On 2014 Nov 05, Fillip Port commented:

This paper reports on the use of the H1 promoter to drive expression of gRNAs for CRISPR/Cas genome engineering. The authors demonstrate that gRNAs expressed from H1 can efficiently modify the genome of cultured human cells in conjunction with Cas9 endonuclease. Interestingly, gRNA expression levels from H1 are lower than from the commonly used U6 promoter. Although this can negatively effect mutagenesis rates at the on-target site, it can also increase CRISPR/Cas specificity, as high activity is more likely to lead to off-target effects. This makes the H1 promoter a potentially useful tool for CRISPR/Cas genome engineering in human cells.

However, the authors suggest that their results have much more general significance by expanding the CRISPR/Cas genome targeting space. This is because the U6 promoter initates transcription at a G nucleotide (although see comment below and reference therein), which according to the authors constrains genomic target sites to GN19NGG. This assertion is surprising as it is common practice in the CRISPR field to target sites that do not start with a G by simply adding a (often mismatched) G to the corresponding gRNA or to replace the first nucleotide with a G to create a gRNA that is mismatched at the first position. The authors acknowledge these strategies in the first paragraph of their discussion, but cite six papers as providing evidence that 5’ extensions or truncations reduce gRNA efficiency. However, these papers in fact provide evidence that 5’ extensions or truncations of a single nucleotide often have no effect on activity and when they do the effect is usually minor (modified gRNAs usually retain >80% activity). Furthermore, the authors do not cite another study that shows that small gRNA truncations retain high activity in the majority of cases and have reduced off-target effects (Fu Y, 2014). Therefore, much of the published evidence suggests that extending or truncating gRNAs by a single nucleotide is well tolerated by the great majority of gRNAs. As a result it is in principle possible to target any genomic site adjacent to a PAM motif with gRNAs expressed from a U6 promoter and hence the H1 promoter, although potentially useful, does not expand the CRISPR target space.

On 2014 Aug 12, Haoquan Wu commented:

The preferred initiation nucleotide for U6 promoter are A and G, as we have shown in our study which can be found http://www.nature.com/mtna/journal/v3/n5/full/mtna201412a.html. We also confirmed that gRNAs starting with A can disrupt the target gene at similar level as gRNAs starting with G in the study. H1 promoter is much weaker than U6, which might significantly lower the CRISPR-Cas9 system efficiency.

It is amazing that the commonly accepted conceptions about U6 promoter initiation and termination are not complete although U6 promoter has been the most commonly used promoters to drive small RNA expression. The commonly accepted conception that the continuous Ts is the termination signal for U6 promoter is also not correct.

On 2014 Aug 12, Haoquan Wu commented:

The preferred initiation nucleotide for U6 promoter are A and G, as we have shown in our study which can be found http://www.nature.com/mtna/journal/v3/n5/full/mtna201412a.html. We also confirmed that gRNAs starting with A can disrupt the target gene at similar level as gRNAs starting with G in the study. H1 promoter is much weaker than U6, which might significantly lower the CRISPR-Cas9 system efficiency.

It is amazing that the commonly accepted conceptions about U6 promoter initiation and termination are not complete although U6 promoter has been the most commonly used promoters to drive small RNA expression. The commonly accepted conception that the continuous Ts is the termination signal for U6 promoter is also not correct.

On 2014 Nov 05, Fillip Port commented:

This paper reports on the use of the H1 promoter to drive expression of gRNAs for CRISPR/Cas genome engineering. The authors demonstrate that gRNAs expressed from H1 can efficiently modify the genome of cultured human cells in conjunction with Cas9 endonuclease. Interestingly, gRNA expression levels from H1 are lower than from the commonly used U6 promoter. Although this can negatively effect mutagenesis rates at the on-target site, it can also increase CRISPR/Cas specificity, as high activity is more likely to lead to off-target effects. This makes the H1 promoter a potentially useful tool for CRISPR/Cas genome engineering in human cells.

However, the authors suggest that their results have much more general significance by expanding the CRISPR/Cas genome targeting space. This is because the U6 promoter initates transcription at a G nucleotide (although see comment below and reference therein), which according to the authors constrains genomic target sites to GN19NGG. This assertion is surprising as it is common practice in the CRISPR field to target sites that do not start with a G by simply adding a (often mismatched) G to the corresponding gRNA or to replace the first nucleotide with a G to create a gRNA that is mismatched at the first position. The authors acknowledge these strategies in the first paragraph of their discussion, but cite six papers as providing evidence that 5’ extensions or truncations reduce gRNA efficiency. However, these papers in fact provide evidence that 5’ extensions or truncations of a single nucleotide often have no effect on activity and when they do the effect is usually minor (modified gRNAs usually retain >80% activity). Furthermore, the authors do not cite another study that shows that small gRNA truncations retain high activity in the majority of cases and have reduced off-target effects (Fu Y, 2014). Therefore, much of the published evidence suggests that extending or truncating gRNAs by a single nucleotide is well tolerated by the great majority of gRNAs. As a result it is in principle possible to target any genomic site adjacent to a PAM motif with gRNAs expressed from a U6 promoter and hence the H1 promoter, although potentially useful, does not expand the CRISPR target space.