Arabidopsis thaliana ambient temperature responsive lncRNAs
Long non-coding RNAs (lncRNAs) have emerged as new class of regulatory molecules in animals where they regulate gene expression at transcriptional and post-transcriptional level. Recent studies also identified lncRNAs in plant genomes, revealing a new level of transcriptional complexity in plants. Thousands of lncRNAs have been predicted in the Arabidopsis thaliana genome, but only a few have been studied in depth.
Here we report the identification of Arabidopsis lncRNAs that are expressed during the vegetative stage of development in either the shoot apical meristem or in leaves. We found that hundreds of lncRNAs are expressed in these tissues, of which 50 show differential expression upon an increase in ambient temperature. One of these lncRNAs, FLINC, is down-regulated at higher ambient temperature and affects ambient temperature-mediated flowering in Arabidopsis.
A number of ambient temperature responsive lncRNAs were identified with potential roles in the regulation of temperature-dependent developmental changes, such as the transition from the vegetative to the reproductive (flowering) phase. The challenge for the future is to characterize the biological function and molecular mode of action of the large number of ambient temperature-regulated lncRNAs that have been identified in this study.
KeywordsLong non-coding RNA (lncRNA) Ambient temperature response Flowering time Arabidopsis thaliana FLINC
ABRE BINDING FACTOR 4
CYTOCHROME P450 FAMILY 78, SUBFAMILY A, POLYPEPTIDE 8
days after sowing
EARLY FLOWERING MYB PROTEIN
EARLY FLOWERING IN SHORT DAYS
FLAVIN-BINDING, KELCH REPEAT, F Box 1
FLOWERING LOCUS C
FLOWERING LOCUS M
long intergenic RNAs
Long non-coding RNAs
MADS AFFECTING FLOWERING
natural antisense transcripts
Next Generation Sequence.
PHYTOCHROME INTERACTING FACTOR 4
Plant long non-coding RNA database
HOMOLOGUE OF CYANOBACTERIAL RBCX 1
rosette leaf number
SUPPRESSOR OF OVEREXPRESSION OF CO 1
SHORT VEGETATIVE PHASE
TERMINAL FLOWER 1
LncRNAs represent a new class of recently discovered regulatory molecules. A key role for lncRNAs in the transcriptional and post-transcriptional regulation of a plethora of biological processes is emerging. LncRNAs are defined as long RNA molecules (longer than 200 bp) that do not code for a protein. They have been divided in different categories based on their location in the genome relative to protein-coding genes: natural antisense transcripts (NATs) are lncRNAs transcribed on the opposite strand of a protein-coding gene; long intergenic RNAs (lincRNAs) are transcribed in intergenic regions; intronic lncRNAs (iLncRNA) are transcribed from an intron in a protein-coding gene and; promoter lncRNAs (pLncRNAs), are transcribed from the promoter region of a gene [1, 2].
In the past years, genome-wide studies have identified many lncRNAs in the Arabidopsis thaliana (Arabidopsis) genome, as well as in other plant species, including rice and maize ; however the biological function of most of these lncRNAs is not known . Currently, only a small number of lncRNAs has been functionally characterized and the emerging picture shows that lncRNAs have several modes of action, ranging from transcriptional interference to chromatin looping and mRNA splicing [1, 2, 4]. Moreover, these lncRNAs act in a wide range of biological processes. Some examples include the lincRNA APOLO, which is involved in auxin response , ELENA1, which functions in plant immunity , Enod40, which plays a role in regulation of sucrose utilization in nodules  and IPS1, which acts as a mimic of a miRNA target in the control of Pi homeostasis .
Vernalization, a long period of cold that is required by many temperate plant species to make the transition from the vegetative to the reproductive phase, is regulated by lncRNAs. The first characterized lncRNAs in plant were the lncRNAs transcribed from the FLOWERING LOCUS C (FLC) locus: the NAT COOLAIR , the iLncRNA COLDAIR , and more recently the pLncRNA COLDWRAP  and NAT ASL . These molecules regulate the epigenetic silencing of FLC during vernalization in response to winter cold. This, together with the fact that the key regulators of flowering time in the ambient temperature pathway, FLOWERING LOCUS M (FLM) [13, 14], MADS AFFECTING FLOWERING 2 (MAF2) , MAF3, MAF4 and MAF5 , all belong to the FLC clade , prompted us to speculate a possible role for lncRNAs in ambient temperature-mediated development.
How plants are able to respond in a fast and dynamic manner to continuous changes in ambient temperature is a long-standing question in plant biology research. It is also important, in light of global warming, to understand how temperature affects plant development, as temperature influences different aspects of plant biology and development, including hypocotyl elongation, leaf shape determination, flowering time control and pathogen responses [18, 19]. Current research on the role of ambient temperature in plant development is mainly focused on identification and characterization of protein coding genes involved in this process. For example, MADS-box transcription factor genes, the bHLH transcription factor PHYTOCHROME INTERACTING FACTOR 4 (PIF4) [20, 21, 22, 23], the TEMPRANILLO genes  and the MYB transcription factor EARLY FLOWERING MYB PROTEIN (EFM)  have been shown to regulate ambient temperature-mediated flowering. PIF4, together with the photoreceptor phytochrome B (phyB), also plays an important role in thermo-morphogenesis [26, 27, 28]. Furthermore, there is an important role in ambient temperature-mediated responses for various circadian clock genes [29, 30, 31].
Despite the clear role for lncRNAs in cold temperature signaling, their potential role as mediators of ambient temperature responses remains unexplored.
Identification of lncRNAs in Arabidopsis thaliana.
We used RNA-seq transcriptome analysis to identify lncRNAs in shoot apical meristem (SAM)-enriched tissue that might play a role in ambient temperature-mediated developmental processes. We focused on transcripts carrying a polyA tail and followed a TAIR10-guided assembly approach using cufflinks. This approach allowed the identification of both known lncRNAs and annotation of novel transcripts . Strand-specific libraries were generated from shoot apical meristem (SAM)-enriched tissue of 5 weeks-old Arabidopsis Col-0 plants grown in short-day conditions, and therefore still in the vegetative stage. The SAM of plants grown in short-day conditions is larger than the SAM of those grown in long-day conditions , allowing the presence of a good proportion of SAM tissue in the sampled material. Using this approach we aimed to identify low-expressed transcripts in the SAM. To annotate the transcripts, the “cuffcompare” function of cufflinks was implemented, which classifies transcripts base on their relationship with the TAIR10 reference genome annotation. A similar approach was recently used successfully for the identification of lncRNAs in the tomato genome .
Natural antisense transcript
long intergenic ncRNA
Only a few lncRNAs are annotated in the TAIR10 annotation. Therefore, we retrieved an extensive list of 14,954 RNAs annotated as lncRNAs from the Plant long non-coding RNA database (PlncDB)  and compared these with our identified putative lncRNA transcripts. In total 644 of the 2132 lncRNA identified in our annotation were also annotated in the PlncDB, 22 iLncRNA, 92 NATs and 530 lincRNAs (Additional file 1: Table S1).
Next, we compared the lncRNAs identified in our analysis with the ones reported in the CANTATAdb 2.0 . We found 321 transcripts to be also present ion the CANTATA database, 10 iLncRNA, 38 NATs and 273 lincRNAs (Additional file 1: Table S1).
A new annotation of the Arabidopsis genome, Araport11, was released recently . This new annotation is based on the analysis of a collection of RNA-seq experiments performed on different tissues from plants grown under different conditions. This approach allows us to identify new lncRNA transcripts expressed in specific conditions. We investigated the overlap between our lncRNA annotation and Araport11 transcripts and found 457 lncRNAs in both data sets (Additional file 1: Table S1).
Remarkably, 1323 lncRNAs were uniquely identified in our analysis, highlighting the power of our approach in isolating potentially new lncRNAs. Only a small number of these are likely to be falsely assigned due to sequencing or annotation errors. The identification of this relatively large number of novel transcripts is likely due to the fact that lncRNA are generally lowly expressed and therefore, their identification is dependent on the experimental setup used and tissues sampled for the RNA isolation.
Because one of the key developmental processes regulated by ambient temperature is flowering time control, we initially compared the genomic location of the lncRNAs in relation to the localization of key flowering time genes. A thorough investigation revealed that 22 lncRNAs are located nearby a flowering time gene , such as the lincRNA AtLnc1134 located in ABRE BINDING FACTOR 4 (ABF4), the AtLnc1860, AtLnc198 and AtLnc236, located respectively in the promoters of MADS AFFECTING FLOWERING 2 [15, 39, 40], FLAVIN-BINDING, KELCH REPEAT, F Box 1(FKF1) [41, 42] and EARLY FLOWERING IN SHORT DAYS (EFS)  (Additional file 2: Table S2).
LncRNA expression is modulated by ambient temperature
Subsequently, co-regulation was investigated for NATs and lincRNAs (Additional file 4: Table S4). In the case of NATs we compared their expression level to the expression of their corresponding sense transcript while for the lincRNAs to their two direct flanking protein coding genes. In general, we observed a positive Pearson correlation between sense and antisense transcript and we found the expression of the sense transcript was either not changed or changed in the same direction as the NAT (Additional file 4: Table S4). We also observed correlation (with prevalence of positive correlation) for most of the lincRNAs with at least one of the flanking genes, suggesting co-regulation between the lncRNAs and the protein coding genes located nearby (Additional file 4: Table S4). For example, AtLnc1444 showed a positive correlation with both the two direct flanking protein coding genes, C-REPEAT/DRE BINDING FACTOR 1 (CBF1) and CBF3, genes involved in response to low temperature [48, 49, 50, 51], while AtLnc488 showed a negative correlation with the transcription factor TEOSINTE BRANCHED1/CYCLOIDEA/PCF 15 (TCP15) [52, 53].
AtLnc428 plays a role in temperature-mediated flowering
We selected five ambient temperature responsive lncRNA for further functional characterization. We focused on lncRNAs for which a T-DNA insertion line is available in the predicted transcript or promoter region. We hypothesized that the promoter T-DNA insertions will affect the lncRNA expression level or pattern. We investigated the role of the temperature responsive AtLnc2, AtLnc120, AtLnc213, AtLnc428 and AtLnc1524 in the regulation of temperature-mediated flowering. Expression of AtLnc2 and AtLnc120 is up-regulated upon temperature increase while expression of AtLnc213, AtLnc428 and AtLnc1524 is down-regulated (Additional file 3: Table S3).
FLINC is an intergenic lncRNA located in chromosome 1 between At1g56233, a gene coding for a defensin-like (DEFL) family protein and At1g56240, a gene coding for a phloem protein 2-B13. At1g56233 and At1g56240 are located respectively at 1.5 kb and 3.2 kb from FLINC (Additional file 6: Figure S2A). The effect of T-DNA insertion in FLINC on expression of FLINC was examined by qPCR by comparing wild-type and mutant plants in a segregating population (Additional file 6: Figure S2B). No FLINC expression was observed in plants with the T-DNA insertion, suggesting that the flinc mutant is a full knock-out. We also examined the effect of the FLINC T-DNA insertion on the expression of the genes flanking FLINC in all above ground tissues of 2 weeks-old plants in the vegetative stage of development. At1g56233 expression could not be detected in either wild type Col-0 or flinc plants, and we did not observe any significant change in the expression of At1g56240 (Additional file 6: Figure S2C). This result indicates that the T-DNA insertion does not affect the expression of the genes flanking FLINC in the investigated material.
Our results indicate a role for FLINC in the regulation of ambient temperature-mediated flowering. To confirm this observation, we generated plants overexpressing the lncRNA under control of the constitutive CaMV35S promoter and analyzed their flowering behavior at different ambient temperature conditions (Fig. 2d). In contrast to the flinc T-DNA insertion line, we found that plants overexpressing the lncRNA transcript are more sensitive to a temperature change than wild-type plants, and providing additional evidence for the role of FLINC in the control of ambient temperature-mediated flowering.
In our RNA-seq experiment, FLINC was down-regulated upon the temperature change under short day conditions (Additional file 3: Table S3). Because flowering time investigations were performed under long day conditions, we examined the effect of an increased ambient temperature on FLINC expression under these conditions. This analysis revealed a similar response upon an increase of temperature from 16 °C to 25 °C as was observed in the initial experiment under short day conditions (Additional file 6: Figure S2E), suggesting that regulation of FLINC expression by temperature is not day-length dependent.
Since a number of genes involved in the ambient temperature pathway are circadian-regulated or code for circadian clock components , we monitored the expression of FLINC in a 24 h-time course. We found that FLINC expression is not influenced by the circadian rhythm (Additional file 6: Figure S2F).
We also examined the FLINC expression profile during Arabidopsis development and in different Arabidopsis tissues (Additional file 6: Figure S2G and Figure S2H). We found that FLINC is expressed broadly in the plant, suggesting that FLINC plays a more extensive role during development than just the control of flowering.
To shed some light on the molecular basis of the flinc phenotype, we investigated the expression of selected flowering-related genes in wild type and 3 weeks-old flinc plants grown in long day conditions at 16 °C. These include the flowering repressors FLC and FLM, the florigen FLOWERING LOCUS T (FT) and the floral integrators APETALA 1 (AP1), SUPRESSOR OF COSTANS 1 (SOC1), LEAFY (LFY) and TERMINAL FLOWER 1 (TFL1) (Additional file 7: Figure S3A). We detected higher expression of FT in flinc than in the Col-0 wild type, in agreement with the early flowering phenotype of flinc (Fig. 2a).
LncRNAs have been show to exert their effect in numerous ways, including nucleotide pairing with mRNAs  of protein-coding genes or miRNAs. Therefore, we searched for regions homologous to FLINC within the Arabidopsis genome. We found ten regions with partial sequence homology to FLINC, of which five are located within protein coding genes. However, for none of these five genes differential expression was found in flinc in comparison to Col-0 wild type (Additional file 8: Table S5).
To obtain insight in conservation of FLINC, we searched for FLINC-related lncRNAs in genomes of other plant species using Blast and found various similar genes (e-value < 0.001), e.g. in the close Arabidopsis relatives Arabis alpina and Brassica rapa, but also in less-related species such as Solanum pennelli and Vitis vinifera (Additional file 7: Figure S3C). These results point to positive selection on FLINC. Further research is needed to confirm the proposed role of FLINC in ambient temperature-mediated flowering time control in Arabidopsis and to explore the function of the closely related genes in other species.
The discovery of lncRNAs as regulatory molecules revealed a new layer of complexity in gene expression regulation. So far, only few of the thousands of identified lncRNAs in plants have been functionally characterized. Nevertheless, these studies revealed that lncRNAs play a role in many different biological processes, from lateral root development to photomorphogenesis [1, 2]. We identified new Arabidopsis lncRNAs in SAM-enriched tissue and show ambient temperature-sensitive expression for a small subset of these lncRNAs. Furthermore, we showed that one of these lncRNAs, FLINC, regulates temperature-mediated flowering in Arabidopsis. Previous studies showed that expression of the lncRNAs transcribed from the FLC locus, COOLAIR , COLDAIR , ASL  and COLDWRAP , is regulated by a prolonged period of cold (4 °C). By contrast, FLINC expression is affected rapidly by a milder change of temperature. FLINC is an intergenic lncRNA and at this moment it is not clear which genes are regulated by FLINC. Thus, further research is needed to elucidate FLINC’s molecular mode of action. The observation that FT is higher expressed at low temperature in flinc mutant plants, suggests that FLINC acts prior to the integration of the environmental signals by FT. Next, the two transcription factors with a known role in temperature-mediated flowering time control, SOC1  and AGAMOUS-LIKE 15 (AGL15) , bind to the FLINC locus suggesting a regulatory interaction. Moreover, the presence of sequence conservation in different plant species points to positive selection and a potential conserved role for FLINC in ambient temperature-mediated flowering time control.
Advances in RNA sequencing technology allow a more precise identification of lncRNA transcripts. Thanks to the use of strand specific RNA-seq data, it is now possible to identify transcripts derived from the opposite stand and therefore, get insight into expression profiles of NATs. The current sequence depth that can be obtained facilitates identification of low abundant transcripts, including as numerous lncRNAs. LncRNA expression appeared to be heavily responsive to environmental conditions, including temperature (this study) or drought , suggesting a role for these molecules in plant adaptation to the environment. The tissue-specific and environmental regulation of lncRNAs illustrates the need for defined experiments to identify all lncRNA-encoding genes in the genome.
The discovery of lncRNA as regulatory molecules revealed a new layer of complexity in the regulation of gene expression, and functional studies in plants revealed roles for lncRNAs in various biological processes. Here we show that lncRNA expression is influenced by ambient temperature changes and identified a significant role in ambient temperature-mediated flowering time the FLINC lncRNA. The challenge for the future is to decipher the biological function and molecular mode of action of ambient temperature responsive lncRNAs.
RNA-seq data were generated previously by Pajoro et al., 2017 . In brief, plants were grown under short day conditions (8 h light, 16 h dark) on rock-wool in growth cabinets with LED lamps with light intensity of 200 μmol m-2 s-1 and 75% relative humidity at 16 °C for 5 weeks. Nutrients were supplied by sub-irrigation with Hyponex. After 5 weeks, plants were either left at the same temperature or moved to 25 °C (6 h after lights on). Three biological samples were generated for each condition. For each sample, tissue from ca. 10 plants was collected 4 hours after light on. Using jeweler’s forceps, leaves were removed to obtain SAM-enriched tissue. Total RNA was extracted using the Invitek InviTrap Spin Plant RNA Mini Kit (REF: 1064100300) according to the manufacturer’s protocol. DNase treatment was performed to remove genomic DNA. DNase I digestion was performed on total RNA using Turbo DNase from Ambion according to the manufacturer’s protocol. RNA integrity was checked on a by agarose gel electrophoresis after DNase I treatment. Samples were prepared for Illumina sequencing using the Illumina TruSeq Stranded mRNA Sample Prep kit (REF: 15032613) according to the manufacturer’s protocol, which generates polyadenylated transcripts. Libraries were analyzed on the Bioanalyzer and quantified using a Qubit fluorometer before pooling for sequencing on an Illumina HiSeq2500. Two lanes were used on a 125 bp paired-end (PE) run.
RNA-seq data analysis
RNA-seq reads were mapped against the Arabidopsis genome version TAIR10 (www.arabidopsis.org) using TopHat2 . To identify un-annotated transcripts, reference based full length transcript reconstruction was performed separately for each sample using Cufflinks with strand-specific awareness . Cuffmerge, which is part of the cufflinks package, was finally used for merging the individual cufflinks results into an overall set of full-length transcripts. The function “Cuffcompared” was used to compare the newly annotate transcripts with the TAIR10 reference annotation. We selected transcripts that belong to the class code “x”, “i” and “u”. The class code “x” defines “Exonic overlap with reference on the opposite strand”, to this class belong e.g. the Natural Antisense Transcripts (NAT). The class code “i” defines “transcribed genomic fragments falling entirely within a reference intron”, to this class belong the intronic lncRNAs (iLncRNA). Finally, the class code “u” defines the “unknown intergenic transcript”, to this class belong e.g. the long intergenic lncRNA (LincRNAs). To retrieve differentially expressed transcripts upon the temperature changes, the number of fragments mapping to the newly annotated transcripts was determined using HTseq count . Differentially expressed genes were detected using DESeq2 .
Flowering time assay
lncRNAs T-DNA insertion lines
Expression profile by qRT-PCR
Plant were grown under long day conditions (16 h light, 8 h dark) at 21 °C as described above. Three biological samples were generated for each condition tested, unless stated otherwise. For each sample, tissue from ca. 10 plants was collected using jeweler’s forceps. Total RNA was extracted and DNAse-treated as described above. cDNA was synthetized using the iScript kit from BIORAD (Cat. No. 1708890) using one μg of total RNA. Gene expression was measured by quantitative RT-PCR (qPCR) using the iQ™ SYBR® Green Supermix from BIORAD (Cat. No. 1708885). The qPCR primers are listed in Additional file 9: Table S6. Relative expression was calculated using TIP41 as reference gene for normalization and the formula 2-(ΔCttarget – ΔCtTIP41). Normalized fold expression was calculated by setting the wild-type (WT) value to one.
Generation of FLINC overexpression lines
Gateway cloning was used to generate the FLINC overexpression construct. The FLINC transcript was amplified by PCR from cDNA and cloned into the pDONR201 entry vector using a BP reaction. The entry vector was then recombined with the destination vector pB7FWG2 carrying the CaMV35S promoter. The resulting expression vector with the p35S:FLINC construct was transformed into Arabidopsis Col-0 plants by floral dip. Transformed plants were selected on media with 10 μg/ml of phosphinothricin.
Identification of protein coding genes with sequences homologous to FLINC
We used the BLASTN software  with default parameter to identify Arabidopsis protein coding genes with sequence similarity to FLINC. The genome coordinates that matched the FLINC sequence were compared to the coordinates of the Arabidopsis gene models using the Bedtools intersect software  and the gene coordinates as reference and –F 1 parameter. To identify sequences similar to FLINC in other plant species, we used Blastn software (NCBI)  with the parameter optimized for ‘somewhat similar sequences’. Only species with an e-value < 0.001 were considered.
The authors would like to thank Kim Boutilier for critical proofreading of the manuscript.
Availability of data and material
The datasets analysed during the current study are available in the Gene Expression Omnibus (GEO) database, Series GSE85281. https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE85281
This work was supported by an NWO grant (849.13.005) for the project ERACAPS13.012, FLOWPLAST. The RNA-seq experiments were supported by a ZonMw Enabling Technology Hotel project (40–43500–98-054). Funding body did not play any roles in the design of the study and collection, analysis, or interpretation of data neither in writing the manuscript.
PA designed the study, performed the experiments, analyzed the data and wrote the paper. ES and LF analized the data. SJ, VF, JB and MB preformed experiments. YZ and FB generated material. GCA and RGHI conceived the study and revised the manuscript. All the authors read and approved the final manuscript.
Ethics approval and consent to participate
Consent for publication
The authors declare that they have no competing interests.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
- 32.Trapnell C, Williams BA, Pertea G, Mortazavi A, Kwan G, van Baren MJ, Salzberg SL, Wold BJ, Pachter L. Transcript assembly and quantification by RNA-Seq reveals unannotated transcripts and isoform switching during cell differentiation. Nat Biotechnol. 2010;28(5):511–U174.CrossRefPubMedPubMedCentralGoogle Scholar
- 56.Immink RGH, Pose D, Ferrario S, Ott F, Kaufmann K, Valentim FL, de Folter S, van der Wal F, van Dijk ADJ, Schmid M, et al. Characterization of SOC1's central role in flowering by the identification of its upstream and downstream regulators. Plant Physiol. 2012;160(1):433–49.CrossRefPubMedPubMedCentralGoogle Scholar
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.