SIT1 is a betaine/proline transporter that is activated in mouse eggs after fertilization and functions until the 2-cell stage

Betaine (N,N,N-trimethylglycine) is actively taken up by 1-cell embryos via an unidentified saturable transport mechanism(Anas et al., 2007; Hammer and Baltz, 2002). Previous work has shown that betaine can remarkably improve the viability of preimplantation (PI) mouse embryos, at least under in vitro conditions. In particular, the addition of betaine to the PI embryo culture medium protects embryo development against stress imposed by alterations in culture medium composition or increased osmolarity(Biggers et al., 1993; Dawson and Baltz, 1997). Embryos cultured with betaine also maintain more normal intracellular Na⁺ and K⁺ concentrations(Biggers et al., 1993) and their pattern of protein synthesis more closely resembles that of in vivo embryos (Anbari and Schultz,1993). Betaine might also function in vivo, as it is present in mouse oviducts (Anas et al.,2007).

One-cell embryos possess a single route for betaine transport that is Na⁺- and Cl^--dependent, also transports proline and is inhibited by several methylamino acids and proline derivatives(Anas et al., 2007). This betaine/proline transporter has not, however, been identified at a molecular level and is unlike any transporter previously described in 1-cell mouse embryos (Anas et al., 2007; Hammer and Baltz, 2002). Its transport characteristics do not resemble those of the betaine transporter BGT1 (encoded by Slc6a12) (Hammer and Baltz, 2002; Yamauchi et al., 1992), nor of the proline transporters PAT1 and PAT2(Slc36a1 and Slc36a2, respectively)(Broer, 2008).

One function of betaine in mammalian cells is as an organic osmolyte, one of an array of neutral organic compounds accumulated by specialized transporters to counter increased external osmotic pressure without the deleterious effects of increased ionic strength(Kwon and Handler, 1995). Based on its osmoprotection of PI mouse embryos in culture, Biggers et al. proposed that betaine functions as an organic osmolyte in PI embryos(Biggers et al., 1993). Cleavage-stage embryos are very sensitive to moderately increased osmolarity,and they accumulate large amounts of glycine as an organic osmolyte to balance the osmolarity of their normal in vivo environment(Steeves et al., 2003). Under in vitro conditions, betaine is as effective as glycine at osmoprotection in 1-cell embryos (Anas et al.,2007; Biggers et al.,1993; Dawson and Baltz,1997; Hammer and Baltz,2002) and thus might also have a function in cell volume homeostasis. Another established function of betaine is to serve as a donor of methyl groups that are ultimately made available to a wide array of methyltransferases, a function known to be important mainly in the liver(Finkelstein and Martin, 1984; Selhub, 1999).

Based on the dependence of betaine/proline transport in 1-cell mouse embryos on Na⁺ and Cl^-, we hypothesized that the embryonic betaine/proline transporter is a member of the neurotransmitter transporter family (NTT; the Slc6 gene family), which contains all known mammalian Na⁺- and Cl^--dependent transporters of organic substrates (Chen et al., 2004; Hoglund et al., 2005). Two possible candidates have been proposed(Anas et al., 2007): the brain proline transporter (PROT; encoded by Slc6a7) and the intestinal imino acid transporter (SIT1, also known as IMINO and XTRP3s1; encoded by Slc6a20a). PROT is a reportedly brain-specific proline transporter with an inhibitor profile (Fremeau et al.,1992; Shafqat et al.,1995) similar to that of the embryonic betaine/proline transporter. However, to our knowledge, whether PROT actually transports betaine is not known. SIT1 was recently identified as the long-sought intestinal proline transporter (Kowalczuk et al., 2005; Takanaga et al.,2005). SIT1 has kinetic properties and an inhibitor profile very similar to those of the embryonic transporter and it accepts betaine as a substrate (Kowalczuk et al.,2005; Takanaga et al.,2005). It was unknown, however, whether either of these transporters is expressed or active in 1-cell mouse embryos.

All chemicals were obtained from Sigma (St Louis, MO) unless otherwise noted. Brefeldin A (BFA) was prepared as a 1000× stock in ethanol, and used at 5 μg/ml (18 μM). Culture media were based on KSOM medium(Lawitts and Biggers, 1993),modified by omitting glutamine and replacing BSA with polyvinyl alcohol (PVA)(1 mg/ml; cold-water soluble, 30-70 kDa). KSOM was equilibrated with 5%CO₂ in air at 37°C. Similarly modified HEPES-KSOM (pH 7.4) was used for oocyte and embryo collection and where specified(Lawitts and Biggers,1993).

Oocytes and embryos were obtained from female CF1 mice (4-7 weeks old;Charles River Canada, St-Constant, PQ, Canada) that had been superovulated by intraperitoneal injection of equine chorionic gonadotropin (eCG; 5 IU). For ovulated metaphase (M) II oocytes and PI embryos, human chorionic gonadotropin(hCG; 5 IU) was administered by intraperitoneal injection ∼47 hours post-eCG. For embryos, the female mice were caged overnight with BDF₁ males (Charles River) immediately after hCG for mating. All procedures were approved by the Animal Care Committee of the Ottawa Health Research Institute.

Germinal vesicle (GV) stage oocytes were obtained by mincing excised ovaries ∼44-46 hours post-eCG, collecting fully grown oocytes, and removing adherent cumulus cells by repeated pipetting. MI oocytes were similarly obtained ∼4 hours post-hCG, and MII oocytes at ∼15 hours post-hCG. One-cell stage embryos were collected (±2 hours) at ∼22 hours post-hCG (or, as specified, between 19 and 29 hours), 2-cell at ∼44 hours, 4-cell at ∼56 hours, 8-cell at ∼67 hours, morulae at ∼76 hours and blastocysts at ∼94 hours. MII oocytes and 1- through 8-cell embryos were removed from excised oviducts by flushing with HEPES-KSOM(containing 300 μg/ml hyaluronidase for MII oocytes and 1-cell embryos to facilitate removal of cumulus), while morulae and blastocysts were flushed from the uterotubular junction and uterus, respectively. Oocytes or embryos were washed through four drops of HEPES-KSOM and then briefly maintained in KSOM according to standard techniques in microdrop cultures(Lawitts and Biggers, 1993)under mineral oil (Sigma-Aldrich, Milwaukee, WI) at 37°C in 5%CO₂ in air until they were used.

For parthenogenetic activation with Sr²⁺, ovulated MII oocytes were incubated for 2 hours with 10 mM SrCl₂ in KSOM(CaCl₂ omitted) as described(Phillips et al., 2002). In one experiment, eggs were instead activated using cycloheximide alone (50μg/ml in KSOM from 1000× stock in water), which inhibits general protein synthesis and thus activates eggs by preventing cyclin B synthesis(Moos et al., 1996; Phillips et al., 2002; Siracusa et al., 1978). The time at which Sr²⁺ or cycloheximide was introduced was designated t=0.

[³H]betaine ([methyl-³H]betaine, 85 Ci/mmol, 1 mCi/ml) and [³H]proline (l-[2,3-³H]proline,45 Ci/mmol, 1 mCi/ml) were obtained from Amersham Biosciences (Arlington Heights, IL). [³H]betaine was custom-synthesized by Amersham as previously described (Anas et al.,2007). Stocks were in 2% ethanol in water, stored at-80°C.

For determining betaine transport activity, groups of 5-12 oocytes or embryos were incubated with 1 μM [³H]betaine for 30 minutes,except after parthenogenetic activation (Sr²⁺ or cycloheximide) or for in vivo-maturing oocytes at defined times post-hCG, where incubation with 2 μM [³H]betaine for 10 minutes was used instead to permit more precise timing. Proline transport was measured using 1 μM[³H]proline for 10 minutes in the presence of 10 mM alanine to eliminate the betaine-resistant, alanine-sensitive component of proline transport in PI embryos (Anas et al.,2007). Oocytes or embryos were then immediately processed for measurement of intracellular ³H.

To determine the intracellular content of ³H, oocytes or embryos were removed from [³H]betaine- or [³H]proline-containing medium, washed in ice-cold HEPES-KSOM, and dissolved in 4 ml scintillation fluid (Scintiverse BD; Fisher Scientific, Pittsburgh, PA). ³H was detected using a 2200CA TriCarb liquid scintillation counter (Packard Instruments, Downer's Grove, IL). Background was measured in equivalent volumes of each final wash drop, and the counts obtained were subtracted from the paired embryo sample. The total amount of each compound in oocytes or embryos was calculated using standard curves constructed for each set of experiments. Transport rates were expressed as fmol/oocyte or embryo/minute(normalized to 1 μM substrate when 2 μM [³H]betaine was used).

Groups of 50 1- or 2-cell embryos, isolated from about five females, were washed extensively with HEPES-KSOM, placed into microcentrifuge tubes with minimal medium, air dried and stored at -20°C. Total betaine was measured using a modification of the high performance liquid chromatography(HPLC)-based method described previously(Storer et al., 2006; Storer and Lever, 2006). Briefly, 25 μl methanol was added to each tube, vortexed, ultrasonicated for 10 minutes, and then 225 μl acetonitrile added prior to derivatization with phenanthrenacyl triflate. HPLC separation was performed using a silica column (3 μm) with a mobile phase of 5 mM dimethylbutylamine, 10 mM succinate and 5% water in acetonitrile at a flow rate of 0.8 ml/minute for a 30-minute run time at 40°C. The sample injection volume was 50 μl. Detection was by fluorescence. Background was determined using a similar amount of medium from the final wash drop. Betaine was quantified by comparison to external standards and the data are reported as pmol/embryo.

RNA was extracted from two independent sets of MII oocytes and PI embryos,with kidney and brain as control tissues, using the RNeasy Micro Kit (Qiagen,Mississauga, ON) and reverse transcribed using the Retroscript Kit (Ambion,Austin, TX). Primer pairs were designed (OligoPerfect, Invitrogen, Carlsbad,CA) using mouse mRNA reference sequences spanning introns. The SIT1(Slc6a20a) forward (sense) primer was 5′-AGCCACCAATGGCCTGATGT-3′ (nt 683-702 of NM_139142) and the reverse was 5′-AGCGATCAGGCTGCCAAAAC-3′ (nt 790-809), yielding a 127 bp amplicon. The PROT (Slc6a7) forward primer was 5′-GTGGCAACTGGTGGAACACG-3′ (nt 651-679 of NM_201353) and the reverse was 5′-CACTCCTCGAACCAGCAGCA-3′ (nt 957-976), yielding a 317 bp amplicon. We used, as a positive control, histone 2A family member Z(H2afz), previously shown to have a particularly stable expression pattern in PI embryos across different stages and culture conditions(Mamo et al., 2007). The same primer pair as in the original report (202 bp amplicon) was used. Each PCR reaction contained 0.1 embryo equivalent of cDNA. PCR was carried out with Hotstar Taq polymerase (Qiagen) in 20 μl reaction volumes in a Mastercycler thermocycler (Eppendorf, Hamburg, Germany): 95°C for 15 minutes, followed by 40 cycles of 94°C for 60 seconds, 60°C for 30 seconds, and 72°C for 60 seconds. Amplicons were visualized by agarose gel (2%) electrophoresis with ethidium bromide staining using a Typhoon 8600 Phosphorimager (Amersham,Piscataway, NJ). Amplicons derived from MII eggs and kidney for Slc6a20a and Slc6a7 were sequenced by the Ontario Genomics Innovation Centre, Ottawa, Canada.

Mouse SIT1 (Slc6a20a) cDNA was cloned into a pGEM-He-Juel vector as described previously (Kowalczuk et al.,2005). Mouse PROT (Slc6a7) in a pYX-Asc vector was purchased from Open Biosystems (Huntsville, AL). Slc6a7 and Slc6a20a cRNA were prepared from NotI-linearized cDNA templates (Qiagen T7 mMessage Machine). Stage VI Xenopus laevisoocytes were obtained as previously described(Liu and Liu, 2006). Briefly,sexually mature female Xenopus (NASCO, Fort Atkinson, WI) were stimulated with eCG (50 IU) 3-7 days before oocyte retrieval. Oocytes were released from excised ovaries by incubation in OR2 medium (82.5 mM NaCl, 2.5 mM KCl, 1 mM CaCl₂, 1 mM Na₂HPO₄, 5 mM HEPES,pH 7.8; CaCl₂ omitted) with 1.5 mg/ml collagenase and 1 mg/ml soybean trypsin inhibitor, and then maintained in OR2 containing 0.1 mg/ml gentamicin. Protocols were approved by the Animal Care Committee of the Ottawa Health Research Institute. Oocytes (in CaCl₂-free OR2) were injected with cRNA (625 μg/ml in water) into the cytoplasm, using injection volumes of 20 nl for SIT1 and 10 nl for PROT, or control injections of water. Injected oocytes were incubated in OR2 for 48 hours (PROT) or 72 hours (SIT1)at 16°C before use.

To assess transport, oocytes were washed in ND96 medium (96 mM NaCl, 2 mM KCl, 1 mM MgCl₂, 5 mM HEPES, 1.8 mM CaCl₂, pH 7.8),rapidly transferred through two 0.5 ml aliquots of ND96 containing 1 μM[³H]proline or [³H]betaine, and then incubated for 20 minutes at room temperature in 1 ml ND96 with the radiolabeled substrate. Oocytes were then washed in cold ND96, placed in 4 ml of scintillation fluid and vortexed. The amount of substrate in each oocyte was determined as described above for mouse oocytes. Where indicated, potential competitive inhibitors were present during incubation at 5 mM.

A morpholino antisense oligonucleotide (MO) (Gene Tools LLC, Philomath, OR)of the sequence 5′-CACTGAGGCCGTGCCTTCTCCATGT-3′ (start codon underlined) directed against mouse Slc6a20a (NM_139142.2;confirmed unique in mouse with BLASTN) and a standard control MO (humanβ-globin β-thalassemia mutation) were used. GV mouse oocytes were injected with 5 pl of Slc6a20a or control MO (0.5 mM in water). Where indicated, Xenopus oocytes were injected with 10 nl of MO (1 mM; MO was co-injected with Slc6a20a cRNA, above, where indicated; owing to inclusion of plasmid sequence, the final MO base and cRNA were mismatched). Final MO concentrations in both mouse GV and Xenopus oocytes were∼10 μM. Injected and uninjected mouse GV oocytes from the same cohort were cultured overnight (∼20 hours) in Minimal Essential Medium αwith l-Gln (MEMα; Invitrogen) with 1 mg/ml PVA added, during which time they spontaneously matured to MII eggs, and then were parthenogenetically activated using Sr²⁺ and maintained in culture in KSOM for a further 8.5 hours. The rate of [³H]betaine transport was measured as described above. Injected Xenopus oocytes were incubated for 72 hours and [³H]proline measured as described above.

Data are expressed as mean±s.e.m. Graphs were produced using SigmaPlot 8.02 (SPSS, Chicago, IL). Statistical analyses were performed using InStat (GraphPad, San Diego, CA).

The rates of saturable betaine transport in germinal vesicle (GV) and metaphase (M) I oocytes, mature MII eggs, and PI embryos at the 1-, 2-, 4- and 8-cell, morula and blastocyst stages were obtained by comparing the total rate of transport of 1 μM [³H]betaine with the non-specific rate measured in the presence of 5 mM unlabeled betaine(Fig. 1A). Substantial saturable transport of betaine was found only at the 1- and 2-cell stages,with a small but significant component persisting into the 4-cell stage. Blastocysts exhibited a high rate of non-saturable betaine transport that was likely to be the result of accumulation of betaine in the blastocoel cavity,as it was not seen in blastocysts that had been mechanically collapsed (not shown). Subtracting the rate of non-specific transport from the total rate yielded specific, saturable transport at each stage and showed that activity is maximal at the 1-cell stage and decreases at the 2-cell stage(Fig. 1B).

Two routes for proline transport were previously identified at the 1-cell stage: one that is inhibited by alanine and the other being the betaine/proline transporter (Anas et al.,2007). We determined the rate of saturable proline transport in the presence of 5 mM alanine at key developmental stages to reveal proline transport by the betaine/proline transporter. The pattern of proline transport activity (Fig. 1C) was similar to that of betaine (Fig. 1B),with a peak at the 1-cell stage, decline during the 2-cell stage, and no activity in oocytes or 8-cell stage embryos. Thus, the activity of the betaine/proline transporter was essentially restricted to the 1- and 2-cell stages.

We used parthenogenetic activation with Sr²⁺ to obtain precisely timed activation of mouse MII oocytes with which to investigate the appearance of betaine/proline transport after egg activation. Betaine transport initially appeared at ∼4 hours after Sr²⁺-induced activation and reached a maximum by 10 hours (Fig. 2). Parthenogenotes in the early 2-cell stage (24 hours post-activation) exhibited maximal betaine transport, but activity declined somewhat by 30 hours and was lost by the 4-cell stage. MII oocytes maintained under identical culture conditions exhibited only low betaine transport activity, indicating that egg activation was required.

To confirm that betaine transport was initiated with a similar time course in fertilized eggs, and thus that Sr²⁺-activated parthenogenotes provided an adequate model, we measured the rate of betaine transport by fertilized oocytes immediately after removal from the oviduct at the specified times post-hCG. At 19 hours post-hCG (2-4 hours after fertilization, when fertilization can be first confirmed by second polar body emission), the rate of betaine transport was still low (Fig. 3). By 21 hours, an increased rate was observed, reaching a plateau by 23 hours. Thus, betaine transport appears with a similar time course in vivo.

The appearance of betaine transport activity after fertilization or egg activation could reflect de novo synthesis or the activation of pre-existing transporters. To distinguish between these, we parthenogenetically activated MII oocytes using the protein synthesis inhibitor cycloheximide, and then maintained the parthenogenotes in the continuous presence of cycloheximide for 12 hours. Cycloheximide efficacy was confirmed by (1) greater than 80%parthenogenetic activation and (2) ∼90% inhibition of ³⁵S-methionine incorporation into the TCA-insoluble fraction (not shown). Cycloheximide-activated parthenogenotes maintained for 12 hours in the continuous presence of cycloheximide developed betaine transport activity, the rate of which was not significantly different from that of Sr²⁺-activated parthenogenotes(Fig. 4). This indicated that the appearance of betaine transport activity after oocyte activation did not require protein synthesis, and showed that an independent method of parthenogenetic activation also resulted in the appearance of betaine transport.

We used the fungal toxin brefeldin A (BFA), which has been widely employed to inhibit Golgi-based membrane vesicle fusion(Dinter and Berger, 1998), to determine whether the appearance of betaine transport requires an intact Golgi. We previously confirmed that BFA causes the physical disruption of Golgi aggregates in mouse oocytes (Wang et al., 2008), consistent with previous findings(Moreno et al., 2002). Betaine transport was measured in fresh MII oocytes (t=0) and in parthenogenotes (Sr²⁺-activated) cultured in the continuous presence or absence of BFA for up to 30 hours. A sample was removed from BFA at 7 hours and then cultured to 12 hours in its absence(Fig. 5, arrow a), and another was removed from BFA at 10 hours and then cultured to 30 hours(Fig. 5, arrow b). The effect of BFA was reversible, as transport activity developed in parthenogenotes after removal from BFA (Fig. 5,arrows). BFA also did not directly inhibit betaine transport because the rate of transport in 1-cell embryos was not significantly decreased upon 4 hours exposure to BFA (not shown). Since, in many systems, Golgi-derived vesicles are transported by cytoskeleton-dependent mechanisms, we assessed the effect of several cytoskeleton-perturbing agents. However, there was no effect on betaine transport activity at 10 hours post-Sr²⁺ activation of the continuous presence of demecolcine or nocodozole (microtubule depolymerization), cytochalasin D (F-actin depolymerization), or jasplakinolide (F-actin stabilization) (not shown).

To determine whether mouse embryos contain endogenous betaine at the developmental stages at which we have found betaine transport activity, in vivo-derived 1- and 2-cell embryos were isolated from oviducts, washed extensively, and then groups of 50 placed in tubes and dried for betaine measurements. The total endogenous betaine contents of three independent samples each of 1- and 2-cell embryos were determined. Very little betaine(0.05±0.03 pmole per embryo equivalent of medium, mean±s.e.m., n=6) was found in wash drop samples, indicating that there was no significant contamination from extracellular betaine (not significantly different from 0 by one-sample Student's t-test, P=0.20). The embryo samples (after background subtraction), by contrast, contained betaine at 0.98±0.23 pmole/1-cell embryo and 1.30±0.23 pmole/2-cell embryo (both significantly different from 0, P<0.05).

Two candidates had been identified that might mediate betaine/proline transport activity in 1-cell embryos: the brain proline transporter (PROT,encoded by Slc6a7) and the intestinal imino acid transporter (SIT1,encoded by Slc6a20a). To determine whether mRNA for either transporter was present in PI embryos, we carried out RT-PCR on RNA isolated from MII eggs, PI embryos from the 1-cell to blastocyst stages, kidney and brain, with primers designed to amplify Scl6a7, Slc6a20a and H2afz (control). Surprisingly, we found that both Scl6a7 and Slc6a20a mRNAs were present in MII eggs and 1-cell embryos(Fig. 6). Amplicons derived from MII eggs and kidney for Slc6a20a and Slc6a7 were excised from gels and sequenced to confirm their identity. We had previously confirmed the H2afz amplicon by sequencing. Slc6a7 and H2afz, but not Slc6a20a, were detected in brain as expected(not shown). We also used independent sets of primers for each gene(Slc6a20a and Slc6a7) in 1-cell embryos and confirmed their expression (not shown). Thus, both PROT and SIT1 remained as possible candidates for the betaine/proline transporter, as their mRNAs were present before transport activity appears.

In order to determine whether either of the candidate transporters had proline and betaine transport characteristics that resembled those in 1-cell mouse embryos (Anas et al.,2007), we expressed mouse SIT1 and PROT in Xenopusoocytes, and determined the characteristics of betaine and proline transport by each. Expression of SIT1 in Xenopus oocytes resulted in a∼9-fold increase in the uptake of 1 μM [³H]proline over 20 minutes (Fig. 7A). Induced proline transport was reduced to the level in control (water-injected) oocytes by 5 mM unlabeled betaine, and thus was saturable. Proline transport was also eliminated in the presence of 5 mM 2-methylaminoisobutyric acid (MeAIB) but was unaffected by 5 mM histidine. SIT1-expressing Xenopus oocytes similarly showed an ∼8-fold increase in [³H]betaine uptake,which was reduced to the level in control oocytes by 5 mM unlabeled proline,was eliminated in the presence of MeAIB, but unaffected by histidine(Fig. 7B).

Expression of PROT in Xenopus oocytes resulted in a ∼4-fold increase in proline uptake (Fig. 7C). In contrast to that in SIT1-expressing oocytes, proline transport in PROT-expressing oocytes was not inhibited by the presence of either 5 mM unlabeled betaine or 5 mM MeAIB, but was reduced to the level in control oocytes by 5 mM histidine. Uptake of betaine by PROT-expressing Xenopus oocytes was not greater than that of control oocytes(Fig. 7D), indicating that PROT does not transport betaine at detectable levels.

To determine whether knockdown of SIT1 expression affected endogenous betaine transport in mouse parthenogenotes, we used a morpholino (MO) designed to block Slc6a20a translation. De novo protein synthesis is apparently not required for the appearance of betaine transport activity after egg activation (see above, Fig. 4). Therefore, we injected GV mouse oocytes and then in vitro matured and parthenogenetically activated them, hypothesizing that the presence of the Slc6a20a MO for the ∼28 hours required for oocyte maturation and activation might interfere with the maintenance of SIT1 protein. GV oocytes that were in vitro matured and then parthenogenetically activated transported betaine at a rate similar to parthenogenetically activated in vivo matured oocytes, indicating that in vitro maturation itself did not affect the post-activation development of betaine transport. In oocytes injected with the Slc6a20a MO, however, transport was significantly reduced (to approximately the level of non-specific transport;see Fig. 1A), whereas the control MO had no effect (Fig. 8A). Therefore, the appearance of betaine transport after egg activation is substantially inhibited when SIT1 protein translation is specifically blocked in mouse oocytes.

To confirm that the MO we used could inhibit functional expression of SIT1,we injected it into Xenopus oocytes expressing exogenous mouse SIT1(as described above). Induced [³H]proline transport was significantly inhibited by the Slc6a20a MO, but not the control MO(Fig. 8B), confirming that the Slc6a20a MO decreased functional expression of mouse SIT1.

Virtually all transporters, including a number of amino acid and metabolite transporters, that have been found to be active in 1-cell embryos are also similarly active in MII eggs and thus do not change upon fertilization(Harding et al., 1999; Van Winkle et al., 1988; Van Winkle et al., 1990). Indeed, to our knowledge, none of the diverse array of transporters of biologically important organic compounds has been found to be activated (or inactivated) shortly after fertilization in mammalian eggs, although several ion channels and intracellular pH-regulatory mechanisms change in activity during the immediate post-fertilization period(Day et al., 1998; Lane et al., 1999; Phillips and Baltz, 1999), and a cysteine transporter is reportedly activated shortly beforehand, during meiotic maturation (Van Winkle et al.,1992). Here we have found that the betaine/proline transporter previously identified in mouse 1-cell embryos is quiescent in oocytes prior to fertilization and then becomes activated within several hours of egg activation, providing the first example of such a transporter that is activated upon fertilization in mammals. Once betaine/proline transport is activated, it persists for only a little more than one cell cycle before becoming inactive again. We found high transporter activity only at the 1-cell and 2-cell stages, both in embryos that had developed in vivo and in parthenogenotes that developed in vitro. We are not aware of any other transporters, the functional expression of which is restricted to so short a developmental period of very early PI embryogenesis.

The appearance of betaine/proline transport after egg activation apparently involves pre-existing transporters, rather than de novo protein synthesis, as parthenogenotes developed the same level of betaine transport in the continuous presence of cycloheximide as in Sr²⁺-activated eggs. We further found that the Golgi disruptor BFA reversibly prevented the development of betaine transport activity in activated eggs, while not affecting activity that had already developed in 1-cell embryos. Thus, we infer that insertion of presynthesized proteins into the plasma membrane is required for activation. One route for membrane protein insertion into the egg plasma membrane after egg activation is cortical granule exocytosis. However,this occurs very rapidly after egg activation and is blocked by perturbation of F-actin (Sun and Schatten,2006), unlike the betaine/proline transporter. A better candidate is a distinct process of protein trafficking that has been reported to be more slowly induced by egg activation in mammals and is blocked by BFA but insensitive to disruption of F-actin or microtubules(Clayton et al., 1995), similar to our findings here for betaine/proline transport. This mechanism mediates the appearance of the cell adhesion molecule uvomorulin (E-cadherin) in the plasma membrane over a period of ∼6 hours following egg activation(Clayton et al., 1995), and occurs during the same period that the fragmented, inactive Golgi in MII oocytes becomes reorganized following fertilization(Payne and Schatten,2003).

It was previously proposed (Anas et al.,2007) that the betaine/proline transporter in 1-cell embryos was most likely a member of the neurotransmitter transporter (Slc6) gene family,and two candidates were identified based on their acceptance of proline as a substrate:PROT (Slc6a7) and SIT1 (Slc6a20a). We unexpectedly found mRNA for both in eggs and 1-cell embryos. In addition, gene array data for PI embryo development including MII oocytes(Zeng et al., 2004) deposited in the Gene Expression Omnibus database(http://www.ncbi.nlm.nih.gov/projects/geo)show similar expression patterns for each gene to those we report here (GEO accession/array ID/gene name: GDS813/1422899_at/Slc6a20a and GDS814/1455469_at/Slc6a7). Thus, we could not rule out either candidate based on mRNA expression.

We expressed each candidate in Xenopus oocytes and tested their ability to transport betaine or proline and whether, like the betaine/proline transporter in 1-cell embryos, this was inhibited by MeAIB. In addition,inhibition by histidine was tested because it had previously been reported to block proline transport by PROT (Fremeau et al., 1992). The results appear to rule out PROT because although Xenopus oocytes expressing PROT showed increased transport of proline, there was no increase in betaine transport. Also, proline transport was blocked by histidine but not inhibited by MeAIB and, most importantly, it was not inhibited by excess betaine. Although we cannot rule out the possibility that accessory proteins required for normal PROT function are absent from Xenopus oocytes, the close correspondence of proline transport characteristics in PROT-injected Xenopus oocytes to those previously reported in mammalian cells(Fremeau et al., 1992) makes this unlikely. By contrast, the transport characteristics of mouse SIT1 in Xenopus oocytes closely matched those of betaine/proline transport in 1-cell mouse embryos (Anas et al.,2007). Both betaine and proline were transported, inhibited by MeAIB, and resistant to histidine. In addition, the rate of transport of 1μM proline was approximately twice that of 1 μM betaine, which is the same relationship as that found in 1-cell mouse embryos(Anas et al., 2007). Taken together, these results implicate SIT1 as the transporter that is likely to mediate betaine and proline transport in 1-cell mouse embryos.

This conclusion was supported by the results obtained with a MO targeted against Slc6a20a, which showed a substantial loss of betaine transport by 1-cell mouse embryos in which the Slc6a20a MO had been present from the GV stage. Since inhibition of global protein synthesis with cycloheximide did not block the appearance of betaine/proline transport activity at 12 hours post-egg activation, the efficacy of the MO when present from the GV stage implies SIT1 synthesis during meiotic maturation. Further investigations into the mechanism underlying the appearance of transport activity after fertilization are clearly needed. Immunolocalization of the transporter protein in mouse oocytes and early embryos would help address this question. However, thus far we have not been able to successfully use the available antiserum raised against SIT1 (a kind gift of Dr F. Verrey) for immunocytochemistry in mouse oocytes.

The activation of betaine/proline transport after fertilization and its presence only during a very short period of development implies that it serves an important function during PI embryo development. Although, in vitro, SIT1 transports betaine and proline, we propose that its physiological function in the early embryo is to mediate betaine transport. Betaine is apparently accumulated to relatively high levels in vivo. Assuming a volume of cleavage-stage embryos of ∼180 pl, the measured amounts of endogenous betaine in 1- and 2-cell embryos correspond to intracellular concentrations of∼6-7 mM. This is comparable to the average concentration of ∼4 mM in liver, which is the tissue with the highest known endogenous betaine level,and much higher than the ∼0.1 mM in rodent blood plasma(Slow et al., 2008) or∼0.5 mM in mouse oviducts (Anas et al.,2007). The apparent accumulation of betaine in vivo would appear to imply that a major function of SIT1 is to mediate betaine accumulation or retention by early PI embryos.

The betaine/proline transporter in PI embryos that we have identified here as SIT1 is the sole route of betaine transport in PI embryos, whereas there are at least two routes of proline transport(Anas et al., 2007). The second proline transport route in 1-cell embryos does not correspond to PROT despite the presence of its mRNA, but instead resembles the classical transport system ASC (Anas et al., 2007). Although proline is likely to be transported into early embryos, the importance of SIT1-mediated uptake relative to other routes is not known.

The mechanism by which betaine exerts a beneficial effect on PI embryo development is not yet known. As discussed above, two established functions for betaine in mammals are as an organic osmolyte and as a methyl group donor. We (Dawson and Baltz, 1997; Hammer and Baltz, 2002) and others (Biggers et al., 1993)have established the ability of betaine to protect PI embryo development against increased osmolarity in vitro, although it is not yet certain that it performs this role in vivo. Previous work has established glycine as a major organic osmolyte in cleavage-stage PI mouse embryos. Glycine is transported by the specific transporter GLYT1 (SLC6A9)(Steeves et al., 2003), and endogenous glycine is present in 1-cell mouse embryos at high levels [∼25 mM free glycine in freshly obtained 1- and 2-cell embryos; our unpublished data, measured as described by Steeves et al.(Steeves et al., 2003)]. Thus,betaine present at ∼6-7 mM might play an important role along with glycine in osmoprotection in vivo.

An alternative hypothesis is that intracellular betaine might instead serve as a methyl pool during PI embryogenesis. One crucial set of events occurring during PI embryogenesis is the global DNA demethylation and remethylation between fertilization and implantation, which mediates the switch from parental to embryonic epigenetic marking, while maintaining DNA methylation of uniparentally imprinted genes (Lucifero et al., 2004). Although the folate pathway is usually thought to provide the methyl groups for the large array of methyltransferases in most cells outside liver, it might be that betaine participates in PI embryos because late PI embryos appear to express betaine-homocysteine methyltransferase (BHMT), the key enzyme in liver that mediates the transfer of methyl groups from betaine (our unpublished data). Further work, however,is needed to explore this possibility, including determining whether betaine accumulated at the 1- to 2-cell stage can be retained long enough to be utilized in the peri-implantation period.

In summary, we have found that the betaine/proline transporter is expressed for only a restricted period of development starting a few hours after fertilization and continuing through the 2-cell stage. This transport activity is likely to correspond to SIT1, arising by activation of pre-existing transporter proteins after fertilization. Although betaine has been clearly shown to have beneficial effects on PI embryos, acting as an organic osmolyte and promoting in vitro development past the 2-cell stage, further work is needed to fully establish the physiological functions of the betaine that is accumulated by early PI mouse embryos.

Supported by Canadian Health Research Institutes (CIHR) operating grant MOP62730 to J.M.B. M.-K.I.A. was supported by a Lalor Foundation Postdoctoral Fellowship. M.B.L. and a portion of the work were supported by the Program on Oocyte Health, Healthy Gametes and Great Embryos Strategic Initiative, CIHR Institute of Human Development, Child and Youth Health, grant HGG62293.