VitelloTag: a tool for high-throughput cargo delivery into oocytes

Delivering molecular tools into oocytes is essential for developmental and reproductive biology. Microinjection, the conventional method, is equipment-intensive, often technically challenging, and low-yield, and is impractical in species with delicate oocytes or restricted spawning seasons. To overcome these limitations, we developed VitelloTag, a cost-effective, high-throughput system using vitellogenin-derived fusion proteins to enable efficient cargo delivery via receptor-mediated endocytosis. We demonstrate its utility by delivering Cas9/sgRNA complexes in two distantly related species for gene knockout.


Introduction
CRISPR/Cas9 technology has revolutionized functional studies not only in established model systems (Friedland et al., 2013;Gratz et al., 2013;Hwang et al., 2013;Wang et al., 2013), but also in a range of emerging research organisms (Munro et al., 2023;Oulhen et al., 2022;Perillo et al., 2023;Pickett and Zeller, 2018).Nonetheless, a major bottleneck in many experiments is the delivery of CRISPR reagents into oocytes for the germline transmission of mutations and analysis of mutant phenotypes in F0 embryos.
Typically, this involves microinjecting gene-editing reagents directly into the oocyte or fertilized zygote.However, microinjection is often a substantial or insurmountable technical hurdle in many species: oocytes are often too fragile to withstand the injection process or possess tough outer chorions, and some species, like many species of coral that spawn only one night per year, can have very limited spawning seasons making gene manipulation by microinjection impractical.Moreover, microinjection often has a steep learning curve and requires specialized training, involves expensive equipment, and is labor-intensive.
To overcome these limitations, we developed a cell-penetrating technique utilizing a natural receptor-mediated endocytic pathway to facilitate specific and efficient trafficking into oocytes.Our approach focuses on harnessing vitellogenesis, the process of yolk accumulation in growing oocytes (Fig. 1a).We targeted Vitellogenin (Vtg), a widely conserved yolk protein that is found in all metazoan phyla except ctenophores (Byrne et al., 1989;Lebouvier et al., 2022), to develop a versatile tool applicable in a phylogenetically diverse range of species.The N-terminal region of Vtg contains a receptor-binding domain (RBD) that specifically interacts with the Vitellogenin receptor (VtgR) to facilitate uptake (Fig. 1b) (Li et al., 2003).We hypothesized that the Vtg RBD could be used to generate a fusion tag sufficient to enable trafficking of protein cargos into oocytes.

Results and Discussion
Using the sea star, Patiria miniata, as a representative model widely used for reproductive and developmental biology studies, we tested whether the RBD region of Pm Vtg1 was sufficient to transport protein cargo into oocytes.To identify required RBD regions, we recombinantly fused the N-and C-terminal halves of the RBD with Superfolder GFP (sfGFP) and a nuclear localization sequence (NLS)(Fig.1c), and tested their endocytosis into oocytes in vitro (Fig. 1d).Both Vtg fragments were successfully endocytosed with equivalent efficiency (Fig 1e, f).GFP-containing vesicles predominantly localized to the cell periphery.Consistent with this acting through the vitellogenesis pathway, uptake required dynamin-mediated endocytosis (Supp.Fig. 1a,   b) (Macia et al., 2006).Because Vtg is typically targeted to acidified vesicles, endosome escape is crucial for effective cargo delivery into the cytoplasm.We found that the endosome escape reagent chloroquine induced the release of Vtg-sfGFP, confirming their import into the nucleus via their NLS (Supp.Fig. 1c, d).Importantly, chloroquine was well-tolerated during development and did not disrupt embryogenesis (Supp.Fig. 2).Collectively, these results demonstrate the potential of Vtg protein fusions for cargo delivery en masse in simple in vitro oocyte cultures.
Our observations indicated that both halves of the Pm Vtg1 RBD contain binding motifs that are sufficient for oocyte transport.To identify these sequences computationally, we used a modeling approach with AlphaFold Multimer to predict interactions between short fragments of Pm Vtg1 and Pm VtgR (Jumper et al., 2021;Mirdita et al., 2022)(Fig.1g).By focusing on fragments predicted to have the highest number of interactions with Pm VtgR (Supp.Fig. 3a) and utilizing a per-residue interaction score (Supp.Fig. 3b) in conjunction with structural modeling (Supp.Fig. 3c,   d), we identified two short (~10-15 amino acid) "best hit" motifs in the N-and C-terminal halves.To assess whether these minimal sequences, which we termed VitelloTag A (residues 28-36: SITIHRNTP) and VitelloTag B (residues 197-209: LMTILNVTKVRDL), could serve as minimal oocyte transport tags, we cloned sfGFP fusions and subjected them to the same in vitro oocyte transport assay.Notably, sfGFP fusions of VitelloTag A and B showed equivalent uptake efficiency to the larger parent fragments (Fig. 1h, i).

Development • Accepted manuscript
To determine the utility of VitelloTags for transporting gene editing reagents, we generated NLS-eSpCas9(1.1) fusion proteins with each VitelloTag.We first tested for successful internalization of the VitelloTagCas9 constructs using immunostaining.The constructs were successfully internalized into the oocyte cytoplasm and showed comparable nuclear localization to Cas9 mRNA injected oocytes, along with the anticipated endosomal signal near the cortex due to its endocytosis (Supp.Fig. 4).We then tested both Vtg-Cas9 proteins for the generation of F0 knockout embryos by targeting the delta locus (which has been extensively studied and shows a clear phenotype) with three synthetic sgRNAs (used together).By 3 days post fertilization, delta knockout embryos displayed strong defects consistent with previously published delta knock-down and knockout phenotypes in P. miniata (Fig. 2a) (Cary et al., 2020;Perillo et al., 2023).Specifically, we scored the number of mesenchyme cells, which is significantly increased compared to Cas9 alone controls in both VitelloTagA:Cas9 and VitelloTagB:Cas9 (Fig. 2b).The frequency of phenotypic embryos was 32% and 63%, respectively (see Methods; Fig. 2c).We also performed HCR in situ hybridizations against gcm, whose overexpression is a marker for delta knockout in P. miniata embryos (Zueva and Hinman, 2023).We saw a greater number of cells expressing gcm as compared to the untreated controlled embryos (Supp.Fig. 5).Genotyping analysis of individual embryos by TIDE spectral decay, a method that determines the percentage of sequences amplified from a genomic locus that contain an indel ("total efficiency") (Brinkman et al., 2014), confirmed disruption of the delta allele (30.4% and 43.9%, respectively, for VitelloTag-Cas9 versus microinjection of Cas9 mRNA) (Supp. Fig. 6).VitelloTag:Cas9 is thus an effective reagent for high-throughput generation of knockout embryos, though knockout efficiency is mosaic based on TIDE analysis and phenotypic penetrance.
To evaluate the potential of VitelloTag:Cas9 for genome editing in other more distantly related organisms, we next tested it in the hemichordate enteropneust, Saccoglossus kowalevskii(Fig.2d).As Saccoglossus is challenging to microinject, VitelloTag could significantly enhance the experimental possibilities in this organism.
Our experiments showed that both VitelloTag A and B Cas9 fusions successfully induced indel mutations knockout phenotypes in Saccoglossus, albeit with varying Development • Accepted manuscript penetrance (10% and 48%, across two target genes, for VitelloTag A and B, respectively; Fig. 2e).VitelloTag-B:Cas9 was particularly effective, generating penetrant knockout phenotypes for two target genes, bmp2/4 and beta-catenin, at frequencies approaching the efficiency seen in P. miniata (38% and 47%, for bmp2/4 and betacatenin, respectively.N = 45 in 3 replicates of 15 embryos)(Fig.2f, g).The observed phenotypes were consistent with zygotic gene knockout: embryos treated with gRNAs targeting bmp2/4, which is active after zygotic genome activation, phenocopied the previously published RNAi phenotype of radialization of the dorsal-ventral axis (Fig. 2f) (Lowe et al., 2006), whereas embryos treated with gRNAs targeting beta-catenin, which has both maternal and zygotic effects, displayed only the zygotic phenotype of anteriorization of the anterior-posterior axis (Fig. 2g, Supp.Movie 1) (Darras et al., 2011;Darras et al., 2018).These results represent the first knockout phenotypes in this important group of animals.
Our results establish the feasibility of cell-penetrant CRISPR/Cas9 genome editing in oocytes for comparative developmental studies.We show that VitelloTags enable high-throughput trafficking of active Cas9-sgRNA ribonucleoprotein complexes in two species, overcoming multiple limitations of standard microinjection protocols.
While similar strategies have been developed in arthropods (e.g., ReMOT Control in mosquitos (Chaverra-Rodriguez et al., 2018)), they exploit non-conserved proteins, and are therefore limited in their cross-species utility.Thus, our results set the stage for broader use of vitellogenin-based transgenesis in diverse emerging model systems.We identify VitelloTag B as a potential cross-reactive oocyte transport tag that is effective in two distantly related species.Sequence conservation analysis indicates that VitelloTag B resides in a highly conserved region of the Vitellogenin receptor binding domain (Supp.Fig. 7), which suggests that it may be functional in a broader array of organisms.
Not only does Vitellotag potentially open up gene editing approaches to a wider phylogenetic range of animals, it may also facilitate functional studies in more remote and poorly equipped field stations where most biodiversity is accessible but microinjection is impractical.These simple methods also lend themselves well to undergraduate teaching laboratories.We note that this version of VitelloTag does not achieve complete knockout efficiency.This may be due to incomplete endosome Development • Accepted manuscript escape or the short half-life of Cas9 protein in animal cells, two parameters that will require further work to optimize (Kim et al., 2014).Importantly, this system may be potentially adapted for the delivery of nucleic acids and other proteins, further expanding the functional tools available in non-canonical research organisms.

Resource Availability
Plasmids for bacterial production of VitelloTag proteins will be made available for academic use on Addgene.

Research Organisms
Wild-caught Bat Stars (Patiria miniata) were obtained from South-Coast Bio LLC (San Diego, CA) and Monterey Abalone Company (Monterey, CA).Age of the animals cannot be determined.Individuals were randomly selected for oocyte and/or sperm isolation.
Acorn worms (Saccoglossus kowalevskii) were field-collected in Waquoit Bay (Falmouth, MA).For Saccoglossus, natural spawning of oocytes and isolation of sperm was performed as previously described (Lowe et al., 2004).

Gonad Extraction
Intact ovary and testis fragments were surgically extracted using forceps through small slits on the oral side of an arm of the sea star.Ovary pieces were kept in sterile seawater containing 10 μg/mL trimethoprim and 50 μg/mL sulfamethoxazole at 15°C.
The ovary fragments were transferred to fresh seawater with antibiotics every other day and only fresh oocytes teased from the ovary were used for experiments (as described in (Swartz et al., 2019)).

Oocyte and Embryo Culture with VitelloTag
Isolated oocytes were cultured in sterile seawater plus antibiotics for a maximum of 24 hours.In all experiments involving Vtg:GFP uptake, treatment and control groups were incubated with 1 μg/mL Vtg:GFP and GFP respectively, overnight.For endosomal escape trials, oocytes were incubated with defined concentrations of chloroquine overnight in addition to Vtg:GFP.Dynasore trials were incubated with defined concentrations of Dynasore in addition to Vtg:GFP, alongside vehicle controls (DMSO 1:1000 dilution).For all experiments, oocytes were kept at 15°C.
For chloroquine toxicity assays, oocytes were incubated overnight (16-18 hours) at 15°C in varying concentrations of chloroquine.On the next day the oocytes were stimulated with 1-methyladenine at a final concentration of 10μM to induce meiotic resumption.For fertilization, extracted sperm was added at a 1:500,000 dilution prior to emission of the first polar body.Fertilized oocytes were transferred to fresh seawater without antibiotics and kept at 15°C.

Sample Preparation and Imaging
For live imaging, oocytes were mounted on a glass dish in approximately 10μL of seawater under a glass coverslip.For analysis of mesenchyme cells, P. miniata embryos were fixed in 2% PFA/SW for 30 minutes, washed in PBST and stained with Phalloidin (1:300) and DAPI.For HCR in situ hybridization, P. miniata embryos were fixed in 4% PFA/MOPS Fixative buffer and then processed with an echinoderm HCR protocol published previously (Formery et al., 2023).For Cas9 immunofluorescence, P. miniata oocytes were fixed in 2% PFA/SW for 1 hour, washed with PBST and methanol, and then stained with anti-Cas9 antibody (Sigma, #SAB4200701-25UL; 1:500 dilution) and Hoechst.S. kowalevskii embryos were fixed in 4% formaldehyde for 1 hour, as described previously (Lowe et al., 2004).Embryos were then either stained with antiacetylated Tubulin antibody (Sigma, #T7451; 1:500 dilution) and DAPI, or prepared for Development • Accepted manuscript fluorescent in situ hybridization using the HCR method (Molecular Instruments) following the manufacturer's instructions.

Image quantification and statistical analysis
Statistical analyses were performed using Prism (10.Quantification of nuclear and cytoplasmic signal intensity was performed using Fiji/ImageJ (Schindelin et al., 2012).A maximum intensity z-projection of the first 31 slices of each image was performed to measure cytoplasmic GFP signal intensity (Fig. 1e, h and Supp.Fig. 1a).Projections were then background subtracted with the rolling ball (radius = 50 pixels) and sliding paraboloid function.A 120-micron circular region of interest was used to measure the pixel intensity in the cytoplasm.Quantification was obtained using the RawIntDen parameter in the Measure function.To measure pixel intensity in the nucleus (Supp.Fig. 1c), a single z-slice which was captured inside the nucleus of the oocyte was chosen.A 50-micron circular region of interest was placed in the nucleus, another 50-micron circular region of interest was placed in the background.
Pixel intensities of both regions were obtained using the RawIntDen parameter in the Measure function.The RawIntDen value of background was subtracted from that of the nucleus to obtain the final quantification of pixel intensity for a given image.For measuring Cas9 nuclear signal (Supp Fig. 4), shallow z-projections of 15 slices each were taken for all samples.These projections then underwent background subtraction with the rolling ball (radius = 50 pixels) and sliding paraboloid function.A 50-micron circular region of interest was then placed in the nucleus of each oocyte.Pixel intensities of Cas9 signal were obtained using the RawIntDen parameter in the Measure function.The "multi-point" tool on ImageJ was used to count the number of individual gcm positive cells in an embryo (Supp.Fig. 5).

Protein Structure Prediction and Interaction Analysis
Overlapping tiled sequence fragments of Pm Vtg1 RBD (residues 1-315) were generated using a 50AA sliding window with a step size of 25AA, and then used as input for AlphaFold Multimer along with the N-terminal region of Pm VtgR (LDL domains 1-3, residues .Interactions between chains within 4 Angstroms were counted per fragment, and then weighted by multiplication with the average pLDDT score across the fragment.Per-residue normalized interactions was calculated as the average interaction score between overlapping fragments, per residue position.Best peaks from both N-and C-terminal halves were targeted for in-depth structural analysis, which was then used to pinpoint contiguous regions with a high number of predicted protein-protein interactions.

Molecular Cloning
Amplicons were generated by PCR using Q5 proof-reading polymerase (NEB) using primers reported in Supplementary Table 1.Constructs were then generated using Gibson Assembly Master Mix (NEB) following manufacturers instructions.

Protein Expression and Purification
Vtg-sfGFP fusion proteins were expressed as 6x His-tagged constructs in BL21(DE3) E.
coli.Protein expression was induced with 0.1 mM IPTG overnight at 18°C.Vtg-Cas9 fusions that were toxic to BL21 cells were produced using BL21-AI E. coli cells.Here, protein expression was induced with 0.1% Arabinose overnight at 18°C.Proteins were purified from 3 liters of bacterial culture with Ni-NTA Agarose Beads (GoldBio), eluted with imidazole (Sigma) and purity was confirmed by SDS-PAGE analysis.
A step-by-step protocol is provided in the Supplementary Materials.In brief, a solution of VitelloTagA:Cas9 (7 mg/ml) or VitelloTagB:Cas9 (4 mg/ml) used 1:250, plus 0.8 µl each of 100 µM delta gRNAs 198, 248 and 519 (target sequences as in (Perillo et al., 2023), and are reported in Supplementary Table 2) and NEB buffer r3.1 in a final volume of 20ul was left for 20 min at 25ºC followed by 5 min at 37ºC to form a Cas9-gRNA ribonucleoprotein (RNP) complex.The RNP solution was then added to oocytes together with 25 µM chloroquine and antibiotics 1:1000 in 500 µl of filtered sea water.
Oocytes were kept at 15ºC overnight, and then matured and fertilized.Fertilized oocytes were washed with antibiotics-free fresh filtered seawater to remove chloroquine, antibiotics and VitelloTag:Cas9.After gastrulation, The delta knockout phenotype was scored by counting the number of mesenchyme cells present in the embryo; phenotypic embryos were defined as those having cell numbers at least two standard deviations higher than the mean value for Control embryos.

Individual Larva genotyping and TIDE Spectral Decay Analyses
Individual larvae were genotyped with PCRs of individual embryos which were sequenced and used for TIDE analysis using primers oZS1032 and oZS1033.
3.0 GraphPad Software).Details of the statistical test used, sample size, and significance values are included in figure legends.Images comparing the same signal across groups are scaled linearly and equivalently.
8 µM, followed by incubation at 25ºC for 20 minutes, then 37ºC for 5 minutes.Unfertilized Saccoglossus oocytes were collected in 500 µL of filtered seawater, then incubated with RNPs diluted 1:50 to a Development • Accepted manuscript final concentration of 36 nM at 22ºC in the presence of 25 µM chloroquine.After 4 hours, eggs were fertilized with 1 drop of diluted sperm solution, and kept in the same solution containing RNPs and escape reagent overnight for ~12 hours.Embryos were screened for phenotypes at 36 hours and 72 hours, for β-catenin and bmp2/4, respectively.Penetrance was quantified as the percentage of embryos displaying an unambiguous knockout phenotype.

Fig. S7 .
Fig. S7.Sequence conservation analysis of Vitellogenin a. Crystal structure of Vitellogenin from Lamprey (PDB: 1LSH), showing the receptor binding domain (RBD, magenta).b.Secondary structure and amino acid conservation of the RBD.Positions of VitelloTag A and B are indicated (black lines).Red box highlights the region shown in (c).c.Multiple sequence alignment of the VitelloTag-B region across representative species from 9 phyla, including vertebrates.Sequence conservation histogram and consensus logo are shown below.Red asterisks indicate positions with 100% amino acid sequence conservation.