Karyopherin α2 is a maternal effect gene required for early embryonic development and female fertility in mice

The nuclear transport of proteins plays an important role in mediating the transition from egg to embryo and distinct karyopherins have been implicated in this process. Here, we studied the impact of KPNA2 deficiency on preimplantation embryo development in mice. Loss of KPNA2 results in complete arrest at the 2cell stage and embryos exhibit the inability to activate their embryonic genome as well as a severely disturbed nuclear translocation of Nucleoplasmin 2. Our findings define KPNA2 as a new maternal effect gene.


INTRODUCTION 28
Karyopherins are soluble transport factors mediating the nuclear import of proteins. The 29 classical nucleocytoplasmic transport complex involves importin/karyopherin  (KPNB) and 30 one of the six (in mice; in humans and rats one of the seven) known importin/karyopherin  31 (KPNA) paralogues. The utilization of different KPNA paralogues for nuclear import processes 32 2 suggests that substrate specificities of KPNAs for different cargoes are existing; on the other 33 hand, dependent on the cell type or tissue, different expression profiles of KPNAs have been 34 found (Kamei et al., 1999;Kohler et al., 1997;Kohler et al., 1999). The differential expression 35 of KPNA paralogues combined with their specificity for distinct substrates can lead to severe 36 effects if one paralogue is depleted (Choo et  The expression profiles of all known KPNA paralogues have been studied earlier on mRNA 39 level in murine oocytes and the results have shown that only a subset of KPNAs is expressed 40 during oocyte maturation (Mihalas et al., 2015). However, a presence of a distinct KPNA 41 paralogue in the developing oocyte does not automatically stand for its essentiality, as KPNA3 42 is highly expressed in murine oocytes and its depletion has no effect on oocyte maturation or 43 embryonic development (Rother et al., 2011). However, studies have shown, that the 44 presence or absence of specific proteins in the oocyte has the potential to affect not only the 45 regular development of this highly specialized cell type, but can also impact processes linked 46 to fertilization, zygote formation or start of transcription from the embryonic genome 47 (Mitchell, 2022). Oocyte proteins that have an impact on these processes are called maternal 48 effect proteins, because their expression from the maternal genome is essential for the 49 further development of the embryo independent of the genotype of the embryo. To date, 50 around 70 genes are known to encode for maternal effect proteins and KPNA6/importin 7 51 has been shown to be one of them (Mitchell, 2022;Rother et al., 2011). 52 53

RESULTS 54
Depletion of KPNA2 impacts the ability to give birth to healthy offspring. 55 We analyzed oocytes and zygotes of C57Bl/6 mice to compare the presence of different 56 karyopherin  paralogues. Western blot analysis revealed a strong presence of KPNA2, KPNA3 57 and KPNA6, while KPNA1 und KPNA4 were not detectable (Fig. 1A). While the absence of 58 KPNA6 in oocytes had been shown to present with a strong phenotype resulting in early 59 embryonic developmental stop, depletion of KPNA3 had been found to be dispensable for 60 embryonic development (Rother et al., 2011) To elucidate the role of KPNA2, which is also 61 strongly expressed in murine oocytes and zygotes, for fertility and embryonic development 62 To analyze the expression and localization of KPNA2 in WT preimplantation mouse embryos, 152 we performed immunohistochemistry for KPNA2 in embryos at different developmental 153 stages (Fig. 5). In GV oocytes, KPNA2 was moderately expressed and localized in the GV as well 154 as in the ooplasm. In MII oocytes, KPNA2 was redistributed in the ooplasm with no specific 155 localization to the MII spindle, while in zygotes, the protein strongly localized to both 156 pronuclei. After first cleavage into 2cell embryos, KPNA2 expression was found in cytoplasm 157 and nuclei with a particular localization at the nuclear membrane. This strong localization to 158 the nuclear membrane was also seen in later developmental stages until morula. In 159 blastocysts, KPNA2 displayed a dotted signal and distributed in the cytoplasm and nuclei with 160 6 no specific enrichment. The strong pronuclear localization of KPNA2 in zygotes led to the 161 conclusion for a very specific function of KPNA2 at this developmental stage. 162 163

Absence of KPNA2 impairs the zygotic genome activation (ZGA). 164
To test, whether the activation of the embryonic genome is affected in embryos derived from 165 KPNA2 KO females, we performed RT-PCR for several genes that are known to be markers of 166 ZGA (Hamatani et al., 2004;Zeng and Schultz, 2005). 2cell embryos derived from KPNA2 KO 167 females displayed a massively reduced expression of the eukaryotic translation initiation 168 factor 1A (EIF-1a) and no expression of KPNA1 and nucleoside diphosphate kinase 2 (Nme2, 169 Previous studies have shown that Oct4 possesses a nuclear localization sequence (NLS) which 179 is essential for its nuclear import and that the transcription factor binds to KPNA2 (Li et al., 180 2008;Pan et al., 2004). We thus hypothesized that Oct4 localization could be affected in 181 embryos from KPNA2 KO females. However, immunohistochemistry with an antibody against 182 Oct4, detecting both isoforms revealed a regular localization of the transcription factor in the 183 GV, pronuclei and nuclei of WT and KO oocytes and embryos (Fig. 7A). Thus, the absence of 184 maternal KPNA2 did not lead to a defective import of Oct4 in these embryos. To further 185 investigate the reason for an undisturbed Oct4 localization in KPNA2 KO embryos, we 186 performed a binding assay to test, whether other KPNA2 paralogues were able to bind to Oct4. 187 Interestingly, all tested KPNA paralogues and importin  bound to Oct4 (Fig. 7B) we analyzed its nuclear localization in zygotes and 2cell embryos. Interestingly, we noticed a 199 high variability in nuclear RSL1D1 signals in WT and KO zygotes with some zygotes showing 200 weaker nuclear enrichment than others ( KPNA2 KO embryos, we tested the localization of multiple known factors that have been 208 described to act as maternal effect proteins. The transcription factor Sall4 which is known to 209 interact with Oct4, the transcriptional activator Brg1/Smarca4, which modifies the chromatin 210 structure, the germ cell specific protein Stella/Dppa3 known to be involved in the epigenetic 211 chromatin reprogramming in the zygote following fertilization, as well as HDAC1, whose 212 proper function prevents a 2cell developmental stop in mouse embryos were analyzed and 213 none of these factors showed an altered spatial distribution in early embryos which would 214 explain the developmental arrest (Fig. S3). Moreover, the methylation of histone H3 at lysin9 215 which has been shown to be important for coordinated ZGA did not reveal any differences 216 between WT and KO zygotes (Fig. S3). 217 In contrast, we found nucleoplasmin 2 (NPM2) to be differentially located in KPNA2 KO 218 embryos. NPM2 is a histone chaperone involved in chromatin reprogramming, especially 219 during fertilization and early development and it is highly abundant in oocytes and zygotes 220 where it has been shown to act as a maternal effect protein in mice, pigs, cattle and zebrafish. 221 Our studies revealed that nuclear localization of NPM2 was severely disrupted in fertilized 222 eggs and 2cell embryos of KPNA2 KO females, while no effect was found in immature GV 223 oocytes ( Fig. 8B,C).

SDS PAGE and western blot 254
For western blot of mouse embryos, protein isolates of 80 oocytes or zygotes were loaded on 255 a 10% SDS gel. For analysis of mouse tissues, 30 g of tissue protein extracts were loaded on 256 9 a 10% SDS gel. After transfer of proteins, the PVDF membrane was blocked by Odyssey 257 blocking solution (LiCor, Bad Homburg, Germany) and subsequently incubated with primary 258 antibodies at 4°C overnight. On the next day, the membrane was incubated with an IRDye-259 coupled secondary antibody for 1 h at room temperature and detection was performed using 260 the Odyssey Infrared Scanner (LiCor, Bad Homburg, Germany). The generation of C-terminal 261 and N-terminal antibodies against importin α7 was accomplished using standard protocols 262 and has been described previously (Kohler et al., 1999). The generation of the KPNA6 antibody 263 and the KPNA3 antibody is described elsewhere (Rother et al., 2011) The complete list of 264 antibodies and conditions can be found in supplementary data (Table S3). 265

RNA isolation and RT-PCR 267
Total RNA was isolated from mouse tissues or mouse preimplantation embryos using the 268 TRIZOL method. Briefly, the material was homogenized in 1 ml Trizol reagent, 0.   (Wang et al., 2023). In their study, Wang and colleagues observed a 357 developmental 2cell stage arrest of embryos derived from KPNA2 KO females and suggested 358 that the entry of the protein RSL1D1 into the pronuclei was impaired and the reason for the 359 developmental arrest. Interestingly, in our investigation we noticed that the pronuclear 360 localization of RSL1D1 was highly variable in KPNA2 KO zygotes, suggesting that even if RSL1D1 361 nuclear import was not completely abolished at this stage, a retarded nuclear transport could 362 be an underlying mechanism. However, extending our analyses to 2cell stage embryos, the 363 stage at which the developmental arrest finally occurs, did not reveal any differences in 364 nuclear localization of RSL1D1 between WT und KO embryos anymore. Thus, although we do 365 not rule out that RSL1D1 contributes to the embryonic arrest, we conclude, that further 366 factors could be involved in the KPNA2-dependent phenotype. oocytes. In zygotes, maternal KPNA2 is translated from stored maternal mRNAs, defining 388 KPNA2 as a maternal effect protein. Furthermore, a de novo transcription of KPNA2 from the 389 embryonic genome is found in 2cell embryos and this transcription is reduced in 2cell embryos 390 derived from KPNA2 KO females. Transcription of KPNA2 at the 2cell embryonic stage has 391 been found in an earlier microarray analysis (Hamatani et al., 2004). Thus, we conclude, that 392 KPNA2 is a maternally and embryonically expressed protein with maternal KPNA2 directly 393 influencing the transcription of embryonic KPNA2. Our study shows that maternal KPNA2 is 394 essential for embryonic development, while embryonic KPNA2 is dispensable, as KPNA2 KO 395 embryos from heterozygous breedings survive. 396 In GV oocytes and zygotes maternal KPNA2 strongly localizes to the nuclei. Beginning at the 397 2cell embryo stage, we found varying localization of KPNA2 within the cell cycle, with phases 398 of localization at the nuclear membrane and temporally strong nuclear accumulation. 399 Currently it is unclear, if the localization of KPNA2 at the nuclear membrane, which can be 400 found at each developmental stage until blastocyst, represents a short period of concerted 401 entry of KPNA2 into the nucleus or if KPNA2 exerts a distinct function at the nuclear 402 membrane. Interestingly, in addition to its function as a nuclear import mediator, KPNA2 has 403 been shown to be involved in mitotic spindle formation (Guo et al., 2019). The abnormalities 404 of the MII spindle in oocytes could indicate a role for KPNA2 also during meiotic spindle 405 formation. Recent studies suggested that KPNA2 is a critical regulator of the spindle assembly 406 factor TPX2 during mitosis (Guo et al., 2021). In this model KPNA2 sequesters TPX2 in the 407 cytosol and timed phosphorylation of KPNA2 releases TPX2 to facilitate mitotic spindle 408 formation. TPX2 has also been found to regulate the formation of the meiotic spindle and its 409 premature presence interferes with proper spindle assembly (Brunet et al., 2008). Thus, the 410 observed malformation of meiotic spindles in KPNA2 KO oocytes could be explained with the 411 absence of regular KPNA2 levels affecting the temporally critical availability of TPX2. 412 We have shown that KPNA2 is a critical regulator of ZGA, an event that marks the de novo 413 transcription from the embryonic genome. In mice, a first transcriptional wave can be 414 observed beginning in late zygotes, while the majority of ZGA genes is transcribed in mid-to-415 late 2cell embryos (Aoki, 2022). Interestingly, the first wave of ZGA meanwhile has been 416 14 regarded as a global low-level expression of thousands of genes throughout the embryonic 417 genome, while the second wave represents a selective expression pattern. While dozens of 418 genes have been shown to affect ZGA, the exact trigger of transcription initiation is not known 419 yet, making it difficult to link KPNA2 and ZGA on a molecular level (Aoki, 2022). Besides KPNA2, 420 KPNA6 has also been shown to be critically for ZGA (Rother et al., 2011). We therefore 421 hypothesize that specific factors relying on nuclear transport could be the trigger of embryonic 422 transcription initiation. KPNA2 and KPNA6 belong to different subfamilies, sharing only 48% 423 identity, yet, both of them are essential for preimplantation development. On the other hand, 424 KPNA2 shares 46% identity with KPNA7, the former being the closest relative of the latter, but 425 KPNA7 is dispensable for development of preimplantation embryos (Wang et al., 2023 In search for KPNA2-dependent targets we analyzed a number of maternal effect factors, 433 proteins that are maternally derived and indispensable for the transition of the oocyte into an 434 embryo. The transcription factor Oct4 is such a protein and has been found to be a master 435 activator of ZGA in zebrafish (Lee et al., 2013;Leichsenring et al., 2013). In mice, Oct4 is a key 436 regulator of the first wave of ZGA and knockdown in mouse oocytes leads to developmental 437 stop at 4cell stage (Foygel et al., 2008;Tan et al., 2013). Considering the fact that Oct4 has to 438 enter the nucleus, that its nuclear localization sequence is essential for the nuclear import and 439 that it has been found to bind to KPNA2 in Y2H screens, pulldown assays and Co-IP 440 experiments, we hypothesized that impaired Oct4 nuclear import could be the underlying 441 mechanism for the developmental arrest in KPNA2 KO derived embryos (Li et al., 2008;Pan et 442 al., 2004). However, by immunohistochemistry we could not find any differences in Oct4 443 localization between WT and KO. Using KPNA binding assays we furthermore showed that 444 Oct4 binds to KPNA 1,2,3,4, and 6, suggesting, that the remaining KPNA paralogues are 445 sufficient to transport Oct4 into the nucleus in GV oocytes, zygotes and 2cell embryos. Our 446 binding assays are contrary to studies by Yasuhara et al. which have shown that in nuclear 447 import assays Oct4 is transported by KPNA1, 2, and 4 only (Yasuhara et al., 2007), but not by 448