Multiplexed activation in mammalian cells using dFnCas12a-VPR

The adoption of CRISPR systems for the generation of synthetic transcription factors has greatly simplified the process for upregulating endogenous gene expression, with a plethora of applications in cell biology, bioproduction and cell reprogramming. In particular the recently discovered Cas12a systems offer extended potential, as Cas12a is capable of processing its own crRNA array to provide multiple individual crRNAs for subsequent targeting from a single transcript. Here we show the application of dFnCas12a-VPR in mammalian cells, with FnCas12a possessing a shorter PAM sequence than As or Lb variants, enabling denser targeting of genomic loci. We observe that synergistic activation and multiplexing can be achieved using crRNA arrays but also show that crRNAs expressed towards the 5’ of 6-crRNA arrays show evidence of enhanced activity. This not only represents a more flexible tool for transcriptional modulation but further expands our understanding of the design capabilities and limitations when considering longer crRNA arrays for multiplexed targeting.

It is important to note that Cas9 represents only one of a variety of known CRISPR systems, 61 with others possessing divergent and useful properties. In particular Cas12a/Cpf1, similarly 62 to Cas9, functions as an RNA guided homing endonuclease. However, unlike Cas9, Cas12a 63 can be targeted by a single crRNA (~40 nt) as opposed to requiring a combined crRNA and tracrRNA (~100 nt) (Zetsche et al., 2015). Furthermore, in contrast to Cas9, Cas12a 65 possesses RNase activity and can recognise and process an array of adjacent crRNAs within a 66 single transcript to enable targeting of the protein to multiple unique loci (Fonfara et al., 67 2016 (Zetsche et al., 2015). However subsequent work by Kim et al. showed it did possess 74 activity in mammalian cells (Kim et al., 2016). FnCas12a was initially characterised as 75 having a shorter PAM sequence than AsCas12a or LbCas12a in vitro (Zetsche et al., 2015). 76 Subsequent cleavage assays in mammalian cells has shown a PAM sequence 77 'K(G/T)Y(C/T)TV(A/C/G)' enables optimal targeting for FnCas12a (Tu et al., 2017), 78 compared to 'TTTV' for both AsCas12a and LbCas12a (Kim et al., 2017). 'KYTV' can on 79 average be found every 21 nt. This targeting density is highly comparable to the targeting 80 capacity of SpCas9 which has a PAM sequence 'NGG', which can on average be found every 81 16 nt. In contrast, the As/LbCas12a PAM sequence 'TTTV' can only be found on average 82 every 85 nt. 83

84
The ability to target more synthetic transcription factors to a specific genomic region 85 becomes essential in cases where narrow windows of targeting are optimal and in particular, 86 when carrying out multiplexed targeting. One clear example is the case of gene network 87 manipulation, where; 1) there is a 350 nt window within the promoter region where optimal 88 transactivation is observed (Gilbert et al., 2014), 2) multiple promoters will be 89 simultaneously targeted and 3) multiple studies including this work show that targeting more 90 than one copy of the synthetic transcription factor to the same promoter can enable enhanced 91 transactivation (Maeder et al., 2013;Tak et al., 2017). 92

93
In the following work we show that FnCas12a can be engineered and applied as a synthetic 94 transcription factor in mammalian cells, before subsequently exploring whether dFnCas12a-95 VPR shows orthogonality when screened alongside dAsCas12a-VPR and dLbCas12a-VPR. 96 We then test whether single crRNAs are sufficient for gene activation and look for 97 synergistic transactivation when multiple crRNA target a single promoter. We further explore multiplexed activation from single crRNA arrays. Finally, we look into the role of position of 99 targeting crRNA within 6-crRNA arrays on the capacity to transactivate targeted genes.

126
The top strand of the promoter region contains six repeated binding sequences for the dCas12a constructs 127 (dark blue) with adjacent PAM sequences that can be recognised by all three variants 'TTTC' (orange).

128
The bottom strand of the promoter region contains 6 repeated binding sequences for dCas9-VPR (light 129 blue) with adjacent PAM sequences that can be recognised by dCas9-VPR (purple). B) Diagrammatic 130 representation of the dual luciferase reporter assay. Alongside the targeted Firefly luciferase reporter 131 plasmids, a non-targeted Renilla luciferase plasmid was delivered to enable normalisation of the relative 132 Firefly luciferase activity between test and control conditions. If a putative synthetic transcription factor 133 was able to transactivate the targeted Firefly luciferase gene, then the ratio of Firefly to Renilla luciferase 134 activity would be increased compared to the negative control condition. C) Testing the three dCas12a-VPR 135 variants alongside dCas9-VPR using the dual luciferase assay. Each construct is delivered with a targeting crRNA/gRNA and the resulting ratio is normalised to the ratio when delivered without a crRNA/gRNA.

137
The results represent three biological replicates and the error bars display the SEM. 138 139

-Orthogonality observed between dCas12a-VPR variants 140 141
We subsequently sought to test for orthogonality between the three dCas12a-VPR variants 142 ( Figure 2A). The activity of each variant when delivered with crRNAs from each of the 143 variants or none was measured using the dual luciferase assay previously described (  Analysis by qRT-PCR showed the mean increase in mRNA abundance for the co-transfected 207 condition was consistently higher than the most active individual crRNA ( Figure 4A). When 208 a two tailed t-test between the co-transfected and most active individual crRNA conditions 209 was performed, we saw a significant increase in mRNA abundance for; HBB crRNA 1 + 2 (P 210 = 0.017), ASCL1 crRNA 2 + 4 (P = 0.001), and IL1RN crRNA 4 + 6 (P = 0.005). We also 211 observed a non-significant increase in mRNA abundance for HBB crRNA 4 + 5 (P = 0.0875). 212 For one of the tested crRNA pairs (IL1RN crRNA 2 and 5) the spacer sequences had partial 213 complementarity (12 nucleotides). This may explain the small decrease in mRNA abundance 214 observed when comparing co-transfection to IL1RN crRNA 2 individually (non-significant, P 215 = 0.117). As a result, the IL1RN crRNA 2 + 5 pair was excluded from subsequent analysis. 216 217 As one of the advantages of Cas12a is the capacity to process crRNA arrays, we sought to 218 test whether 2-crRNA arrays, consisting of a pair of crRNAs in tandem, could be utilised by 219 dFnCas12a-VPR for transactivating target genes and whether these short arrays would enable 220 increased or synergistic activation compared to the delivery of individual crRNAs. To test 221 this, 2-crRNA arrays were constructed using the most active crRNA pairs from the preceding 222 experiments. Three days after transfection into HEK293 cells, the fold upregulation induced 223 using these arrays was tested compared to the co-transfected crRNAs and a non-targeting 224 crRNA control. We consistently observed that the crRNA arrays performed as well if not 225 better than the co-transfected crRNAs, with a higher mean fold upregulation for the arrays 226 compared with the co-transfected crRNAs in all cases ( Figure 4B (ASCL1 crRNA 2) within the array could be tested ( Figure 6A). The capacity of the each 295 crRNA array to up-regulate ASCL1 expression was then measured as previously described. 296 The results showed a reduction in transactivation of ASCL1 as the targeting crRNA was This translates to being able to target on average every 21 nt as opposed to on average every 317 85 nt, highly comparable to the Cas9 PAM sequence 'NGG' which enables targeting on 318 average every 16 nt. This allows much denser targeting, with more potential targets within 319 any given length of DNA. This is of particular interest for transactivation of target genes as 320 we have also highlighted that delivery of multiple active crRNAs targeting the same promoter 321 region further enhances up-regulation. 322

323
We found that dFnCas12a-VPR can be used for multiplexed transactivation of three different 324 genes from a single transcript. We have consistently observed a reduction in activity for 325 crRNAs positioned closer to the 3' of crRNA arrays expressed from the U6 pol 3 promoter. 326 This information can inform design constraints when targeting multiple genes for upregulation from a single array. In particular, arrays can be designed to express crRNAs 328 closer to the 5' end of an array where they target genes that are challenging to upregulate or 329 where higher overexpression is desired. Conversely, crRNAs targeting genes that are easier 330 to up-regulate or where lower over-expression is desired can be positioned closer to the 3' 331 end of an array.