|Conference and seminar notes|| Main project page|
Previous entry Next entry
Notes for talks with lab members and miRNA conference 3e Jeudis de Génétique - Necker
MicroRNAs – 3e jeudis de Necker (Génétique) – 20 novembre - general intro
miRNAs have both a functional part but also a structural part that links up to the cytoskeleton.
See Mattick 2005 – with increasing body complexity, lots of non-coding RNA in a pool of total transcripts (up to 40%). Exons to mRNA but also ncRNA, but introns to miRNAs and snoRNAs after splice processing, which feed back.
2004 there were 719 miRNAs listed over all species, think possible up to 1% of ncRNAs. Products 22nt, from precursors of 70-100nt hairpins.
miR7 in Drosophila – N signaling miR196 in patterning developmental; miR1/181/others for cardiac development.... etc.
Viral resistance in plants
Argonaut complex associates miRISC – if perfect match leads to mRNA cleavage but if imperfect, leads to translation inhibition anyhow but still transcript stability. siRNA are endogenous in most organisms but can come from exogenous sources – eg. the virus genome infecting a plant.
Arabidopsis has 4 Dicers. Antiviral response in Drosophila mouth also against certain viruses. Does such a natural immune response exist in mammals? Group made Dicer hypomorphs in mouse that can survive to adult stage. These are indeed hypersensitive to VSV virus (vesicular…) cf Otsuka et al 2007. Implicates miRNAs from cells – four complementary sequences miR379, miR93, miR146, miR24 will regulate the translation of some of the viral genes. Not involving viral siRNAs.
EBV codes for 23 of its own small miRNAs. (Pfeffer et al 2004) – adapts silencing machinery to help regulate either host or own genome translation. Other herpes and polyoma viruses have microRNAs as well. Grundhoff et al, 2006. Uses particularly Karposi sarcoma virus = HHV8. Infects endothelial or B lymphocytes. In immunocompromised patients, leads to lymphomas from those lymphocytes after initial black lesions in the skin.
In the 5’ part of miRNA, nucleotides 2-8 are most important for recognition. Even if recognition not perfect, can lead to sequestering or decapping of RNAs in compartments called P bodies. Looked for KSHV viral miRNA targets to begin – transducted using lentiviruses for stable transduction in DG75 lymphoma and Ea.Hy endothelial cell lines, then compared for seed-matches to miRNAs potential – looked for changes of 1.4-4x transcription. 64 high confidence targets in DG75 amd Ea.Hy 211 targets – leads to 4 common targets, and 23 lower-confidence targets.
Which targeted pathways? Kegg pathway analysis in any case Tgfb signaling, more generally cell cycle. How validate targets - using a luciferase psCHECK2 Vector with the 3’UTR after the luciferase gene. pCMV-pcDNA5-FRT-TO to activate. Checked that caspase-3 seemed to be targeted by 5 of the miRNAs from KSHV (miR-K12-1, -3, -4-3p, -12-10a, 12-2).
Phenotypically lead to disaggregation of cells in culture after viral infection, miRNAs reproduce this phenotype.
Changes in cell surface markers? Why didn’t these come out of the transcriptome screen? Chose proteomic approach using external protein label and 2D gels – this is underway. Also can do IP of Argonaut protein complexes using mAbs against Argo1 or 2. Beitzinger et al., 2007 (Gunter Meister) – this is the approach that Alex’s student is taking. CD69 is enriched 20-fold in Ago2-specific IP vs. control IP and has seed-matches against miRK12-1, 12-4-3P etc. It induces TGFb expression. So maybe the virus via its miRNA inhibits CD69 translation.
Some other miRNAs seem to be involved in viral replication as well. Looking for targets.
3 silencing pathways using small RNAs – piRNAs (24-26nt – Argonaute3 and Aubergine and Piwi, made from “selfish elements” in the pericentromeric repetitive elements with sense and antisense orientations and using processing without Dicer protein – used mostly in germinal lineage), siRNAs (21nt - Argonaute 2 via Dicer 2 [in Drosophila], activated by endogenous [recent data from inverse repetitions] or exogenous double-stranded RNAs) and miRNAs (approx 22 nt – Argonaute 1, hairpin structure etc.).
Endo si-RNAs used for maintenance of heterochromatin structure as well as piRNAs.
Studying with a copper-inducible promoter construction the miR5 and miR-6 within the intron of the construct (I don’t know what the little gene is that fuses with the GFP), upstream of the 2nd GFP-tagged exon. Make a mutagenesis to make a MiG1 and miG2 that are complementary to luciferase.
Ago2-RISC activity is involved in self-silencing of target – get cleavage of target by Ago2 with perfect complementary and Dicer2 used to load the miRNA into Ago2. Cf that there is sorting – Tomart, Du and Zamore Cell 130 (2007).
miG1 construct is better loaded into Ago2 RISC complex than miG2. “AutomiR system” reporting for miRNA biogenesis and activity through Ago2 only. However, the miGs can also act through Ago1. Use the automiR as an in vivo reporter in the fruitfly.
Ubiquitin promoter – bit of 1st exon – miG1 miG2 –bit of 2nd exon-GFP. Lots of GFP silencing. Went to a Drosophila RNAi screening center to inactivate all 18K annotated genes. Let to lots of target genes notably involved in RNA processing and translational regulation. But ¼ unknown.
Carcinomas are frequent and malignant, adenomas less frequent (more in women though, leading to hemorrhagia) and benign otherwise. Some because of activating beta-catenin, others inactivating HNF1a, or 30-35% through inflammation with an activating mutation of gp130 (recent data, in Nature).
Carcinoma risk factors – cirrhosis or alcohol, other aggressions of liver, obesity. Six different transcriptional profiles, correlated with genetic and functional classes, associated with particular gene mutations (eg TP53). Looked then at miRNA profiles. Profiles = b-catenin, alcohol, HNF1a, inflammatory, hepatitis B virus, etc.
Used qRT-PCR with 250 miRNA primers from Applied Biosystems comparing 28 HCCs, 13 adenomas, 5 HNF and 4 normal livers. Then validate 22 miRNAs in other clinical samples.
Some deregulated = miR-224, miR-122a, miR-422b. First is upregulated, others are downregulated. miR-200c and -203 are upregulated in adenomas. Three others upregulated in in malignant hepatocarcinomas.
In b-catenin-mutated adenomas AND HCC, downreg of miR-375, meaning beta-catenin somehow regulates it? Apparently did something functional but the evidence went by way too fast.
While miR-107 in HNF1a- tumors downregulated. Strong correlation of the expression of both.
In HBV-carcinomas, miR-96 overexpressed, but only when carcinogenesis not just in hepatitis-infected livers.
miR-126* underexpressed when HCC developed on an alcoholic background.
Other people are interested – cf. Varnhol et al., 2008 meta-study of eight studies. But they are not very comparable because of either mixed or different subgroups of HCC. 10-55 samples per study.
Lab discussions of progress 19-20 novembre 2008
For overexpression – use pcDNA 3.1 version C (but also A and B for frame shift) eukaryotic expression vector. MCS use BamHI for ISL1 or HindIII for GATA4/TBX20 (see article in Development and Disease). She has designed primers for use on cDNA after RT-PCR with a cutting site integrated in the tail, like the T7 tail for RNA primers. At the 3’ EcoRI end, there are two versions – one with the stop codon integrated so as to let the expression vector not make a fusion protein with myc/his, and another one without the stop codon.
Then this tagged fusion protein is useful for Western blotting using an anti-myc. Can also get rid of the myc tag by excision with an unusual restriction enzyme and use an anti-His for further use. Can also use for stable transfection under neomycin selection (current Fugene transfection is non-stable).
Check out the roles of TBX1/5/20 – do they directly or indirectly bind DNA, have a degenerate consensus site?
Chris thought problems perhaps due to cells used (HeLa, then tried locally available 293T cells). These both had high and higher activation respectively of the non enhancer-containing luciferase reporter gene when transfected with her truncated ISL1 construct. She noticed in papers dealing with ISL1 function, that many authors used the mesenchymal potential cell line from mouse 10T1/2.
So rather than change from pGL3 to pGL4, she is making a 3rd construct in which the enhancer is not before the promoter or after the luciferase but between them, after the promoter but before the coding sequence (not in MCS therefore). (In pGL3.) Will try in these new cells, obtained from CR in M. Buckingham’s group at Pasteur.
Studied the sequence of the 1kb fragment she had inserted, which is flanked with two ISL1 sites in the same direction as transcription of the FGF10 gene (on the chromosomal negative strand). In addition, there are, closer to the first exon (at the 5’ end) two additional sites at 50bp upstream, oriented in the antisense direction and separated by 3bp. This distribution and orientation TAATGNNNTAATG…(50bp)…TAATG is identical to that in the promoter of the amylin (APP) gene on chromosome 12 and tested in pancreatic cells to indeed be activated by ISL1.
In addition, at 50bp within the studied fragment from the 3’ sense orientation ISL1 site, there is an antisense-oriented perfectly consensual GATA4 site.
Christelle also made a different positive control. Instead of putting 6 repeats of the consensus ISL1 in a row before pGL3 to check activation, as was done in the paper that showed that the protein needs to lose its LIM domain to be active, is designing another one with the TAATGNNNTAATG…(50bp)…TAATG enhancer ahead of the 5’ promoter.
Sophie and Candice
– are deriving neural crest from chicken. However, even though trypsin works, it always kills their cells when they re-seed. This is just bizarre. Maybe they can just scrape out and let proliferate a week before trying to passage? Already done. Tried an EDTA-containing PBS which is a little better for survival, but not ideal. Also diluted 2x trypsin doesn’t like.
Decided to order some quail eggs because this is just weird. With Yves are also going to get mouse NCC; if they are easy to passage will work with them for the time being.
The idea is to finally be able to test re-injection into the chicken embryo!
All three will need to go to Broussais in rotation to bring back some embryos between now and Christmas.
Meanwhile, only our hNCC cells have become contaminated yet again with “champignons” and get apoptotic and unhappy. Lots of other occupants of the etuve. Only had been disinfected with Biocidal – more effective? No one else suffering, though, even other Petri dishes. Had been doing pretty well perhaps with the melanocytic differentiation. Hurry up with the tyrosinase stain, then, and perhaps co-stain with CNPase if possible (though TuJ1 would be interesting also, perhaps only on the N medium cultures?).
3rd or 4th time! New medium was made so that’s not what introduces contamination. I suggested Biocidal on outside of ampoules as thaw.
I must contact Zeiss or (better) Leica for demonstration of image capture software to purchase with an attached PC. We have a Wild M10 equipped with a Sony DXC-390P color CCD3 camera.
Send back to MPE two counters for radioactivity and certification. These are mini-monitor g.m meter type 5.10 serial number 022759 and the Numelec hand-held MCB1 serial number 1303. The latter takes LR1 batteries and I told Stephanie Tome where she can find them in the VWR catalog or maybe just at the supermarket. Left it in radioactivity room.
Meanwhile there are two counters that work pretty well: the MCB21 réf 18429/305 (étiquette INSERM D1819970030) which just needs to not be shaken up, is one of them. Will send to MPE the le MCB21 réf 18429/306 and /307 (étiquettes INSERM D1819970028 and D1819970029) instead of the really beat-up mini-monitor; they are fresher and were revised in 2001.
Discussion with Alexandra: of the SAGE tags in bank, 60% are annotated. Any of these precursors to miRNAs? I think not but check.
Of the remaining 40%, will these be small enough a file to open in Excel as a list (or, just make a text file)? If only available through Access, which Alexandra doesn’t have, is there an equivalent Open Office that allows one to open these .mdb files?
Last, I should be able somehow to map these 40% of 21nt tags to the genome and get back genomic coordinates for them, as if the sequencer had just given FASTA results (not the ELAND files). Can that be done with CisGenome? It would be easier for her (of course).