CRISPR-Cas is a prokaryotic immune system built from capture and integration of invader DNA into CRISPR (Clustered Regularly Interspaced Short Palindromic Repeats) loci, termed ‘Adaptation’, which is dependent on Cas1 and Cas2 proteins. In Escherichia coli, Cascade-Cas3 degrades invader DNA to effect immunity, termed ‘Interference’. Adaptation can interact with interference (‘primed’), or is independent of it (‘naïve’). We demonstrate that primed adaptation requires the RecG helicase and PriA protein to be present. Genetic analysis of mutant phenotypes suggests that RecG is needed to dissipate R-loops at blocked replication forks. Additionally, we identify that DNA polymerase I is important for both primed and naive adaptation, and that RecB is needed for naïve adaptation. Purified Cas1-Cas2 protein shows specificity for binding to and nicking forked DNA within single strand gaps, and collapsing forks into DNA duplexes. The data suggest that different genome stability systems interact with primed or naïve adaptation when responding to blocked or collapsed invader DNA replication. In this model, RecG and Cas3 proteins respond to invader DNA replication forks that are blocked by Cascade interference, enabling DNA capture. RecBCD targets DNA ends at collapsed forks, enabling DNA capture without interference. DNA polymerase I is proposed to fill DNA gaps during spacer integration.

SummaryTo interrogate the alternative fates of pancreas and liver in the earliest stages of human organogenesis, we developed laser capture, RNA amplification, and computational analysis of deep sequencing. Pancreas-enriched gene expression was less conserved between human and mouse than for liver. The dorsal pancreatic bud was enriched for components of Notch, Wnt, BMP, and FGF signaling, almost all genes known to cause pancreatic agenesis or hypoplasia, and over 30 unexplored transcription factors. SOX9 and RORA were imputed as key regulators in pancreas compared with EP300, HNF4A, and FOXA family members in liver. Analyses implied that current in vitro human stem cell differentiation follows a dorsal rather than a ventral pancreatic program and pointed to additional factors for hepatic differentiation. In summary, we provide the transcriptional codes regulating the start of human liver and pancreas development to facilitate stem cell research and clinical interpretation without inter-species extrapolation.

Mitchell-riley syndrome (MRS) is caused by recessive mutations in the Regulatory Factor X, 6 (RFX6) gene and is characterised by pancreatic hypoplasia and neonatal diabetes. To determine why MRS patients specifically lack pancreatic endocrine cells, we micro-CT imaged a 12-week old foetus homozygous for the nonsense mutation RFX6 c.1129C>T, which revealed loss of the pancreas body and tail. From this foetus, we derived iPSC and show that differentiation of these cells in vitro proceeds normally until generation of pancreatic endoderm, which is significantly reduced. We additionally generated an RFX6HA reporter allele by gene targeting in wild-type H9 cells to precisely define RFX6 expression and in parallel performed in situ hybridization for RFX6 in the dorsal pancreatic bud of a Carnegie Stage 14 human embryo. Both in vitro and in vivo, we find that RFX6 specifically labels a subset of PDX1-expressing pancreatic endoderm. In summary, RFX6 is essential for efficient differentiation of pancreatic endoderm, and its absence in MRS patients specifically impairs formation of endocrine cells of the pancreas head and tail.