Retaining pluripotency and exogenous mRNA introduction in planarian stem cell culture

Planarians possess naturally occurring pluripotent adult somatic stem cells (neoblasts) required for homeostasis and whole-body regeneration. However, no reliable neoblast culture methods are currently available, hindering mechanistic studies of pluripotency and the development of transgenic tools. We report the first robust methods for neoblast culture and delivery of exogenous mRNAs. We identify optimal culture media for maintaining neoblasts in vitro, and show via transplantation that cultured stem cells retain pluripotency for two days. We developed a new procedure that significantly improves neoblast yield and purity by modifying standard flow cytometry methods. These methods enable the introduction and expression of exogenous mRNAs in neoblasts, overcoming a key hurdle impeding the application of transgenics in planarians. The advances in cell culture reported here create new opportunities for mechanistic studies of planarian adult stem cell pluripotency, and provide a systematic framework to develop cell culture techniques in other emerging research organisms.

reported formulations (e.g., IPM and TPP), and dilutions of these media that 84 better match osmolarity suitable for planarian cells (~120 mOsm/kg) 85 (Supplementary Table 1 To measure viability, cells cultured for one day were observed and stained 87 with propidium iodide (PI), which labels the DNA of dead cells. We determined 88 the percentage of PI negative cells by flow cytometry. Cells cultured in CMFB 89 with or without 5% CO 2 modification displayed irregular cell surface morphologies 90 accompanied by sizeable cellular debris, suggesting poor viability (Fig. 1d). 91

Consistent with microscopic evaluation, cells cultured in CMFB showed poor 92
survival with or without 5% CO 2 modification (>60% dead cells) (Supplementary 93 Fig. 1a). In contrast, cells in all other conditions, such as IPM with or without 5% 94 CO 2 , had normal morphology, suggesting high viability (Fig. 1d). Among all 95 media conditions, seventeen formulations yielded viability higher than 60% 96 ( Supplementary Fig. 1a). Notably, cells in Leibovitz's L-15 medium (L15) without 97 5% CO 2 extended long protrusions that were visible even after six days of culture 98 and smedwi-1-cell morphology in the isolated populations showed that smedwi-175 1+ cells were generally larger than smedwi-1-cells (Fig. 2b). To discriminate 176 between small and large cells in the SiR-DNA+ population, the cytoplasmic dyes 177 Cell Tracker Green (CT) and Calcein AM (CAM) were tested in combination with 178 SiR-DNA for neoblast isolation (Fig. 2c, d). This dual dye staining strategy 179  irradiated planarians (Fig. 2g). Importantly, no noticeable differences in colony 187 size were observed at 7 dpt among unstained X1(FS), single (SiR-DNA)-, or 188 double (SiR-DNA/CT)-stained populations (Fig. 2g) Fig. 4a). In addition, we observed no positive effects 208 of co-culturing differentiated X1(FS) cells with SirNeoblast (Supplemental Fig.  209 4b). Based on these findings, we concluded that SiR-DNA/CT dual label-based 210 cell sorting could be used to isolate clonogenic, pluripotent neoblasts. Moreover, 211 these isolated SirNeoblasts can be maintained in primary culture and serve as 212 donor cells in transplantation assays. 213

Exogenous mRNA delivery by electroporation 214
Following the optimization of in vitro culture conditions to maintain neoblast 215 pluripotency, we next tested different conditions for the delivery of exogenous 216 (nucleic acid) molecules into neoblasts to attempt the genetic transformation of 217 planarians. To this end, we first used dextran-FITC as a fluorescent indicator of 218 membrane permeation to screen for the most suitable electroporation conditions 219 of Hoechst 33342-stained neoblasts (Supplemental Fig. 5a). We tested 52 220 electroporation programs and 10 different buffers using X1 cells (Hayashi et al., 221 2006;Reddien et al., 2005) and found that dextran-FITC was most effectively 222 delivered into neoblasts using IPM buffer with electroporation at 100-120V 223 (Supplementary Table 2 and Supplemental Fig. 5b-d). Similarly, applying this 224 electroporation method to X1(FS) cells rather than Hoescht 33342-sorted X1 225 cells showed that dextran-FITC+ populations could only be detected following 226 electroporation at 110V and 120V. However, less than 6% of dextran-FITC+ 227 X1(FS) cells were smedwi-1+ neoblasts, and virtually no smedwi-1+ cells could 228 be detected after one day of culture in KnockOut DMEM with 5% CO 2 229 (Supplemental Fig. 5e). Consistent with the drastic reduction in smedwi-1+ cell 230 viability post-electroporation, none of the donor X1(FS) cell populations subjected 231 to more than 100V formed colonies following transplantation into lethally 232 irradiated donors (Supplemental Fig. 5f). We hypothesized that this failure was 233 likely due to the low proportion of smedwi-1+ neoblasts in total X1(FS) cells, 234 which was even further reduced after electroporation. 235 Based on these findings, we then sought to identify the optimal electroporation 236 conditions to retain the viability of SirNeoblasts following the introduction of 237 foreign genetic material (Fig. 3a). Consistent with previous experiments, 238 electroporation at 110V-120V was required for dextran-TMR internalization into 239 SirNeoblasts (Fig. 3b, c). Different from that in X1(FS), smedwi-1+ cells were 240 more abundant in the electroporated (110V and 120V) SirNeoblasts compared to 241 X1(FS) cells, and the electroporated SirNeoblasts persisted for one day in culture 12 ( Fig. 3d). In addition, the electroporated SirNeoblasts were capable of forming 243 colonies and rescuing lethally irradiated hosts upon transplantation (Fig. 3e, f). 244 However, 120V electroporation resulted in SirNeoblast rescue of relatively fewer 245 irradiated hosts, suggesting that high voltages may negatively impact SirNeoblast 246

viability. 247
To assess whether exogenous mRNA could be delivered into SirNeoblasts by 248 electroporation, tdTomato mRNA was added to the electroporation reaction along 249 with Dextran-FITC. Dextran-FITC-positive SirNeoblasts were sorted and cultured 250 in KnockOut DMEM with 5% CO 2 . To confirm the successful delivery of mRNA, 251 we probed cells via FISH at 20 hours after electroporation and found detectable 252 tdTomato mRNA signal in cells electroporated at either 110V or 120V (Fig. 3g). 253 However, co-staining with smedwi-1+ revealed that not all tdTomato mRNA+ 254 cells retained neoblast identity in culture. The number of SirNeoblasts positive for 255 both tdTomato mRNA and smedwi-1 expression was significantly higher after 256 110V electroporation than after 120V, which was a similar response to 257 electroporation to that observed in X1 and X1(FS) cells (Fig. 3h). These findings 258 indicated that 110V electroporation was the most suitable condition for 259 introducing exogenous, charged molecules such as RNA into neoblasts, while 260 maintaining their viability and pluripotency. 261 Unfortunately, tdTomato expression was undetectable by either microscopy 262 or antibody staining in cultured neoblasts. This effect was likely due to two 263 potential contributing factors: 1) The culture conditions did not support the 264 translation of the delivered mRNA, or 2) an unknown mechanism prevented 265 suggested that further optimization to culture conditions and mRNA sequence 288 was necessary for reliable transformation. A recent study reported that Nanoluciferase (NanoLuc) mRNA could be 292 expressed in somatic planarian cells through Viromer or TransIT transfection 293 (Hall et al., 2021). We therefore suspected that the high sensitivity and low 294 autofluorescence background of the NanoLuc reporter could provide a tractable 295 approach for visualizing neoblast transgene expression. To confirm that NanoLuc 296 was indeed translated in cultured SirNeoblasts, we transfected SirNeoblasts with 297 NanoLuc mRNA using the TransIT system (Fig. 4a). Culture medium 298 supplements (sodium pyruvate, vitamin, and amino acids, see Supplemental 299 Table 1) were also included in the modified KnockOut DMEM to determine 300 whether they could enhance the mRNA expression (Hall et al., 2021). We found 301 that NanoLuc expression levels were higher in SirNeoblasts cultured in modified 302 KnockOut DMEM with 5% CO 2 modified atmosphere than in cells grown under 303 the same conditions without supplements (Fig. 4b). To confirm that neoblasts 304 could also be maintained in modified KnockOut DMEM with 5% CO 2 , 305 SirNeoblasts were stained and their proportion was compared to that of smedwi-306 1+ cells after one or three days of culture in modified KnockOut DMEM with 5% 307 CO 2 , KnockOut DMEM with 5% CO 2 , and Iso-L15 under ambient conditions. The 308 results showed that SirNeoblasts could be maintained on modified Knockout 309 DMEM as well as the Knockout DMEM, but not the Iso-L15 (Fig. 4c, d), which 310 combined with earlier findings that higher CO 2 was required for consistently high 311 SirNeoblast activity (Fig. 1e), led us to use modified KnockOut DMEM with 5% 312 CO 2 modification in subsequent experiments. We next sought to compare the 313 efficiency of mRNA delivery between the TransIT system and electroporation and 314 discovered that the NanoLuc signal was only detectable following TransIT 315 transfection (Fig. 4e). This finding suggested that TransIT provided higher 316 efficiency and more robust mRNA delivery into the neoblasts. 317 To confirm this methodology with other mRNAs, we used TransIT to deliver 318 mRNA encoding NanoLuc, smed-histone3.3-2xflag, mCherry, or NanoLuc-319 mCherry. Surprisingly, none of these proteins were detectable by Western blot 320 signals. However, NanoLuc-mCherry transcripts exhibited a measurably lower 332 signal due to a relatively longer coding sequence for reduced transfection 333 efficiency, which suggested that the signal was indeed due to NanoLuc transcript 334 expression (Fig. 4m). 335 To investigate if this low ratio was attributable to either a low transfection 336 efficiency, a low translation efficiency, or a combination of issues, we next 337 detected the presence of NanoLuc mRNA in SirNeoblasts after transfection. The 338 transfection efficiency was unexpectedly high, with close to 92.03% of 339 SirNeoblasts carrying the Nanoluc transcripts ( were pre-coated with poly-D-lysine (50µg/ml, BD Biosciences). 463

In situ hybridization and antibody staining 464
Whole-mount in situ hybridization was carried out as previously described 465  Fig. 2a, b). Considering the rate of ~10% cell death in culture, 496 we cultured 5 X 10 4 X1(FS) cells for each test condition to ensure that sufficient 497 viable cells were available at the time of transplant. 498

mRNA synthesis and electroporation 499
According to the protocol, the capped mRNA with poly(A) tail was transcribed 500 in vitro via mMESSAGE mMACHINE T7 ultra kit (ThermoFisher Scientific, 501 AM1345). tdTomato mRNA was transcribed from the linearized plasmid 502 pcDNA3.1(+)-tdTomato. The PCR product used as a template was amplified by 503 loader. Several slides were prepared at once and then loaded and processed 532 automatically using a combination of Nikon Elements Jobs for all robot and 533 microscope control and Fiji for object-finding and segmentation. Slides were 534 imaged at low magnification, and objects were identified before re-imaging tiled 535 z-stacks using a Plan Apo 10X 0.5NA air objective. Tiled images were stitched, 536 projected, and smedwi-1+ puncta were counted using custom macros and 537 plugins in Fiji. 538 Generation of optimized mCherry sequence 539 mCherry candidate sequences were generated using a custom python script. 540 Amino acid sequences were back-translated to 21 nucleotide sequences from 7 541 amino acid words at a time. Each potential nucleotide sequence was screened 542 against a list of known piRNAs to generate the sequence with the fewest piRNA 543 matches. A piRNA match consists of no more than a single G/T mismatch in the 544 six nucleotide seed region (positions 2-7 of a piRNA) (Zhang et al.,  545   2018). Additional G/T mismatches were scored as .5 and other mismatches as 1. 546 Only the first 21 base pairs of the piRNAs were aligned. The highest scoring 547 piRNA determined the score for that potential nucleotide sequence. The 21 548 nucleotide sequence with the lowest score was retained. The script was run with 549 four alternate coding tables. The "all" coding table contained all possible codons 550 for each amino acid., The "highestgc" contained only those codons with the most 551 G or C nucleotides. All 15 modified mCherry sequences are included in 552 Supplemental Table 3.

Data availability 615
All codes used for plugins in Fiji are available at:https://github.com/jouyun. All 616 original data underlying this manuscript can be accessed from the Stowers 617 Original Data Repository at: http://www.stowers.org/research/publications/libpb-618 1281. All reagents are available from the corresponding author upon reasonable 619 request. 620

Statistical analyses 621
Microsoft Excel and Prism 6 were used for statistical analysis. Mean ± s.e.m. 622 is shown in all graphs. Unpaired two-tailed Student's t-test was used to 623 determine the significant differences between the two conditions. P < 0.05 was 624 considered a significant difference. 625

Acknowledgments 626
We thank I. Wang and P. Reddien for assistance with the transplantation 627 technique. We thank all members of Sánchez Lab, especially J. Jenkin and C. 628 Guerrero, for animal maintenance and irradiation assistance, L. C. Cheng and E.