Cultured pluripotent planarian stem cells retain potency and 1 express proteins from exogenously introduced mRNAs 2 3

Planarians possess naturally occurring pluripotent adult somatic stem cells (neoblasts) required for homeostasis and whole–body regeneration. However, methods for culturing neoblasts are currently unavailable, hindering both mechanistic studies of potency and the development of transgenic tools. We report the first robust methodologies for culturing and delivering exogenous mRNA into neoblasts. We identified culture media for maintaining neoblasts in vitro, and showed via transplantation that the cultured stem cells retained pluripotency. By modifying standard flow cytometry methods, we developed a new procedure that significantly improved yield and purity of neoblasts. These methods facilitated the successful introduction and expression of exogenous mRNAs in neoblasts, overcoming a key hurdle impeding the application of transgenics in planarians. The tissue culture advances reported here create new opportunities to advance detailed mechanistic studies of adult stem cell pluripotency in planarians, and provide a systematic methodological framework to develop cell culture techniques for other emerging research organisms.


Abstract: 23
Planarians possess naturally occurring pluripotent adult somatic stem cells 24 (neoblasts) required for homeostasis and whole-body regeneration. However, 25 methods for culturing neoblasts are currently unavailable, hindering both 26 mechanistic studies of potency and the development of transgenic tools. We report 27 the first robust methodologies for culturing and delivering exogenous mRNA into 28 neoblasts. We identified culture media for maintaining neoblasts in vitro, and 29 showed via transplantation that the cultured stem cells retained pluripotency. By regenerative abilities and a seemingly limitless capacity for tissue homeostasis. Of 50 the many freshwater planarian species known to exist, Schmidtea mediterranea 51 has become one of the most widely studied 9 . Planarians thus provide a unique 52 context in which to explore how nature has solved the complex problem of 53 maintaining stem cell pluripotency in a long-lived adult animal. 54 Expression of conserved genes regulating pluripotency have been identified in 55 planarian neoblasts and functionally studied using RNA interference 10-13 . However, 56 due in part to the lack of methodologies for cell culture, exogenous gene 57 expression, and transgenesis in planarians, the mechanisms regulating the 58 pluripotency of these adult stem cells in vivo are poorly understood. Therefore, 59 developing planarian transgenesis is of great significance 14 . A review of the history 60 of cell culture methodologies and attempts to develop transgenics in planarians 61 cultured for 1 or 2 days efficiently proliferated in vivo, except for those cultured in 200 dGrace's medium + 5% CO2 (Fig. 2b-d). By comparing the number of smedwi-1+ 201 neoblasts in each transplant, X1(FS) cells cultured for 1 day in either IPM or 202 KnockOut DMEM formed the largest colonies in vivo (Fig. 2b, d). X1(FS) cells 203 cultured for 2 days displayed decreased expansion potential in all conditions, but 204 all were still capable of forming colonies in vivo with the exception of those cultured 205 in dGrace's medium + 5% CO2. X1(FS) cells cultured for 3 days were largely 206 incapable of forming colonies following transplantation, though small colonies 207 formed from cells cultured in dSchneider and dL15 media (Fig. 2c, d). In summary, 208 IPM and KnockOut DMEM performed best following the first day in culture, but 209 performed similarly to dKnockOut DMEM, dSchneider's, dL15, and dDMEM after 210 two days of culture. In addition, we observed that clonogenic capacity of X1(FS) 211 neoblasts diminished greatly following three days in culture, regardless of the 212 media used. These results suggest that IPM, KnockOut DMEM, dL15, dKnockOut 213 DMEM, dSchneider's, and dDMEM are capable of maintaining the proliferation 214 potential of neoblasts for up to two days in culture in the presence of 5% CO2. 215

Cultured neoblasts can rescue stem cell-depleted planarian hosts 216
To evaluate the functional pluripotency of neoblasts cultured in these six media 217 (IPM, KnockOut DMEM, dKnockOut DMEM, dL15, dSchneider's, and dDMEM), 218 rescue was assessed following bulk-cell transplantation. Genotyping PCR and 219 restriction fragment length polymorphism (RFLP) assays were performed to test 220 whether sexual hosts had been transformed into the asexual biotype following 221 transplantation of the asexual neoblasts ( Supplementary Fig. 5a) 8 . For non-222 cultured, freshly collected X1(FS) cells, 30-50% of the lethally irradiated (6,000 223 rads) sexual S. mediterranea hosts were rescued (Fig. 3b, c, and Supplementary 224 lethally irradiated hosts. These data indicate that of all conditions tested, KnockOut 232 DMEM +5% CO2 is the best one for maintaining pluripotent neoblasts in culture for 233 up to 2 days. IPM and dL15 medium were also capable of maintaining pluripotent 234 neoblasts in culture for up to 2 days albeit with reduced rates of irradiate animal 235 rescue after transplantation ( Fig. 3c and Supplementary Fig. 5e). 236 In summary, we found that after screening 23 media followed by assaying 5 237 criteria (i.e., viability, smedwi-1 expression, cell division, clonogenic capacity and 238 rescue efficiency of irradiated animals), three types of media (KnockOut DMEM, 239 IPM, and dL15) were capable of maintaining pluripotent neoblasts in vitro. Of these 240 three different media, KnockOut DMEM produced cultured neoblasts with the 241 strongest performance across the multiple measured criteria, with IPM and dL15 242 medium performing slightly less well (Fig. 3d). 243

Electroporation delivers fluorescent dextran into neoblasts 244
Following the successful optimization of in vitro culture conditions for the 245 maintenance of pluripotent neoblasts, we next attempted to test conditions for the 246 delivery of exogenous molecules into neoblasts, the next step required for 247 developing transgenic methods for planarians. We first used dextran-FITC as a 248 fluorescent indicator to screen suitable electroporation conditions for neoblasts 249 labeled by Hoechst 33342 staining (Fig. 4a). We tested 52 electroporation 250 programs and 10 different buffers using X1 cells 25,26 , and found that dextran-FITC 251 was delivered into neoblasts most efficiently in IPM buffer with electroporation at 252 100-120V (Supplementary Table 2  cell populations subjected to more than 100V formed colonies following 260 transplantation into lethally irradiated donors (Fig. 4f). We reasoned that the failure 261 was likely due to the low purity of smedwi-1+ neoblasts in X1(FS), which was even 262 further reduced after electroporation. Therefore, it was necessary to develop a new 263 strategy for neoblast isolation that would result in both higher clonogenic and 264 pluripotent smedwi-1+ cell enrichment than the X1(FS) sorting protocol. 265 We also tested whether X1(FS) can express exogenously delivered mRNA in 266 current culture conditions. We cloned a planarian endogenous gene, Smed-267 histone3.3 and fused with two copies of flag tag (2´flag). After electroporation and 268 one day of culture, cells electroporated with Smed-histone3.3-2´flag mRNA had 269 more anti-FLAG staining positive cells (9.7 ± 1.4%) than electroporated cells 270 without mRNA (1.2±0.7%) (Fig. 4g). Even though the anti-FLAG antibody stained 271 enucleated cells, nuclear localization signal in nucleated cells suggested 272 successful expression of Smed-histone3.3-2´flag mRNA ( Fig. 4g and 4h). Even smedwi-1+ neoblasts to a ratio ~60% (Fig. 5a, b, f and Supplementary Fig. 7). 285 Comparison of smedwi-1+ and smedwi-1-cell morphology in the isolated 286 populations showed that smedwi-1-cells were generally smaller than smedwi-1+ 287 cells (Fig. 5b). To discriminate between small and large cells in the SiR-DNA+ 288 population, the cytoplasmic dyes Cell Tracker Green (CT) and Calcein AM (CAM) 289 were tested in combination with SiR-DNA in neoblast isolation (Fig. 5c, d). Using SirNeoblasts and those cultured in KnockOut DMEM +5% CO2 for one day were 303 capable of rescuing lethally irradiated planarians at frequencies comparable to 304 those seen with X1(FS) cells ( Fig. 3c and Fig. 5h). Based on these results, we 305 conclude that SiR-DNA/CT dual labeling-based cell sorting can be used to isolate 306 clonogenic, pluripotent neoblasts that can be maintained in primary culture and 307 serve as donor cells in transplantation assays. To further characterize the 308 SirNeoblasts, we stained SirNeoblasts with Hoechst 33342 to analyze their cell 309 cycle. However, co-staining of SiR-DNA and Hoeschst 33342 resulted in a failure 310 to detect SiR-DNA staining. We then tested whether Hoeschst 33342 can stain 311 SiR-DNA stained cells, and found that Hoechst 33342 can replace the staining of 312 SiR-DNA, which showed the cell cycle distribution of SirNeoblasts consisted of 313 ~17.89% at G1, 13.02% at S, and ~69.09% at G2/M cell cycle phases 314 ( Supplementary Fig. 7). This reversible staining of SiR-DNA may also explain the 315 reason why SirNeoblasts can proliferate after staining unlike Hoechst 33342 316 stained X1 cells. 317 Next, we determined conditions to optimize electroporation efficiency and 318 viability for SirNeoblasts (Fig. 6a). Consistent with previous studies, 319 electroporation at 110V-120V was required for dextan-TMR entry into SirNeoblasts 320  smedwi-1 expression was significantly higher after 110V electroporation than after 342 120V (Fig. 6h). This result indicates that under the conditions tested, 110V 343 electroporation may be the most suitable to both introduce exogenous, charged 344 molecules such as RNA into neoblasts, while maintaining their viability and 345

potency. 346
Unfortunately, expression of tdTomato was not detected by either microscopy 347 or antibody staining. Two possibilities were suspected: 1) The culure condition is 348 not good enough to support the translation of the delivered mRNA; 2) There is an 349 unknown mechanism that prevents the translation of the delivered mRNA. A recent 350 discovery in C. elegans indicated that endogenous piRNAs can target on the 351 exogenous transgene sequences and prevent their translation 30 . Similarly, 352 planarian neoblasts contain abundant PIWI and piRNAs. We thus hypothesize that 353 a similar piRNA targeting mechanism may exist in planarian neoblassts, which may 354 prevent the translation of the delivered mRNAs. We tested this hypothesis by 355 synthesizing multiple mRNAs encoding the fluorescent protein mCherry in which 356 conservative nucleotide substitutions were introduced in order to minimize 357 potential pairing of the exogenous mRNA with endogenous piRNAs, as was 358 recently described in C. elegans 30 . The synthetic mCherry mRNAs were tested 359 via electroporation into SirNeoblasts (Fig. 7a). Significantly, we found one mCherry 360 mRNA construct that resulted in robust mCherry+ cultured SirNeoblasts ( Fig. 7b-361 e, twice with high expression, five times with medium/low expression, ten times 362 without expression). Even though we have yet to fully comprehend the 363 mechanisms that may be underpinning piRNA targeting in neoblasts, the 364 successful expression results indicated that the culture and electroporation

Planarian care and irradiation treatment 521
Asexual (Clone CIW4) and sexual (Clone S2F1L3F2) strains of Schmidtea 522 mediterranea were maintained in Montjuïc water at 20ºC as previously described 523 8,20 . Animals were starved for 7-14 days before each experiment. Animals exposed 524 to 6,000 rads of g rays were used as transplant hosts 8 . After transplantation, hosts 525 were maintained in Montjuïc water with 50 µg/ml Gentamicin (GEMINI, 400-100P). 526 spinning disk head and a Prior PLW20 Well Plate loader. Several slides were 610 prepared at once and then loaded and processed automatically using a 611 combination of Nikon Elements Jobs for all robot and microscope control and Fiji 612 for object-finding and segmentation. Slides were imaged at low magnification and 613 objects identified before re-imaging tiled z-stacks using a Plan Apo 10X 0.5NA air 614 objective. Tiled images were stitched, projected, and smedwi-1+ puncta were 615 counted using custom macros and plugins in Fiji. 616 Generation of optimized mCherry sequence 617 mCherry candidate sequences were generated by means of a custom python 618 script. Amino acid sequences were back translated to 21 nucleotide sequences 619 from 7 amino acid words at a time. Each potential nucleotide sequence was 620 screened against a list of known piRNAs to generate the sequence with the fewest 621 piRNA matches. A piRNA match consists of no more than a single G/T mismatch 622 in the 6 nucleotide seed region (positions 2-7 of a piRNA) 30 . Additional G/T 623 mismatches were scored as .5 and other mismatches as 1. Only the first 21 624 basepairs of the piRNAs were aligned. The highest scoring piRNA determined the 625 score for that potential nucleotide sequence. The 21 nucleotide sequence with the 626 lowest score was retained. The script was run with four alternate coding 627 tables. The "all" coding table contained all possible codons for each amino acid. 628 The "smed" coding table contained only those codons known to be most used in 629 S. mediterraea 51 . "lowgc" contained only those codons with the fewest G or C 630 nucleotides. "highgc" contained only those codons with the most G or C 631 nucleotides. The "highgc" sequence is shown in Figure 7. The other three 632 sequences as well as 5 additional sequences generated by shuffling the four 633 generated sequences and one sequence generated by backtranslating the amino 634 acid sequence with sms failed to show fluorescence 52 . 635

Data availability 636
All codes used for plugins in Fiji are available at:https://github.com/jouyun. All 637 original data underlying this manuscript can be accessed from the Stowers Original 638 Data Repository at: http://www.stowers.org/research/publications/libpb-1281. All 639 reagents are available from the corresponding author upon reasonable request. 640

Statistical analyses 641
Microsoft Excel and Prism 6 were used for statistical analysis. Mean ± s.e.m. 642 is shown in all graphs. Unpaired two-tailed Student's t-test was used to determine 643 the significant differences between two conditions. p < 0.05 was considered a 644 significant difference. 645

Acknowledgments 646
We thank I. Wang and P. Reddien for assistance with the transplantation 647 technique. We thank all members of Sánchez Lab, especially J. Jenkin and C. 648 Guerrero for animal maintenance and irradiation assistance, L. C. Cheng and E.   following culturing in indicated media + 5% CO2 for 1, 2, or 3 days. Ten to twelve animals 10 assayed per condition.