Live transcribed / unedited notes:
Notes:
Nicole King (Coenos)
Many cell-cell interaction genes present in Coenos.
‘evo-cell’ how did specialized animal cell first evolve?
Role of bacteria in shaping evolution of development
Earh 4.6Ga, first life ~3.6Ga, Multicell 1.0Ga – evolves in a bacterial environment
Conserved genes between fungi/plants/animals are house-keeping, not cell-cell interaction (evolved multi-cellularity independently).
Spounges generate substantial water-flow through an identical mechanism of chanocytes.
Shared extracellular matrix + cytoplasmic bridges
Colonial in nature but mostly single-cells in lab culture. Difference is bacterial diversity. Adding bacteria recovers colonies, killing them with antibiotics removes colonies. Only one particular species is actually inducing the colony formation.
Sphigolipids from bacteria induce rosettes. Actually a sulfonolipid (has sulfur rather than serine headgroup). named RIF1.
Andy McMahon Role of chromatin and gene regulatory networks in shh-mediated neural development
shh concentration gradient determines 5 different neural cell types. FP, V3, MN, V2, V1, V0.
Different concentrations of shh induce explant cells to different cell types. Requires temporal integration, can reprogram (partially?) differentiated cells – MN to V3.
‘class 1’ response genes (repressed by Shh), ‘class 2’ activated by shh.
Embryo bodies (from ESCs) + Retinoic acid > dorsal neural cells. + Shh > all ventral types.
Compare to DNAse hypersenstivity for the region?
Nkx enhancers missing repressor? (seems to be on in the floorplate cells)?
Nkx2.2 loose expression entirely in low expressing cells. Is there a repressor?
Increasing Gli1 affinity at a single site in FoxA2 enhancer expands domain.
FoxA2 autoregulation.
Using smRNA FISH of patched to read out shh activation, see a low, then spike in signaling, then drops. Naive cells express proteins that increase sensitivity to shh, which are repressed upon shh induciton. Shh also upregulates
why have FoxA2 on this delicate switch? “Want to switch on fast, don’t want to expand”.
Sox2 binds very different genes in ESC cells vs. neuralized embryoid body (NEBs).
Sox2 precedes Gli1: adds H3Kme2 at enhancers?
Ostrander (National Human Genome Research Inst.)
Dog breeds by official breed registry are genetically isolated (like a few rare human populations).
Some severe population bottlenecks.
Genetics of 40 fold skeletal variation. Portugese water-dog (2 fold variation), 10,000 of these in the US from 30 dogs 50 years ago. Identify shared haplotypes with very large and very small dogs. Hit: insulin like growth factor1 IGF1. Look for evidence of selective sweep. Heterozygosity near selected genes plummets. Different small breeds share mostly common haplotype.
Other ‘small’ linked genes: STC2, GHR (growth hormone receptor), SMAD2, HMGA2.
Important X-linked locus in giant dogs.
Large allele for IGF1 → 50lbs +. less than 15 lbs, need all 5 small genes.
Short-legged dogs (parker et al Science 2009).
Retrogene of FGF4 (no introns or regulatory info, inserted in to the genome at a different place). Premature closure of growth plates. Single SNP between original gene and retrogene.
PCA on data from skull landmarks. Find co-regulated shape features.
BMP3 linked to ‘pug nose’ phenotype. Mutation looks like a loss of function missense mutation. Can compress both upper and lower jaw. See changes at 32 hr of development and after.
Most traits controlled by few number of genes (probably reflects way of breeding/ population structure).
Disease associations all in non-coding regions. (phenotypic selection in coding regions).
Good Afternoon and Welcome to the Hilde-Mangold Postdoctoral symposium.
I’m Alistair Boettiger and along with James Dutko I will be co-chairing the session.
We have an excellent line up for you this afternoon, which spans a broad range of exciting questions and innovative approaches developmental biology. This includes two top selected guest speakers from the European society.
Oct4 sox2 cluster, division or signalling
FGF RA vs wnt acti
Do you see crowding in the basal layer as a result of overproliferation.
Initaitor tRNA extra copies faster growth, larger animal, more protein synthesis (40%).
Effect on fertility # of progeny
increased tRNA synthesis in fat body are downstream of tor.
Effect on longevitiy?
Eph ephrin signaling mediates contract repulsion.
In simulating contact repulsion, do you model 180 degree change in direction,
disrupt ephrin, more holes in distribution of cells.
Ortholog of tailess leads to mulitple anchor cells that fail to invade.
DNA licensing factor are a markers of quiescence, move around in nuc during replication.
Why can’t proliferating cells invade? Cytoskeleton difference? No invadipodia?
Neural crest delamination, snail twist.
Genetic component separating these two vs. mechanical effect
how similar is the skeleton of two different wildtype embryos ?
characterize number of connections, length of connections?
Distribution of connections
Benoit Bruneau (McGill)
Epigenetic regulation of heart development
PRC2 (EzH2, SuZ(12) ) → H3K27me3. Delee EzH2 in early heart development, mice live to birth and have massively enlarged hearts. PRC2 represses Six1 to turn off skeleton muscle genes, fibrosis genes, hypertrophy genes.
Make heart cells in a dish. Do chip-seq at different steps.
Unsupervised clustering based on chromatin marks to ID co-regulated genes.
Claim cardiac genes are pre-activated – go from blank/naked, to silent but H3K4me1, K27ac (promoter and enhancer), then later turn on.
Konrad Hochedlinger (Harvard) Understanding Cellular Reprogramming and Pluripotency
Yamanaka: retrovival infection with TFs get induced pluriplotent stem cells at very low efficiency .1-3%. 12 days for stable reporgramming → self renewing pluripotent state. Retrovival genes get silenced in iPS cells. Silent X gets reactivated, mortal cells become immortal.
Can get iPS cells to re-encorporate into embryo and mature to fully developed mouse
Develop mechanism to induce iPS cells without viral mediation (using TetO/dox).
Get abberrant silencing of some genes due to aberrant dna methylation of IG-DMR. Dlk1-Dio03 locus. Silenced transcriptionally early and by DNA methulation late. Paternally imprinted.
Dnmt3a (invovled in paternal imprinting) silences IG-DMR. If knocked out get a non-silenced reprogrammed cell.
Adding more Vit C (ascobric acid) It maintains activating histone marks at dlk1-Dio3, prevents DNA methylation.
Should derive iPSC cells in presence of AA to keep these key alleles on in reprogramming fibroblasts.
Can take B cells and reprogram them into functional iPS cells.
Tetrapoid embryo complimentation: inject ES cells into mouse, tetraploid cells make placenta, ES cells make mouse. Most iPS cells fail in this assay. Most ES cells don’t.
Zeitlinger (recruitment of poised Pol II is regulated over developmental time).
Examples of paused genes (twist – doesn’t pass Vivek’s GD7 gro-seq, but maybe in meso?)
is pausing ‘preparing genes for activation’ (presumably the argument for heat-shock genes).
DAPI and wildfield sorting for early embryos.
Very few genes before mid-blastula transition. Poised genes are activatd later (well it looks like everyone is activated at MBT).
Poised polymerase is a stage specific
sim paused well before it turns on. Snail ‘not paused’ early on (fully loaded early on? But what about off tissue adding in?). Claim no time for pausing in early transitions (it doesn’t need to slow you down).
Paused genes / regulated genes have clean start sites. Paused genes more TATAless. Genes induced without pausing stronger TATA, less DPE. ‘paused later’ do have TATA.
Housekeeping genes / sloppy start site genes have nice periodic nucleosomes downstream, depleted just upstream.
Poised have nucleosome at 0 or -1. TATA don’t. TATA have very low K4me3 (few histones to start with?)
Denis Duboule
Hox gene transcription during limb development.
Wolpert: domains of hox gene expression follows order of the genes. Interested in understanding this topology.
Extremity of hoxD cluster required to build extermity of the limb. Central part (including 2 overlapping genes) build fore-arm. Megabase desert on either side. Induce inversion in distal megabase + more flanking sequence (towards centromere) loose digits expression (i.e. caught dstial enhancers. Other flanking region has the forarm regulation.
Eileen Furlong. Martina Rembold
Mef2 tin htl expression lost in sna ‘previously known’
snail binds to mesodermal enhancers (well it looks just like twist in all of these examples). Claim 35% are co-bound regions.
Look at tin intronic locus, do in vitro luciferase. Tin + snail in vitro activates more than tin alone. (snail alone does nothing).
Tin enhancer is lost in snail mutant (as is tin endogenous). Still consistent with direct repression.
Identical in vitro for Mef. Mutate snail sites, twist response still present, snail response is gone.
(knowning snail activates itself may strengthen this study).
More positive genes: CycE, stumps htl vepD, RhoL NetB, cenG1A. 50% of snail ‘targets’ go down in sna null. With 23 motifs can distinguish 83% of positive vs. negative.
Some other protein binding the TCF site.
Occupancy of co-repressors.
Twist no snail in vitro assay – this is a snail null, why do you see expression?
Cis regulatory modules (cal tech)
most genes have multiple CRMs. One CRM at a time or multiple CRMs at once.
Overlap? (combinatorial model) vs. non-overlap (simple model)
Nam and Davidson, approach for ID 130 CRMs (in single embryo?) Have temporal but not spatial information. 60% of time points overlap.
Test 2 respond to same perturbation. ‘CRM responded vs gene responded to each perturabation’.
CRM and Gene response identical in only a small number of cases (8 of ~70).
(Davidsons binary version of gene expression clearly getting in the way).
Univin spatial temporal overlap intronic and 5p enhancer.
ABA causes stoma to close (leaves to coil and go verticle). Functions in both predictable drought (winter) and unpredictable drought.
Maria Barna
Live confocal imaging of labeled transgenic populations of chick embryos.
Mesenchymal long philopodia / cytonemes (Actin rich extensions).
shh binds patched, patched releases smoothened, smoothened migrates to primary cilia and starts signaling. Cdo and Boc also are co-receptors of Shh.
Extensions carry ssh particles that move along particular philopodia. (move both ways but net away). Niether patched nor smoothened in philopidia of responding cells.
Don’t see transfer of the particle from the philopodia tip of sending to receiving cell.
Savo Lavic (Ian Scott lab).
First heart field differentiates and forms a beating tube. Second heart field migrates to the tube later and then differentiates.
Use photoconveratble KickGFP for lineage tracing of heart cells. (Turn first differentated cells in heart tube red. ID new cells arrive green.
Degradation of the photo-converted red form replaced by newly synthesized green form?
Nppa marks econd heart field cells? (upregulated in adult heart disease. Knock out in mouse not strong effect.
Truncated Npp marks SHF progenitors.
(Stowers)
Neural crest migration
R4 cell migration. Flanking regions inhibited by semaphorins/ephrins repress lateral migration. Attractant signal unknown (?)
VEGF is expressed in surface ectoderm along which neural crest cells migrate. VEGF beads appear to attract NC cells. VEGF cells can attract NC cells into flanking regions.
Surface of migration is growing while NCs migrate.
Model solution: cells absorb/degrade local VEGF (creating a gradient). Only front cells get out (trailers don’t see enough of a gradient).
Compare gene expression profiles between ‘leader’ and ‘trailer’ cells.
Transplanted ‘trailer’ cells ‘become leaders’ continue to migrate to target cite.
Dividing neural crest cells during migration. Migratory cells slow down/pause but don’t change neighbors. Have random division orientation. Variable times / positions away from neural tube midline after delamination. Proliferation accelerates substantially when cells reach target (bracheal arch). Mostly quiscent, most trailing cells exit cell cycle.
Will they go the wrong way if you transplant the whole population to the other end?
Garcia-Castro
Rabbit
neural crest cells have properties of what we call ‘ectoderm’ (skin, nerve producing) and mesoderm (muscle, skeletal producing) cell types.
Investigating neural crest in mouse. Characterizing first appearance TFs
in sox- pax- see only a delay (not a loss) in activation of the downstream factors. Mouse embryo at this stage very difficult to manipulate, order of magnitude smaller than chick, connicle not flat structure
Rabbit is flat (like human) and much larger (~1/2 size of chick).
Characterize temporal unfolding of dev. TFs.
Crest cells specified in cultured explants in the absence of mesodermal factors.
Electerpolate specific cells in cultured explant tissue (using a GFP marker and signaling antagonist).
Rabbit NCs specified by FGF.
John Wallingford
Genomic control of development and function in multi-ciliated cells.
Multi-ciliated cells on the amphibian epidermis.
Massive duplication of centrioles (multicilin). Scattered cell population (via Notch). Coordinated beating (planar-cell-polarity).
DAF19 ‘master regulator’ of ciliogenesis in C. elegans. (different splice forms have different fxns). Yeast family govern cell cycle exit.
Cilia play several differentiation roles : left-right asymmetry (9+0 motile), directional fluid flow (9+2 motile), hedgehog signalling (9+2 non-moltile).
Can grow many frog sub tissues in vivo. Very high reproducibility.
Use RNA seq data to infer genome (assembly not yet released).
RFx2 has cilia dev errors (e.g. dorsal closure). See defects in transport in cillia.
ID new genes which cluster with cilia beating genes. Knock these down, most cilia don’t beat, others beat slowly. Flow is directionally correct but much slower.
Most Rfx2 targets are probably not ‘ciliogenesis’. Maintenace of cilia cells requires insertion of cell from basal precursor into the epithelium followed by apical expansion and differentiation.
Insert at vertices not at edges. Rf2x weak knockdown apical expansion no cilia. Stronger, extrudes cilia, but apical expansion never follows.
Owen Tamplin (Leonard Zon lab) [from discussion last night]
HSC stem cell migration: migrate, expand, migrate again, become new stem cell population (for blood). Runx1+23 enhancer marks stem cell population: cells progressively migrate in waves from aorta out into circulation and end up in tail niche. Associate with endotheial cells
in 10 min can collect a falcon tube of zebrafish embryos (6,000 +) array into 96 well plates and screen against thousands of chemical compounds. Test effects by in situ.
Zinzen
H3K4me1 not enriched on mesodermal enhancers. K27acme3 79me3 and PolII enriched on enhancers. Do contain H3K4me1
Is the PolII associated on enhancers reflective of a structural proximity of the enhancer and the gene body, or a transcribed enhancer? PolII.
100-1,000 fold difference in binding site occurance and binding site occupancy in ChIP data. Ttk, zelda snail Dorsal enriched in early binding sites.
Why is twist not binding twist sites of late enhancers?
Rather than mutate sites to lose enhancer activity.
Shows that twist binding goes down in ChIP when Zelda sites are mutated (cact). Zelda does bind its own sites.
six1 homeobox gene (sine oculis in Drosophila)
T. Gregor
When is a gene fully on? Segregation by the nubmers
Shawn Little +
‘what does on/off’ mean at the level of transcriptional activity. Looking at high transcription.
Resolve ‘contradiction’ of noisy bursty transcription vs. precise disposition of profiles.
20-150 20mers. 3-4 adajcent slices (airy
dim spots noise, cytoplasmic spots -single mRNA, bright spots. 3 orders of intensity at means.
Multicolor label
not distinguishing individual spots at high concentration. Use qPCR to FISH comparison – agree to 10%.
Measure relative intensity of nascent transcripts. Not there yet. –
doubled spots are G2 cells sister chromatids.
Average transcription site activity ‘much noiser’, ‘individual spots ‘more variable’ –
missed detection.
Cytoplasmic counts. ‘Noise in transcription sites ~23%. Need spatial averaging to reduce to 7% variation in
‘presence or absence’. Fully on (not TF limited) – limited by PolII arrival, expect poisson distributed.
250 RNA per cell in 15 min of nc13. ‘no degredation’ because it increases during cell cycle.
max load 10, ‘on’ half-maximal load.
4 copies (when does DNA replication occur during cc13? It’s not 4 copies the whole time). – estimate 2-3 min by fraction of embryos with 4 dots.
Hb molecules per nuc cc13 (~600, range: 500-700). drops nicely to zero.
Kr vs Hb molecule counts. Slope 1.15. in their maximally loaded regions are made at the same rate. Claim all gap genes are made at the same rate in their maximally on regions.
[I measured peak around levels sloping up from 500 to 700 with local range of around 100, slightly more slanted but essentially the same counts.
At TF saturation an additional stochastic stpe (enhancer looping, RNAP binding, chromatin opening/closing rate).
Ms2loop at 5prime end.
Med14
Stas
morophogenesis by folding. Fate map, order of folds, appropriate force. Orgami rhino.
morphogenesis is live egg chamber originally developed by Berg lab).
Broad enhancers: early then late enhancer. Mirr represses early enhancer and activates late enhancers. Broad is required for late expression of broad. Broad early drives broad to inhibit a repressor of broad.
Propose Br represses rho. Rho activates Spitz activates EGFR, activates Pnt which represses broad.
For Alex: if K4me3 is poly deposited, how can people propose it’s a ‘poising’ factor in ‘bivalent’ systems.
Chd1 talk. (morning session talk 4)
Ser2p RNA pol gH2Ax (replicative stress response? High in E5.5 stage). Do bivalent domains exist in vivo? Cell heterogeneity? Strongest evidence is all ipSC. Evidence ‘against’ this in ESCs from blastocyst. Sperm have nulceosomes on only 2-4% of genome. Do see signs of bivaliancy in sperm cells. Not poised to do anything though?
Examine E11.5 PGCs see evidence for bivalent domains.