs and 917 lncRNAs (Table S3; Fig. 4AB). Most miRNAs, especially the conserved ones, could target several lncRNAs, such as miR399a, miR5303c, miR172b and miR156e had at least three targets. Analysis of the expression patterns of miRNAs and their corresponding target lncRNAs, the results showed as Fig. 4B, all of the target lncRNAs exhibited the opposite expression trend with their corresponding miRNAs, indicates the expression patterns of lncRNAs are indeed regulated by those miRNAs (Fig. S1).

Fig. 4.

Modes of direct post-transcriptional interaction among lncRNAs and miRNAs. Notes: (A) Some lncRNAs are degraded by miRNAs, as described for TCONS00144482 (degraded by miR156a), TCONS00024234 (by miR172b), TCONS00038296 (by miR172b); indicates the alignment between the miRNA and target lncRNA; (B) The expression patterns of miRNAs and their target lncRNAs; (C) The secondary structure of miR172b precursor sequence; (D) The secondary structure of TCONS00181308, and the result indicate that lncRNAs also produce microRNAs; (E) The expression patterns of miRNAs and their precursor lncRNAs

On the other hand, some lncRNAs were identified as endogenous target mimics (eTMs) of miRNAs, previous studies have proved that lncRNA act as eTMs could regulate miRNA functions through blocking the interaction between miRNA and its authentic target. In this research, psRobot (http://omicslab.genetics.ac.cn/psRobot) was used to predict eTMs of miRNAs, and parameters set as follows: maximal number of permitted gaps: 3, position after which with gaps permitted: 11. The result found that TCONS_00079589 were predicted to be the potential eTM of miR3627, and the expression pattern of the eTM is consistent with the miRNA authentic target C88C12H4G092610 (Fig. 5A-C). In addition, we also found TCONS_00099412, TCONS_00166633 could act as eTMs of miR172b and miR482b respectively, moreover, these transcripts showed differential expression in the Atlantic and Huansong 66 tubers, indicate that they may jointly participate in the synthesis of anthocyanins. Interestingly, we discovered that miR172b can interact with lncRNAs in different ways and participate in regulating the functional mode of lncRNAs.

Fig. 5.

Functional prediction of potato lncRNAs as miRNA target mimics. Notes: (A) Predicted base-pairing interaction of miR3627-TCONS 00079589; (B) Predicted base-pairing interaction of miR3627 with its target gene C88C12H4G092610; (C) The expression patterns of miRNAs, target gene and target mimics

Based on the above analysis, we constructed an lncRNA-mRNA-miRNA interaction network. Among them, TCONS_00195762 and TCONS_00181309 serve as the precursor sequence of miR172b, TCONS_00024233 acts as the endogenous target mimic of miR172b, while some lncRNAs, such as TCONS_00191232, TCONS_00218960, TCONS_00223307 and so on, are also the target genes of miR172b (Fig. 6). The co-expressed or co-localized target genes of these LncRNAs are mostly transcription factors of the MYB, bHLH and WDR families, which involved in anthocyanin synthesis regulation. Moreover, an interaction exists between miR156a and miR172b (Fig. 6), the target gene StSPL9 of miR156 is also targeted by TCONS_00021404. TCONS_00021404 was showed higher expression patterns in Huasong 66 rather than in Atlantic, and its target genes StUFGT (C88_C10H3G081620) and StF3’5’H (C88_C11H1G019590), which encode flavonoid 3’,5’-hydroxylase and anthocyanidin 3-O-glucosyltransferase, also showed the same expression patterns. This regulatory network revealed that lncRNAs function as precursor sequences, target genes, or endogenous target mimics of miRNAs. Additionally, the target genes of both miRNAs and lncRNAs, including various transcription factors or anthocyanin biosynthesis structural genes, have been demonstrated to regulate anthocyanin synthesis in different crops. These results fully demonstrate that the lncRNA-mRNA-miRNA interaction network may synergistically regulate anthocyanin biosynthesis in potato.

Fig. 6.

A co-regulatory network of lncRNA-mRNA-miRNA involved in anthocyanin synthesis. Notes: The circles represent protein coding genes participated in the regulation of anthocyanin synthesis, magenta circles represent transcription factors, purple circles represent structural genes; The green triangles represent the lncRNAs which targeted by miRNAs, the blue triangles represent the lncRNAs act as the precursor sequence of miRNAs, and orange triangle is a endogenous target mimics of miR172b; The pink diamonds represent the different expression miRNAs between white and purple potatoes

Identification and verification of anthocyanin synthesis related lncRNAs

To identify lncRNAs which participate in anthocyanin synthesis in potato, the expression levels between each development stage and sample were calculated and compared using FPKM. A total of 9,156 lncRNAs were significantly differentially expressed (Fig. 7A). Among them, there were 7,390 ones showed different expression patterns between the same development stages of the yellow-skinned and white-fleshed potato Atlantic and purple-skinned and purple-fleshed potato Huasong 66. During the period of tuber formation, 3,259 of 4,665 lncRNAs were up-regulated in H1 relative to D1, and the other 1,406 lncRNAs were down-regulated (Fig. 7B). Similarity, our results indicated that there were 3,385 lncRNAs showed different expression patterns between the period of tuber bulking (H2 vs. D2), and approximately 59% (1,993/3,385) lncRNAs presented a significantly up-regulated pattern while only about 41% lncRNAs were significantly down-regulated in H2 relative to D2 (Fig. 7C). When the tubers reached maturity, a total of 4,613 different expression lncRNAs (DElncRNAs) were found betweent Atlantic and Huasong 66 (H3 vs. D3) (Fig. 7D). To validate the reliability of the RNA-sequencing results and whether these differentially expressed lncRNAs are engaged in potato tubers, 9 lncRNAs were randomly selected and detected using qRT-PCR. Overall, quantitative RT-PCR measurements were generally in agreement with RNA sequencing (R² = 0.7473; Fig. 8A), although the relative expression levels of all lncRNAs by RNA-Seq were greater than those by qRT-PCR (Fig. 8B), providing further proof that the prediction accuracy was sufficient. Among the 7,390 different expression lncRNAs, 1,831 ones were showed differential expression patterns simultaneous across the different tuber development periods between Atlantic and Huasong 66, either at the formation, bulking or the maturity stage (Fig. 7EF). In addition, 108 of 1,831 lncRNAs were present higher expression patterns with FPKM ≥ 10 (Table S4). Considering the higher expression abundance, and both differential expression patterns, further predicting that these lncRNAs are likely related to anthocyanin metabolism and play important roles in anthocyanin synthesis in potato.

Fig. 7.

Identification of lncRNAs responding to anthocyanin synthesis. Notes: (A) The statistical results of the number of differentially expressed genes between different samples; (B) The volcano plot of DElncRNAs between the white potato and purple potato tubers during the formation period; (C) The volcano plot of DElncRNAs between the white potato and purple potato tubers during the bulking period; (D) The volcano plot of DElncRNAs between the white potato and purple potato tubers during the maturity period; (E) Venn diagram of DElncRNAs under three development stages; (F) Expression pattern of 1,831 lncRNAs which common differentially expressed in the formation, bulking and maturity stages between Huasong66 and Atlantic

Fig. 8.

Confirmation of the expression patterns of lncRNAs using quantitative RT-PCR. Notes: (A) Correlation of the lncRNAs expression profile between the results of sequencing and qRT-PCR; (B) The expression pattern of lncRNAs under different materials and stages. The values shown are the means±standard deviation of three replicates. StACTIN was used as the reference gene; “*” and “**” represents the significant differences at P < 0.05 and P < 0.01 levels, respectively

Functional analysis of LncRNA targets and screening candidate lncRNAs related to anthocyanin formation

Due to the lower conservation of lncRNAs between different plant species, up to now, there are few functionally characterized about lncRNAs in potato. Recent researches showed that lncRNAs could regulate the expression pattern of adjacent genes or overlapping genes. To investigate the possible functions of potato lncRNAs more clearly, we predicted the potential targets of lncRNAs in cis and trans. We screened known protein-coding genes located within 100 kb upstream and downstream of 1,831 different expressed lncRNAs and collected the trans target based on their expression correlation coefficient. A total of 40,478 associated target protein-coding genes were identified, and most lncRNAs had at least twenty target genes. Expression correlation analysis indicated that lncRNAs were present diverse correlations with their corresponding target protein-coding genes, including positively (Fig. S2, I, III, IV, VI) and negatively (Fig. S2, II, V). Gene Ontology (GO) analysis of targets indicated that 1,649 terms were mainly involved in the biological process, cellular components and molecular function, include catechol oxidase activity, photosynthetic membrane, cellular glucan metabolic process, cellular polysaccharide metabolic process, L-phenylalanine biosynthetic process, flavin-containing compound metabolic process, and sulfur compound binding (Fig. 9A). These results demonstrated that lncRNAs may play an critical role in the tuber development and flavonoid and anthocyanin synthesis. KEGG pathway analysis showed that these target genes were enriched in citrate cycle (TCA cycle), phenylalanine metabolism, anthocyanin biosynthesis, flavone and flavonol biosynthesis, phenylalanine, tyrosine and tryptophan biosynthesis (Fig. 9B). Further through analyzing the differential expression patterns between H1 vs. D1, H2 vs. D2, H3 vs. D3, screen the genes involved in the pathways related to anthocyanin synthesis, and select genes with expression levels greater than 1 (FPKM > 1), we finally identified 120 genes that might be related to the synthesis of anthocyanins (Fig. 9C). These 120 genes are regulated by 30 lncRNAs through co-localization, including TCONS00021404, TCONS00046283, TCONS00116538, TCONS00116538 and so on (Fig. 9D). Among them, TCONS00021404 showed higher expression patterns in Huasong 66 rather than in Atlantic, not only during the tuber formation and bulking stages, but also during the maturity stage (Figs. 8B and 9D). Interestingly, the target genes of TCONS00021404, include C88_C10H3G081620 and C88_C11H1G019590, also showed similar expression pattern with TCONS00021404, and they are encode anthocyanidin 3-O-glucosyltransferase and flavonoid 3’,5’-hydroxylase respectively, and participated in flavone and flavonol biosynthesis, anthocyanin biosynthesis process (Fig. S3). These results demonstrated that TCONS00021404, C88_C11H1G019590 and C88_C10H3G081620 synergistically controlled anthocyanin synthesized by participating in flavonoid biosynthesis (Fig. S4-S5).

Fig. 9.

Functional analysis of target genes of differentially expressed lncRNAs. Notes: (A) Gene Ontology analysis of the lncRNA target genes, red letters represent the GO terms related to anthocyanin synthesis; (B) KEGG pathway classification of DElncRNA correlated differentially expressed genes. The y-axis corresponds to KEGG pathway with a q-value ≤ 0.05, and the x-axis shows the enrichment ratio between the number of DEGs and all unigenes enriched in a particular pathway. The color of the dot represents q value, and the size of the dot represents the number of DEGs mapped to the reference pathways. C, D The expression pattern of candidate lncRNAs (C) and their target genes (D) during different stages in white and purple potatoes, the red dotted line represents the targeting relationship of TCONS00021404 and its target genes, C88_C11H1G019590 and C88_C10H3G081620

Differentially expressed genes regulate anthocyanin synthesis by participating in the flavonoid biosynthesis process

Flavonoids are a major class of plant secondary metabolites that serves a multitude of functions including pigments and antioxidant activity. Flavonoids are synthesized from phenylpropanoid derivatives by condensation with malonyl-CoA. After a series of modifications yield a variety of structural forms including chalcones, flavanones, dihyroflavonols, flavans, anthocyanins, flavones and flavonols, and isoflavonoids. In our study, we found a lot of different expression lncRNAs and their inherent target genes involved in flavone and flavonol biosymthesis, and further regulated anthocyanin biosynthesis. At the initial stage of anthocyanin synthesis, TCONS_00016186-C88_C06H4G129910 (C4H2) / C88_C06H4G114900 (C4H1), TCONS_00102221-C88_C09H4G115180 (CHS) / C88_C09H1G027380 (CHS) and other lncRNA-genes interact model catalyzed the synthesis of dihydrokaempferol, dihydroquercetin, and dihydromyricetin from naringenin, eriodictyol and dihydrotricetin respectively. Then, these substances are regulated by various genes to form flavonoids, or they directly synthesize anthocyanins. In these processes, TCONS_00259508-C88_C08H3G075390 (Leucoanthocyanidin dioxygenase), TCONS_00139884-C88_C11H2G045540 (Flavonoid 3’, 5’-hydroxylase), TCONS_00139885-C88_C08H1G021710 (Leucoanthocyanidin dioxygenase), TCONS_00102221-C88_C11H3G060800 (Flavonol synthase/flavanone 3-hydroxylase), TCONS_00219250-C88_C08H4G101770 (Leucoanthocyanidin dioxygenase), TCONS_00021404-C88_C11H1G019590 (Flavonoid 3’, 5’-hydroxylase) and TCONS_00265446-C88_C02H3G088330 (Dihydroflavonol 4-reductase) play important roles. Moreover, these genes present significant differential expression pattern between white and purple tuber potatoes, and the corresponding lncRNAs (TCONS_00102221-C88_C11H3G060800, TCONS_00259508-C88_C08H3G075390, TCONS_00021404-C88_C11H1G019590) also showed similar expression pattern and maybe directly participate in the regulation of anthocyanin synthesis too (Fig. 10).

Fig. 10.

Expression levels of key lncRNAs and genes involved in potato anthocyanin synthesis pathway

Discussion

LncRNA identification and analysis during white and purple potatoes

The potato ranks as the fourth most significant staple crop globally. The tuber of the potato is rich in starch, protein, fat, and crude fiber, and it contains an abundance of essential minerals such as calcium, phosphorus, iron, and potassium, along with various vitamins. Potato is broadly categorized into two main types: conventional potato and colored potato. Conventional potatoes typically exhibit white or yellow flesh, whereas colored potatoes are characterized by their deep purple, violet, or red flesh. Colored potatoes not only encompass all the nutrients found in conventional potatoes but are also particularly rich in anthocyanins [48]. Given their extensive potential applications, research on anthocyanins in colored potatoes has emerged as a prominent area of interest in recent years.

The importance of lncRNAs involved in anthocyanin synthesis has been extensively described in some plant species, including apple [49–51], strawberry [52, 53], capsicum annuum [45] and melastoma candidum [54]. However, to date, there is no lncRNA related to potato anthocyanin synthesis has been reported. With the potato genome sequences publication and the RNA-seq data accumulation, make genome-wide identification and analysis of lncRNAs in potato feasible. In this study, we mainly focused on the contribution of lncRNA in anthocyanin synthesis during three stages of white potato Atlantic (D1, D2, D3) and purple potato Huasong 66 (H1, H2, H3). A total of 15,191 lncRNA loci were identified by analysing 220.28 Gb clean data based on strict criteria. Although the strict criteria pipeline for identification of potato lncRNAs in present research is similar to that used in previous studies [55, 56], there are several unique ways to predict lncRNAs. (I) Three computational approaches include CPC (Coding Potential Calculator), CNCI (Coding-Non-Coding Index) and CPAT (Coding-Potential Assessment Tool) were combined to sort non-protein coding RNA candidates from putative protein-coding RNAs in the unknown transcripts. Only the transcripts with scores < 0 in all three approaches were remained [39–41]. (II) In our study, the transcripts which with the open reading frame (ORF) encoding more than 100 amino acids were exclude. However, some previously studies proved that lncRNAs have the potential to encode peptides more than 100 amino acids, such as HOTAIR with 106 amino acids, XIST with 136 amino acids and KCNQ1OT1 with 289 amino acids [57]. These type of lncRNAs could not meet the relatively strict criteria used in this work and were filtered out in the screening. (III) We required lncRNA abundance estimate greater than 0.1 FPKM in at least one of the samples [58]. Although some lncRNAs might be missed due to the strict screening criteria, all potato lncRNAs provided in this study is robust and reliable. During these 15,191 lncRNAs, according to the locations relative to the nearest protein coding genes, we observed there are 8,124 lncRNAs are lincRNAs, and the rest 1,959, 2,302 and 2,806 lncRNAs are divide into antisense-lncRNA, intronic-lncRNA and sense overlapping respectively.

The characteristic of distribution of transcript length, exon length and exon number is similar with the lncRNAs in cotton flower bud [58]. Potato lncRNAs were have lower densities in the pericentromeric heterochromatin regions than in the euchromatin, this phenomenon is similar with the result in citrus and rice lncRNAs [15, 59]. But it is inconsistent with a model microorganism Chlamydomonas reinhardtii, which lncRNAs distributed evenly across chromosomes [60]. LncRNAs are considered poorly conservation in various species. In the present study, there is lower conservation among distantly related species, such as potato and maize, potato and rice, and potato and Homo sapiens. Besides the poorly conservation relative to mRNA, lncRNAs are with shorter transcripts length and fewer exons than mRNAs [61]. In addition, lncRNAs exhibit lower levels of expression and higher natural expression variation in contrast with mRNAs [12].

LncRNAs are involved in the anthocyanin synthesis in potato

The mechanism of anthocyanin synthesis and its regulation has been elucidated in many plant species, and it represents a branch of the flavonoid pathway [5]. The catalytic enzymes include those involved in early anthocyanin biosynthesis, such as chalcone synthase (CHS), chalcone isomerase (CHI), flavanone 3-hydroxylase (F3H), and flavonoid 3’-hydroxylase, and late biosynthesis, such as dihydroflavonol 4-reductase (DFR), anthocyanidin synthase (ANS), and flavonoid 3-O-glycosyl-transferase (UFGT) [5]. Although the research about lncRNAs function in the regulation of development in plants is far behind in mammals, there are some plant lncRNAs have been proved to play important roles in regulating anthocyanin synthesis [53, 55, 57]. LncRNAs are considered highly tissue-specific in various species [62], especially in the colored tissues, it is thought that contained more tissue-specific lncRNAs than other tissues. This characteristic of lncRNAs might imply that they could function in organs colored development, further exploration of those lncRNAs may achieve a breakthrough leading to a better understanding of the molecular control of anthocyanin synthesis in plants. In our study, we found total of 1,831 lncRNAs were simultaneous differentially expressed in three developed stages of tuber between the white potato and purple potato. These lncRNAs showed stable and differential expression patterns in tissues of different colors, suggesting that they may be related to color formation. LncRNAs can regulate the nearby PCgenes expression either in a cis or trans manner [62]. Functional analysis of their target genes indicated that most of them participated in L-phenylalanine biosynthetic process, flavin-containing compound metabolic process, phenylalanine metabolism, anthocyanin biosynthesis, flavone and flavonol biosynthesis, phenylalanine, tyrosine and tryptophan biosynthesis and other biological processes related to anthocyanin occurred. Previous study have demonstrated that genes involved in flavone and flavonol biosynthesis in strawberry [54] and Capsicum annuum [56] are closely related to anthocyanin synthesis. And these genes regulated by 214 lncRNAs which exhibited significant differential expression patterns in white and purple potatoes through co-expression (Fig. S6), may simultaneously affect the synthesis of anthocyanins.

Recently, lncRNAs were demonstrated to be acted as a endogenous target mimics (eTMs) involved in miRNA regulation in Arabidopsis [63–65]. Several lncRNAs include IPS1, ath-eTM160-1, osa-eTM160-3, and nta-eTMX27 have been proved play an effective inhibition role in the corresponding miRNA activities in A. thaliana and tobacco [63–64, 66]. miR858 targeted MsMYB62-like and negatively regulated its expression. Overexpression of miR858 significantly increased the anthocyanin content of crabapple flowers under low nitrogen conditions. Further studies showed that MsMYB62-like inhibited the expression of MsF3’H, which was the main reason for the significant increase in anthocyanin content after overexpressing miR858 [50]. In addition, this study demonstrated that two lncRNAs, eTM858-1 and eTM858-2, can act as endogenous target mimics of miR858, which bind to miR858 and prevent MsMYB62-like from cleavage, thereby negatively regulating anthocyanin biosynthesis [50]. In our previous research, we have analyzed the correlation between miRNAs and anthocyanin synthesis in potato (Unpublished). Here, we predicted the eTMs of miRNAs in potato tubers, and identified some potential eTMs for different miRNAs. miR172 was found to be well conserved among different crops, and the target gene AP2 has been proved involved in flavonoid and anthocyanin metabolism in many species [67–69]. TCONS_00099412 was acted as eTMs of miR172, and they were also present different expression patterns between white potato and purple potato. So we speculated that TCONS_00099412 participated in anthocyanin synthesis through inhibitory ability on miR172 regulating their targets. In addition, we also found part of lncRNAs could act as targets or precursors of miR172, and they may jointly regulate the metabolism of anthocyanins.

In summary, combined with genome-wide screening, tissue-specific identified, different expression pattern compared, miRNA-lncRNA interaction and functional analyzed, we identification of a set of lncRNAs which may reveal a more general mechanism for the regulation of the anthocyanin metabolism process in potato.

Anthocyanidin 3-O-glucosyltransferase (UFGT) and Flavonoid 3’,5’-hydroxylase (F3’5’H) in tuber is essential to anthocyanin synthesis in potato

Flavonoids, including anthocyanins, flavones and flavonols, are important secondary metabolites present throughout the plant kingdom, involved in a wide range of biological functions, such as defence against pathogen attack, legume nodulation [70]. The biosynthesis of anthocyanins in plants is the largest subclass of the flavonoid metabolism pathway [5]. Two kinds of structural genes participate in the flavonoid pathway have been extensively studied in several plants. One encodes UDP-glucose: flavonoid 3-O-glucosyltransferase (UFGT), which catalyzes the transfer of the glucosyl moiety from UDP-glucose to the 3-hydroxyl group of anthocyanidins and this glycosylation step is the key for their stability and water solubility [71]. The other encodes UDP rhamnose anthocyanidin 3-O-glucoside rhamnosyltransferase (RT) that adds a rhamnose to the 3-O-bound glucose of the anthocyanidin 3-O-glucoside molecule to produce the anthocyanidin 3-O-rutinosides, catalyzes the glucosylation at the 3-hydroxy position of the anthocyanidin molecule and produces the first stable colored metabolites in anthocyanin biosynthesis [72]. It is the first anthocyanidin-modifying enzyme ever described and has been isolated from Petunia hybrida (petunia). The expression of the gene coding for UFGT was consistently associated with pigmented anthocyanins in plants. Comparison of biochemical and expression data revealed a positive correlation between the amount of anthocyanins in fruit peel and leaves and the UFGT1 expression level in pepper [73]. In S. tuberosum, the expression profiling analysis revealed that StUFGT expression was induced correspondingly by gibberellic acid and sucrose, suggesting that UFGT might play a regulatory role in anthocyanin biosynthesis at the transcriptional level [74]. In order to reveal the mechanism underlying the formation of its picotee bicolor pattern in the petal, the results showed that the defect of anthocyanidin 3-O-glucosyltransferases (UFGTs) accumulation resulted in the failure of the glycosylation of anthocyanidin is the main reason [75]. Expression analysis of UFGT genes in white grape cultivars and red-skinned sports revealed that UFGT gene was present, but not expressed in the white cultivars [76]. Flavonoid 3’,5’-hydroxylases (F3’5’Hs) play a key role in biosynthesis of blue coloured anthocyanin complexes in plants. In the flavonoid biosynthesis pathway, the flavonoid 3’ -hydroxylase (F3’H) and F3’5’H are generally known to catalyze the hydroxylation of the B ring leading to the biosynthesis of hydroxylated flavonoids [77]. In safflower, the constitutive expression of the CtF3’5’H1 gene was validated at different flowering stages, indicating their diverse transcriptional regulation through flower development. In addition, overexpressed CtF3’5’H1 in Arabidopsis, the transgenic plants exhibit the significantly increased anthocyanins and flavonoid content [78]. In tea plants, expression analysis found that the expression of F3’5’H was highly correlated with epigallocatechin gallate accumulation [79]. The results of Cao et al. [80] found that overexpression of MrF3’5’H from Morella rubra in tomato resulted in high accumulation of flavonols in fruit up to 3.2 g kg^{− 1} by upregulating expression of genes involved in the phenylpropanoid pathway including SlPAL, SlC4H, Sl4CL, SlCHS1, SlCHI-like, SlF3H, and SlFLS. In our study, C88_C10H3G081620 and C88_C11H1G019590, which encode anthocyanidin 3-O-glucosyltransferase (UFGT) and flavonoid 3’,5’-hydroxylase respectively, were showed higher expression patterns in purple potato Huasong 66 rather than in white potato Atlantic, not only during the tuber formation and bulking stages, but also during the maturity stage. Interestingly, a long non-coding RNA TCONS00021404 also showed similar expression pattern, and it target to C88_C11H1G019590 and C88_C10H3G081620. Comprehensive all results, TCONS00021404 and two genes participated in flavonoid metabolism pathway, may be co-regulated the anthocyanin synthesis in potato.

Conclusions

Through comparing the lncRNA expression patterns during the three development stage tubers, include formation, bulking and the maturity, of Atlantic (D1, D2, D3) and Huasong66 (H1, H2, H3). we found 990 lncRNAs were preferentially expressed in the Huasong 66 and 679 lncRNAs were preferentially expressed in the Atlantic. A total of 1,831 lncRNAs were common differential expression across three tuber development periods between Atlantic and Huasong 66. And partly of their target genes mainly participate in phenylalanine metabolism, anthocyanin biosynthesis, flavone and flavonol biosynthesis, phenylalanine, tyrosine and tryptophan biosynthesis. LncRNA TCONS00021404 and its target genes, C88_C11H1G019590 and C88_C10H3G081620, which encode flavonoid 3’,5’-hydroxylase and anthocyanidin 3-O-glucosyltransferase, were all showed higher expression patterns in Huasong 66 rather than in Atlantic, not only during the tuber formation and enlargement stages, but also during the maturity stage. This result provides a solid foundation for further research on the synthesis mechanism of anthocyanins in potatoes, and is helpful to create the potatoes with high anthocyanin content.

Authors’ contributions

YM conceived and designed the research. HN and RC performed the experiments. DZ (IMAU), DZ, JM, XW, YM(Yu MA), CS, EY and NL analysed the data and wrote the manuscript. All of the authors read and approved the final manuscript.

Funding

This study was supported by grants from National Natural Science Foundation of China (32560141), Natural Science Foundation of Inner Mongolia Autonomous Region of China (2024QN03070), Research Initiation Project for Introduction of High-level/Excellent Doctoral Talents of IMAU (NDYB2022-48) to H NIE. The funders had no role in the design of the study; the collection, analysis, and interpretation of the data; or the writing of the manuscript.

Data availability

All the sequencing data have been uploaded to the NCBI database with the accession number PRJNA1440173.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.