Download Supplemental Material 1 Simultaneous isolation of mRNA, miRNA

Supplemental Material 1
Simultaneous isolation of mRNA, miRNA and proteins from tissues
mRNA and proteins were isolated from frozen tissue and select cell lines (wild-type SMCs
and DICER-/-) using an AllPrep® DNA/RNA/Protein Mini Kit (QIAGEN AG, Basel,
Switzerland) according to the published protocol. Enriched miRNAs were extracted and
purified using the mirVana miRNA Isolation Kit (Ambion, Austin, TX), and their
concentration and quality were determined with the NanoDrop ND-1000 spectrophotometer
(Thermo Scientific, Rockland, IL). The quality of RNAs was determined by the Agilent
BioAnalyzer 2100 (Agilent Technologies, Santa Clara, CA).
IPA
To better understand the biological function and/or diseases that were most relevant to the
data sets and facilitate understanding beyond a functional link to intracranial aneurysms (IA),
ingenuity pathway analysis (IPA) was used to compare different types of cellular interactions,
including gene-gene interactions and gene-miRNA interactions, with the genes analyzed
above. IPA calculates the significance value of a given canonical pathway or biological
function as the probability that the pathway or function is associated with the dataset by
random chance. The dataset containing miRNA gene identifiers and corresponding fold
changes was uploaded into the Pathway Studio, and each gene identifier was mapped to its
corresponding gene object. The p value was calculated using the right-tailed
Benjamini-Hochberg multiple testing correction with values of p < 0.05 considered
significant. The “Core, Metabolomics Comparison Analysis and Path eExplorer” option of
IPA was used to connect all of the identified proteins, mRNA and miRNAs associated with IA
and its angiogenesis.
Double immunofluorescence and confocal analyses
Cells grown on coverslips were fixed with 4% formaldehyde for 20 min at room temperature,
followed by 0.5% Triton X-100 treatment for 5 minutes and 3% BSA blocking. The slips
were then incubated with the corresponding primary and secondary antibodies along with
DAPI staining for visualization of nuclei. Fluorescence images were acquired with a Zeiss
710 META microscope.
Plasmids
1
The 3’UTRs of MYOCD, ARHGEF12, FGF12, and ADCY5 were amplified by PCR from
the genomic DNA of SMC and cloned into pcDNA3.1(+) according to standard protocols.
The MYOCD 3’UTR was then subcloned into psiCHECK-2 using XhoI and NotI restriction
sites. MYOCD 3‘UTR was also subcloned into the pMS2 vector for RIP analysis. Primer
sequences were as follows: MYOCD, 3‘UTR-F GGACCTCACTCCGCCAAAT, MYOCD,
3’UTR-R
TATTCCTCCACATCCCAC;
TGTTAGTATATTCTTTTTCTTAATA,
ARHGEF12,
ARHGEF12,
3‘UTR1-F
3’UTR1-R
ACCAGTCATTTAGCTTGAAAGAGGG;
FGF12
3‘UTR2-F
AAATGTCTGCAGTTCAAGAAAAGTT,
FGF12,
3‘UTR2-R
GTGTAAAGTGAGGGAAAC; and ADCY5, 3‘UTR-F CAAAGGCGAGATGATGAC, and
ADCY5, 3’UTR-R CTTCACTTCTGGTCCCTACTCAGCT.
Cycle Stretch
A custom-made test apparatus was designed and manufactured for the application of cyclic
tension capable of operating inside an incubator with a strain range of 0-25% and frequency
range of 1-3 Hz. The tensile device consists of electrical and mechanical units. The electrical
unit includes a PLC+HMI, a stepper motor driver, and a power supply. Data acquisition is
performed using a computerized program. The mechanical unit contains a step motor,
ballscrew, connector rod, an encoder and mobile and fixed grippers for stretching an elastic
membrane according to a programmed pulsatile algorithm. The uniformity of applied strain
was tested along the length of membrane by measuring the strains of stained locations on the
stretched membrane. A maximum deviation of 0.5% of the measured strain value across the
membrane ensured adequate manufacturing quality[2].
Reagents
The following antibodies were used: anti-MYOCD antibody sc-33766 (Santa Cruz
Biotechnology, Inc., Dallas, TX); anti-ARHGEF12 antibody 89-201-312 (ABNOVA, Taipei,
Taiwan); anti-FGF12 antibody ab58157 (ABCAM, Cambridge, MA); anti-SM22 antibody
sc-51442 (Santa Cruz Biotechnology, Inc.); and anti-ADCY5 antibody CAB-2602MH
(Creative Biomart, Shirley, NY). The following siGENOME siRNA reagents were used: for
nontargeting,
2
(siNC);
MYOCD
(siMYOCD);
ARHGEF12(siARHGEF12);
FGF12(siFGF12); ADCY5(siADCY5); and Dharmafect 1 (Dharmacon). The following
2
reagents and kits were used: Lipofectamine 2000, Trizol reagent, Dulbecco’s modified Eagle
medium (DMEM), Opti-MEM reduced serum media, and fetal bovine serum (FBS)
(Invitrogen Life Technologies, Carlsbad, CA); psiCHECK-2 vector and dual-luciferase
reporter assay (Promega, Madison, WI); and RNeasy mini kit, DNeasy blood and tissue kit,
and Qiaprep spin miniprep kit (Qiagen, Manchester, UK).
IA model and survival analysis
The left common carotid artery and the posterior branches of both renal arteries of 30 male
7-week-old C57BL/6J mice (200-300 g) were ligated to induce cerebral aneurysms. These
procedures were performed with the mice under intraperitoneal zoletile anesthesia (30 mg/kg)
with xylazine (10 mg/kg) and additional injections if necessary. After the operation, 1%
normal saline was substituted for the drinking water to enhance the degree of hypertension.
Then, the mice were intraperitoneally injected with 3 doses of 1-8 mg of tamoxifen (T5648,
Sigma-Aldrich Co. LLC., St. Louis, MO) freshly dissolved in 100 μl corn oil (C8267,
Sigma-Aldrich Co.) at 10 mg/ml, 20 mg/ml, 40 mg/ml or corn oil ehicle[3]. The
subcutaneous injection site was sealed with a drop of Vetbond tissue adhesive (3M).
Following tamoxifen administration, the mice were housed individually for 5-10 days before
being analyzed for Cre-recombinase-mediated activity[4]. All animal care and experimental
designs in this study complied with community standards on the care and use of laboratory
animals.
The survival analyses were performed on C57BL/6J mice IA models of wild-type MYOCD
(n = 28) and MYOCD-/- (n = 13) mice, the high-expressing MYOCD (n = 13) or ceRNA
group (ARHGEF12 n = 16; FGF12 n = 13; ADCY5 n = 12) and the low-expressing MYOCD
(n = 15) or ceRNA group (ARHGEF12 n = 12; FGF12 n = 15, ADCY5 n = 16) according to
the median within the wild-type group. Respectively, the survival analyses were performed
on secondary hemorrhages of IA patients of the high-expressing MYOCD (n = 26) or ceRNA
group (ARHGEF12 n = 29; FGF12 n = 23; ADCY5 n = 20) and the low-expressing MYOCD
(n = 24) or ceRNA group (ARHGEF12 n = 21; FGF12 n = 27, ADCY5 n = 30) according to
the median of gene expression levels, see also Fig. 7B, C.
ceRNA criteria according to Tay et al [1]:
We took a list of 5 miRs that were known from the literature to regulate MYOCD and they
3
generated a list of other, likely targets for these miRs using rna22. First they discard all
predicted targets that have fewer than 2/5 miRs in common with MYOCD. Next, they filter
these likely targets by a few other criteria to bolster confidence that some of the targets
compete with MYOCD for miR binding. Gene X is predicted to be a ceRNA of MYOCD if:
1) Genes X and MYOCD have many miRs in common. This criteria is expressed as (the
number of MYOCD miRs that are predicted to target gene X’s 3’UTR) / (the total number of
miRs that target the MYOCD 3’UTR). The more miRs the two transcripts are predicted to
share, the more this ratio approximates a value of 1. As the number of predicted, shared miRs
between two transcripts increases, so does the likelihood that at least some of the miRs
actually regulate both transcripts. Furthermore, if two transcripts are to exhibit a crosstalk
phenomenon, then sharing more miRs means that the two transcripts have more ways they
can regulate one another.
2) Shared miRs are predicted to target gene X over a short region. For a given miR, miR-n,
gene X has a high ratio of (MREs for miR-n)/(distance in bases between left-most and
right-most MREs of miR-n). This consideration arises because, longer 3’UTRs cause target
prediction algorithms to return more spurious target predictions. This is approximated by the
formula: E=K∗L1∗L2∗exp(−λS), where:
• λ= A constant determined by the scoring system.
• K = A constant determined by the database used.
•
L1 = The length of the 3’UTR.
•
L2 = The length of the MRE.
• E = The mean of a P
Poisson distribution of the number of matches between L1 and L2.
As the length of 3’UTR under consideration (L1) increases, the higher the mean of the
distribution becomes. Thus, target predictions across the entire length of a 3’UTR can be
noisy. This criterion prioritizes targets with shorter islands because they are less likely to
contain spurious predictions.
3) There is an even distribution of miR targets within target islands in gene X. These criteria
are expressed as (distance between leftmost and rightmost MRE for miR-n)^2 / (sum of
squared distances between miR-n MREs). This criterion does two things. First, it penalizes
4
targets where most but not all MREs aggregate in a small neighborhood. In this way, longer
target islands are penalized because they may introduce spurious predictions. Second, it
rewards instances that maximize the minimum distance between the same MREs. The MREs
that are targeted by miR-n must be spaced out enough that miR-n+RISC can sit on the mRNA
target without steric interference by another miR-n+RISC complex. By maximizing the
minimum distance between MREs, we increase confidence that miR binding to all these
MREs is possible simultaneously.
4) Fewest possible miRs give rise to a large number of MREs in gene X. This can be
expressed as [(All MREs in transcript X for all miRs predicted to target X’s 3’UTR) - (# of
miRs predicted to target X’s 3’UTR + 1)] / (MREs in X’s 3’UTR for all considered miRs).
This criteria rewards transcripts for having a higher number of predicted MREs for each
targeting miR. As the number of predicted MREs in target X for a particular miR increases,
so does the likelihood that the miR actually regulates the target in at least one place.
Once 1-4 above are calculated for all the miRs validated to regulate a gene of interest, the
values can be multiplied out for every protein-coding transcript in the transcriptome to
generate MuTaME scores. The transcriptome can then be rank-ordered based on the
likelihood each gene functions as a ceRNA of your gene of interest. To review, the MuTaMe
scoring approach takes validated regulatory miRs of a gene of interest and uses rna22 to
generate lists of other genes likely regulated by those same miRs. This list is then
rank-ordered by applying the criteria above which prioritize potential ceRNAs based on the
number of miRs they share with a gene of interest, attributes of ceRNA target islands, and
how likely miRs are to regulate ceRNA 3’UTRs.
References:
[1]. Tay, Y., et al., Coding-independent regulation of the tumor suppressor PTEN by competing endogenous
mRNAs. Cell, 2011. 147(2): p. 344-57.
[2]. Ghazanfari, S., M. Tafazzoli-Shadpour and M.A. Shokrgozar, Effects of cyclic stretch on proliferation of
mesenchymal stem cells and their differentiation to smooth muscle cells. Biochem Biophys Res Commun, 2009.
388(3): p. 601-5.
[3]. Wicksteed, B., et al., Conditional gene targeting in mouse pancreatic ss-Cells: analysis of ectopic Cre
transgene expression in the brain. Diabetes, 2010. 59(12): p. 3090-8.
[4]. Wirth, A., et al., G12-G13-LARG-mediated signaling in vascular smooth muscle is required for
salt-induced hypertension. Nat Med, 2008. 14(1): p. 64-8.
5
6