Optimization of cryopreservation and in vitro fertilization techniques for the African turquoise killifish Nothobranchius furzeri

Over the last decade, the African turquoise killifish, Nothobranchius furzeri, has emerged as an important model system for the study of vertebrate biology and ageing. However, rearing this fish in captivity can pose challenges, due to the short window of fertility, inbreeding problems, and the continuous maintenance of different strains and transgenic lines. To date, the main means of long term strain maintenance is to arrest embryos in diapause, a poorly understood and unreliable method. To solve these problems, we developed a robust protocol to cryopreserve sperm and to revive them for in vitro fertilization (IVF), as a better option for long term storage of N. furzeri lines. We tested a variety of extender and activator buffers for sperm in vitro fertilization, as well as cryoprotectants to achieve maximal long term storage and fertilization conditions tailored to this species. Our optimized protocol was able to preserve sperm in a cryogenic condition for months and to revive an average of 40% upon thawing. Thawed sperm were able to fertilize nearly the same number of eggs as natural fertilization, with an average of ~25% and peaks of ~55% fertilization. This technical advance will greatly facilitate the use of N. furzeri as a model organism.


43
Over the last few years the African killifish, Nothobranchius furzeri, has emerged 44 as important model system for the study of vertebrate ageing. The life cycle of 45 these fish is characterized by a fast growth rate, reaching sexual maturity by 4-5 46 weeks, and a maximum lifespan of 6,5-7 months (as mean lifespan of the most 47 long-lived 10% of a given cohort) [1], [2], making them among the shortest-lived 48 vertebrate species bred in captivity and a unique platform for the rapid 49 exploration of aging and age-associated diseases [3]. 50 Unfortunately fast growth and aging features also carry some drawbacks that 51 make the maintenance in captivity of this species quite challenging. Rapid  [9] and CRISPR-mediated mutagenesis [10] in this fish means that 60 more genetically engineered lines require continuous maintenance and space 61 usage. Breeding to preserve a line takes considerable effort and is fraught with 62 risk of accident or infection that can result in strain loss. Furthermore, N. furzeri 63 husbandry requires a large amount of space, since this species is optimally 64 grown in captivity when single housed because of fish-to-fish harassment and 65 food competition [8]. 66 Given these constraints, it is essential to develop protocols to maintain stocks 67 without constant breeding. Currently, the only known way to "freeze" a 68 generation is through diapause, a state of arrested development [2]. However, 69 diapause itself is quite variable and a topic of intense investigation [11], [12]. It is 70 difficult, to date, to induce and release diapause in a controlled and synchronized 71 manner from a large pool of embryos. As well, it is challenging to retain 72 permanence in this stage. Even eggs in diapause need periodic maintenance, 73 their medium or substrate must be checked, cleaned and changed. As the 74 number of strains, species or lines to preserve increases, this method becomes 75 quickly untenable at larger scales. 76 To solve these problems, researchers of other fish species rely on in vitro 77 fertilization and sperm cryopreservation techniques. Through cryopreservation, it 78 is possible to create sperm banks that can store fish genetic pools with minimum 79 maintenance effort for years. Upon revival, the thawed sperm can usually fertilize 80 eggs with a success rate from10-80% [13]. Specific protocols have been 81 established to preserve and activate the sperm of both production fish species 82 (salmonids, sturgeons, carps and catfishes) and research species (Zebrafish, 5 83 Medaka) [14]. Despite this, there is no protocol available to date for in vitro 84 fertilization or sperm cryopreservation of any killifish species.

85
Here we optimized a protocol for Killifish sperm cryopreservation and in vitro 86 fertilization. Our protocol will greatly facilitate the husbandry and the usefulness 87 of N. furzeri as a model organism for research.

90
The optimum conditions for a suitable fertilization environment can vary greatly 91 for each fish species, as well as the protocols, buffers and setups for 92 spermatozoa cryopreservation. For most freshwater fish, sperm motility can be 93 initiated by hypotonic osmolalities [15] and/or by alteration of ion concentration 94 such as potassium or calcium [16], [17]. Once activated, the sperm usually have 95 a short period of motility (30 s to 5 min), depending on species [15]. 96 Protocols from other fish species avail of an "extender" and an "activator" 97 solution. The extender is usually a saline-buffered solution that is mixed with the 98 extracted sperm that keeps it in a stable and inactive condition. This is made 99 possible mostly by the high molality of the extender solution, which is 100 comparable or higher than the ion concentrations inside the gonad [ Iwamatsu solution at different dilutions and molalities (Fig 1).

126
Immediately afterwards, the gonads were removed from the medium, 10 µl of the 127 mixture was put in a hemocytometer chamber and sperm movements were  Tank water, the natural medium where fish biologically breed, was able to 132 activate ~40% of the sperm. Deionized water was able to trigger a very initial  Thus, our second goal was to establish conditions for long-term sperm 178 cryopreservation. Our target was to find an extender solution, which combined 179 with the proper cryoprotectant, could preserve the collected sperm in the inactive 180 state in a frozen condition. Since BSMIS and FBS worked best to activate sperm 181 when diluted (Fig 1B), we focused our studies on these.

182
In the first experiment we mixed the selected extender with an arbitrary    (Fig 4B), dropping slightly on higher concentrations, probably due to 215 increasing toxicity. Any concentration of methanol was able to moderately protect 216 sperm from cryodamage even though the revival rates were inferior compared to 217 DMSO by 5% to 20% less. Other combinations were not able to protect sperm 218 cells efficiently. Less than 10% of sperm were able to revive after freezing in any 219 concentration of DMA or DMF (Fig 4B). Methanol as cryoprotectants to perform further optimization of cryopreservation.

224
Apart from the application of cryoprotectants, the freezing procedure is also a 225 crucial step for cryopreservation. A slow freezing rate can produce large ice 11 226 crystals and damage cellular ultrastructure, whereas a rapid freezing rate induce 227 only small intracellular ice crystals that are less likely to prompt damage [26]. 228 We optimized sample freezing rate with various freezing setups. We placed the DMSO had higher activation than that cryoprotected with 10% methanol (Fig 4C), 241 similar to above. Therefore, we selected DMSO as the final cryoprotectant for 242 our cryopreservation protocol.

243
To assess the proper thawing rate for the frozen sperm vials, we thawed the 244 vials at 1) 30 o C in a water bath, 2) room temperature and 3) 4 o C in the fridge to 245 achieve rapid, medium, or slow thawing rates, respectively. We then revived 246 sperm using BSMIS 1:4 and observed the activation rate. Our results showed 247 that rapid thawing achieves the highest survival and activation of sperm (Fig 4D).

Egg fertilization and survival
12 249 Finally, we tested if the active sperm obtained with our protocol was also able to 250 fertilize eggs obtained from N. furzeri females. To perform this experiment in the 251 most comparable way, we set up several aquarium tanks with 5 female and 2 252 male specimens per tank and allowed them to naturally breed for 2 days. We 253 then collected the eggs generated from the natural breeding and monitored their 254 survival rate until mid somitogenesis. After the natural breeding, the males were 255 separated from the females, kept alone for 2 days, then sacrificed for gonad 256 extraction. The females were anesthetized, dried carefully (Fig 5A), and their 257 unfertilized eggs were gently pushed out from their belly (Fig 5B).  Fig 3). 266 After a variable time period (from 1 day to 2 months), the frozen sperm were 267 thawed, activated and used to fertilize another pool of eggs, following the same 268 procedure (Fig 5). 269 All the eggs fertilized in both ways were monitored until the stage of mid 270 somitogenesis ( Fig 6C) or later ( Fig 6D) and some up to the point of hatching 271 (Fig 6E). Survival rates were scored (Fig 6F).

279
For the detailed fertilization protocol, females were anesthetized and carefully 280 dried from any residual water to prevent spontaneous egg cortical reaction or 281 activation (Fig 5A). Laying the fish on an open hand, a gentle pressure was 282 applied with a finger on the female belly, pushing gently from the middle toward 283 the anus (Fig 5B). 5 to 35 eggs were usually expelled. Those eggs were 284 collected using forceps in an Eppendorf tube containing the extender-sperm 285 solution (Fig 5D). In the case of frozen sperm, the tube was thawed immediately 286 before in a 30 o C water bath (Fig 5C). These actions were performed by two 287 people, with one person thawing the sperm as soon as the other one began to 288 expel eggs from the female. Eggs were placed at the edge of the tube and gently 289 pushed to the bottom (Fig 5D). The best yields of fertilization were achieved 290 using aliquots of sperm-extender of 60 ul and no more than 35 eggs per aliquot.

291
Once the eggs were completely immersed, the tube was gently flicked for 10-20 292 seconds (Fig 5E), allowing the mix to homogeneously distribute around all the 293 eggs. The activator solution was then pipetted into the tube letting drops slide 294 over the tube's border ( Fig 5F) and then mixed with the extender by gently 295 flicking the tube for 20-30 seconds (Fig 5G). The activated sperm was left with 296 the eggs for 10 minutes and the tubes standing open on a bench at room 297 temperature (Fig 5H). At this step, 10ul of the mixture were pipetted under the 298 microscope to evaluate sperm motility. To avoid damage due to DMSO toxicity, 299 the embryos were transferred after 10 minutes to a petri dish using a pipette and 14 300 methylene blue buffered tank water (Fig 5I). The water was replaced twice and 301 the petri dish incubated at 28 o C (Fig 5J). 302 We noticed that the cortical reaction, the earliest process of development where 303 the distance between the yolk and chorion membrane increases, was not 304 correlated with fertilization success in N. furzeri. Once in contact with an 305 aqueous medium, a large number of eggs were able to spontaneously undergo 306 the cortical reaction in the absence of sperm, while a smaller percentage 307 remained blocked in the pre-cortical reaction stage (Fig 6A).  (Fig 6C) over the total initial number of eggs used for fertilization 318 (Fig 6F). Moreover, several embryos were allowed to develop and were followed 319 post-somitogenesis (Fig 6D), post-hatching and up to adulthood (Fig 6E). The 320 growth rates were normal and no defects were detected. Those fish were fertile 321 and able to produce viable embryos.

322
In conclusion, we found that under our conditions, rates of fertilization with frozen 323 sperm ranged from 15 to 25% and were only slightly below fresh IVF or natural 15 324 fertilization. FBS and BSMIS were better at in vitro fertilization than BSMIS alone 325 (Fig 6F). A detailed protocol for the entire procedure is found in the 326 Supplemental methods.

328
In this study we optimized sperm activation, cryopreservation and in vitro 329 fertilization in the species N. furzeri, aiming to establish protocols to obviate proper growth, health, and age of the fish were among the most important 344 features influencing sperm and egg quality. Sperm or egg pools derived from fish 345 too young or too old, or fish that did not grow properly or those that presented 16 346 the early stages of a disease, led to very low performance in sperm activation or 347 fertilization. 348 We achieved the best results when using fish 9-11 weeks of age. Even though 349 the natural breeding in this species ensues prior to this age, the gonads largely 350 grow in size between week 6 to week 10, allowing a greater amount of 351 collectable sperm. Also females produce significantly more eggs at week 10 352 compared to week 6, probably because they are bigger and can store more in 353 their belly.

354
It is important to emphasize that it was not possible to extract sperm by  In these studies, fertilization were performed with an average of 60.000 sperm 365 per ul. We did not systematically modulate sperm concentration as a variable to 366 maximize the fertilization efficiency. Nevertheless, during our tests, several 367 sperm concentrations arose often due to different gonad dimension or to different 368 volumes of the extender, yet we never observed remarkable differences in sperm 17 369 activation. We suggest therefore that within a range between 15,000 -200,000 370 sperm/ul, fertilization occurs at comparable rates.

371
Another technical point is the use of eppendorf tubes versus cryovials. We found 372 that the conical 1.5ml eppendorf tubes ensured rapid sperm thaws and a better       In vitro fertilization by thawed sperms 477 Females were quickly anesthetized in a Tricaine methanesulfonate solution 478 (0.5mg/ml) and carefully dried using towel paper to prevent any residual water on 479 the fish surface. Eggs were extracted from the female by gently massaging and 480 slightly pushing their belly. Eggs were laid over a glove and collected through a 481 forceps over the side of a freshly thawed eppendorf with sperm-extender (60l).

482
A pipette tip was used to push the eggs into the sperm mixture and the the microscope to check the sperm activity. The eggs were finally transferred to 488 a petri dish filled with tank water and the water was replaced twice to wash away 489 any residual cryoprotectant. The petri dish were incubated at 28 o C and the eggs 490 were monitored under a microscope for any morphological changes for 4 days. in case of mean of the means. Graphs were produced from these data using 497 excel. 499 Tracking images were acquired from videos using Imaris snapshot function and 500 brightfield images were acquired using a Leica M80 microscope equipped with a 501 Leica MC170 HD camera. Images were enhanced in brightness, contrast and 502 saturation using GIMP to improve the visual quality.

503
Graphics and drawings were realized using paint, GIMP and power point.