Loss of all 3 Extended Synaptotagmins does not affect normal mouse development, viability or fertility.

The extended synaptotagmins, E-Syt1, 2 and 3, are multiple C2 domain membrane proteins that are tethered to the endoplasmic reticulum and interact in a calcium dependent manner with plasma membrane phospholipids to form endoplasmic reticulum - plasma membrane junctions. These junctions have been implicated in the exchange of phospholipids between the 2 organelles. The E-Syts have further been implicated in receptor signaling and endocytosis and can interact directly with fibroblast growth factor and other cell surface receptors. Despite these multiple functions, the search for a requirement in vivo has been elusive. Most recently, we found that the genes for E-Syt2 and 3 could be inactivated without effect on mouse development, viability, fertility or morphology. We have now created insertion and deletion mutations in the last of the mouse E-Syt genes. We show that E-Syt1 is specifically expressed throughout the embryonic skeleton during the early stages of chrondrogenesis in a pattern quite distinct from that of E-Syt2 or 3. Despite this, E-Syt1 is also not required for mouse development and propagation. We further show that even the combined inactivation of all 3 E-Syt genes has no effect on mouse viability or fertility in the laboratory. However, this inactivation induces an enhancement in the expression of the genes encoding Orp5/8, Orai1, STIM1 and TMEM110, endoplasmic reticulum - plasma membrane junction proteins that potentially could compensate for E-Syt loss. Given the multiple functions suggested for the E-Syts and their evolutionary conservation, our unexpected findings suggest that they may only provide a survival advantage under specific conditions that have as yet to be identified.

Introduction

The E-Syts (E-Syt1, 2 and 3) are C2-domain membrane proteins that bind membrane phospholipids in a Ca2+ dependent manner. They were initially implicated in glucose transport and fibroblast growth factor (FGF) signaling. Subsequently, they were shown to enhance the formation of contact sites between the endoplasmic reticulum (ER) and the plasma membrane (PM) in response to Ca2+ signaling, see , and for reviews. The E-Syts are broadly conserved throughout the animal kingdom and are analogous to the tricalbins (Tcb1, 2 and 3), which perform analogous functions in ER-PM tethering and phospholipid transport in yeast. By their ability to form ER-PM junctions, the E-Syts were shown to facilitate the transport of phospholipids to the plasma membrane (PM). In this way they can act to sustain homeostasis of PM phosopholipids and particularly of phosphatidylinositol (PI) and phosphatidylinositol 4,5 biphosphate (PtdIns(4,5)P2 or PIP2). PIP2 is a major substrate in cell signaling and its hydrolysis by phospholipase C (PLC) activates Ca2+ signaling via the release of inositol 1,4,5-trisphosphate (IP3). Since activation of PLCγ and PIP2 hydrolysis is a major pathway of FGF signaling, the ability of the E-Syts to replenish plasma membrane PIP2 would naturally enhance and prolong signaling through this pathway, as was suggested for E-Syt2 during early Xenopus development.

The ability of the E-Syts, and particularly E-Syt2, to induce tight ER-PM junctions would logically also promote the bidirectional transfer of phospholipids between the PM and ER via both the ORP5/8 and Nir2/3 exchange proteins, e.g. see. This clearly has broader implications for the maintenance of PM phosphoinositide homeostasis, as shown recently by the demonstration that E-Syt2 dependent ER-PM junctioning enhances the removal of PI(4)P via ORP5/8 and the Sac1 phosphatase. However, targeted inactivation of all 3 E-Syt genes in HeLa cells had no effect on PtdIns(4,5)P2 dynamics following activation of muscarinic receptor, though it did affect the rate of removal of DAG from the PM. This latter effect is consistent with a reduction in the rate of DAG to PA conversion or more likely PA removal via the Nir2/3 exchange pathway, see also.

Given the apparent importance of the E-Syts in the dynamics of membrane lipid composition and in Ca2+, FGF and potentially other signaling pathways, it is essential that we understand their role in vivo. It was, therefore, surprising to find that inactivation of the mouse E-Syt2 and 3 genes had no detectable phenotypic effect, and that esyt2 mice were both viable and fertile. Here we have extended these in vivo studies to the last of the E-Syts, E-Syt1. We find that the E-Syt1 gene is discretely expressed in the developing embryonic skeleton, but its functions are apparently also not essential for mouse viability. Indeed, we show that esyt1 and even esyt1 mice are viable and fertile and display no overt functional or morphological anomalies.

Results

Insertional mutation of the E-Syt1 gene

Using a recombinant vector from the EUCOMM consortium, we generated mouse embryonic stem (mES) cells carrying a potentially conditional “flox-neo” insertion in the E-Syt1 gene. Lox recombination sites were placed in intron 2 and intron 7, and a ß-Gal marker gene and a neo selection gene flanked by FRT (Flipper) sites were inserted within intron 2 (Fig. 1A). Nine independent ES cell clones heterozygous for this esyt1 allele were isolated and 2 of these (B11 and D7, Fig. 1B) were used to generate mouse lines carrying this “knock-out first” inactivating mutation (Fig. 1C). The mice were then crossed with FLPo and Sox2-Cre recombinase expressing mice to generate esyt1, and esyt1 insertion and esyt1 (esyt1) deletion alleles (Fig. 1D). Finally, the Cre and Flipper transgenes were removed by backcrossing.

Figure 1.

Targeted disruption of the esyt1 gene in mouse. A) Maps of the EUCOMM esyt1 Knockout First vector PRPGS00088_C_F09, the wild type (WT, +) mouse esyt1 gene locus, the initial targeted (flox-neo) gene insertion and the subsequent ßGal insertion and exon 3 to 7 deletion (Δ) products generated by recombination respectively of the inserted FRT and Lox sites. Positions of relevant restriction sites, genotyping PCR primers (thick arrows) and hybridization probes are indicated. B) Southern analysis of the insertion mutant ESC clones used to generate mouse lines. C) Typical example of PCR genotyping of mouse lines carrying the insertion (flox-neo, ßGal) and deletion (Δ) mutant esyt1 alleles in comparison with the WT allele. D) Typical example of PCR genotyping of homozygous esyt1, 2 and 3 wild type and esyt1, 2 and 3 null mouse lines. In (C) and (D), the PCR primers used for esyt1 were those shown in (A), and those for esyt2 and esyt3 were as previously described .

E-Syt1 is strongly expressed in the embryonic skeleton and in adult lung and spleen

Insertion of a β-galactosidase (βGal) gene into the esyt1 locus permitted the expression pattern of this gene to be determined during early mouse development. E-Syt1 expression was first detected at 11.5 dpc within the developing mouse skeleton and this expression became rapidly stronger and more striking, such that by 13.5 dpc it was observed in a very distinct pattern throughout the developing thoracic skeleton, the cervical, thoracic and lumbar vertebrae, and within the limb buds (Fig. 2). The detail of staining within the lateral mesodermal plate at 13.5 dpc suggested that E-Syt1 was initially expressed at the outer edges of the mesenchymal condensations that presage chondrogenesis (see enlargements in Fig. 2). In fact, its pattern of expression at this stage somewhat resembled that of the transcription factor Runx2, a master regulator of osteoblast differentiation implicated in the process of mesenchymal condensation.

Figure 2.

Expression pattern of the esyt1 gene in early mouse embryos carrying the β-galactosidase (βGal) allele (Fig. 1A). Expression was determined at stages 11.5, 12.5 and 13.5 dpc by conversion of X-Gal (blue-green), see Materials and Methods. Enlarged panels below the whole mount views of the 13.5 dpc embryos show detail of the lateral mesodermal plate and anterior limb-bud. The images are of embryos homozygous for the βGal insertion allele which lack a functional esyt1 gene. The pattern staining in these embryos was indistinguishable from that of esyt1 mice but was somewhat stronger consistent with a gene dosage effect.

Our previous data showed that E-Syt1 gene was strongly expressed in the lung and spleen of adult mice. Given that this pattern was quite distinct from that previously published for adult human tissues, we repeated the analysis of E-Syt1 expression, and in parallel E-Syt2 and 3, in wild type adult mouse tissues (Fig. 3A). The results concurred with our previous data for all 3 E-Syts, showing strong E-Syt1 expression only in lung and spleen. To further confirm these data, we repeated the extraction of total RNA from kidney, lung, spleen, stomach, testes and heart tissues of wild type mice and determined E-Syt expression levels by Real-Time qRT-PCR (Fig. 3B). Again here, E-Syt1 expression was found to be strong in lung and spleen, weak in stomach and insignificant in kidney, testes and heart, while the data for E-Syt2 and 3 expression closely followed that obtained in Figure 3A and previously.

Figure 3.

A) RT-PCR analysis of E-Syt1, −2 and −3 mRNA levels in wild type adult mouse tissues. B) Real-Time qRT-PCR analysis to verify E-Syt1, −2 and −3 mRNA levels from a selected subset of wild type adult mouse tissues. C) RT-PCR analysis of E-Syt1, −2 and −3 mRNA levels in E-Syt1-null (esyt1) adult mouse tissues. D) E-Syt1 protein levels in lung and spleen of mice carrying homozygous or heterozygous deletions (−/− and +/−) of E-Syt1 as compared to the tissues from wild type (+/+) mice. See Materials and Methods for primers.

The E-Syt1 gene is not essential in mouse

Given the very striking skeletal pattern of E-Syt1 expression, we were very surprised to find that mice homozygous for the esyt1 “knockout first” insertion mutation developed normally and were viable and fertile, suggesting that the E-Syt1 gene was not essential (Table 1A). Similarly, mice homozygous for the esyt1 deletion allele (esyt1Δ) displayed no evident phenotype or any significant effects on viability, fertility or indeed longevity (of the 75 esyt1 mice used for breeding, 47 were sacrificed at > 5 months, 27 at > 6 months, 14 at > 7 months and 7 > 8 months, none died of natural causes or suffered from any obvious disorder). Further, the esyt1 allele displayed Mendelian inheritance (Table 1A). To verify that the E-Syt1 gene had in fact been inactivated, we analyzed a broad range of tissues from esyt1 and wild type mouse siblings for E-Syt1, 2 and 3 mRNA levels, and both spleen and lung for E-Syt1 protein (Fig. 3C and D). The esyt1 mice expressed no functional E-Syt1 mRNA, and no E-Syt1 protein could be detected in either tissue. Thus, esyt1 represented a true null allele. These data further showed that loss of E-Syt1 mRNA was not compensated by an enhancement in the level of E-Syt2 or 3 mRNAs (Figure A and C). It was therefore concluded that E-Syt1 is not essential for mouse viability, fertility or longevity.

Table 1.

Genotype analysis of the progeny born A) from esyt1+ and esyt1+/− crosses, and B) from esyt1/esyt2+/−/esyt3 crosses. “n” indicates the number of offspring.

A)
	esyt2+/+/esyt3+/+/
Cross	esyt1+	esyt1+/−	esyt1−/−
esyt1target /+ x esyt1target /+	23%(n = 17)	50%(n = 37)	27%(n = 20)
esyt1+/− x esyt1+/−	42%(n = 8)	37%(n = 7)	21%(n = 4)
Total	27%(n = 25)	47%(n = 44)	26%(n = 24)
B)
	esyt1−/−/esyt3−/−/
Cross	esyt2+	esyt2+/−	esyt2−/−
esyt1−/−/esyt2+/−/esyt3−/−x esyt1−/−/esyt2+/−/esyt3−/−	28%(n = 17)	48%(n = 29)	23%(n = 14)

Combined deletion of all 3 E-Syt genes has no effect on viability and fertility

Given the surprising finding that the E-Syt1 gene was not essential, we asked if its function was redundant in the presence of the other 2 E-Syt genes. We had previously generated mice lacking functional E-Syt2 and E-Syt3 genes and showed them also to be viable and phenotypically normal. Thus, we attempted to generate mice in which all 3 E-Syt genes were inactivated. Unexpectedly, we found that mice lacking all 3 genes were viable and the null alleles displayed Mendelian inheritance. For example, crosses of esyt1 mice showed Mendelian inheritance of the esyt2 allele (Table 1B). Further, mice lacking all 3 E-Syt genes (esyt1) displayed no overt phenotype, displayed no premature mortality and were fertile. Indeed, when E-Syt-null (esyt1) mice were crossed they generated typical sized litters and offspring displayed no evident phenotype. Thus, not one of the 3 E-Syt genes is essential for normal mouse survival or fertility and mice lacking all 3 genes are phenotypically normal.

Loss of the E-Syts may be partially compensated by upregulation of other ER-PM tethering proteins

The data from yeast suggest that loss of the E-Syt analogs Tcb1 to 3 is in greater part compensated by the actions of other the ER-PM tethering proteins. Hence we asked if loss of the mouse E-Syts increased expression of other tethering proteins. The Osh homologs Orp5/8, like the E-Syts, are ER membrane proteins that bind PM phospholipids and induce ER-PM tethering. The PM Ca2+ channel protein Orai interacts with the ER membrane protein STIM1 during Store Operated Calcium Entry (SOCE) and also tethers ER to PM, an action stimulated by TMEM110. The genes for Orp5, Orai1 and TMEM110 were all strongly expressed in lung, coinciding with strong expression of E-Syts 1 to 3 (Fig. 3A and B and Fig. 4). Expression of Orp5, Orai1 and TMEM110 was found to be 1.5 to 2 times higher in lung from E-Syt-null mice as compared with wild type mice of the same background. E-Syt 1 and 2 expression in spleen corresponded with the expression of Orp8, Orai1, STIM1 and TMEM110, and in each case expression was also significantly increased in E-Syt-null mice. Thus, enhanced expression of a range of tethering proteins may at least in part have compensated for the loss of the E-Syts, providing a potential explanation for the lack of an E-Syt-null phenotype.

Figure 4.

Analysis of Orp5/8, Orai1, STIM1 and TMEM110 tethering protein expression levels in tissues from wild type and E-Syt-null mice. See Materials and Methods for primers.

Discussion

We previously generated mice lacking the E-Syt2 and 3 genes and found that they were viable, fertile and displayed no morphological abnormalities. This finding was already very surprising given the various studies implicating these proteins in plasma membrane homeostasis and key cell signaling pathways. Here we have shown that inactivation of the E-Syt1 gene, and even the combined inactivation of all 3 E-Syt genes, also has no discernable phenotypic effect, esyt1 first (F1) and even second generation (F2) mice being viable, fertile and displaying normal survival. However, the present study would not have revealed subtle changes in mouse tissue structure, physiology or behavior. This said, despite the greatly increased complexity of the mouse system, our findings are in fact in line with data on the inactivation of the Tcbs of yeast. ER-PM junctions in yeast were only significantly affected when all 6 ER-PM tethering proteins (Ist2, Scs2 and 22, and Tcbs 1, 2 and 3) were inactivated, and even then the growth rate in rich medium remained unaffected. It is, therefore, likely that also in mammals the E-Syts function redundantly with other ER-PM tethering proteins. Indeed we observed a significant enhancement in the expression of the Osh homologs Orp5/8 and the SOCE associated proteins Orai1, STIM1 and TMEM110 known to tether the ER to the PM. As we recently discussed, the paradoxical finding that E-Syt2 knockdown in early Xenopus embryos affects FGF signaling may be related to the size of these embryos, the complex nature of their membranes and the need for extremely rapid lipid transport during cleavage divisions. Thus, it may be necessary to study extreme situations of ER or PM functioning or stress, or other behavioral anomalies, before we can identify an in vivo requirement for the mammalian E-Syts. One possible function for the E-Syts, one that could explain their broad evolutionary conservation, is in the regulation viruses-cell interactions such as has been observed for the Arabidopsis synaptotagmin SYTA protein.

Materials and methods

Generation of the E-Syt1 conditional mutation

The esyt1 gene was mutated using the targeted KO-first, conditional ready, lacZ-tagged mutant vector PRPGS00088_C_F09 vector from the EUCOMM targeting project 69413. The vector was linearized and used to electroporate WW6 ES cells, which were then selected with G418. Resistant clones were amplified and analyzed by 5′ and 3′ Southern blotting to identify correctly recombined clones (Fig. 1A and B). These clones were then used to generate 2 independent mouse lines using the services of the McGill Transgenic Core Facility. The subsequent crossing to induce recombination of FRT and Lox sites used the mouse strains codon-optimized FLP recombinase (FLPo) (#012930) and Sox2-Cre (#004783) from the Jackson Laboratory. The mice were housed and manipulated according to the guidelines of the Canadian Council on Animal Care and experiments were approved by the institutional animal care committee.

Genotyping of mice

Genotypes of animals derived from the various crosses were routinely determined by PCR amplification of genomic DNA, (Fig. 1C and D). Primers used for esyt1 were A (5′-CTCTTTCGATGCCTTTCCAC-3′), B (5′-AGGGTCCCAGAATCATGAAG-3′), C (5′-CCACAACGGGTTCTTCTGTT-3′), D (5′-TCTGGCTGAGCTCGAATTTGT-3′), E (5′-ATCCTGGGCAGAGGTTCAGA-3′) and F (5′-CGGTCGCTACCATTACCAGT-3′)

For esyt2, primers were: A (5′-CCAATCAGCAGTCTTACCAT), B (5′-CGTCTCAAGGGAAGGAAATAA) and C (5′-CGCCATACAGTCCTCTTCAC). For esyt3, primers were A (5′- CTGAAGCCTCCCAGTAGGTG), B (5′-CCATCACCCCTAGTTGTTGC), and D (5′-GAGGCTCCAGGCCTTAGTTT).

Gene expression analysis by RT-PCR and real-time RT-PCR

Total RNA was extracted from mouse tissues using Trizol (Invitrogen) and analyzed by RT-PCR in the linear amplification range (qRT-PCR) as previously described, or by Real-Time qRT-PCR. For Real-Time qRT-PCR, the RNA was reverse transcribed as for RT-PCR, but amplification was carried out in triplicate in 20 μl reactions containing 10 μl QuantiFast SYBR Green and 1μM of each primer. 35 reaction cycles of 10 s at 95°C and 30 s at 58°C were carried out on a Multiplex 3005 Plus (Stratagene/Agilent). The primers used for the E-Syts were: mE-Syt1.FOR (5′-TGGGATCCTGGTATCTCAGC), mE-Syt1.REV (5′-CTGGGAGATCACGTCCATTT), mE-Syt2.FOR (5′- CGAATCACCGTTCCTCTTGT), mE-Syt2.REV (5′- GCTCTGGAAGATTTGGTTGC), mE-Syt3.FOR (5′- CAAGCCCTTCATAGGAGCTG), mE-Syt3.REV (5′- AGCAAATGGACTCGGATCAC), mGAPDH.FOR (5′- AACTTTGGCATTGTGGAAGG), mGAPDH.REV (5′ ACACATTGGGGGTAGGAACA). Amplicons were of the expected sizes of 296 bp for E-Syt1, 192 bp for E-Syt2, 246 bp for E-Syt3 and 223 bp for GAPDH. Products were sub-cloned and sequenced to confirm their specificity. Primers for other tethering proteins were taken from Primer Bank (https://pga.mgh.harvard.edu/primerbank/) ; Orai1.FOR (5′-GATCGGCCAGAGTTACTCCG), Orai1.REV (5′-TGGGTAGTCATGGTCTGTGTC), Orp8.FOR (5′-ATGGAGGCAGCCTTAGCAGA), Orp8.REV (5′-CAAATGCTGAGGTTCGTCACT) STIM1.FOR (5′-TGAAGAGTCTACCGAAGCAGA), STIM1.REV (5′-AGGTGCTATGTTTCACTGTTGG), TMEM110.FOR (5′-GCGCTCATGCACAGTTTCG), TMEM110.REV (5′-ACAGTGAACAAGGGTCCTCTT), Orp5.FOR (5′-TTCTGGGCTGCGAAAATGAG), Orp5.REV (5′-GTCAGATCCATTGCATAGCCTG), GAPDH.FOR (5′-AGGTCGGTGTGAACGGATTTG), GAPDH.REV (5′-TGTAGACCATGTAGTTGAGGTCA)

X-gal Staining

Mouse embryos were isolated at E11.5 to E13.5 and fixed for 30 minutes in 1% Formaldehyde, 0.2% Gluteraldehyde, 0.02% NP-40 in 1 x PBS, washed 3 times 20 min. each in Wash Solution (2 mM MgCl2, 0.02% NP40, 1 x PBS). Embryos were protected from light and incubated overnight at R/T in the Staining buffer solution (5 mM potassium ferricyanide, 5 mM potassium ferrocyanide and 1 mg/ml X-gal in Wash Solution). Embryos were rinsed 3 times, 20 min. each, in 1 x PBS. Clarification was performed with “Scale” solution as described previously.