Introduction

Gene expression can be regulated in many ways and one of them is mediated by microRNAs (miRNAs), which are short single-stranded RNAs of 21–24 nt in length that regulate gene expression in eukaryotes at the post-transcriptional level (Brodersen and Voinnet 2006). The understanding of gene regulatory mechanisms that involve miRNAs has had a significant impact on the knowledge about the regulation of plant developmental processes such as leaf morphogenesis and polarity, floral organ identity and differentiation, organ development, signaling pathways, and the response to stresses (Válóczi et al. 2006; Pulido and Laufs 2010; Rubio-Somoza and Weigel 2011).

miRNAs are grouped into families according to their sequence and they regulate their target genes through mRNA cleavage (Dugas and Bartel 2004) or translational repression (Brodersen et al. 2008). In plants, miRNAs perfectly or near-perfectly bind to their target mRNAs, and mRNA cleavage is the most common mechanism (Mallory et al. 2004). Based on the diversity and abundance of plant miRNAs, it may be possible that the regulation of all physiological and biochemical processes in plants involve the action of miRNAs (Jones-Rhoades et al. 2006; Chen 2009). Fruit development is a complex process and many genes are under dynamic control to ensure the proper establishment of the different fruit tissues and structures (McAtee et al. 2013). There is an impressive diversity of fruit developmental programs across the plant kingdom and it is to be expected that miRNAs also play important roles in these. miRNA studies in fruits are relatively scarce still. Currently, miRNAs were identified in fruits such as tomato (Pilcher et al. 2007; Din et al. 2014), grape (Carra et al. 2009; Wang et al. 2012, 2014), citrus (Song et al. 2009; Xu et al. 2010), melon (Gonzalez-Ibeas et al. 2011), apple (Xia et al. 2012), and strawberry (Xu et al. 2013; Ge et al. 2013), among others. Highly conserved miRNAs such as miR156, miR159, miR164, miR166, and miR172 have been identified in fruit tissues. Several studies demonstrated that conserved and specific miRNAs regulate genes involved in fleshy fruit development (e.g., Moxon et al. 2008; Karlova et al. 2011; Mohorianu et al. 2011).

For the prickly pear cactus (Opuntia ficus indica), no information on miRNAs is available. Currently, the genome sequence is unavailable; however, the use of the knowledge on conserved miRNAs of other plant species (Chavez-Montes et al. 2014) could be an approach to identify miRNAs in prickly pear cactus. The prickly pear cactus fruit is a simple berry, botanically speaking an accessory fruit formed from an inferior ovary adhering to the receptacle, covered with spines (Fig. 1) and is considered a non-climacteric fruit (Campos-Guillén et al. 2012). An important physiological characteristic of the prickly pear cactus fruit is the variability in fruit ripening time. The time that elapses from the onset of the floral bud to a ripe fruit ranges from 45 to 154 days (Fig. 1) in different varieties, therefore, they have been classified into early, middle, and late ripening fruits (Campos-Guillén et al. 2012). Interestingly, precise transcriptional and post-transcriptional gene regulation during prickly pear cactus fruit development has been observed (Campos-Guillén et al. 2012), suggesting that miRNAs might participate. Altogether, this makes prickly pear cactus fruit an attractive species for miRNA analysis.

Fig. 1
figure 1

Diagram of prickly pear cactus fruit development. a Prickly pear cactus fruit at different stages of development. The images illustrate major physiological events, their timing, and the sampling time points. b The phases represent: I, floral development and fruit-set; II, cell division during young fruit development; III, cell expansion; and IV, fruit ripening

Full size image

In the present study, bioinformatics and molecular biology approaches were combined to obtain miRNA expression profiles during prickly pear cactus fruit development. Furthermore, the spatio-temporal expression of miR164 during prickly pear cactus fruit development was analyzed in detail.

Materials and methods

Plant material

Prickly pear cactus fruits (Opuntia ficus indica) were collected at INIFAP (Mexican National Institute of Forestry, Agriculture, and Livestock Research) Campo Experimental Norte de Guanajuato, in San Luis de la Paz, Gto., Mexico. The tissues were grouped into different stages of prickly pear cactus fruit development: young floral bud (not fertilized yet), floral bud, flowering bud, young fruit, green fruit, and ripe fruit (Fig. 1). After harvesting, the tissues were washed, frozen in liquid nitrogen, and stored at −80 °C until further use. For RNA extraction, the samples were sliced and ground to a fine powder in a mortar with liquid nitrogen. For tissue print analysis, the samples were cut immediately and printed on nylon membranes.

Low molecular weight (LMW) RNA extraction and sRNAs analysis in polyacrylamide gel

Low molecular weight (LMW) RNA was extracted from different stages of prickly pear cactus fruit development according to our previously reported protocol (Rosas-Cárdenas et al. 2011). In summary, LiCl extraction buffer and phenol pH 8.0 was used for extraction, the samples were incubated for 5 min at 60 °C and centrifuged for 10 min at 13,400g at 4 °C. The supernatant was re-extracted with 600 μl of chloroform-isoamyl alcohol and centrifuged. For the separation of high molecular weight (HMW) RNA, the supernatant was incubated for 15 min at 65 °C and precipitated with 50 μl of 5 M NaCl and 63 μl of 40 % polyethylene glycol 8000 (w/v), followed by incubation on ice for at least 30 min and centrifuged. The supernatant that contained LMW RNA was extracted with 500 μl of phenol–chloroform-isoamyl alcohol and centrifuged. LMW RNA was precipitated with 50 μl of 3 M sodium acetate pH 5.2 and 1 ml of absolute ethanol, incubated overnight at −20 °C and recovered by centrifugation, dried and resuspended in RNAse-free water. LMW RNA was visualized in polyacrylamide gels. RNA concentration and quality was quantified in a Nanodrop ND-1000 spectrophotometer (NanoDrop Technologies Inc.), measuring absorbance to calculate the A260/A280 and A260/A230 ratios.

Northern blot miRNA analysis

A total of 2 µg of LMW RNAs of different developmental stages of prickly pear cactus fruit were separated according to their size in 12.5 % denaturing polyacrylamide gels using a mini-vertical electrophoresis gel system (GE Healthcare). A semidry trans-blot system (Biorad) was used to transfer the sRNAs from gel to Amersham Hybond-N membrane (GE Healthcare) in 0.5× TBE buffer for 1 h at 10 V. The membranes were dried at room temperature, fixed with 12 ml EDC cross-linking solution (0.16 M 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide (EDC; Sigma-Aldrich) prepared in 0.13 M 1-methylimidazole at pH 8.0, incubated for 1.5 h at 65 °C, and rinsed twice with water. The membranes were dried at room temperature and stored at −20 °C till further use. The northern blot analysis was performed according to a previously reported protocol with some modifications (Pall and Hamilton 2008; Rosas-Cárdenas et al. 2011). The membranes were pre-hybridized with 15-ml hybridization solution (Rapid-hyb Buffer Amersham; GE Healthcare) for 1-h at 37 °C, followed by the addition of the labeled probe of interest, and then incubated for 24-h at 37 °C. Each probe used was labeled with [γ-32P] ATP to detect the respective miRNA and control U6 (small nucleolar RNA; 5′-AGGGGCCATGCTAATCTTCTC-3′). The membranes were washed twice with wash solution (2× SSC/0.1 % SDS), first for 4 min, and then a second time for 2 min at room temperature, followed by exposure to a storage Phosphor Screen System (Amersham Biosciences) for approximately, 48 h.

Reverse northern blot hybridization

Fifty-three synthetic DNA oligonucleotides corresponding to the reverse complementary sequences of selected mature miRNAs were used to print the miRNA arrays or used as probe for northern blot hybridization (Suppl. Table S1). Each probe (5 µl of each oligonucleotide stock 100 µM) was manually spotted on a 2 × 2 cm Amersham Hybond-N nylon membrane (GE Healthcare), dried at room temperature, and fixed with 12 ml EDC cross-linking solution, as described previously. Printed membranes were wrapped in aluminum foil and stored at −20 °C until they were used for hybridization. miRNA fractions isolated from different tissues were used for miRNA array hybridization. A total of 16 µg of LMW RNAs of each tissue (divided over several lanes) were resolved in 12.5 % denaturing polyacrylamide gels using a mini-vertical electrophoresis gel system (GE Healthcare) and the sRNA fraction corresponding to 18–24 nt was excised from the polyacrylamide gel and stored at −20 °C until the purification.

The gel slices with the sRNA fractions were placed in two or three 1.5 ml tubes, crushed with a glass rod, and then 200 μl sterile nuclease-free water was added and continued to crush the gel into fine slurry. The sample was placed at 70 °C for 10 min. Following the manufacturer’s recommendations, one column with a gel filtration matrix (DTR column; EDGE Biosystems, Gaithersburg, MD, USA) was prepared for each gel slice. The gel slice slurry was vortexed for 30 s and the entire volume transferred onto the DTR column and centrifuged at 850g for 3 min. Then 3 μl of 10 mg/ml glycogen, 25 μl of 3 M NaOAc (pH 5.2), and 900 μl of ice cold 100 % EtOH was added to the flow-through, mixed by inversion and placed at −80 °C for 30 min, subsequently the tubes were centrifuged at 13,400g for 10 min and the supernatant discarded. Finally, the RNA was air dried. The purified fractions with the sRNAs were resuspended in water (max. 20 μl), samples from same tissue were pooled, and stored at −20 °C until tested.

A total of 40 ng of each previously purified sRNA fraction corresponding to 18–24 nt was dephosphorylated with Antarctic Phosphatase (New England Biolabs) for 1 h and then radioactively labeled with T4 polynucleotide kinase (Invitrogen) and 1.5 µl [γ-32P]-ATP 222 TBq/mmol (370 MBq/ml; AccesoLab, Mexico D.F., Mexico) for 1 h at 37 °C. The labeling reaction was stopped by incubation at 65 °C for 5 min and then directly incubated on ice for 3 min. The miRNA arrays were pre-hybridized in 15 ml of Rapid-hyb Buffer Amersham (GE Healthcare) for 1 h at 37 °C. The labeled miRNA samples were hybridized in Rapid-hyb Buffer Amersham (GE Healthcare) for 15 h at 37 °C. After hybridization, the miRNA arrays were washed with wash solution (2× SSC/0.1 % SDS) for 10 min at 37 °C. The membranes were then exposed using a storage Phosphor Screen System (Amersham Biosciences) for 2 days. The phosphor screen was scanned in a Storm 860 Gel and Blot Imaging System (Amersham Biosciences). Signals were quantified using the ImageQuant TL software (Amersham Biosciences).

Data analysis of reverse northern blot arrays

The signal intensity of each spot on the array was quantified using ImageQuant TL using the array module (Amersham Biosciences). Background correction was performed with the Spot Edge Average method to determine the average background intensity around each spot outline. Normalization of intensity data between arrays was performed using a normalization factor calculated based on the expression profile of miR166 of a traditional northern blot (Suppl. Fig. S1). First, the traditional northern blot of miR166 was normalized using the intensity values of U6 determined using ImageQuant TL (Suppl. Fig. S2). For each stage, this miR166 normalized expression value was compared to the miR166 value of the array, and the ratio between these two values was used to normalize the values of each miRNA on the respective array, allowing the comparison of miRNAs between arrays. An expression profile matrix was built representing the level of expression for each stage for each miRNA detected in at least one stage. The expression profile matrix was imported into the Multiple ArrayViewer program (MeV; Saeed et al. 2003) and hierarchical clustering was performed using the following options selected: ‘gene tree’, ‘sample tree’ ‘optimize gene leaf order’, ‘Pearson correlation’, and ‘complete linkage clustering’.

miRNA detection by microarray hybridization

The expression profiles of miRNAs detected by reverse northern blot were verified by microarray hybridization. Briefly, total RNA of young fruit, green fruit, and ripe fruit was extracted with Trizol (Invitrogen). RNA of each condition was labeled with biotin and hybridized to separate arrays, GeneChip® miRNA 2.0 Array (Affymetrix) according to the manufacturer. In detail, first a hybridization was done with a pool of RNAs from each stage and second, each sample of each stage was hybridized in an independent way as second repetition (data not shown; readers can obtain details from the authors). The Affymetrix fluidic station 450 was used for the wash and stain operation of the arrays. Images were captured with the Scanner Unit model “3000 7G”. These procedures were performed at the ‘Unidad de Genotipificación y Análisis de Expresión’ of the Instituto Nacional de Medicina Genómica (Inmegen, Mexico). The R software, available from the Bioconductor Project (http://www.bioconductor.org), was used to normalize data before comparison and to evaluate differential expression between pool samples. Data normalization was made using the Robust Multichip Average (RMA) method (Irizarry et al. 2003). It consists of three steps: a background adjustment, quantile normalization, and finally summarization. All differential expression values were selected from at least >2×-fold change using the Oligo (de Carvalho et al. 1992) and Limma (Smyth 2005) packages.

miRNA detection by stem-loop RT-PCR

In brief, total RNA was isolated using Trizol (Invitrogen). The stem-loop RT-PCR method used is described in Li et al. (2009). The oligonucleotide sequences used are as follows: miR164 RT loop, 5′-TCAACTGGTGTCGTGGAGTCCGGCAATTCAGTTGAGTGCACGTG-3′, miR164 forward, 5′-ACACTCCAGCTGGGTGGAGAAGCA-3′, and miR164 reverse, 5′-AACTGGTGTCGTGGAG-3′. The final PCR products were separated in a 4 % agarose gel.

miRNA detection by tissue printing and northern blot hybridization

Tissue printing provides a simple and rapid method to analyze the localization and expression of miRNAs at the tissue level. This is especially convenient for large tissues, such as prickly pear cactus fruit. For the spatio-temporal analysis miR164 was selected. The prickly pear cactus fruits were washed with water and dried at room temperature. The materials were cut in longitudinal and transverse sections. Transverse sections were made from basal, medial, and apical regions of prickly pear cactus fruits. After cutting the samples, they were immediately placed with its cut surface face down on Amersham Hybond-N membrane (GE Healthcare). The different sections were firmly pressed on the nylon membrane for 30 s. The membranes were treated in a similar way as described above, were dried at room temperature and stored at −20 °C till further use. To analyze the miRNA localization and expression, the miRNA detection was carried out as the northern blot analysis described above, at 37 °C. An oligonucleotide for U6 RNA (5′-GGGGCCATGCTAATCTTCTC-3′) was used as positive control and the M13 forward oligonucleotide (5′-TGTAAAACGACGGCCAGT-3′) as negative control. Duplicate analyses assured reproducibility of the results.

Results

miRNA selection for expression analysis

Some miRNAs are highly conserved across the plant kingdom (Chavez-Montes et al. 2014), making it possible to detect them in many plant species using the available miRNA sequences. Based on this principle, we took the approach of screening for conserved miRNA expression profiles during prickly pear cactus fruit development. We generated miRNA arrays with a selection of miRNA sequences from different fruit bearing species (Suppl. Fig. S3). In summary, 53 miRNAs were selected based on in silico analysis, of which 49 are expressed in at least one fruit crop, 14 miRNAs are expressed in all fruit analyzed, and 4 are not expressed in the analyzed fruit crops (http://smallrna.udel.edu/; (Chavez-Montes et al. 2014) (Suppl. Fig. S3). To analyze which of the miRNAs are expressed in prickly pear cactus fruit, we selected four developmental stages: young floral buds, young cactus fruits, green cactus fruits, and ripe cactus fruits. These samples represent the most marked changes during fruit development, which is depicted in Fig. 1. The first sample, young floral bud, represents the phase I (ovary development). The young fruit sample represents the phase II (cell division and seed formation), the green fruit sample represents the phase III (cell expansion and embryo maturation), and the ripe fruit sample the phase IV (ripening fruit) (Fig. 1) (Gillaspy et al. 1993; Reyes-Agüero et al. 2006). The sRNA fractions of these different tissues were obtained and analyzed by spectrophotometric analysis and polyacrylamide gels, showing well-defined 5S and tRNA bands (Suppl. Fig. S2), suggesting a good quality of LMW RNA. Each sRNA sample was labeled and used for hybridization with the miRNA array.

Detection of miRNAs at different stages of prickly pear cactus fruit development

To elucidate the potential roles of miRNAs in prickly pear cactus fruit development, global miRNA expression patterns at different fruit stages were studied by northern blot hybridization, reverse northern blot hybridization (arrays; Suppl. Fig. S4), and GeneChip® miRNA 2.0 microarrays (Suppl. Fig. S5). As a first approach to detect miRNA presence in prickly pear cactus we performed sRNA northern blot hybridizations with five miRNAs previously identified as highly conserved in plants (Chavez-Montes et al. 2014). As expected, all five conserved miRNAs (miR159, miR166, miR167, miR168, and miR172) were detected during different developmental stages (Fig. 2). We observed that these miRNAs were highly expressed during early floral development and less expressed during fruit development.

Fig. 2
figure 2

Northern blot hybridization analysis of miRNA expression at different stages of prickly pear cactus fruit development. 2 µg per sample of LMW RNA were electrophoresed in 12 % polyacrylamide gels under denaturing conditions, blotted, and hybridized with 32P-labeled oligonucleotide probes complementary to the indicated miRNAs. The U6 panel shows equivalent loading. AL Arabidopsis leaves (positive control), YFB young floral bud, FB floral bud, YF young cactus fruit, FG green cactus fruit, FR ripe cactus fruit

Full size image

The following experiment performed was a reverse northern blot analysis. For this, we prepared arrays with spotted oligonucleotides corresponding to sequences of 53 different miRNAs (see “Materials and methods” for details). sRNA fractions of the four developmental stages were individually hybridized to the array. Interestingly, a total of 34 different miRNAs were detected in prickly pear cactus: 22 miRNAs in floral buds, 26 in young fruit, 25 in green fruit, and 17 in ripe fruit (Figs. 3, 4; Suppl. Table S2). A good correlation was observed between the expression obtained by the miRNA reverse northern blot array experiment and the northern blot analysis for miR159, miR166, and miR168 (Fig. 2). In addition, miR167 and miR172, which were not detected by our array hybridizations, were detected by northern blot hybridization. However, they showed very low expression levels (Fig. 2), suggesting that different hybridization efficiencies may cause some lowly expressed miRNAs to be undetected by the array approach. On the other hand, several miRNAs previously reported as highly conserved and highly abundant in plants, as miR172, were expressed at very low levels in prickly pear cactus. This indicates that not all conserved miRNAs present a similar expression pattern in prickly pear cactus versus other species.

Fig. 3
figure 3

Distribution of miRNAs detected at different stages of prickly pear cactus fruit development. The Venn diagram shows shared and specifically expressed miRNAs at different developmental stages of prickly pear cactus fruit. Note: miR393 and miR398 are not depicted in the figure because their expression overlap is only at the floral bud with the green fruit stage; for miR159 the northern blot result was also taken into account

Full size image
Fig. 4
figure 4

Cluster analysis of miRNA expression at different stages of prickly pear cactus fruit. Heatmap of normalized miRNA expression levels during prickly pear cactus fruit development; gene-normalized data were used. Each column represents a stage and the intensity of the red color indicates the relative expression level at that stage. Hierarchical clustering resulted in three major groups marked with color bars on the right vertical axis. The histograms simulating the expression patterns of the corresponding clusters are presented on the right. Hierarchical clustering was performed using MeV with the following options selected: ‘gene tree’, ‘sample tree’ ‘optimize gene leaf order’, ‘Pearson correlation’, and ‘complete linkage clustering’. YFB young floral bud, YF young cactus fruit, GF green cactus fruit, RF ripe cactus fruit

Full size image

Finally, as third experiment, a preliminary microarray hybridization experiment was performed for floral bud, green fruit, and ripe fruit (see “Materials and methods” for details). Comparing this data with the expression profiles of the detected miRNAs by reverse northern blot hybridization (Fig. 4), 18 miRNAs show a similar profile (Suppl. Fig. S5). The rest of the miRNAs show some differences in their expression profile, most of them being miRNAs that are lowly expressed and therefore perhaps having more variation (Suppl. Fig. S5).

Expression profiles of miRNAs during prickly pear cactus fruit development

The reverse northern blot experiment revealed that 14 miRNAs (miR159, miR164, miR168, miR165, miR166, miR390, miR395, miR482, miR535, miR822, miR1524, miR2119, miR3635, and miR3636) could be detected in all developmental stages (Fig. 3). Other miRNAs were detected in two or three stages. miR397 was expressed both at the floral bud and the young fruit stage, while the expression of miR319, miR399, and miR3623 was detected at the young fruit and the green fruit stage. Moreover, miR156, miR157, and miR3633 were found to be expressed at the young fruit, green fruit, and ripe fruit stage. Furthermore, we also found miRNAs that were only expressed at a single stage. For instance, miR408 and miR394 were expressed specifically in floral buds. At the young fruit stage, specific expression was detected for miR824, miR828, miR894, miR1916, and miR3954. miR169, miR170, miR846, and miR3629 were specific to the green fruit stage. Interestingly, we did not detect any of the selected miRNAs to be specifically expressed at the ripe fruit stage.

Although the miRNAs are present in one, some, or in all stages analyzed, the expression levels were different through different stages (Suppl. Table S2). miR159 and miR408 were the most abundant miRNAs in the floral bud sample; and miR164, miR390 and miR1524 were the most abundant miRNAs in the young fruit. A few miRNAs were expressed at high levels in ripe fruit, such as miR2119 and miR395. In general, the miRNA expression levels ranged from very low to high as fruit development progressed. However, stage specific miRNAs were expressed at relative low levels, with the exception of miR408.

Hierarchical cluster analysis of miRNA expression profiles

To identify trends among the detected miRNAs during development, we performed a hierarchical cluster analysis (Fig. 4). The hierarchical clustering resulted in three major groups of miRNA expression profiles, named as: Group I, II, and III (Fig. 4). miRNAs in the same group might be involved in a similar process during development. Group I contains 7 miRNAs: miR159, miR165, miR166, miR394, miR397, miR398, and miR408, which are highly expressed in floral buds and then decrease their expression during development, though miR159 and miR166 show a slight increase in ripe fruit. Group II is the largest group, containing 16 miRNAs, and can be divided into two subgroups. The first subgroup shows no or very low expression in young floral buds, a high expression in young fruit and a significant decrease in expression in green and ripe fruit (miR535, miR824, miR828, miR1916, and miR3954). The other subgroup contains miRNAs that maintain their expression relatively high, also after the young fruit stage: miR156, miR164, miR168, miR319, miR390, miR399, miR482, miR822, miR3633, and miR3635. Finally, Group III contains 11 miNAs that are mainly expressed at the green fruit stage, namely miR157, miR169, miR170, miR393, miR395, miR846, miR1524, miR2119, miR3623, miR3629, and miR3636.

Detailed expression analysis of miR164

Development is regulated by spatially and temporally coordinated regulatory networks that include regulation by miRNAs. For the prickly pear cactus fruit developmental process, we were interested in miRNAs of group II, which showed an increase in expression during fruit development and a decrease in ripe fruit. We confirmed the expression profile of one of the miRNAs in this group, miR164, in the four developmental stages of the prickly pear cactus fruit (Fig. 5) by stem-loop primer based RT-PCR, which is a highly sensitive and specific method for miRNA detection (Chen et al. 2005; Li et al. 2009). This temporal pattern was also previously reported for miR164 in tomato (Mohorianu et al. 2011). However, while this data provides information about the time of expression of this miRNA, it provides little information about its spatial pattern of expression, because tissues are homogenized to isolate RNA.

Fig. 5
figure 5

Stem-loop RT-PCR of miR164 at different stages of prickly pear cactus fruit development. RT-PCR products were electrophoresed in 4 % agarose gel. FB floral bud, YF young cactus fruit, GF green cactus fruit, RF ripe cactus fruit

Full size image

Until now, the spatio-temporal expression of miRNAs, which includes their localization in specific tissues has not been studied in fleshy fruits. The use of common in situ hybridization techniques provides information about the spatial pattern of expression, but it might be challenging for the tissues of large fruits. Therefore, we used tissue printing followed by miRNA hybridization to observe the spatio-temporal expression of miR164 during prickly pear cactus fruit development (Fig. 6; Suppl. Fig. S6). We found that miR164 was expressed in all analyzed developmental stages (Fig. 7). We also observed that, at early stages of floral development, miR164 was expressed in meristematic zones such as the floral meristem and the meristems producing the glochids (i.e., hair-like spines) (Fig. 6a). In open flowers, miR164 was highly expressed in fusion zones such as the carpel base, within the stamens and receptacle, and the compitum zone (Fig. 6b). High miR164 expression was observed in the ventral carpel vascular bundle (Fig. 6c). Moreover, in middle-ripe and ripe fruits, miR164 was mainly homogenously expressed in the complete fruit (Fig. 7). The observed hybridization patterns correlate well with the miRNA expression profile obtained with the reverse northern blot hybridization and the stem-loop RT-PCR experiment.

Fig. 6
figure 6

miR164 spatio-temporal expression pattern during early stages of prickly pear cactus fruit development. Tissue print of prickly pear cactus followed by northern blot hybridization was performed. a miR164 expression in meristems. b miR164 expression at the compitum of prickly pear cactus. c miR164 expression in prickly pear cactus placenta. Scale bars indicate 0.5 cm

Full size image
Fig. 7
figure 7

Overview of the localization of miR164 expression during different stages of prickly pear cactus fruit development. Tissue print was performed for prickly pear cactus tissues at different developmental stages followed by northern blot hybridization. Scale bars indicate 1 cm

Full size image

Discussion

In plants, miRNAs participate in diverse biological processes, including fruit development (Moxon et al. 2008; Karlova et al. 2013). High-throughput sequencing to identify plant miRNAs has greatly advanced our knowledge about their functions. However, without the availability of a genome or transcriptome, the identification of miRNAs in non-model plant species using high-throughput sequencing strategies is challenging. Fortunately, a high number of identified miRNAs are conserved among plant species, and even their target genes are also conserved (Axtell and Bartel 2005). Based on these observations, we performed an expression analysis of selected miRNAs throughout four stages of fruit development in the cactus species Opuntia, followed by a detailed analysis of the spatial pattern of expression of miR164 during fruit development.

Prickly pear cactus contains at least 34 miRNA families

miRNA expression analysis revealed that at least 34 (64 %) out of 53 studied miRNAs are expressed during prickly pear cactus fruit development. This indicates that the genome of the cactus species Opuntia also encodes for conserved miRNAs and has at least 34 miRNAs. Furthermore, it suggests that the detected miRNAs may have a role in prickly pear cactus fruit development. 39 % of the detected miRNAs were present in all four analyzed prickly pear cactus stages and 32 % of the detected miRNAs appeared to be stage-specific (Fig. 3), suggesting both general and specific roles during fruit development, respectively. The lowest expression of miRNAs was detected in the ripe fruit, which could suggest that miRNAs are less important or that specific miRNAs, not included in this analysis, participate at this stage. Of the 34 detected miRNAs, 14 miRNAs are conserved with cucurbits (Jagadeeswaran et al. 2009), 18 miRNAs are conserved with tomato, 21 with chili (http://smallrna.udel.edu/; Chavez-Montes et al. 2014), and 24 miRNA families with grapevine (Pantaleo et al. 2010; Wang et al. 2012). Some miRNAs such as miR156, miR159, miR164, miR165, miR166, miR168, and miR390 are conserved between prickly pear cactus and all above-mentioned species, suggesting deeply conserved roles in these species. Based on phylogenetic analysis, grapevine is the closest related model species to prickly pear cactus (Tree of Life; http://www.tolweb.org), which is also reflected in the fact that prickly pear shares the highest miRNA family conservation with grapevine. Accordingly, it was previously reported that miR3636 and miR3635 were grapevine specific miRNAs (Pantaleo et al. 2010), but we detected them also in prickly pear cactus fruit (Fig. 3).

miRNA expression profiles during prickly pear cactus fruit development

Hierarchical cluster analysis of the detected miRNAs resulted in three groups: early developmental stage-associated miRNAs (Group I), middle late stage-associated (Group II), and late stage-associated miRNAs (Group III) (Fig. 4). This result supports the idea that the observed miRNA expression changes are not random, but that the miRNAs are highly regulated during fruit development, leading to specific target gene regulation that contributes to proper fruit development.

Group I contains seven miRNAs, which in general decrease their expression during fruit development. One of them is miR159, targeting GAMYB-related transcription factors involved in floral transition and anther development (Achard et al. 2004). A similar expression profile was reported in tomato for this miRNA that targets the ACS gene, which is involved in ethylene biosynthesis, a hormone important for fruit ripening (Zuo et al. 2012). Other members of this group are miR165/166 that regulate diverse aspects of plant development, including shoot apical meristem development, flower development, leaf polarity, and vascular development (Chitwood et al. 2007; Jung and Park 2007). Furthermore, several other miRNAs (miR397, miR398, and miR408) are involved in biotic and abiotic stress responses, and nutrient deprivation (Sunkar et al. 2007; Kruszka et al. 2012). The majority of these miRNAs showed low expression, as had been previously reported for other species (Pantaleo et al. 2010; Korir et al. 2013). Interestingly, high expression was observed for miR408, and this expression was specific to floral buds (Fig. 4), as reported for tomato (Moxon et al. 2008). In M. truncatula, this miRNA is expressed in flowers and fruits, and is upregulated in response to water deficit (Trindade et al. 2010). In poplar trees, miR408 expression is induced by tension and compression stresses in xylem tissues, suggesting that this miRNA has a critical role in the structural and mechanical fitness of woody plants (Lu et al. 2005). Furthermore, miR408 is upregulated under copper deficiency (Abdel-Ghany and Pilon 2008). In summary, the miRNAs present in this group are involved in general plant developmental processes as well as responses to stresses, which is quite interesting considering that a characteristic feature of cactus species is their capability to cope with environmental stresses.

Group II contains various moderately expressed miRNAs, with the highest expression in young fruit. miR824 targets AGL16 in Arabidopsis, which is involved in stomatal development (Kutter et al. 2007), and in strawberry is predicted to regulate ent-kaurene synthase, which is involved in gibberellin synthesis (Ge et al. 2013). miR828 targets various MYB transcription factors involved in anthocyanin biosynthesis, furthermore, it targets the Trans-Acting SiRNA Gene 4 (TAS4) (Luo et al. 2012), and it also has been reported to regulate the Ethylene-insensitive 2 (EIN2) gene in tomato (Zuo et al. 2012). miR390 also targets a TAS gene, TAS3. TAS3-derived siRNAs target the AUXIN RESPONSE FACTORS2 (ARF2), ARF3, and ARF4 (Yoon et al. 2010; Marin et al. 2010). In Arabidopsis, ARF3 is also known as ETTIN, and is important for gynoecium and fruit development (Sessions and Zambryski 1995).

Moreover, Group II contains miRNAs with low or undetectable expression in floral buds but with increased expression during fruit development. This subgroup includes miR156, miR164, and miR168. miR156/miR157 target the tomato transcription factor COLORLESS NON-RIPENING (CNR), a member of the squamosa-promoter binding protein (SBP) family, which plays a pivotal role in fruit ripening (Manning et al. 2006; Moxon et al. 2008). The overexpression of miR156 in tomato negatively affects the number and weight of fruits (Zhang et al. 2011). Another important regulator of ripening in tomato is APETALA2a (AP2a), which cross-talks with CNR (Karlova et al. 2011), and is negatively regulated by miR172 (Karlova et al. 2013). In our miRNA array experiment we did not detect miR172 expression, however, by northern blot hybridization we detected low expression in floral buds and floral buds after fertilization (Fig. 2). A similar pattern was reported in tomato, though still moderate expression was detected in developing fruits (Mohorianu et al. 2011; Lopez-Gomollon et al. 2012). Another known miRNA is miR164, which targets NAC domain transcription factors (Mallory et al. 2004) (further discussed below). miR164 showed high expression during fruit development in prickly pear cactus, as also reported for tomato. However, in tomato its expression starts at a low level and then increases during fruit development (Mohorianu et al. 2011). Furthermore, we also detected expression of miR168, which regulates ARGONAUTE1 (AGO1) that is involved in miRNA-directed transcript cleavage (Vaucheret et al. 2004). miR319 targets TCP (TEOSINTE BRANCHED/CYCLOIDEA/PROLIFERATING CELL FACTORS) transcription factors, which are involved in various developmental processes, like leaf morphogenesis and senescence (Palatnik et al. 2003), but also involved in salt and drought tolerance (Zhou et al. 2013).

Group III contains 11 miRNAs that have the highest expression at the green and ripe fruit stages. miR157 and miR169 are well-studied, which are involved in fruit ripening (Moxon et al. 2008), and miR169 also in stress responses such as those elicited by nutrient deficiencies, salt, and drought (Zhao et al. 2009, 2011; Ni et al. 2013). The reported expression pattern for miR169 in tomato shows lower expression in buds compared to developing fruits (Moxon et al. 2008). Furthermore, miR170 belongs to this group, and is known to target SCARECROW-LIKE (SCL) transcription factors (Rhoades et al. 2002). Notably, miR395 is known to be induced by sulfate deficiency (Kawashima et al. 2009). It will be interesting to investigate whether there is a connection with fruit ripening.

The correlation observed between miRNA expression profiles obtained in prickly pear cactus fruit development and in other fruit crops suggests that these miRNAs have a conserved role in controlling, or are being controlled by related events during fruit development. Furthermore, clustered miRNAs may be involved in the same or similar developmental processes, which will be interesting to further investigate.

miR164 plays different roles during prickly pear cactus fruit development

The first attributed function to miR164 was the regulation of the Arabidopsis CUC1 and CUC2 genes, which belong to the NAC transcription factor family and are involved in tissue differentiation (Laufs et al. 2004; Nikovics et al. 2006; Sieber et al. 2007). Furthermore, it has been reported that miR164 also targets ORESARA1 (ORE1), which is involved in the regulation of aging-induced cell death and senescence (Kim et al. 2009). During prickly pear cactus fruit development, miR164 expression was observed at meristematic zones in floral buds, in open flowers at places of organ fusion (carpel base with stamens and receptacle), and homogenously in the complete ripening fruit. These results suggest that miR164 is highly regulated and furthermore, that it probably performs different functions during fruit development by regulating different NAC transcription factors (Karlova et al. 2013). In addition to boundary morphogenesis and leaf senescence, NAC genes are involved in different developmental processes such as biotic and abiotic stress responses and ripening (Delessert et al. 2005; He et al. 2005; Liu et al. 2009; Greco et al. 2012).

Figure 8 shows a schematic representation of the miR164 spatio-temporal expression during prickly pear cactus fruit development. We speculate that miR164 has three functions during prickly pear cactus fruit development: (1) in early stages miR164 may be associated with regulation of CUC-like genes in meristematic regions, a role demonstrated in other species (Válóczi et al. 2006; Peaucelle et al. 2007; Adam et al. 2011), (2) before fertilization, high expression in the compitum and carpel base might suggest that miR164 prevents premature abscission of the style, stamens, and petals, and (3) high miR164 expression in the complete fruit could be involved in controlling aging-induced cell death, ripening, and senescence.

Fig. 8
figure 8

Schematic representation of the miR164 spatio-temporal expression pattern during prickly pear cactus fruit development. miR164 expression, based on tissue printing and northern blot hybridizations, is indicated by red coloring. The color intensity represents miR164 expression levels, with intense color corresponding to high expression. Scale bars indicate 1 cm

Full size image

Conclusions

To our knowledge, this is the first study that detected expression of conserved miRNAs during prickly pear cactus fruit development. The expressed miRNAs revealed dynamic expression patterns through prickly pear cactus fruit development and could be clustered in three groups, suggesting involvement of miRNAs in different processes. Detailed analysis of the spatio-temporal expression of miR164 suggests that it participates in different processes during prickly pear cactus fruit development. Based on these results, we suggest that the conserved miR164 is a regulator of fleshy fruit development.

Author contribution

FFRC performed most of the experimental work. JCP and XGR performed the stem-loop RT-PCR and the microarray hybridization experiment. FFRC, NMM, ACH, and SDF conceived the project and designed the experiments. FFRC, NMM, and SDF drafted the manuscript. All authors read and approved the final manuscript.