SEARCH

SEARCH BY CITATION

Keywords:

  • miRNA;
  • microarray;
  • Notch;
  • Hedgehog;
  • cyclopamine;
  • DAPT

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

microRNAs (miRNAs) are small (∼22 nucleotide) non-coding RNAs that regulate gene expression at the post-transcriptional level, typically by inhibiting translation. The genes encoding these small RNAs are estimated to comprise approximately 2–3% of animal genomes yet potentially regulate a majority of protein-coding genes including those involved in cell specification and development. A key remaining question is to identify target mRNAs regulated by microRNAs. As a means to identify potential targets, we designed a sensitive microarray to analyze global miRNA expression patterns at twelve developmental stages in zebrafish. Further, we conducted arrays on zebrafish embryos treated with small molecule inhibitors of the Hedgehog and Notch signaling pathways to enable identification of differentially expressed miRNAs that target genes controlling key developmental pathways during early embryogenesis. Developmental Dynamics 236:2172–2180, 2007. © 2007 Wiley-Liss, Inc.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

microRNAs are small (∼22nt), non-coding RNA molecules that show remarkable conservation among eukaryotes. Direct cloning strategies and bioinformatic prediction based on the presence of conserved hairpin structures and sequences have suggested that animal genomes encode hundreds, perhaps thousands, of microRNAs (Lagos-Quintana et al.,2001; Lau et al.,2001; Lee and Ambros,2001; Lai et al.,2003; Lim et al.,2003a; Lim et al.,2003b; Kim and Nam,2006). These small RNAs regulate early development, cell specification, differentiation, and proliferation (Lee et al.,1993; Pasquinelli et al.,2000; Reinhart et al.,2000; Ambros,2003; Bartel,2004; Griffiths-Jones,2004; Hobert,2004; Du and Zamore,2005; Hatfield et al.,2005; Morris and McManus,2005; O'Donnell et al.,2005; Carthew,2006; Esquela-Kerscher and Slack,2006; Naguibneva et al.,2006; Plasterk,2006; Slack and Weidhaas,2006; Voorhoeve et al.,2006).

Primary microRNA transcripts are synthesized from heterogeneous loci by RNA polymerase II, capped, and polyadenylated (Lee et al.,2004). Approximately 50% are encoded within the introns of other genes (Rodriguez et al.,2004) whereas the remainder are synthesized as either mono- or poly-cistronic transcripts (Kim and Nam,2006). Primary transcripts (pri-miRNAs) are first cleaved by the RNase III-like enzyme, Drosha, into ∼70 nt hairpins (pre-miRNA) that are subsequently exported from the nucleus via Exportin 5 (Yi et al.,2003; Lund et al.,2004). Once in the cytoplasm, another RNase III enzyme, Dicer, cleaves pre-miRNAs into mature miRNA duplexes, one strand of which is then unwound and incorporated into RNA Induced Silencing Complexes (RISC), the formation of which appears to be linked to target identification and pairing (Gregory et al.,2005; Matranga et al.,2005; Rand et al.,2005).

A central problem toward complete understanding of miRNA function is identifying the target genes regulated by individual miRNAs (Enright et al.,2003; Brennecke et al.,2005; Krek et al.,2005). Most miRNAs do not pair with perfect complementarity to their targets such that bioinformatic prediction is difficult and experimental validation is required. As a first step toward target identification, global miRNA expression patterns are needed, both temporal and spatial (Wienholds et al.,2005). Integrating miRNA and target expression data at specific stages of development will help to refine lists of possible targets for specific miRNAs. Here, we designed a sensitive (∼0.1–0.7 fmol) microarray to expand expression analysis to 346 vertebrate miRNAs. We utilized this array to study differences in miRNA expression during normal zebrafish development as well as in embryos treated with DAPT (N-[N-(3,5-difluorophenacetyl)-L-alanyl]-S-phenylgly cine t-butyl ester) and cyclopamine, pharmacologic inhibitors of the Notch and Hedgehog signaling pathways, respectively.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Microarray Design and Verification

Using sequences from the miRNA Registry as well as published reports (Lim et al.,2003a; Griffiths-Jones,2004), we designed DNA oligonucleotides containing 2 complete regions of complementary to 346 known and predicted vertebrate miRNAs (see Supplemental Fig. 1A,B, which can be viewed at www.interscience.wiley.com/jpages/1058-8388/suppmat). DNAs were synthesized containing C6 amine modified amino termini for coupling to slides. As controls, 6 positive and negative control oligonucleotides were included, complementary to zebrafish mRNAs (APC, b-actin, and b-catenin), zebrafish 18S rRNA, tomato lycopene synthase, and a bacterial dehydrogenase (PsA-NAD-dehyd). All oligonucleotides were spotted in duplicate or triplicate on glass mirror slides (Genexprex) by a 16-bit spotter in the Vanderbilt Microarray Shared Resource facility.

miRNAs are estimated to account for approximately 1–2% of cellular RNA (Lau et al.,2001). To enrich for microRNAs from zebrafish embryos, total RNA was isolated using TRI reagent and then passed over mirVana miRNA isolation kits (Ambion) or sucrose gradients (Supplemental Fig. 1C). Small RNAs were fluorescently labeled by attachment of Cy5 to the N7 position on G residues using LabelIT (Mirus) (Supplemental Fig. 2). This leaves all base pairs that participate in hydrogen bonding unaffected, and since only three vertebrate miRNAs are known to lack even a single G residue (miR-197, miR-467a, and miR-467b), the majority of miRNAs should be labeled. Fluorescently labeled RNAs were then hybridized to array slides. For analysis of miRNA expression, local background fluorescence levels were subtracted as well as signals from negative control spots containing oligonucleotides complementary to a bacterial dehydrogenase (PsA-NAD-dehyd). Subsequently, signal intensities for individual miRNAs were determined at multiple RNA concentrations. Signals were too close to background to provide reliable data until 50–100ng of purified RNA were used for labeling whereas signal intensities appeared to plateau at concentrations above 4 μg (Supplemental Fig. 2C). Based on these results, we used 2 μg of labeled purified small RNA for all subsequent arrays.

Microarray sensitivity was determined by spiking hybridizations with exogenously added, fluorescently labeled 22nt RNAs complementary to tomato lycopene synthase mRNA. Background and negative control fluorescence intensities were again subtracted and the resulting signals were quantified and plotted (Supplemental Fig. 2A,B). Dose-dependent increases in fluorescent signals were observed but, importantly, as little as 1 pg of lycopene synthase RNA could be detected at levels sufficiently above background. This suggests that the array is able to detect individual miRNAs within small RNA pools down to 0.1–0.7 fmols (1–5 pg).

To analyze specificity, we utilized oligonucleotide probes that contained 1, 2, or 6 nucleotide mismatches. When probes containing 2 or 6 mismatches were tested, little or no hybridization could be detected due to the altered sequence (Supplemental Fig. 2D–F). For single nucleotide changes, expression patterns from different members of the let-7 family were compared where let-7a, c, f, and g all differ by just one nucleotide (Supplemental Fig. 2D). Completely different patterns of expression between these let-7 members were detected during zebrafish development (Supplemental Fig. 2D,F). If cross-hybridization was occurring, similar, if not identical, expression patterns would have been observed.

Developmental Expression of miRNAs

Using the methodology described above, we analyzed global miRNA expression patterns by isolating total RNA from 12 specific stages of zebrafish development and carrying out hybridizations at each time point (Fig. 1). Local background and negative control signals were subtracted as above and the data from each stage were obtained from 3 independent hybridizations. As shown in Figures 1 and 2, the overall pattern of miRNA expression increased during development. Such changes are indicated by an increase in red color as development proceeds. This agrees with the hypothesis that miRNAs play an important role in differentiation and development (Chen et al.,2005; Wienholds et al.,2005; Giraldez et al.,2006; Naguibneva et al.,2006; Schratt et al.,2006; Voorhoeve et al.,2006). The one notable exception to relatively limited expression of miRNAs during early development is the pattern observed at sphere stage embryos where there appears to be significantly greater expression compared to other early time points (Fig. 1). Sphere stage coincides with the mid-blastula transition signaling the beginning of zygotic transcription and coincides with silencing or degradation of maternal RNAs in the zygote. The miR-430 family has been shown to be involved in this process (Giraldez et al.,2006).

thumbnail image

Figure 1. Wild-type zebrafish developmental array. Small RNAs were isolated from 12 different stages during zebrafish embryonic development, fluorescently labeled, and hybridized to an array containing 346 miRNAs. The increase in expression across developmental stages is shown with a blow-up of the region containing miRs-34a, -27b, -129a, -20, -206 to illustrate individual miRNAs. Blue indicates background with red indicating high expression.

Download figure to PowerPoint

thumbnail image

Figure 2. miRNA expression patterns during zebrafish development. Heat maps as in Figure 1 illustrate the changes in expression for individual miRNAs, denoted at the right.

Download figure to PowerPoint

The array shown in Figures 1 and 2 analyzed 346 known miRNAs. To verify the array results, Northern blots were performed using probes against 3 different miRNAs at 5 different stages of development (Supplemental Fig. 3). As expected, the array was more sensitive but, importantly, the two different techniques showed very similar results. Previous analyses of miRNA expression during zebrafish development examined 154 different miRNAs (Wienholds et al.,2005). Here we expand the analysis to 346 miRNAs and analyze expression at specific developmental stages. A consistent difference between the data shown in Figures 1 and 2 and previous work is our finding that a significant number of miRNAs were detected at sphere stage. One potential explanation could be related to developmental timing as the previous reports harvested RNA based on hours post fertilization whereas we specifically isolated sphere stage embryos.

miRNA Expression and Hedgehog Signaling

Beyond adding to the number of vertebrate miRNAs examined across different stages of development, we also sought to determine expression patterns under conditions where developmentally important signaling pathways were disrupted. For this, we used cyclopamine and DAPT to inhibit the Hedgehog and Notch signaling pathways, respectively.

Hedgehog gradients are necessary to establish proper cell specification during embryogenesis (Krauss et al.,1993; Currie and Ingham,1996; Chen et al.,2001; Lewis and Eisen,2001; Varga et al.,2001). In vertebrates, Hedgehog binds to Patched, which inhibits Smoothened and allows the activation of Gli proteins. The absence of Hedgehog allows Smoothened to be internalized and promotes Gli3 processing and transcriptional repression of downstream genes. Cyclopamine inhibits the Hedgehog signaling pathway by binding to Smoothened inhibiting its localization to cilia structures (Chen et al.,2002; Corbit et al.,2005). In order to study the effects of inhibition of Hedgehog signaling on miRNA expression during somite and neuronal development, we added cyclopamine to zebrafish embryos at 90% epiboly, RNA was isolated beginning approximately 1 hr later at the tailbud stage and then continued through 1 dpf. Microarrays were performed across these time points and Figure 3 and Supplemental Figure 4 show the detailed expression data.

thumbnail image

Figure 3. Differential miRNA expression due to inhibition of hedgehog signaling. Cyclopamine was added at 90% epiboly and RNA was extracted at the indicated stages, fluorescently labeled, and hybridized to an array contain 346 miRNAs. A blow-up of several differentially expressed miRNAs is shown below. Complete expression analysis for individual miRNAs is included as Supplemental Figure 4.

Download figure to PowerPoint

In order to examine changes in miRNA expression that are directly related to the inhibition of the Hedgehog pathway rather than possible indirect downstream effects, we narrowed our analysis to the tailbud stage. Fold changes were calculated by dividing expression levels after treatment with cyclopamine with wild-type values from untreated tailbud stage embryos. Table 1 shows the most dramatic up- and down-regulated miRNAs as well as a partial list of predicted targets for those miRNAs. Interestingly, the number of up-regulated miRNAs was much greater than those down-regulated.

Table 1. miRNA Expression Changes After Cyclopamine Treatmenta
NameFold ChangePPredicted targets
  • a

    Fold changes associated with cyclopamine addition to wild-type embryos are listed with potential targets as predicted using miRanda and miRBase.

miR-214−1.600.0367klf2, mmp14a-b, hes5, sufu, disp2, nrarpa, plplb, hoxbla
miR-380-3p21.290.0154mmp12, geminin, or216, nox4, wisp3, ul, col8al, eif4ebp3
miR-42921.340.0841rab24, nog, tubb, o5le1, mam12
miR-20321.470.0003nusap1, mmp2, myoD, nrf1, pax6a, ppab, fgf18, six4.1, her4-9, fhl, wnt8, pin
miR-37921.500.0162exp1, mapk15, rab30, fgf17, wisp2, nrf1, fgfrs3
miR-130b21.610.0001hoxd3a-2b, or5.3-9.2, myog, mmp14b, afgf, cry-dash, six4.1, nkx2.2, notch1b
miR-8621.630.0000clk2, ptr14, hox6, rcor1, plc, runx3
miR-37321.660.0001egfrk, myohd1, casp6, ipp, srp54, foxq1
miR-34c21.940.0001dll1, dcp1a, aqp8, ubp1, myf5
miR-38122.120.0000eif3s10, smarcb1, maf1, sfrp2, rhoH
miR-41023.960.0677ptx3, cdc16, tbx5, stab2, hoxdl1
miR-41224.050.0049stab2, fgfr2, pcfl1, igf2r, eif5A, wnt6
miR-37425.220.0600msx2, noggin, slit3, rab10, tbp1, igfbp-1, sox6
miR-8426.290.0000clk2, css2, h1.1, nubp1, runx3, fzd3

miRNA Expression and Notch Signaling

A second major signal transduction pathway known to control multiple steps during early embryogenesis is the Notch signaling cascade. Notch signaling is required to establish proper neuronal specification and somitogenesis in developing embryos (Blader et al.,1997; Dornseifer et al.,1997; Appel and Eisen,1998; Haddon et al.,1998; Takke and Campos-Ortega,1999; Takke et al.,1999). Interaction between the Notch receptor and any of several ligands promotes cleavage mediated by the γ-secretase complex. The Notch intracellular domain then translocates to the nucleus where it mediates transcriptional activation or repression of target genes. The role of Notch signaling in establishing neuronal outgrowth and differentiation is fairly well characterized (Teo et al.,2005; Salama-Cohen et al.,2006), whereas the involvement in somite formation is less well known. Notch signaling has been implicated in proper polarization of developing embryos with a loss of somite boundaries being associated with Notch mutants (Holley,2002).

DAPT is a specific γ-secretase inhibitor that prevents cleavage of the Notch intracellular domain. We, therefore, treated wild-type zebrafish embryos with DAPT at the 90% epiboly stage, and again isolated RNA beginning 1 hr later at the tailbud stage, continuing through 1 dpf. Figure 4 and Supplemental Figure 5 show the detailed array data. As above, fold changes were determined and Table 2 shows those miRNAs whose expression was most dramatically up- or down-regulated as well as a partial list of predicted targets for each miRNA.

thumbnail image

Figure 4. Differential miRNA expression due to inhibition of notch signaling. DAPT was added at 90% epiboly and RNA was extracted at the indicated stages, fluorescently labeled, and hybridized to an array contain 346 miRNAs. A blow-up of several differentially expressed miRNAs is shown below. Complete expression analysis for individual miRNAs is included as Supplemental Figure 5.

Download figure to PowerPoint

Table 2. miRNA Expression Changes After DAPT Treatmenta
NameFold changePPredicted targets
  • a

    Fold changes associated with DAPT treated wild-type embryos are listed with potential targets as predicted using miRanda and miRBase.

miR-214−12.760.0303klf2, mmp14a-b, hes5, sufu, disp2, nrarpa, plplb, hoxbla
miR-430b−8.320.0646cap1, bambi, twist1, mycn, map4k5, lhx8, klf4, btg4
miR-207−5.980.0721lass4, prp19, tubb, srf, hsp70, efs, hoxc10, stat6
miR-197−5.220.0686Rac3, cilp, drosha, egfr, atp6v0a2, nkx2.8-9
miR-328−5.060.0418tusc4, chrnb2, foxa2, aqp10, nxph3, mapk15
miR-343−4.920.0848sema3fa, kcnf1, aebp1, U1, ctsz, pla2g4f
miR-133a−3.880.0667colla2, elf2, fgf18, lp1, calm2d, ccn, six3a, rab5c, snai1a, ptc2, mmp14b, ngfb
miR-19a3.330.0158lrp2, epn2, ptger3, wnt1, hoxd1, nav3, ci1p, mbd4, ireb2
miR-29a3.370.0028solo, bgl3, gpm6ab, lmnb1, colla2, tp731, cav1, mmp14a
miR-2a3.420.0003msxd, E(spl), mcd31, nipsnap1, mcm10, bwa
miR-29b3.620.0001lin7a, hbfg1, piwil1, colla2, gbp, rab5c, vtg1, gpm6ab, mmp14a
miR-603.650.0017mcm7, dsl3, cif5, plpc1, ins1, dvl2-3
miR-877.070.0000ptc2, spare, wnt2, smn1
let-7d10.820.0168punc, hoxc6a, cmyc, foxh1, crx, dvr1rpb, snai1a, otx2
miR-32620.310.0873myl3, mapk15, pla2, smad14, pax8, srp68

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Deciphering global miRNA expression patterns during development is a necessary first step toward identifying functional miRNA targets. Temporal expression patterns can rapidly be determined using microarray approaches and coupling such assays with the ability to analyze spatial patterns using in situ hybridization (Wienholds et al.,2005) makes zebrafish a powerful model system to examine miRNA expression. Here, we used direct labeling of small RNAs to examine the expression of 346 miRNAs during zebrafish development. Our array was found to be sensitive to as little as 1 pg of miRNA and was able to distinguish between single base mismatches. This provides a powerful approach to examine overall patterns of miRNA expression in wild type or mutant embryos as well as embryos exposed to small molecules, drugs, or other environmental stimuli. While the approach has proven quite sensitive, microRNAs that are expressed in single cells, such as lsy-6 in C. elegans (Johnston and Hobert,2003), will likely go undetected due to a low abundance in whole embryo RNA preparations.

miRNA Expression During Zebrafish Development

The overall profile of miRNA expression during normal development is shown in Figures 1 and 2 and in general agrees with previous work showing that miRNA expression patterns become more complex as development proceeds (Chen et al.,2005; Wienholds et al.,2005). One difference over and above the fact that we expanded the analysis to 346 miRNAs, is that we observed significantly higher expression of miRNAs during sphere stage embryos. In agreement with previous work, the miR-430 family is highly expressed at this early stage where they function to mediate targeted clearance of maternal mRNAs at the mid-blastula transition when zygotic transcription begins (Giraldez et al.,2006). However, we also observed significant expression of many more microRNAs at this time. One explanation is that we developmentally staged embryos prior to RNA isolation whereas previous analyses likely had a slightly asynchronous population since selection was based entirely on hours post fertilization. It is possible that a narrow window exists at exactly the midbastula transition when many genes are initially transcribed even if they then become downregulated until later development. Previous work has suggested that assembly of chromatin versus association of the basal transcription machinery limits transcription prior to the midbastula transition but that specific transcription activators can still access the DNA before this time (Prioleau,1995). After the midblastula transition, the basal transcription machinery joins to activate transcription but subsequent production of other regulators could then establish developmentally regulated transcriptional programs.

Besides the sphere stage, expression patterns detected using our array are mostly in agreement with previous cloning and microarray approaches showing limited early miRNA expression, which increases upon organogenesis and development into adulthood with terminal differentiation of multiple cell types (Chen et al.,2005; Wienholds et al.,2005). For example, muscle-specific miR-206 is initially expressed during somite development and then persists through adulthood. Likewise, miR-181a begins expression early before localization to the nose and eyes in adult fish (Wienholds et al.,2005). However, notable differences were also observed between the approaches as exemplified by analysis of the let-7 family of miRNA where timing differences were noted that are most likely due to differences in specificity since our array was able to detect single base mismatches.

miRNA Expression Following Inhibition of Hedgehog and Notch Signaling

Besides providing a resource for others to identify vertebrate microRNAs whose expression profile might be consistent with regulation of predicted targets, we sought to identify those miRNAs whose expression changes upon inhibition of the Notch and Hedgehog pathways. Given the importance of these two pathways for overall development, we identified a series of microRNAs that are both up- and down-regulated upon treatment with either DAPT or cyclopamine. Treatment with cyclopamine, which blocks the Hedgehog pathway, changed the expression of several miRNAs whose predicted targets are genes known to be involved in Hedgehog signaling (Table 1). For instance, miR-203, which was up-regulated over 20-fold, is predicted to target several brain morphogenic markers that are known to be regulated by Hedgehog signaling including pax6a and six4.1. Likewise, miR-130b is predicted to target nkx2.2 and hoxa2b while let-7e is predicted to target eng2a, foxh1, foxb1.1, and snai1a.

Notch signaling is crucial for both neuronal differentiation and somite development. As shown in Table 2, many of the predicted targets of miRNAs whose expression changes upon exposure to the Notch inhibitor DAPT are involved in exactly these processes and further localize to the brain and neural crest (Wienholds et al.,2005). Specific examples of Notch signaling components that are predicted to be targets of one or more of the differentially expressed miRNAs shown in Figurer 3 and Table 2 include sema3fa (miR-204), which is necessary for neural crest migration (Gammill et al.,2006) as well as the glycoprotein gpm6ab (miR-29b), matrix metalloproteinase mmp14a (miR-29a), neuroepithelial polarity protein lin7a (miR-296), and the midbrain nucleolar protein, midnolin (miR-155).

It will be informative to determine if there is overlap between spatial expression of Hedgehog- or Notch-dependent cell types and the miRNAs identified above. One possibility is that there may be little overlap such that those microRNAs that are up-regulated upon inhibition of either pathway may well serve to ensure precise gene expression by eliminating inappropriately expressed transcripts. This mechanism is consistent with the micromanager model of miRNA action (Bartel and Chen,2004).

Cross-talk and Cross-Regulation of the Notch and Hedgehog Pathways

The Notch and Hedgehog pathways do not act independently during development. Multiple developmental events require both signaling cascades. For example, both pathways are involved in oligodendrocyte development and muscle formation (Cornell and Eisen,2000,2002; Scheer et al.,2001; Park et al.,2002). Neuronal differentiation of motor neurons and oligodendrocytes requires Olig2 expression in the neural plate and Olig2 expression requires both ngn1, a Notch activator, as well as intact Hedgehog signaling (Park et al.,2002). Somite development requires the Notch pathway and proper slow muscle formation requires Hedgehog signaling. To identify key miRNAs that might coordinately regulate both pathways, we looked for miRNAs with predicted targets known to be involved in both pathways and whose expression changes were altered by treatment with both inhibitors. We grouped these miRNAs into subgroups that show those miRNAs whose expression is up-regulated with both treatments, those showing down regulation with both treatments, and those that increase with cyclopamine but decrease with DAPT (Table 3). Future work will be directed toward validation of such predicted targets (see Tables 1–3).

Table 3. Cross-Regulation of the Hedgehog and Notch Signaling Pathways
NameCyclopamine fold changePDAPT fold changePCyclopamine targetsDAPT targets
  1. amiRNAs whose expression was altered by treatment with both DAPT and cyclopamine at the tailbud stage are shown with potential targets as predicted using miRanda and miRBase.

miR-34a14.850.01781.590.0508gli2neurl2
miR-27b8.130.02411.290.0391ptc1notch1b
miR-20a11.721.24E-041.530.0056sufunotch1b
miR-2067.780.0331−1.70.084sufunotch1b
miR-214−1.60.0367−12.760.0303sufunrarpa

EXPERIMENTAL PROCEDURES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Microarrays and Target Prediction

All DNA oligonucleotides were purchased from MWG Biotech and printed on mirror slides (Genexprex) in duplicate or triplicate with two complete arrays per slide. Slides were blocked with a quick wash in 0.2% SDS, a 45-min incubation at 55°C in blocking buffer (5× SSC, 1% BSA, 1% SDS), rinsed with deionized water, and spin-dried. Hybridizations were carried out in 25% deionized formamide, 5× SSC, 0.1% SDS for exactly 16 hr in ArrayIt® hybridization chambers followed by three successive washes (3 min each) in 2× SSC, 0.1% SDS; 1× SSC; and 0.1× SSC. All microarrays were scanned with a GenePix4000B scanner and data were analyzed with GeneSpring 7.0 software. All time points were performed in triplicate. For normalization, hybridizations were spiked with identical amounts of tomato lycopene synthase RNA. Arrays with significant background and/or insufficient lycopene synthase levels were discarded. Predicted mRNA targets were obtained by combining a variety of currently available prediction algorithms (miRanda, miRTar, miRBase). For a complete listing of all oligonucleotide sequences and raw array data gathered during wild type zebrafish development, see Supplemental Tables 1 and 2.

miRNA Isolation

Total RNA was isolated from zebrafish embryos using TRI Reagent (Molecular Research Center). Small RNAs were isolated using 15–45% sucrose gradients or by passage over mirVana miRNA isolation kits (Ambion).

RNA Labeling

For Cy5 labeling, 2.4 μg of small RNA were labeled in a 1:1 (RNA:Cy5) ratio using LabelIT (Mirus) for 1 hr at 37°C. Unincorporated Cy5 was removed by passage over nucleotide removal kits (Qiagen) and diluted to 2 μg per array after purification. Synthetic lycopene synthase RNA was labeled and purified using the above process and was diluted to 50 pg per array.

Northern Blots

RNAs were separated on 12% acrylamide gels and electroblotted to positively charged nylon membranes. DNA oligonucleotides complementary to specific miRNAs were labeled with α-32P-dATP using StarFire labeling kits (IDT). Hybridizations were carried out in 7% SDS and 0.2 M NaPO4, pH 7.2 for 12 hr followed by washes in 2×SSPE-0.1%SDS, 1×SSPE-0.1%SDS, and 0.5×SSPE-0.1%SDS.

Cyclopamine and DAPT Treatments

Cyclopamine and DAPT were added to wild-type embryos at 90% epiboly and analyzed at tailbud, 6-somite, 18-somite, and 1dpf. Final concentrations of cyclopamine and DAPT were 15 and 100 μM, respectively (Wolff et al.,2003; Bae et al.,2005). For a complete listing of all raw array data for both cyclopamine and DAPT treatments plus an expanded table showing P values for fold changes under both conditions, see Supplemental Tables 3–12.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

The Vanderbilt Microarray Resource Core printed the microarrays and provided software and assistance for data collection. We thank Ima Paydar for help with RNA isolation. This work was supported by NIH GM-075790. E.J.T. was supported in part by GM62758 and A.S.F. received support from T32 GM 008554.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

The Supplementary Material referred to in this article can be viewed at www.interscience.wiley.com/jpages/1058-8388/suppmat

FilenameFormatSizeDescription
jws-dvdy.21211.fig1.tif147K Supplemental Figure 1. Microarray Probe Design and RNA PurificationA) Oligonucleotide probes containing 2 (44 nucleotides) or 3 (66 nucleotides) regions of complementarity to eithermiR-10b , ormiR-124awere coupled to mirror slides for initial microarray hybridization experiments. Individual antisense regions are indicated in green, red, or blue. B) Total RNA from zebrafish embryos was fractionated using either MirVana isolation kits or sucrose gradients. RNAs were separated on 12% 19:1 polyacrylamide gels and stained with ethidium bromide. MirVana fractions (Lanes 1, 2) and the top three sucrose gradient fractions (Lanes 3-5) are shown. Northern blot of the same RNAs in B using probes againstmiR-124a . C) Small RNAs isolated from 2 days post fertilization (2 dpf) zebrafish embryos were fluorescently labeled and hybridized to small scale arrays containing the probes described in A and pixel intensities were measured. Despite the increase in base pairing, no significant difference in fluorescence intensity was observed between the long and short oligonucleotides so that final arrays were printed with 2 regions of complementarity.
jws-dvdy.21211.fig2.tif214K Supplemental Figure 2. Microarray Sensitivity and SpecificityA,B) Sensitivity. Increasing amounts of a 22 nucleotide RNA encoding tomato lycopene synthase were spiked into microarray hybridizations and signal intensities were determined at each concentration after background and negative control subtraction. Values are shown in graphical (a) and heat map (b) format. As little as 1pg could be detected well above background. C) RNA Labeling. To optimize for the levels of RNA needed in each hybridization, signal intensities from 9 different miRNAs were determined after microarrays performed using from 0.1 to 4000ng labeled RNA. APC is a probe complementary to the zebrafish adenomatous polyposis coli mRNA. Based on these results, we chose to use 2ug for all subsequent arrays. D) Specificity. Single nucleotide specificity using Let-7 family members and mismatch controls. Four members of the let-7 miRNA family are shown with indicated single nucleotide differences. Two and six mismatches were also incorporated intomiR-124ashort oligonucleotides. E, F) Heat map representation of signal intensities for the miRNAs shown in D. Specificity is demonstrated by the lack of cross-hybridization between probes (Red: high expression values; Blue: low to zero expression values).
jws-dvdy.21211.fig3.tif308K Supplemental Figure 3. Northern blot VerificationMicroarray results were verified for three microRNAs using Northern blots.
jws-dvdy.21211.fig4.tif430K Supplemental Figure 4. Detailed Cyclopamine Expression Profiles Detailed view of expression analysis shown in Figure 3. Blue indicates little to no expression and red indicates high expression.
jws-dvdy.21211.fig5.tif427K Supplemental Figure 5. Detailed DAPT Expression ProfilesDetailed view of expression analysis shown in Figure 4. Blue indicates little to no expression and red indicates high expression.
jws-dvdy.21211.tbl1.xls186K Supplemental Table 1. Detailed Array DesignExact array design and probe sequences for all spots on the array as per MIAME guidelines.
jws-dvdy.21211.tbl2.xls1472K Supplemental Table 2. Raw Array Data During Wild Type Zebrafish DevelopmentF635 Median is the median intensity level of the signal. B635 Median is the median intensity level of the background surrounding the probe. A flag of -50 means that the signal could not easily be distinguished from background, and a flag of 0 means that the signal is present. Each stage is presented on a different sheet within the same excel workbook.
jws-dvdy.21211.tbl3.xls720K Supplemental Table 3. Raw Array Data After Cyclopamine TreatmentF635 Median is the median intensity level of the signal. B635 Median is the median intensity level of the background surrounding the probe. A flag of -50 means that the signal could not easily be distinguished from background, and a flag of 0 means that the signal is present. Each stage is presented on a different sheet within the same excel workbook.
jws-dvdy.21211.tbl4.xls710K Supplemental Table 4. Raw Array Data After DAPT TreatmentF635 Median is the median intensity level of the signal. B635 Median is the median intensity level of the background surrounding the probe. A flag of -50 means that the signal could not easily be distinguished from background, and a flag of 0 means that the signal is present. Each stage is presented on a different sheet within the same excel workbook.
jws-dvdy.21211.tbl5.xls34K Supplemental Table 5. Fold changes after cyclopamine treatment at tailbud stageEmbryos were treated with cyclopamine and microarray signals were determined at the tailbud stage and compared to wild type signals at the same stage. The fold change is as indicated for each microRNA with p-values as indicated.
jws-dvdy.21211.tbl6.xls4K Supplemental Table 6. Fold changes after cyclopamine treatment at 6 somite stageEmbryos were treated with cyclopamine and microarray signals were determined at the 6 somite stage and compared to wild type signals at the same stage. The fold change is as indicated for each microRNA. Since the effects of cyclopamine at later stages of development could be indirect, p-values were not determined for the 6 somite stage.
jws-dvdy.21211.tbl7.xls5K Supplemental Table 7. Fold changes after cyclopamine treatment at 18 somite stageEmbryos were treated with cyclopamine and microarray signals were determined at the 18 somite stage and compared to wild type signals at the same stage. The fold change is as indicated for each microRNA. Since the effects of cyclopamine at later stages of development could be indirect, p-values were not determined for the 18 somite stage.
jws-dvdy.21211.tbl8.xls5K Supplemental Table 8. Fold changes after cyclopamine treatment at 1 dpfEmbryos were treated with cyclopamine and microarray signals were determined at 1dpf and compared to wild type signals at the same stage. The fold change is as indicated for each microRNA. Since the effects of cyclopamine at later stages of development could be indirect, pvalues were not determined for 1 dpf.
jws-dvdy.21211.tbl9.xls35K Supplemental Table 9. Fold changes after DAPT treatment at tailbud stageEmbryos were treated with DAPT and microarray signals were determined at the tailbud stage and compared to wild type signals at the same stage. The fold change is as indicated for each microRNA with p-values as indicated.
jws-dvdy.21211.tbl10.xls4K Supplemental Table 10. Fold changes after DAPT treatment at 6 somite stageEmbryos were treated with DAPT and microarray signals were determined at the 6 somite stage and compared to wild type signals at the same stage. The fold change is as indicated for each microRNA. Since the effects of DAPT at later stages of development could be indirect, p-values were not determined for the 6 somite stage.
jws-dvdy.21211.tbl11.xls4K Supplemental Table 11. Fold changes after DAPT treatment at 18 somite stageEmbryos were treated with DAPT and microarray signals were determined at the 18 somite stage and compared to wild type signals at the same stage. The fold change is as indicated for each microRNA. Since the effects of DAPT at later stages of development could be indirect, p-values were not determined for the 18 somite stage.
jws-dvdy.21211.tbl12.xls5K Supplemental Table 12. Fold changes after DAPT treatment at 1 dpfEmbryos were treated with DAPT and microarray signals were determined at 1dpf and compared to wild type signals at the same stage. The fold change is as indicated for each microRNA. Since the effects of DAPT at later stages of development could be indirect, p-values were not determined for 1 dpf.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.