Breeding of Cav2.3 deficient mice reveals Mendelian inheritance in contrast to complex inheritance in Cav3.2 null mutant breeding

High voltage-activated Cav2.3 R-type Ca2+ channels and low voltage-activated Cav3.2 T-type Ca2+ channels were reported to be involved in numerous physiological and pathophysiological processes. Many of these findings are based on studies in Cav2.3 and Cav3.2 deficient mice. Recently, it has been proposed that inbreeding of Cav2.3 and Cav3.2 deficient mice exhibits significant deviation from Mendelian inheritance and might be an indication for potential prenatal lethality in these lines. In our study, we analyzed 926 offspring from Cav3.2 breedings and 1142 offspring from Cav2.3 breedings. Our results demonstrate that breeding of Cav2.3 deficient mice shows typical Mendelian inheritance and that there is no indication of prenatal lethality. In contrast, Cav3.2 breeding exhibits a complex inheritance pattern. It might be speculated that the differences in inheritance, particularly for Cav2.3 breeding, are related to other factors, such as genetic specificities of the mutant lines, compensatory mechanisms and altered sperm activity.


Results
Ca v 3.2 breeding results and characteristics of inheritance. Ca v 3.2 mutant mice were bred for eight years in different projects of our group (see [62][63][64]67 ). Ca v 3.2 +/+ , Ca v 3.2 +/− and Ca v 3.2 −/− mice were generated using three different breeding schemes, i.e., Ca v 3.2 +/− × Ca v 3.2 +/− , Ca v 3.2 +/− × Ca v 3.2 +/+ , and Ca v 3.2 +/− × Ca v 3.2 −/− . In total, 926 offspring from 164 litters were analyzed. For the Ca v 3.2 +/− × Ca v 3.2 +/− breeding scheme (including both sexes) with 344 offspring from 58 litters, a deviation from Mendelian inheritance was detected with an increase of Ca v 3.2 +/− , and a decrease of Ca v 3.2 +/+ and Ca v 3.2 −/− mice compared to the Mendelian distribution (Fig. 1A I ,  Table 1A, Suppl. Tab. 1A). Interestingly, a sex-specific analysis of the related breeding results did not confirm this non-Mendelian inheritance in both combined sexes ( Fig. 1A  For the Ca v 3.2 +/− × Ca v 3.2 −/− breeding (with 309 offspring from 62 litters), a deviation from the Mendelian inheritance pattern was detected for both sexes as well as in the sex-specific analysis. There turned out to be an increase of Ca v 3.2 +/− and a decrease of Ca v 3.2 −/− mice compared to the Mendelian distribution (Fig. 1C I , C II , C III ,  Table 1A, Suppl. Tab. [1][2][3]. For the three different breeding schemes, a significant alteration in litter size was only observed for both sexes, but not for separate analysis of male and female offspring (Table 2A).
Ca v 2.3 breeding results and characteristics of inheritance. Ca v 2.3 mutant mice were bred for about eight years in different projects of our group (see 28,31,38,39,67  Next, we carried out a sex-specific analysis of the offspring breeding results. Notably, neither in females nor in males, we observed any significant deviation from Mendelian inheritance pattern ( Fig. 2A    www.nature.com/scientificreports/    66 . In addition, a deviation from Mendelian inheritance was also detected for Ca v 3.2 +/− × Ca v 3.2 +/− breeding for both sexes, but not for male and female offspring separately. The latter is in contrast to what has been reported by Alpdogan et al. (2020) 66 .
Currently, there is no scientific evidence that any of the aforementioned genetic aspects could be responsible for the exceptions to Mendelian inheritance in Ca v 3.2 null mutant breeding. One aspect that justifies special attention is the functional involvement of Ca v 3.2 VGCCs in sperm and oocyte physiology.
In many species including mice, molecular, pharmacological and electrophysiological studies suggested that VGCCs are involved in spermatogenesis and sperm function, particularly sperm motility and the acrosome reaction [74][75][76][77][78][79][80][81][82][83][84] . The mammalian acrosome reaction is Ca 2+ dependent and requires a complex spatio-temporal activation of different entities of Ca 2+ influx, i.e., via Ca v 3.2 VGCCs, IP 3 receptors, and TRPC2 channels 85,86 . Early reports suggested the presence of both Ca v 3.1 and Ca v 3.2 VGCCs in sperm 87 . However, the dominant T-type Ca 2+ currents in spermatogenic cells turned out to be related to Ca v 3.2, as Ca 2+ current density in spermatogenic cells was not reduced in Ca v 3.1 −/− mice compared to control animals 87 . Furthermore, studies in testes from immature and adult mice revealed a complex spatio-temporal transcription pattern for Ca v 3.2 VGCCs 88 . The Ca v 3.2 function in murine spermatogenesis, sperm motility, capacitation and acrosome reaction was not further evaluated for the potential consequences on breeding upon Ca v 3.2 ablation 74,89,90 . However, inhibition of spermatogenic T-type Ca 2+ channels by genistein was shown to attenuate mouse sperm motility and acrosome reaction 91 .
Importantly, the spatio-temporal fine tuning of Ca 2+ -influx is also critical in maturing oocytes and eggs and proper mammalian development post fertilization 92 . The mouse egg remains arrested at metaphase of the second meiotic division until fertilization triggers sustained Ca 2+ oscillations 92,93 . These oscillations are critical for the activation of embryonic development in mice [93][94][95][96][97][98] . Bernhardt et al. (2015) demonstrated in mouse eggs that Ca v 3.2 VGCCs are a prerequisite for proper accumulation of Ca 2+ during oocyte maturation, for Ca 2+ influx following fertilization, and for proper egg activation 92 . In Ca v 3.2 +/+ eggs, characteristic T-type Ca 2+ currents were detected which are in accordance with previous studies 99 . As expected, T-type Ca 2+ currents were reduced by 44% in Ca v 3.2 +/− eggs (compared to Ca v 3.2 +/+ eggs) and not measurable in Ca v 3.2 −/− eggs. Thus, Ca v 3.2 VGCCs seem to represent the only functional T-type Ca 2+ channel in mouse eggs with severe impact on Ca 2+ homeostasis and dynamics 92 . Importantly, the Ca v 3.2 −/− mouse line was originally reported to be viable and fertile 60 . Recent analysis of fertility revealed that the number of pups per litter was significantly reduced in Ca v 3.2 −/− females compared to Ca v 3.2 +/+ females 92 . These findings are in accordance with the results of our large-scale breeding studies in which a reduced litter size from Ca v 3.2 −/− females was detected. Also, the results of Bernhardt et al. (2015) are in line with our observation of fewer homozygous mutant mice than expected in the Ca v 3.2 +/− × Ca v 3.2 −/− breeding scheme and a relative increase in Ca v 3.2 +/− mice 92 . As Ca v 3.2 null mutant mice are not completely infertile, it was also suggested that additional Ca 2+ entry mechanisms may act as a partial compensatory mechanism to sustain Ca 2+ oscillations 92 .
Current scientific data point to the fact that the favorite explanation for the observed deviation from Mendelian inheritance in Ca v 3.2 null mutant breeding originates from the important roles of Ca v 3.2 VGCCs during oocyte maturation and following fertilization 92 as well as the implications in spermatogenesis, sperm motility and acrosome reaction 74,[77][78][79]89,90 . As genotyping in our study was carried out at the post weaning state, we do not have information about a potential decrease in null alleles at the pre and post-embryonic stage. Litter size analysis for our breeding schemes revealed alterations for offspring of both sexes, but not for separate analysis of male or female offspring (Table 2A). We cannot comment on knockout and wild-type litter sizes, as we did not breed Ca v 3.2 −/− × Ca v 3.2 −/− or Ca v 3.2 +/+ × Ca v 3.2 +/+ . In summary, transmission ratio distortion with biased genotype distribution and reduced litter size often gives rise to either selective embryonic lethality (impaired embryonic development at the pre-or post-implantation state) or reduced oocyte production (dysgametogenesis) 73 . Whether prenatal lethality-as previously suggested by Alpdogan et al. (2020) 66 -accounts for the reduced number of Ca v 3.2 −/− mice and reduced litter size remains to be proven in the future.  66 . What both lines have in common is that they represent constitutive knockout models breed into C57BL/6 J mice 26  . Therefore, there is no molecular, biochemical or electrophysiological evidence that suggests or even proves the formation of functional Ca v 2.3-like channels based on potential two domain fragments in the model we used. Also, there are no indications that such potential fragments could be cytotoxic and influence the inheritance pattern. Notably, we previously checked for compensatory mechanisms in the Ca v 2.  15 . The N-terminus of Ca v 2 Ca 2+ channels is not only involved in G-protein regulation but also responsible for dominant negative (cross-) suppression of Ca v 2 channels in general 103 . It is essential to note that a reduction/elimination of Ca v 2.3 expression shown by Western blotting using antibodies directed against domain I or domain IV does not rule out the potential existence of such an N-terminal protein fragment in this model ("Schneider Ca v 2.3 model") 15 . However, the existence of such fragments and their potential devastating impact on e.g., gametogenesis (spermatogenesis/oogenesis) remains speculative as well. Given the lack of available micro-array data from this model, compensatory mechanisms that might account for the observed deviation from Mendelian inheritance in Alpdogan et al. (2020) 66  Given the important physiological roles of Ca v 2.3 R-type VGCCs, e.g., in the cardiovascular system and germ cell physiology, it is tempting to hypothesize that ablation of this channels might have severe effects on prenatal development and might thus influence the inheritance pattern. In the heart for example, Ca v 2.3 is involved in the impulse generating and conduction system, but also the autonomic cardiac control 104 . Although a number of cardiac electrophysiological alterations have been described in Ca v 2.3 −/− mice using multi-electrode arrays (MEA) and radiotelemetric electrocardiographic (ECG) recordings, there are no indications that these alterations directly impair the lifespan of Ca v 2.3 deficient mice or cause prenatal lethality [18][19][20]105 .
Another aspect that warrants attention is the expression of Ca v 2.3 VGCCs in sperms. Several publications have suggested the expression of Ca v 2.3 in mature sperms, pachytene spermatocytes and other spermatogenic cells 106,107 . In the Ca v 2.3 null mutant model generated by Tanabe's group, ablation of the Ca v 2.3 Ca 2+ channel resulted in reduced Ca 2+ transients in the sperm head region and impaired sperm motility 107,108 . These findings also suggest that Ca v 2.3 VGCCs contribute to the control of flagellar movement, particularly the asymmetry in flagellar beat and randomized swimming patterns 108 . The latter seems to be based on Ca v 2.3 expression on the proximal segment of the principal piece of mouse sperm and is thus important for chemotaxic orientation 108,109 . Importantly, it turned out that the effect of Ca v 2.3 ablation on flagellar movement was medium-dependent, e.g., on the bicarbonate concentration. Furthermore, the motility of sperms is known to depend on the complex intravaginal/intrauterine environment 110 . We are still lacking information how Ca v 2. www.nature.com/scientificreports/ large-scale breeding. For our breeding schemes, analysis of litter sizes did not reveal any significant alterations, neither for offspring of both sexes, nor for male and female offspring separately (Table 2B). It is essential to note that previous phenotyping studies on Ca v 2.3 null mutant mice did not always reveal consistent findings. Whereas impairment of glucose tolerance and insulin release, for example, was described consistently in both the "Tanabe Ca v 2.3 model" 114 and the "Schneider Ca v 2.3 model" 15 , substantial discrepancies were found for thalamocortical oscillations between the "Shin Ca v 2.3 model" 44 and the "Schneider Ca v 2.3 model" 45 . The same held true for sleep architecture and circadian rhythmicity between the "Schneider Ca v 2.3 model" 32 and the "Miller Ca v 2.3 model" 31 . Differences between the models might thus also affect the reproductive system.

1.
Our results from large-scale breeding studies partially confirm a previous report about a deviation from Mendelian inheritance in the Ca v 3.2 null mutant line 66 . Whether this phenomenon is related to prenatal lethality-as suggested by Alpdogan et al. (2020)-cannot be specified here, as no scientific evidence is yet available to prove this hypothesis. It might be speculated that the described role of Ca v 3.2 VGCCs in spermatogenesis, oogenesis, fertilization and embryonic development is responsible for the observed exceptions to Mendelian inheritance.
2. We cannot confirm a deviation from Mendelian inheritance in Ca v 2.3 null mutant breeding. This discrepancy might be due to the specificities in genetic engineering in both models and related physiological consequences. Although Ca v 2.3 VGCCs are involved in sperm physiology as well, there is no direct scientific evidence that a lack of Ca v 2.3 alters classic inheritance. Importantly, we have no indication of prenatal lethality in the Ca v 2.3 null mutant line that we used in our study.
3. Four different Ca v 2.3 null mutant lines have been generated and there are examples of physiological discrepancies between these models, e.g. in the field of sleep architecture and circadian rhythmicity or in inheritance patterns as outlined in this study. Intrinsic phenomena related to the specificities of genetic engineering and compensatory mechanisms upon gene inactivation might account for such phenotypic variation. Though resource-intensive, our results suggest that physiological studies should be carried out and confirmed in more than one null mutant line if possible.  60,62,63 ). PCRs were carried out using a C1000 thermal cycler (BioRad, Germany) with initial denaturation (94 °C for 3 min), followed by 35 cycles (denaturation, 94 °C for 15 s; annealing, 61 °C for 15 s; extension 72 °C for 15 s) and final extension (72 °C for 1 min). Finally, PCR products were separated using agarose gel electrophoresis and visualized by ChemiDoc Touch (BioRad, Germany). Examples of our genotyping of Ca v 3.2 mutant mice are provided in detail in 62,63 . Note that genotyping of all experimental animals was carried out twice per animal (see supplementary tables 1-3) at the post weaning state. Further molecular details on the mutant Ca v 3.2 line are also described by Chen et al. (2003) 60 . The reduction/absence of the Ca v 3.2 expression in Ca v 3.2 +/− and Ca v 3.2 −/− mice was further proven by our group using the Western blot approach 62 All animal experimentation was carried out according to the guidelines of the German Council on Animal Care, and all protocols were approved by the local institutional and national committee on animal care ( KO forward 5′-ATT GCA GTG AGC CAA GAT TGT GCC-3′. PCR was carried out using the C1000 thermal cycler (Bio-Rad) with an initial denaturation (94 °C for 3 min) followed by 35 cycles (each cycle containing the following steps: denaturation 94 °C for 15 s, annealing 59 °C for 15 s, extension 72 °C for 15 s) and final extension (72 °C for 1 min). Subsequently, PCR products were separated via agarose gel electrophoresis and detected by ChemiDoc Touch (Bio-Rad). For details on the procedure and genotyping results see also 28,31,38,39 . Note that genotyping of all experimental mice was carried out twice per animal (see supplementary tables 4-6) at the post weaning state. Further molecular characterization of the model is provided by Wilson et al. (2000) 100 . The reduction/absence of the Ca v 2.3 expression in Ca v 2.3 +/− and Ca v 2.3 −/− mice was further proven by our group using the Western blot approach 38,39 . Statistics. As widely used in genetics, Pearson's chi-square test was used to check for Mendelian inheritance.

Methods
The procedure applied here was described in detail by Montoliu et al. (2012) 115 (see Table 1). Litter size analysis was carried out using One-Way ANOVA. Statistical analysis and graphical representations were conducted using GraphPad Prism (version 6) for Windows (Graphpad Software, Inc., USA). All data were displayed as mean ± standard error of the mean (SEM).

Data availability
All relevant data are provided within this manuscript and the related supplementary information.