Discovery and comparative profiling of microRNAs in a sweet orange red-flesh mutant and its wild type
© Xu et al. 2010
Received: 10 November 2009
Accepted: 17 April 2010
Published: 17 April 2010
Skip to main content
© Xu et al. 2010
Received: 10 November 2009
Accepted: 17 April 2010
Published: 17 April 2010
Red-flesh fruit is absent from common sweet orange varieties, but is more preferred by consumers due to its visual attraction and nutritional properties. Our previous researches on a spontaneous red-flesh mutant revealed that the trait is caused by lycopene accumulation and is regulated by both transcriptional and post-transcriptional mechanisms. However, the knowledge on post-transcriptional regulation of lycopene accumulation in fruits is rather limited so far.
We used Illumina sequencing method to identify and quantitatively profile small RNAs on the red-flesh sweet orange mutant and its wild type. We identified 85 known miRNAs belonging to 48 families from sweet orange. Comparative profiling revealed that 51 known miRNAs exhibited significant expression differences between mutant (MT) and wild type (WT). We also identified 12 novel miRNAs by the presence of mature miRNAs and corresponding miRNA*s in the sRNA libraries. Comparative analysis showed that 9 novel miRNAs are differentially expressed between WT and MT. Target predictions of the 60 differential miRNAs resulted 418 target genes in sweet orange. GO and KEGG annotation revealed that high ranked miRNA-target genes are those implicated in transcription regulation, protein modification and photosynthesis. The expression profiles of target genes involved in carotenogenesis and photosynthesis were further confirmed to be complementary to the profiles of corresponding miRNAs in WT and MT.
This study comparatively characterized the miRNAomes between the red-flesh mutant and the wild type, the results lay a foundation for unraveling the miRNA-mediated molecular processes that regulate lycopene accumulation in the sweet orange red-flesh mutant.
Sweet orange (Citrus sinensis [L.] Osbeck), one of the most important fruit crops worldwide, is responsible for 75% of citrus production used both as fresh fruit and processed juice . Orange fruit is believed to contain many phytochemicals which are beneficial to human health . Carotenoids are a group of the health-promoting products which may serve as antioxidants, vitamin A precursors, and cancer-preventing effectors . Orange fruit is reported with more than 115 species of carotenoids , but lycopene are absent from common varieties [5–7]. Interest in lycopene metabolism and regulation is growing rapidly because a variety of epidemiological trials have suggested that higher intake of lycopene-containing fruits or foods are associated with decreased cardiovascular disease and prostate cancer risks . So far, three sweet orange mutants with lycopene accumulation in the fruits were reported: Shara , Cara Cara  and the recently reported 'Hong Anliu' . 'Hong Anliu' is a spontaneous bud mutant from its wild type 'Anliu' sweet orange. We found that lycopene in the 'Hong Anliu' mutant was 1000-fold higher than that in its wild type fruits, and in juice sacs the lycopene accumulation was coincided with increased expression of upstream carotenogenic genes and reduced expression of downstream genes . Thereafter, transcriptional study including cDNA microarray in combination with suppression subtraction hybridization (SSH) was used to investigate gene expression changes in the mutant, and a total of 267 differentially expressed genes were detected . More recently, massively parallel signature sequencing was applied to decipher the transcriptome changes between WT and MT; and the results indicated that partial impairment of lycopene downstream flux are critical for the formation of lycopene accumulation trait in the mutant . Meanwhile, comparative proteomic analyses were performed between the mutant and its wild type. And the proteomic data, when considered in combination with transcriptional data, suggest that post-transcriptional regulations are involved in shaping the red-flesh trait of the orange mutant . However, it remains largely unknown what kind of post-transcriptional mechanism was involved in the red-flesh mutant. So far, one interesting case has been reported in Arabidopsis where miRNA was proposed to regulate the expression of APETALA2 transcription factor (designated as RAP2.2) in controlling a key carotenoids biosynthesis gene (phytoene synthase gene) . But to date, very little is known about miRNAs or miRNA-mediated molecular processes that regulate lycopene accumulation in plant fruits.
Small RNAs (sRNAs) are emerging as important regulators of biological processes on post-transcriptional level in most eukaryotes [16, 17]. In plants, sRNAs are involved in a variety of activities that are essential for genome stability, development, and adaptive responses to biotic and abiotic stresses [16–21]. There is still increasing data for the implication of sRNAs in the regulation of other biological processes [22–24].
Based on difference of biogenesis and action, sRNAs have been mainly grouped into two categories, microRNA (miRNA) and short interfering RNA (siRNA) . miRNAs distinguish from siRNAs by their biogenesis. miRNAs are cleaved from stem-loop precursor molecules that derive from single-stranded RNAs, while siRNAs are generated from double-strand RNA precursors . In plants, miRNA regulates gene expression mostly on post-transcription level . The majority of the miRNAs initially isolated from model plant Arabidopsis thaliana was found to be evolutionarily conserved across plant species . Later, many researches suggested that the miRNA repertoire of plant species comprises a set of non-conserved miRNAs besides the conserved ones [27–29]. To date, 1894 miRNA sequences have been identified from 24 plant species, as revealed from miRNA registry database (miRBase v13.0; http://microrna.sanger.ac.uk). However, no citrus miRNAs have been found in this database so far.
Plant miRNAs have been identified mostly by two strategies: traditional Sanger sequencing of the small RNA libraries as for Arabidopsis [19, 26], rice [30, 31], and poplar ; and computational prediction of conserved miRNAs by searching for homologous sequences using EST or genomic sequences [18, 32–36]. However, the computation-based approach is mostly limited to the discovery of conserved miRNAs; and the small-scale traditional sequencing approach identified mainly conserved miRNAs, as suggested by Morin et al.  and Moxon et al. . Recently developed high-throughput sequencing technology has allowed the discovery of several non-conserved or lowly expressed miRNAs as well as quantification of the expression of miRNAs. This strategy has been successfully applied to both model plants [28, 38–44], and non-model plants [23, 37, 45–50].
The criteria for miRNAs characterization and annotation in plants have been advanced since plants have relatively large and complex sRNA populations within which miRNAs are often a minority . The minimum criteria include biogenesis criterion that miRNA precursors can be folded into stem-loop structure, and expression criterion that mature miRNAs should be detected by northern blotting, qRT-PCR or sequencing . Proof of biogenesis requires data of the cloning or detection of miRNA* sequences, which is complementary to miRNA sequences in the precursor molecule [29, 51]. Expression detection can also be done by recently developed qRT-PCR-based methods [53, 54], where each miRNA is reverse-transcribed from total RNA using a specific stem-loop primer, followed by PCR amplification and detection. This method is believed to be a powerful alternative for sensitive, rapid, and cost-effective detection of miRNA [24, 55]. Target prediction can be an ancillary criterion for miRNA annotation. Sequence complementarity between miRNAs and their target genes is very high in plants, which has made the search for plant miRNA target genes a straightforward process . A number of prediction algorithms are available publicly for plant miRNA target prediction, such as miRU , WMD3 , and UEA plant sRNA toolkit .
Here we described the deep sequencing of sRNAs from a red-flesh sweet orange mutant and its wild type. Comparative profiling of the sRNA transcriptome between wild type and mutant revealed a set of 51 known and 9 novel miRNAs showing significant expression changes. The annotation of the potential targets of differential miRNAs indicated that high ranked genes are that implicated in biological processes including carotenoid biosynthesis, transcription regulation, and photosynthesis.
Summary of sRNAs sequences from WT and MT sweet orange
matching protein-coding gene
We identified 85 miRNAs, belonging to 48 familes, in sweet orange by BLAST against the miRBase v13.0 using our combined (WT and MT) sRNA sequences (Additional file 5). The potential conserved miRNAs were used for BLASTN against sweet orange unigene dataset to search precursor sequences (Detail in Methods). Since the orange genome is largely unknown, we used orange unigene dataset (from TIGR, the institute for genomic research) and 1.2× coverage sequence of sweet orange (from NCBI) to predict RNA secondary structures. The sequences of perfect match with sRNA sequences were used for fold-back structure prediction by Mfold program . The miRNA-containing fold-back structures were shown in Additional file 6.
To compare the miRNA abundance in different libraries, the count of each miRNA was normalized to transcripts per million (TPM). In the WT sRNA library, miR168, miR172 and miR156 were the most overrepresented families, with the abundance at 8,927, 4,754 and 3,566 TPM, respectively. These three miRNA families also frequently represented in MT sRNA library, but the difference is that the miR172 abundance ranked third (3,330 TPM) after miR156 (4,397 TPM) and miR168 (8,437 TPM) in the library. It is interesting that in both libraries miR168 was the most frequent miRNA which constituted nearly 32% of the miRNA transcriptome of sweet orange.
Putative targets of miRNAs were determined using publicly available algorithms including miRU  and WMD3 . Of the 85 known miRNAs, 82 had putative targets, with the exception of miR167a, miR845b, and miR1520c (Additional file 7). Many putative targets are transcription factors and homologs of known miRNA target genes in other plant species, such as SBP for miR156, NAC for miR164, bZIP for miR166, AP2 for miR172 and F-box for miR394. One occasional case in grapevine  was also observed in sweet orange: csi-miR535 was predicted to target squamosa promoter-binding protein-like (SPL) gene while not target to a brassinosteroid signaling regulator as that firstly reported in moss .
Potential novel miRNA genes from sweet orange
Precursor Unigene ID
All the 12 novel miRNAs had a predicted target in at least one orange EST sequence (Additional file 9). Putative targets of four miRNAs were transcription factors, including NAC for csi-novel-01 and csi-novel-11, AUX/IAA TF for csi-novel-03 and YABBY for csi-novel-10. It is notable that one miRNA potentially target disease resistance genes, i.e. csi-novel-06 predicted to target on serine/threonine kinase gene.
Summary of small RNAs generated from miRNA-target genes
Producing sense small RNAs
Producing antisense small RNAs
Overlap between sense and antisense
sense or antisense
WT miRNA target genesa
Top 20 differentially expressed miRNAs between WT and MT
Higher expression in MT
Lower expression in MT
To confirm the expression difference of the miRNA, we analyzed the expression of 21 known miRNA genes using stem-loop qRT-PCR in addition to the 12 novel miRNA genes described earlier. The expression level of each gene in the MT and WT was compared with its abundance from the sequencing data of the sRNA library. Results showed that among the 33 miRNAs, 18 known miRNAs and 8 novel miRNAs have the same expression pattern in WT and MT as that from sequencing data (Figure 4B).
List of the important KEGG pathways more than 3 miRNA-target genes affiliated
EY732960, EY732985, EY748192, TC18796, TC9807
EY665527, EY667061, TC11997, TC13950, TC17473, TC24400, EY722705, TC11997
EY732960, EY732985, EY748192, TC18796, TC9807
Fatty acid metabolism
EY732960, EY732985, EY748192, TC18796, TC9807
EY697662, EY742573, TC14650, TC3846, TC14650, TC3846
EY732960, EY732985, EY748192, TC18796, TC9807
Glyoxylate and dicarboxylate metabolism
EY667061, EY708354, EY722705
Linolenic acid metabolism
EY732960, EY732985, EY748192, TC18796, TC9807
EY697129, EY724821, TC20228
Pentose phosphate pathway
TC13950, TC17473, TC15377
EY664657, EY697129, NP13097244, TC20228
Photosynthesis - antenna proteins
EY672715, EY682256, EY701112, EY701374, EY756276
EY681302, EY752996, TC18634
EY730974, TC363, EY752486
Ubiquitin mediated proteolysis
EY656660, EY665372, EY752038
Valine leucine and isoleucine degradation
EY732985, EY732960, EY748192, TC18796, TC9807, TC5927
We have characterized the small RNA transcriptome in a red-flesh mutant that with abnormal pattern of lycopene accumulation and its wild type. With deep sequencing strategy by Illumina platform, it provides us not only large number of conserved miRNAs but also novel miRNAs. To the author's knowledge, this is the first report of cloning and characterization of large scale sRNAs and identification of novel miRNAs from sweet orange. A total of 5 and 7 million sRNAs were obtained from WT and MT, respectively; and the size distribution of small RNAs (Figure 2) is quite similar to those published from phylogenetically close plant species such as Arabidopsis , with the total percentage of 21-24 nt sRNAs being more than 90% and percentage of 24 nt sRNAs over 50%. This suggests that the Illumina sequencing data of sRNA libraries was reliable.
In this study, 85 conserved miRNAs belonging 48 families as well as 12 novel miRNA families were identified in sweet orange. The miRNAs cloned in this study were compared with those from previously reported citrus miRNAs [36, 66, 67], of them four sweet orange miRNAs were not overlap with this study. These results suggest that at least 101 miRNAs exist in the miRNA transcriptome (miRNAome) of sweet orange. We believe that this is a conservative estimate since only fruit samples were used in this study. The miRNA species in wild type and mutant was different, i.e. 69 miRNAs were detected in WT sample while 66 miRNAs were in MT sample. Moreover, the expression levels of each miRNA varied greatly within the transcriptome, i.e. in MT the abundance ranged from 0.9 TPM for csi-miR1858 to 8437 TPM for csi-miR168. All these data implied that miRNAome, as small RNA transcriptome , is distinct and complex though miRNA sequences are believed to be conserved across different plant families [23, 29, 68].
Take a broader view on the small RNA transcriptome of sweet orange, we noticed that the size distribution pattern of sRNAs from sweet orange (the total percentage of 21-24 nt sRNAs > 90% with 24 nt sRNAs > 50%) is quite similar to Arabidopsis as described earlier ; however, the distribution pattern is strikingly different from that in Pinus contorta, a conifer species  where 24-nt RNAs was less frequently (2.5%) and 21-nt RNAs were more abundant (> 50%). The striking difference also existed when compare the sweet orange small RNA transcriptome with another gymnosperm plant species, Chinese yew (Taxus chinensis) . When compared with monocot species rice, the difference is less severe but still exists, i.e. in rice 24-nt sRNAs has the highest percentage (approximately 20%) followed by the 21-nt sRNAs (approximately 10%) . These analyses indicated that the small RNA transcriptome is complicated across plant species and can be significantly different between phylogenetically-distant plant families.
A total of 12 novel miRNAs were identified from sweet orange, as evidenced from the biogenesis characteristics of miRNAs including the stem-loop structure of the pre-miRNAs (Additional file 8) and sequencing of the miRNA* in the libraries. The expression of all the 12 miRNAs was also detected by qRT-PCR analyses. Additional qRT-PCR analyses of miRNA*s revealed that some miRNA* species can accumulated at certain levels in fruits; moreover, the expression pattern of miRNA*s in WT and MT fruits can be different from that of miRNAs. For example, csi-novel-05 has lower expression in MT (30 TPM) than in WT (90 TPM) as confirmed by qRT-PCR analysis; in contrast, csi-novel-05* has higher expression in MT (50 TPM) than in WT (26 TPM) which was also confirmed by qRT-PCR. The differential accumulation of miRNAs and miRNA*s may be due to the different activities of key enzyme in the sRNA-pathway, as shown in rice that the differential expression of AGO genes in different tissues might cause the change of miRNA*s abundance .
The relationship of sRNA genes and the miRNA target genes is one of the hot topics in the phenomenon of "miRNA-associated transitivity" [69–71]. In this study, 13.7% (WT) and 17.5% (MT) of the miRNA putative target genes generated sRNAs. This data implied that miRNA target gene could be hot spots for generating sRNAs. Moreover, high percentage of antisense sRNAs was associated with the miRNA-target genes, with 11.7% and 12.6% for WT and MT, respectively. The antisense sRNAs/transcripts can function in post-transcriptional gene silencing [62, 72]. The post-transcriptional gene silencing is mainly caused by formation of dsRNA duplexes; this could be supported from the observation that 75% of the antisense sRNAs have their cognate sense transcripts in WT and 64% happens in MT sample in this study.
Up to now, miRNA identification in fruits using a cloning approach is rare except in tomato  and grape . In this study, cloning from sweet orange fruit led to the identification of 48 conserved miRNA families. Analysis of them revealed that 6 miRNA families are expressed in fruits of the three plant species; but they are not fruit-specific. Furthermore, the comparison between the two fruit crops (85 miRNAs from sweet orange and 140 miRNAs from grapevine) showed that 8 miRNA families are expressed in fruits of both crops, and only miR535 is specific to the two species. These analyses suggested that expression and function of fruit-specific miRNAs across plant species is lacking, and miRNAs involved in fruit ripening may overlap with that function in other organs, or the same miRNAs in different organs may have different functions.
The characterization and comparative profiling of entire sets of small RNAs (Small RNA transcriptome), especially the microRNA transcriptome, lays the foundation to unravel the complex miRNA-mediated regulatory networks controlling lycopene accumulation in the red-flesh sweet orange mutant. In this study, a number of miRNAs have been identified to be differentially expressed between 'Anliu' sweet orange (WT) and its red-flesh mutant (MT). Compared with WT, 16 known miRNA genes and one novel miRNA genes were expressed exclusively in MT. On the other hand, we found 19 known miRNAs and two novel miRNAs were WT-specific. Moreover, a total of 51 known miRNAs and 9 novel miRNAs were found to exhibit significant expression changes between MT and WT.
Target prediction of these differential miRNAs could provide information on the biological processes regulated by miRNA. Putative targets were predicted by psRNATarget (the updated version of miRU)  and web miRNA designer (WMD3) . This strategy was also used in Arabidopsis for the identification of nutrient-responsive miRNAs and their potential targets , though further direct validation is necessary to elucidate the true in-vivo targets. However, the annotations of these potential miRNA target genes did provide an alternate view of gene regulation of the red-flesh trait formation in MT (Additional file 11). We found that three groups of predicted miRNA-target genes are most possibly involved in the mutant trait formation of lycopene accumulation. The first interesting observation is that two potential targets (EY752486 by csi-miR167 and TC5 by csi-miR1857) are the key enzymes of the carotenoid biosynthesis pathway (Additional file 11). They encode geranylgeranyl pyrophosphate synthase (GGPS), a key enzyme for the production of prerequisite molecules for carotenoids accumulation, and lycopene β-cyclase (LYCb), a rate-limiting enzyme in the conversion of lycopene to downstream cyclic carotenes. The expression of LYCb and csi-miR1857 also indicated possible interaction between them because of the complementary expression profile (Figure 6). And we also detected the GGPS gene expression, but failed due to the incorrect sequence information in the current public database. The second interesting target genes are APETALA2 (AP2) transcription factors, including EY726563 by csi-novel-03, and TC7810 by csi-miR172a. In Arabidopsis, one AP2 protein (designated as RAP2.2) has been experimentally validated to interact with phytoene synthase (PSY) gene, which encodes a rate-limiting enzyme for the synthesis of carotenoids. The RAP2.2 gene was also believed to be regulated by miRNA because of the large discrepancy between mRNA and protein levels in the overexpression lines . The data from this study also supported the proposition, and indicated that RAP2.2 (encoded by TC7810 gene in this study) is a potential target of miRNA172a in sweet orange. The third group of target genes is those involved in photosynthesis or carbon fixation (Table 5). The potential target genes associated with photosynthesis fits well with the GO results with large number of genes that are located in chloroplast or plastid. Moreover, our photosynthesis-related gene expression analysis and in situ detection of the photosynthetic rate confirmed that photosynthesis in MT is stronger than that in WT (Figure 6). The result is quite consistent with our gas chromatographic analyses that the sugars in juice sacs are higher in MT than in WT . Tomato hp-2 mutant which manifested with increased carotenoids, primarily lycopene, showed that genes related with photosynthesis are consistently up-regulated throughout fruit ripening . All these evidences implied the important role of photosynthesis in the accumulation of lycopene. Taken together, the annotation of the potential targets of differentially expressed miRNA genes between WT and MT indicated that several miRNA-directed biological processes including carotenoid biosynthesis, transcription regulation, and photosynthesis may be important for the regulation of lycopene accumulation in the sweet orange red-flesh mutant.
Our study provides a large scale cloning and characterization of miRNAs in sweet orange. The interpretation of the small RNA data uncovered a total of 85 known and 12 novel miRNAs. Meanwhile, comparative profiling of these miRNAs between a spontaneous red-flesh mutant and its wild type revealed that 51 known and 9 novel miRNAs are differentially expressed. Target prediction of the 60 differential miRNAs indicated that two key genes of the carotenoid biosynthesis, geranylgeranyl pyrophosphate synthase (GGPS) and lycopene β-cyclase (LYCb), are under miRNA regulation. Functional annotation of the 418 target genes of differential miRNA revealed that high ranked biological processes are transcription regulation, protein modification and photosynthesis. The expression of miRNAs and target genes involved in photosynthesis and carotenogenesis were further confirmed. Taken together, this study provides new insight into the miRNAome and possible miRNA-directed molecular processes that regulate lycopene accumulation in the red-flesh sweet orange mutant.
The red-flesh mutant 'Hong Anliu' and the wild type 'Anliu' sweet orange (Citrus sinensis L. Osbeck), cultivated at the Institute of Citrus Research located in Guilin, Guangxi Province, China, were used as materials. The mutant 'Hong Anliu' sweet orange (MT) is a spontaneous bud mutation of wild type 'Anliu' sweet orange (WT); and they are with isogenic background as revealed from our previous physiological traits and molecular marker evaluations . Same sampling strategy as previously published  was employed. Fruit samples were harvested at 170 days after flowering (DAF) that is the critical stage for the flesh color transition. Three different trees were used, and 10 representative fruits from each tree were collected. All samples were separated into peel and pulp, immediately frozen in liquid nitrogen and kept at -80°C until use. Total RNA was extracted according to Liu et al. .
Small RNA library construction and sequencing was performed as described by Yao et al. . Briefly, small RNAs with 16-30 nt in length were isolated from the total RNA of WT and MT samples by size fractionation with 15% Tris-borate-EDTA urea polyacrylamide gel (TBU). After extraction from the gel, sRNA was dephosphorylatd by alkaline phosphatase (New England Biolabs Inc.). Then the small RNAs were ligated sequentially to 5' and 3' RNA/DNA chimeric oligonucleotide adaptors (Illumina). The ligation products were purified from 10% TBE urea polyacrylamide gel. Reverse transcription was performed with the ligated sRNAs as templates. The obtained cDNAs were sequenced using the Illumina 1 G Genome Analyzer by Beijing Genomics Institute (BGI) (Shenzhen, China).
Small RNA sequences were extracted from raw reads matching both the last 7 nt of the 5'-adaptor and the first 7 nt of the 3'-adaptor as described by Moxon et al. . After trimming off the adaptors, sequences between 18 and 30 bp in length and reads more than 3 were used to search the Rfam database http://www.sanger.ac.uk/Software/Rfam by BLASTN analysis to discard rRNA, tRNA, snRNA, and snoRNA. The remaining sequences were compared to known plant miRNAs from miRBase v13.0  using BLASTN algorithm to search conserved miRNAs. Sequences in our libraries with identical or with four or fewer mismatches to currently known miRNAs from other plant species were regarded as potential conserved miRNAs. The potential miRNA sequences were used for BLASTN search against sweet orange unigene dataset and 1.2× coverage sequence of sweet orange (from NCBI), the perfectly matched sequences were used for fold-back structure prediction by Mfold program .
To identify non-conserved miRNAs from sweet orange, the remained sRNAs were used for BLASTN search against sweet orange ESTs from TIGR gene index database http://compbio.dfci.harvard.edu/tgi/cgi-bin/tgi/GO_browser.pl?species=Orange&gi_dir=csgi and 1.2× coverage sequence of sweet orange (from NCBI), since the sweet orange genome is not fully sequenced. The sequences of perfect match with sRNA sequences were used for fold-back structure prediction by Mfold program . The sRNAs that have stem-loop precursors were regarded as putative non-conserved miRNAs. Furthermore, we searched miRNA* sequences (complementary to miRNA in the precursor molecule) in the sRNA libraries. Only those with miRNA-miRNA* duplexes were regarded as novel miRNAs.
Significance level of the difference of miRNA frequency and abundance between WT and MT was analyzed using Z-score method according to Kal et al. . The significance of differentially expressed miRNAs were set at the p value < 0.01.
The detection of miRNA expression was performed using stem-loop qRT-PCR method according to Chen et al.  and Varkonyi-Gasic et al. . Stem-loop primers for reverse transcription and primers for quantitative PCR were list in Additional file 12. Total RNA was reverse-transcribed using SuperScript III Reverse Transcriptase (Invitrogen, US) by a pulse reverse transcription program . Real-time quantitative PCR (qPCR) was performed on the ABI 7500 Real Time System (PE Applied Biosystems, Foster City, CA, USA) using actin gene as endogenous control according to Liu et al. . Briefly, the primers for miRNAs and actin were diluted in the SYBER GREEN PCR Master Mix (PE Applied Biosystems) and 20 μl of the reaction mix was added to each well. Reactions were performed by an initial incubation at 50°C for 2 min and at 95°C for 1 min, and then cycled at 95°C for 15 s and 60°C for 1 min for 40 cycles. Output data was generated by the instrument on-board software Sequence Detector Version 1.3.1 (PE Applied Biosystems). The electrophoresis of the PCR products was performed on 3% agarose gel for 1 hr under 70 V voltage. The marker contained two fragments, 50 and 60 bp in length, the sequences information was: GTCGTATCCAGTGCAGGGTCCGAGGTATTCGCACTGGATACGACCCGATG and CGGCGAAGATTGAAGTGAAGGTGCAGGGTCCGAGGTATTCGCACTGGATACGACTTCCGA respectively.
Target prediction of miRNAs from sweet orange was performed based on methods described by Pant et al. . Two publicly available algorithms, miRU  and WMD3  were used in the prediction. The sweet orange EST sequences from TIGR database as described earlier were used to predict the targets. The hybridization energy of each miRNA-predicted target duplex (ΔG) was determined from WMD3 output.
To investigate biological processes possibly regulated by miRNAs, assignments of putative functions to potential target genes of differentially expressed miRNAs were performed using annot8r program, which was run locally to BLAST against a reference database that stores UniProt entries, their associated Gene Ontology (GO), Enzyme Commission (EC) and Kyoto Encyclopaedia of Genes and Genomes (KEGG) annotations . The GO categorization results were expressed as three independent hierarchies for biological process, cellular component, and molecular function . The biological interpretation of the potential targets was further completed by assigning to metabolic pathways using KEGG .
The expression levels of one carotenoid biosynthesis gene (lycopene β-cyclase, LYCb) and one carbon fixation-related gene (phosphoenolpyruvate carboxylase, TC11997) were determined on the ABI 7500 Real Time System (PE Applied Biosystems, Foster City, CA, USA) using actin gene as endogenous control according to Liu et al. . The primers are as following: TC11997: forward, GAGATGGTGCAGGCCAAGTT and reverse GTTGAGGAGGTCGCATGGTT; LYCb: forward, GGCTATATGGTGGCAAGGACTT and reverse, CAGAATTGAGGCTTCGAACGA. The data of in situ rates of photosynthesis in the mutant and the wildtype were extracted from our previous data .
The small RNA sequence data from this study have been submitted to Gene Expression Omnibus (GEO) under accession No. GSE18207 at website: http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?token=bvcnfememqckehi&acc=GSE18207
Days after flowering
Kyoto Encyclopedia of Genes and Genomes
NAM, ATAF, and CUC
Ribulose-1,5-bisphosphate carboxylase oxygenase
Squamosa promoter-binding protein
Small nuclear RNA
Small nucleolar RNA
short interfering RNA
Transcripts per million
We thank Dr. Hanhui Kuang from Huazhong Agricultural University (China) for their helpful suggestions on this study. This work was supported by the Natural Science Foundation of China (NSFC Nos. 30921002, 30830078 and 30800745), Specialized Research Fund for the Doctoral Program of Ministry of Education of China (No.200805041024), and Hubei Provincial Natural Science Foundation (2008CDB083).
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.