zGrad: A nanobody-based degron system to inactivate proteins in zebrafish

The analysis of protein function is essential to modern biology. While protein function has mostly been studied through gene or RNA interference, more recent approaches to degrade proteins directly have been developed. Here, we adapted the anti-GFP nanobody-based system deGradFP from flies to zebrafish. We named this system zGrad and show that zGrad efficiently degrades transmembrane, cytosolic and nuclear GFP-tagged proteins in zebrafish in an inducible and reversible manner. Using tissue-specific and inducible promoters in combination with functional GFP-fusion proteins, we demonstrate that zGrad can inactivate transmembrane, cytosolic and nuclear proteins globally, locally and temporally with different consequences. Global protein depletion results in phenotypes similar to loss of gene activity while local and temporal protein inactivation yields more restricted and novel phenotypes. Thus, zGrad is a versatile tool to study the spatial and temporal requirement of proteins in zebrafish.


38
The study of the consequences of the loss of gene function is a central technique in biology. In 39 principle, loss of gene function can be achieved through gene, mRNA or protein inactivation.

150
Among other tissues, the cxcr4b promoter drives expression in the somites and the posterior 151 lateral line primordium (primordium) (14). Such embryos were heat shocked at 30 hpf for one 152 hour and imaged over 9.5 hours. Compared to control embryos that did not carry the 153 hsp70l:zGrad transgene, H2A-EGFP degradation was discernible in zGrad-expressing embryos 154 within two to three hours post heat shock in all tissues that expressed nuclear EGFP from the 155 cxcr4b promoter (skin, pronephros, somites, neural tube and primordium, Figure 2A-E, Figure 2 156 -Video 1). We quantified H2A-EGFP levels in heat-shocked control embryos and heat-shocked 157 hsp70l:zGrad embryos using the fluorescence intensity of H2A-mCherry as a reference. Since 158 H2A-mCherry was expressed from the same promoter as H2A-EGFP and since H2A-mCherry 159 is not recognized by the anti-GFP nanobody and should not be subjected to zGrad-mediated 160 protein degradation (7), comparing the ratio of H2A-mCherry expression levels to H2A-EGFP 161 should be a measure of H2A-EGFP in the absence of zGrad-mediated degradation. Further, we 162 normalized the H2A-EGFP-to-H2A-mCherry fluorescence intensity ratios between zGrad-163 expressing and heat-shocked control embryos. In the somites, the levels of H2A-EGFP was 164 decreased by 87%, while in the primordium, the levels of H2A-EGFP was decreased by 22% ( Figure 2C). The more efficient degradation of H2A-EGFP in the somites than in the primordium 166 is probably due to the lower levels of H2A-EGFP and the lack of H2A-EGFP production in the 167 somites at the time of zGrad induction.

169
We further characterized the kinetics of zGrad-mediated degradation by determining the time 170 interval between the start of the heat shock to induce zGrad expression and the first observable 171 difference in EGFP/mCherry levels between zGrad-expressing embryos and the control 172 embryos. We termed this time interval as the time for onset of degradation. We also fitted the 173 EGFP/mCherry ratio to a one-phase exponential decay model to extract the half-life of zGrad-174 mediated degradation. Although this is a simplification because the model does not account for 175 EGFP production -a variable that we cannot easily measure -we expect that it gives a rough 176 estimate for the time it takes to degrade a GFP-tagged protein once zGrad is expressed. The  (Table 2).

218
Together, these observations indicate that zGrad efficiently targets nuclear, cytoplasmic, and 219 transmembrane proteins tagged with EGFP or Citrine for degradation.

313
In our study, we sought to develop a tool that allows for the acute inactivation of proteins in 314 zebrafish. We modified the anti-GFP nanobody-based deGradFP system from flies (9) and 315 adapted it to zebrafish and named it zGrad. zGrad efficiently degrades GFP-tagged 316 transmembrane, cytoplasmic and nuclear proteins. It recognizes different GFP versions (EGFP, 317 sfGFP and Citrine, Figure 1 and 2) and targets the tagged proteins for degradation. In contrast to other degron systems (2) and similar to deGradFP (7), zGrad degraded proteins tagged at 319 the N-terminus, embedded within the protein and at the C-terminus in the examples reported 320 here (Figure 1 and 2). We found that zGrad rapidly degrades GFP-tagged proteins with half-321 lives of around 20 min for transmembrane and cytoplasmic proteins ( Figure 2 and Table 2).

322
These kinetics are similar to the kinetics reported for other degron-based systems (5, 11, 13).

323
The degradation of H2A-EGFP was significantly slower displaying a half-life of about 2.5 h 324 ( Figure 2 and Table 2). This is could be due to the long life-time of histone proteins (30) and the 325 possibility that protein degradation is less efficient in nuclei. Importantly, in our system zGrad

346
An important consideration is that zGrad needs to be expressed at high enough levels to 347 deplete proteins to sufficiently low levels to fully disrupt protein function. One way to achieve 348 high tissue-specific zGrad expression is to amplify the production of zGrad through such 349 systems as Gal4/UAS (32) or -to also add temporal control -inducible systems such as the

459
To construct pCS2+-sfGFP-mAID, the sfGFP coding sequence was amplified by PCR and the

465
To construct pCS2+-zif1, the coding sequence of zif1 was codon optimized for zebrafish by 466 gene synthesis (IDT) and inserted into the pCS2+ plasmid by Gibson assembly.

655
The procedures for RNA probe synthesis and whole-mount in situ hybridization were done as 656 previously described (50). The RNA probe against cxcr4b was previously described (38). The 657 template for the synthesis of the in situ RNA probe against vhhGFP4 was amplified from 658 pcDNA3-NSlmb-vhhGFP4 (Addgene plasmid #35579) using the following primer pair:

676
Images for Figure 6E were taken by mounting embryos in 3% Methyl cellulose. cxcr4b mutant 677 embryos were genotyped by PCR after image acquisition. The migrating distance of the 678 primordium was quantified manually using ImageJ.     Figure 2C, 2E, 2H, 2K, the curves were fitted to a one-exponential decay model (Y = 707 Span*exp(-k*X)+Plateau) using Prism 7 (Graphpad). The values of T1/2 and the plateau, which 708 we assumed to be the value of maximal degradation, were extracted from the fitted curves. In 709 Figure 5D (inset, 30 min to 240 min), two data sets were compared by paired-t test using Prism.