METHODS TO WATCH | SPECIAL FEATURE

❯❯CRISPR targets RNA C2c2 nuclease, which has an RNase domain but no known DNase domains. They target- Having revolutionized DNA editing, ed C2c2 by a single 28-nucleotide gRNA and CRISPR turns to RNA. saw single-strand RNA cleavage in the bacte- rial target transcripts ( 353, aaf5573, Adapting the CRISPR bacterial immune 2016). Simple base substitutions converted system to eukaryotic genomes has enabled PAM C2c2 into a catalytically inactive RNA bind- an unprecedented ability to modify DNA ing protein that can now be coupled to differ- from disabling to swapping or tagging ent effector proteins. genes, to introducing specific point muta- Both approaches open the possibility of C2c2 (green) or dCas9 with a PAM-providing DNA tions, to enhancing or repressing select oligo (orange) and their respective guide RNAs influencing the many post-transcriptional genes’ activities and carrying out genome- (gray) bring effector proteins (yellow) to RNA processing steps of RNA. C2c2 is a particu- wide functional screens. In the CRISPR sys- (blue). larly good candidate, since recent work on tem, a short guide RNA (gRNA) molecule its catalytic activity showed that it has two steers an endonuclease to a target sequence groups took different routes. Gene Yeo, in independent RNase activities, one to process complementary to the gRNA and located collaboration with Jennifer Doudna, expand- its own RNA guides and another to cleave next to a protospacer-adjacent motif (PAM). ed on previous findings from the Doudna lab the target RNA (Nature 538, 270–273, 2016), While CRISPR has mostly been applied showing that Cas9 can be targeted to RNA if which will allow multiplexed applications. to DNA, recent advances have expanded the PAM is provided by a separate DNA oli- With the appropriate effector fused to its range to RNA editing. RNA interference gonucleotide that binds the target RNA. The the nuclease, one can envision a myriad has been a mainstay to knockdown gene researchers recently demonstrated that this of uses, such as regulating splicing, direct- expression, but to target particular tran- approach can target specific RNAs; a fusion ing RNAs to particular subcellular local- scripts for imaging or subcellular localiza- between a catalytically inactive Cas9 and izations, attaching or removing chemical tion has required cumbersome design of GFP allowed them to track the subcellular modifications, and affecting degradation, RNA aptamers or proteins such as Pumilio localization of RNA in live mammalian cells to name only a few. RNA-targeted CRISPR for each target RNA. (Cell 165, 488–496, 2016). The teams of Feng will give researchers access to a regulatory To target RNA in a way that can easily be Zhang and Eugene Koonin took a different layer of which we have so far only scratched programmed and scaled up, two independent approach. Instead of Cas9, they used the the surface. Nicole Rusk

❯❯How single cells do it mouse sympathetic ganglion to reveal Single-cell sequencing is poised to specific neuronal populations that inner- vate the muscles behind goosebumps and elucidate how cells contribute to nipple erection (Nat. Neurosci. 19, 1331– tissue function. 1340, 2016). Computational approaches are also being Single-cell sequencing has cracked open developed to infer the gene-regulatory the problem of tissue heterogeneity and changes that drive differences in cell state. enabled the study of new cell types and Pseudotime inference can place single cells rare cell populations. Novel applications along reconstructed developmental tra- and analytical tools are now putting jectories, making it possible to pick out

emphasis on inferring the functional the cells involved in developmental transi- Marina Corral Spence/Springer Nature roles of cells in tissues and developmen- tions. Better methods are needed to extract tal events, as well as the genetic programs the gene regulatory changes that drive Single-cell sequencing helps define a cell’s role and how it is played. that drive them. these transitions and cellular decisions. To Transcriptional similarity is widely understand transitions in the blood cell dif- used to categorize individual cells within ferentiation cascade, one approach profiled Perturbation experiments are also a tissue. The same data can also pro- single-cell gene expression from mixtures promising routes to finding function. vide functional insight into cell states. of blood cells at different states and gener- Combining the CRISPR editing sys- Researchers are increasingly using ated a dynamic model of the underlying tem with single-cell studies will be a ­single-cell data to identify cell-type- factor regulatory networks powerful way to screen for the effect specific markers and then label and map (Nat. Biotechnol. 33, 269–276, 2015). Time- of gene knockouts on single-cell tran- these cell types back in the intact tissue. course experiments will be more feasible as scription and cellular phenotypes. Now For example, one study identified many single-cell sequencing becomes less expen- that single-cell RNA sequencing has rare cell types in the gut that likely func- sive and more accessible, and the added become routine at large scales, we look tion in secretion (Nature 525, 251–255, dimension should help researchers to glean forward to experimental and analytical 2015), while another combined unbiased which changes are causal with respect to developments that shed light on cellular single-cell RNA-seq with imaging of the cell state. functions. Tal Nawy

NATURE METHODS | VOL.14 NO.1 | JANUARY 2017 | 33