Homocysteine transport by systems L, A and y⁺L across the microvillous plasma membrane of human placenta

Abstract

Elevated maternal plasma levels of homocysteine (Hcy) are associated with pregnancy complications and adverse neonatal outcomes, suggesting placental transport of Hcy may impact on fetal development. However, such transport mechanisms have not been defined. In this study we characterise Hcy transport mechanisms across the microvillous plasma membrane (MVM) of the syncytiotrophoblast, the transporting epithelium of human placenta. Three candidate transport systems, systems L, A and y⁺L, were examined utilising competitive inhibition to investigate the effects of Hcy on the uptake of well-characterised radiolabelled substrates for each system into isolated MVM vesicles, and that of model substrates on 10 μm[³⁵S]l–Hcy uptake. System L activity was inhibited by both l-Hcy and dl–Hcy, comparable to model substrates including 2–aminobicyclo[2.2.1]heptane-2-carboxylic acid (BCH). System L constituted the major transport mechanism, with significant BCH inhibition (∼69%) of [³⁵S]l–Hcy uptake. System A activity was also inhibited by l–Hcy and dl-Hcy with a smaller contribution (∼21%) to [³⁵S]l–Hcy uptake. Inhibition by l–Hcy and dl–Hcy of system y⁺L activity was Na⁺ sensitive with a significant inhibition constant (K_i) shift observed following K⁺ replacement; l–arginine reduced [³⁵S]l–Hcy uptake by ∼19%. Kinetic modelling of [³⁵S]l–Hcy uptake resolved two, Na⁺-independent, transport components (K_m 72 μm and 9.7 mm). This study provides evidence for the involvement of systems L, A and y⁺L in placental Hcy transport. Such transport, by competing with endogenous amino acids for transporter activity, could have major implications for syncytiotrophoblast metabolism and function as well as fetal development.

Homocysteine (Hcy) is a thiol group-containing amino acid metabolite formed in the methionine (Met) cycle (Selhub, 1999). As a branch point metabolite, Hcy can be remethylated back to Met or converted to cysteine through the transsulphuration pathway (Selhub, 1999). Perturbation of these metabolic pathways leads to hyperhomocysteinaemia (HHcy), a condition in which the plasma concentration of total Hcy (tHcy), comprising reduced and oxidized forms of Hcy, is elevated. Plasma tHcy is regulated by several factors including nutritional deficiencies in the vitamins that act as co-factors or co-substrates in Hcy metabolism (folate, vitamins B₁₂ and B₆), or genetic defects in the enzymes responsible for Hcy metabolism (Refsum et al. 1998; Selhub, 1999).

Evidence from several epidemiological studies has revealed that HHcy is associated with a diverse array of health complications of considerable public health significance, reflecting the pleiotropic nature of the effects elicited by Hcy. Indeed, such studies have identified Hcy as an independent risk factor for cardiovascular disease, stroke, cognitive dysfunction and osteoporosis (Refsum et al. 2006; Selhub, 2006). Hcy is also an independent risk for obstetric vascular disease. Raised maternal plasma levels of tHcy are associated with various vascular-related complications of pregnancy including pre-eclampsia (Ray & Laskin, 1999; Vollset et al. 2000; de la Calle et al. 2003), placental abruption (Goddijn-Wessel et al. 1996; Ray & Laskin, 1999; Vollset et al. 2000; de la Calle et al. 2003), recurrent pregnancy loss (Ray & Laskin, 1999; Quéréet al. 2001; de la Calle et al. 2003), fetal growth restriction (Vollset et al. 2000; de la Calle et al. 2003) and stillbirth (Vollset et al. 2000), as well as adverse outcomes for the baby such as neural tube defects and congenital malformations (Vollset et al. 2000; de la Calle et al. 2003). Supplementation with folic acid and B group vitamins has proved beneficial in normalising maternal tHcy and improving pregnancy outcome (Quéréet al. 2001; de la Calle et al. 2003) signifying that some of the risks are modifiable.

Maternal plasma tHcy concentration is significantly lowered during pregnancy (Murphy et al. 2004) but this is not accounted for by haemodilution, a decrease in albumin levels to which 70–80% of tHcy is bound, or folic acid supplementation (Murphy et al. 2002). Indeed it has been proposed that this is a physiological adaptation to pregnancy (Murphy et al. 2002). Maternal plasma tHcy concentration is higher than fetal plasma tHcy concentration and is a major determinant of the latter, with a significant, positive correlation between these two variables (Malinow et al. 1998; Molloy et al. 2002; Murphy et al. 2004). These observations are compatible with the concept that Hcy is transported by the placenta, supported further by the evidence of a decreasing concentration gradient of plasma tHcy from maternal vein to umbilical vein to umbilical artery (Malinow et al. 1998). However, Hcy transport mechanisms in the placenta have not previously been explored.

We reasoned that maternofetal transfer of Hcy across the transporting epithelium of the human placenta, the syncytiotrophoblast, would involve amino acid transporters that are intrinsic to both the maternal-facing microvillous plasma membrane (MVM) and the fetal-facing basal plasma membrane of the syncytiotrophoblast which transport neutral amino acids (Jansson, 2001; Cleal & Lewis, 2008). These comprise Na⁺-dependent and Na⁺-independent mechanisms with some transport mechanisms common to both syncytiotrophoblast plasma membranes, including systems A, L and y⁺L (Jansson, 2001; Cleal & Lewis, 2008).

System A activity is mediated by three highly homologous isoforms of the sodium-coupled neutral amino acid transporter (SNAT) family, namely SNAT1 (SLC38A1), SNAT2 (SLC38A2) and SNAT4 (SLC38A4), all of which are expressed in the syncytiotrophoblast and localised to MVM (Desforges et al. 2009). System y⁺L comprises a heterodimeric complex with the CD98 heavy chain linked to either y⁺LAT1 (SLC7A7) or y⁺LAT2 (SLC7A6) light chain (Torrents et al. 1998; Pfeiffer et al. 1999) which confer functional activity and substrate specificity. Both y⁺LAT light chains are expressed in human placenta at the mRNA level (Dye et al. 2004). System L is also a heterodimeric transport system in which CD98 associates with LAT1 (SLC7A5) and LAT2 (SLC7A8) light chains, both of which are expressed in human placenta at the mRNA level (Kudo & Boyd, 2000). LAT1 is particularly enriched in MVM compared to basal plasma membrane (Okamoto et al. 2002) whilst the distribution of LAT2 remains unknown. Functional evidence suggests LAT2 may be present on both MVM and basal plasma membrane (Kudo & Boyd, 2001; Lewis et al. 2007). It has also been proposed that LAT4 (SLC43A2), a monomeric amino acid transporter mediating system L–type transport, and which is highly expressed in placenta (Bodoy et al. 2005), might also mediate placental transport of system L amino acid substrates across the basal plasma membrane (Cleal & Lewis, 2008).

Each of these three transport mechanisms, systems A, L and y⁺L, have the common characteristic of being able to utilise Met as a substrate. Based on the biochemical similarity of Met to Hcy, we hypothesised that these transporters would also support Hcy as a substrate. To test this hypothesis, we investigated whether these candidate transport systems are involved in Hcy transport across the MVM of human placenta, the first plasma membrane barrier to maternofetal amino acid transfer. Isolated MVM plasma membrane vesicles were employed in a dual strategy of examining the effect of (i) unlabelled Hcy on the uptake of well-characterised radiolabelled substrates for each transport system and (ii) unlabelled model substrates for each transport system on radiolabelled Hcy uptake.

Methods

Materials

All chemicals were purchased from either Sigma-Aldrich Co. Ltd (Poole, UK) or VWR International (Lutterworth, UK) unless otherwise stated. l–Hcy was custom-synthesized from l–Met (Department of Chemistry, University of Manchester) by the method of Shiraiwa et al. (2000) with >94% purity confirmed by HPLC analysis and no evidence of oxidized product homocystine. [¹⁴C]α-(methylamino)isobutyric acid (MeAIB; 58.8 mCi mmol⁻¹), [³⁵S]l–Met (>1000 Ci mmol⁻¹) and [³H]l–arginine (Arg; 50 Ci mmol⁻¹) were obtained from Perkin Elmer (Beaconsfield, UK), GE Healthcare (Little Chalfont, UK) and MP Biomedicals (London, UK), respectively.

Tissue acquisition

Placentas were obtained with written informed consent in accordance with ethical approval (Central Manchester Local Research Ethics Committee). Placentas were collected from normal, uncomplicated, singleton pregnancies at term (38–41 weeks) within 30 min of vaginal delivery or Caesarean section. Gestational age was calculated from the first day of the last menstrual period and confirmed by ultrasound scan.

Isolation of MVM vesicles

MVM vesicles were prepared from human placenta as described previously (Method 3 in Glazier et al. 1988). Protein concentration was measured by the method of Lowry et al. (1951). Membrane purity was assessed by measuring the enrichment of alkaline phosphatase (activity in vesicles/initial homogenate), as a marker of MVM (Glazier et al. 1988). Vesicles were suspended in intravesicular buffer (of the following composition: 290 mm sucrose, 5 mm Tris, 5 mm Hepes, pH 7.4) unless indicated otherwise. MVM vesicles were stored at 4°C prior to transport assay which were performed within 48 h of isolation. Alkaline phosphatase activity was enriched 20 ± 1-fold in MVM vesicles (n= 50), of comparable purity to our previous studies (Mahendran et al. 1993).

Transport assays

Strategy for the characterisation of l–Hcy transport mechanisms in MVM of human placenta

To characterise the transport mechanisms for l–Hcy across human placental MVM, we firstly undertook competitive inhibition studies, examining whether l–Hcy inhibited the uptake of radiolabelled model substrates for systems A, L and y⁺L, the three candidate transport systems under investigation. The inhibitory effect of l–Hcy was compared to that of well-characterised amino acid substrates for each transport mechanism; a similar pattern of inhibition was taken to indicate mediation by a common transport pathway. Additionally, the effect of dl–Hcy was included for reference, this being the commercial source of Hcy used previously by others to characterise transport mechanisms (Naggar et al. 2003; Hultberg, 2004). Secondly, in a more direct approach, uptake of radiolabelled [³⁵S]l–Hcy was measured and the effect of unlabelled model substrates for each candidate transport system determined.

Effect of Hcy on system A activity in MVM vesicles

System A activity in MVM vesicles was measured as the Na⁺-dependent uptake of [¹⁴C]MeAIB, a non-metabolisable substrate transported by system A (Mackenzie & Erickson, 2004), as described previously (Mahendran et al. 1993). Briefly, uptakes at initial rate (30 s) were performed by adding 20 μl MVM vesicles to 20 μl extravesicular buffer (EVB) with Na⁺ (EVB-Na⁺; 145 mm NaCl, 5 mm Tris, 5 mm Hepes, pH 7.4) or without Na⁺ (EVB-K⁺; 145 mm KCl, 5 mm Tris, 5 mm Hepes, pH 7.4) containing 0.33 mm[¹⁴C]MeAIB. Competitive inhibition of system A activity by 20 mm l–Hcy and dl–Hcy was compared to that of l–alanine (Ala), l–serine (Ser), l–Met, and MeAIB as paradigm substrates for this transport mechanism. 2-Aminobicyclo[2.2.1]heptane-2-carboxylic acid (BCH), which is not transported by system A (Johnson & Smith, 1988), was included as a negative control. In order to explore the concentration dependency of inhibition, the effect of l–Hcy on Na⁺-dependent [¹⁴C]MeAIB uptake into MVM was examined over a broad concentration range (50 μm–20 mm), and the concentration dependency of inhibition compared to that of l–Met and MeAIB.

Effect of Hcy on system L activity in MVM vesicles

System L activity in MVM vesicles was measured as the Na⁺-independent uptake of [³⁵S]l–Met, a model substrate for system L (Ganapathy et al. 1986; Johnson & Smith, 1988). [³⁵S]l–Met uptakes were performed initially over 15–60 s to establish the initial rate by adding 20 μl MVM vesicles to 20 μl EVB-K⁺ containing 0.4 μm[³⁵S]l–Met. Competitive inhibition by Hcy of system L activity in MVM vesicles was examined at apparent initial rate (30 s; found in the time course experiments to be on the linear part of the uptake) by the addition of unlabelled l–Hcy and dl–Hcy (0.1, 1, 10 and 20 mm), with comparison to the effects of l–leucine (Leu), BCH and l–Met as model substrates for this transport mechanism (Ganapathy et al. 1986; Johnson & Smith, 1988). The effect of dl–Hcy was also examined for reference. MeAIB, an amino acid analogue not transported by system L (Ganapathy et al. 1986), was included as a negative control. To investigate the concentration dependency of inhibition, the effect of unlabelled l–Hcy was examined over an expansive concentration range (0.5 μm–20 mm) and compared to the effect by l–Met.

Effect of Hcy on system y⁺L activity in MVM vesicles

System y⁺L activity was measured as the uptake of 0.2 μm[³H]l–Arg following pre-incubation of MVM vesicles with 4 μm valinomycin for 1 h as described previously (Ayuk et al. 2000). MVM vesicles were suspended in intravesicular buffer (50 mm KCl, 50 mm choline chloride, 100 mm mannitol, 20 mm Hepes-Tris, pH 7.4) and uptakes performed at initial rate (30 s) by adding 20 μl MVM vesicles to 100 μl EVB (50 mm, NaCl, 50 mm KCl, 100 mm mannitol, 20 mm Hepes-Tris, pH 7.4) containing 0.24 μm[³H]l–arginine (Ayuk et al. 2000). Competitive inhibition of system y⁺L activity by Hcy was examined by comparing the effect of unlabelled dl–Hcy to that of cationic (l–Arg and l–lysine (Lys)) and neutral (l–Leu, l–glutamine (Gln) and l–Met) amino acid substrates for this transporter (Devés & Boyd, 1998).

The Na⁺ dependency of inhibition by l–Hcy and dl–Hcy of system y⁺L activity was investigated and compared to that of l–Met as a model neutral amino acid substrate. Uptake of [³H]l–Arg was measured in the presence of competing amino acid (50 μm–10 mm) with and without Na⁺ (with K⁺ replacement) to allow determination of the inhibition constant (K_i) under both conditions. K_i values were determined using a statistical program (Graphpad Prism 4, San Diego, CA, USA) from the equation of Cheng & Prusoff (1973) where K_m= 7.6 μm for system y⁺L was applied, as determined previously in MVM (Ayuk et al. 2000). An increase in K_i when Na⁺ was replaced by K⁺ was taken to indicate transport by system y⁺L (Devés & Boyd, 1998; Ayuk et al. 2000).

Measurement of [³⁵S]l–Hcy uptake into MVM vesicles

[³⁵S]l–Hcy was prepared from [³⁵S]l–Hcy thiolactone (5 mm; 357 μCi μmol⁻¹) by alkaline hydrolysis (Duerre & Miller, 1966) using either 5 m NaOH or KOH (allowing for transport measurements to be performed in the presence or absence of Na⁺, respectively), as described previously (Büdy et al. 2006), but with modifications as described below. Following alkaline hydrolysis for 2 min at 37°C, the mixture was immediately neutralised with 0.1 m TES buffer (pH 7.4) containing 50 mm Tris(2-carboxyethyl)phosphine hydrochloride (TCEP.HCl) as reductant to preserve the thiol group (Cline et al. 2004) and an appropriate volume of 5 m HCl added to achieve a pH of 7.4. The concentration of [³⁵S]l–Hcy was measured by quantification of thiol concentration using a commercial kit according to the manufacturer's instructions (Measure-iT Thiol Assay Kit; Invitrogen, Paisley, UK). Converted [³⁵S]l–Hcy was stored at −20°C between transport assays.

Uptake of [³⁵S]l–Hcy into MVM vesicles was performed as described previously for system A (Mahendran et al. 1993). Uptake was initiated by adding 20 μl MVM vesicles (∼250–300 μg protein) to 20 μl EVB-Na⁺ or EVB-K⁺ containing 20 μm[³⁵S]l–Hcy. Non-specific binding of tracer to MVM plasma membrane was quantified following disruption of vesicle integrity with 0.2% Triton.

To characterise the kinetic properties of [³⁵S]l–Hcy uptake into MVM vesicles, uptakes were performed at 30 s (taken to be initial rate; see Results) in the presence of increasing concentrations of unlabelled l–Hcy (5 μm–10 mm). To examine the involvement of candidate transport mechanisms systems A, L and y⁺L in [³⁵S]l–Hcy transport across human placental MVM, competitive inhibition of [³⁵S]l–Hcy uptake into MVM vesicles was performed in the presence of model substrates for each transport system; MeAIB, BCH and l–Arg for systems A, L and y⁺L, respectively.

Statistical analysis

Data are presented as the mean ±s.e.m., where n is the number of placentas. Competitive inhibition of amino acid uptake into MVM was analysed by Kruskal–Wallis with Dunn's multiple comparison test, two-way ANOVA with Bonferroni post test or Friedman with Dunn's multiple comparison test as appropriate. For each transport mechanism, the half-maximal inhibitory concentration (IC₅₀) was derived by non-linear regression to a one-site transport system model. All statistical analyses were performed using a statistical program (Graphpad Prism 4, San Diego, CA, USA).

Kinetic modelling of l–Hcy competitive inhibition curves was achieved by comparison to Michaelis–Menten equations with up to two saturable components (SIMFIT version 6.0.18; W.G. Bardsley, University of Manchester; http://www.simfit.man.ac.uk). The SIMFIT computer package supports curve fitting of data with discrimination of kinetically distinct transporters by determination of the best-fit to the Michaelis–Menten equation. The data were fitted using a program within SIMFIT which assumes the kinetic transformation process is the same whether the substrate is labelled or not, so if the radiolabelled substrate is fixed ([hot]), the initial rate of uptake (y) will be proportional to the concentration of unlabelled substrate ([cold]) added, allowing isotope displacement kinetics to be modelled to the following equation:

where [hot] is the concentration of radiolabelled substrate, [cold] is the concentration of unlabelled substrate, K_m is the Michaelis constant and V_max is the maximal velocity. The best fit to the data was determined by the F test, comparing the closeness of fit (weighted sum of squares of the variance) against the number of parameters in the model.

Results

Inhibition of system A activity by Hcy

Both l–Hcy and dl–Hcy caused a profound inhibition of system A activity in MVM vesicles, similar to that observed with the model amino acid substrates for this transport mechanism (Fig. 1A). BCH, which is not transported by system A, was not inhibitory, as expected (Fig. 1A). Inhibition of system A activity by l–Hcy was concentration dependent, also observed for the positive controls l–Met and MeAIB (Fig. 1B). The IC₅₀ for l–Hcy (0.53 ± 0.04 mm) was not significantly different from that for l–Met (0.27 ± 0.02 mm) or MeAIB (1.13 ± 0.12 mm; n= 6 for all). However, IC₅₀ for l–Met was significantly lower (P < 0.001) than that for MeAIB, suggesting a higher preference for the former substrate.

Figure 1. Hcy inhibition of system A activity in MVM vesicles.

A, effect of l–Hcy and dl–Hcy on system A activity at initial rate (30 s) compared to the model system A substrates MeAIB, l–Ala, l–Ser and l–Met. BCH served as negative control (20 mm for all). All data are expressed as a percentage of control uptake (no amino acid; 100%). Data are shown as mean ±s.e.m. (n= 9) except for BCH (n= 4). *P < 0.05, **P < 0.01, ***P < 0.001 vs control (Kruskal–Wallis with Dunn's multiple comparison test). B, effect of varied concentrations (50 μm–20 mm) of l–Hcy (▾), MeAIB (▪) and l–Met (▴) on the Na⁺-dependent uptake of [¹⁴C]MeAIB (0.165 mm) at initial rate. Data are mean ±s.e.m. (n= 6). Data from individual experiments were fitted by non-linear regression to a sigmoidal dose–response model.

Inhibition of system L activity by Hcy

In preliminary studies we explored the possibility of using [³H]l–Leu as a paradigm substrate for system L, but observed that uptake equilibrium was achieved extremely rapidly, within 5 s, the earliest time point that could be reproducibly measured (data not shown). We therefore employed [³⁵S]l–Met as a system L substrate (Fig. 2) and initially examined the linearity and Na⁺ independence of [³⁵S]l–Met uptake. Uptake of [³⁵S]l–Met was linear over 15–60 s and unaltered by an inwardly directed Na⁺ gradient (Fig. 2A). Therefore, system L activity in MVM vesicles was measured as the Na⁺-independent uptake of [³⁵S]l–Met with 30 s taken to indicate apparent initial rate of uptake. We use the term ‘apparent initial rate’ here as extrapolation of the regression line to time zero in Fig. 2A yielded an uptake value of ∼0.2 pmol (mg protein)⁻¹, greater than the residual component of uptake following BCH inhibition (∼0.05 pmol (mg protein)⁻¹; Fig. 2B), suggesting there is an extremely rapid initial uptake of [³⁵S]l–Met by a BCH-inhibitable component. The dose-dependent inhibition by both l–Hcy and dl–Hcy of Na⁺-independent [³⁵S]l–Met uptake into MVM was of comparable magnitude to that exhibited by all system L substrates (l–Met, l–Leu, BCH) examined (Fig. 2B), suggesting mediation by a common transport mechanism. As expected, uptake was unaltered in the presence of MeAIB, a specific system A substrate not transported by system L, included as a negative control. The concentration dependency of inhibition was further explored, comparing the effect of l–Hcy to l–Met as reference (Fig. 2C). The observed similarity in the inhibition profiles between these two amino acids, as illustrated in Fig. 2C, lends further support to the concept that these two amino acids are transported by a common transport mechanism. Indeed, the IC₅₀ was not significantly different between l–Hcy (0.29 ± 0.05 mm) and l–Met (0.17 ± 0.07 mm; n= 6), suggesting l–Hcy has a relatively high affinity for system L.

Figure 2. Hcy inhibition of system L activity in MVM vesicles.

A, time-dependent uptake of 0.2 μm[³⁵S]l–Met into MVM vesicles over 15–60 s in the presence (▪) and absence (▴) of an inwardly directed Na⁺ gradient. Data are mean +s.e.m. (n= 8) with linear regression plot shown (r² > 0.9, P < 0.05 for both conditions). There was no significant difference between the slopes of the lines. B, effect of varied concentrations (0.1, 1, 10 and 20 mm) of l–Hcy, dl–Hcy, l–Met, l– l–Leu, BCH and MeAIB on uptake of 0.2 μm[³⁵S]l–Met into MVM vesicles. Data are mean +s.e.m. (n= 6, except at 20 mm where n= 5). *P < 0.05, **P < 0.01, ***P < 0.001 vs control (C) (Kruskal–Wallis with Dunn's multiple comparison test). C, effect of varied concentrations (0.5 μm–20 mm) of l–Hcy (▾) or l–Met (▴) on uptake of 0.2 μm[³⁵S]l–Met into MVM vesicles. Data are mean ±s.e.m. (n= 3). Data from individual experiments were fitted by non-linear regression to a sigmoidal dose–response model.

Inhibition of system y⁺L activity by Hcy

As shown in Fig. 3, l–Hcy (Fig. 3A), dl–Hcy (Fig. 3B) and l–Met included as model substrate for system y⁺L (Fig. 3C), all demonstrated concentration-dependent inhibition of ³[H]l–Arg uptake into MVM under both Na⁺-containing and Na⁺-free (by K⁺ replacement) conditions. Abolition of the inwardly directed Na⁺ gradient across the MVM vesicular membrane resulted in an upward shift in the inhibition curve for l–Hcy (Fig. 3A), dl–Hcy (Fig. 3B) and l–Met (Fig. 3C). This shift was most notably divergent for l–Hcy (Fig. 3A) and l–Met (Fig. 3C). The Na⁺ dependency of inhibition was lost at higher substrate concentrations, a pattern characteristic of system y⁺L (Devés & Boyd, 1998). Table 1 provides confirmation of the Na⁺ dependency of inhibition for all three amino acids as shown by the significantly lower K_i values found in the presence of Na⁺. This implies a higher affinity of neutral amino acid binding in the presence of Na⁺ providing positive identification of system y⁺L activity (Devés & Boyd, 1998; Ayuk et al. 2000). As might be predicted from the preference of system y⁺L for the l–stereoisomer form of neutral amino acids (Devés & Boyd, 1998), the K_i shift for dl–Hcy was less pronounced than for l–Hcy. However, further experiments examining the effect of dl–Hcy revealed a similar inhibitory capacity at high concentration (10 mm) compared to other system y⁺L neutral amino acid substrates (l–Gln, l–Met, l–Leu), although at lower concentrations dl–Hcy was less effective (Fig. 4). Figure 4 also serves to demonstrate that there was a component of [³H]l–Arg uptake that was resistant to neutral amino acid inhibition but which was inhibitable by the cationic amino acid substrates l–Arg and l–Lys; this can be attributed to system y⁺ activity (Ayuk et al. 2000).

Figure 3. Na⁺ dependency of Hcy inhibition of [³H]l–Arg uptake into MVM vesicles.

Effect of l–Hcy (A), dl–Hcy (B) and l–Met (C) as positive control on 0.2 μm[³H]l–Arg uptake by MVM vesicles in the presence (▪) or absence (▴) of an inwardly directed Na⁺ gradient at initial rate (30 s). Data are mean ±s.e.m. (n= 6) and were fitted by non-linear regression to a one-site binding model.

Table 1.

K_i values for amino acid inhibition of 0.2 μm[³H]l–arginine uptake into MVM vesicles

	Ki (μm)
Amino acid	Na+	K+	Ki shift§
l–Hcy	535 ± 184	1419 ± 435*	2.6
dl–Hcy	826 ± 100	1354 ± 216*	1.6
l–Met	177 ± 20	1021 ± 320*	5.8

Data are mean ±s.e.m., (n= 6). *P < 0.05 vs Na⁺ (paired t test). §K_i shift was calculated as ratio (K⁺/Na⁺) of mean K_i value.

Figure 4. Hcy inhibition of [³H]l–Arg uptake into MVM vesicles.

Effect of unlabelled (0.05–10 mm) neutral (dl–Hcy, l–Gln, l–Met, l–Leu) and cationic (l–Lys, l–Arg) amino acids on 0.2 μm[³H]l–Arg uptake into MVM vesicles at initial rate (30 s). Data are mean ±s.e.m., except where this falls within size of symbol (n= 6). *P < 0.05, ***P < 0.001 dl–Hcy vs all other amino acids, except at 10 mm, l–Lys and l–Arg only (two-way ANOVA with Bonferroni's post test).

Uptake of [³⁵S]l–Hcy by human placental MVM vesicles

Time course and effect of inwardly directed Na⁺ gradient

Uptake of 10 μm[³⁵S]l–Hcy into MVM was time dependent and linear over 10–40 s both in the presence and absence of an inwardly directed Na⁺ gradient (Fig. 5). The rate of uptake was not significantly different between these two conditions, suggesting that Na⁺-independent transport mechanisms of [³⁵S]l–Hcy uptake prevail under the conditions of assay. The magnitude of uptake was, however, significantly higher in the presence of Na⁺ (P < 0.05, paired t test). Although these observations would be consistent with there being unidirectional uptake of [³⁵S]l–Hcy across the MVM plasma membrane into an intravesicular space under both conditions, disruption of vesicular integrity by 0.2% Triton indicated there was considerable binding of [³⁵S]l–Hcy to the MVM plasma membrane. The degree of binding measured at 40 s was similar in the presence or absence of Na⁺ (9.6 ± 1.1 and 10.8 ± 0.9 pmol (mg protein)⁻¹, respectively, n= 5). Addition of unlabelled l–Hcy demonstrated this component was largely exchangeable with ‘cold’l–Hcy, as shown by marked inhibitory capacity (Fig. 6). We therefore reasoned that this did not reflect irreversible binding of [³⁵S]l–Hcy to the MVM plasma membrane and took the displacement by exogenous l–Hcy to infer that this component represented binding of [³⁵S]l–Hcy to transport moieties prior to translocation across MVM. All uptake data are therefore shown uncorrected for binding. For all subsequent measurements, 30 s was taken to indicate initial rate of uptake.

Figure 5. Linearity of [³⁵S]l–Hcy uptake into MVM vesicles.

Uptake of 10 μm[³⁵S]l–Hcy into MVM vesicles in the presence (▪) or absence (▴; with K⁺ replacement) of an inwardly directed Na⁺ gradient over 10–40 s. Linear regression plot is shown (r² > 0.97, P < 0.05 for both conditions). Data are presented as mean ±s.e.m. (n= 6).

Figure 6. Inhibition by unlabelled amino acids or analogues of [³⁵S]l–Hcy uptake into MVM vesicles.

Effect of unlabelled amino acids or analogues on uptake of 10 μm[³⁵S]l–Hcy into MVM vesicles at initial rate (30 s) in the absence (A) or the presence (B) of an inwardly directed Na⁺ gradient shown as percentage of control uptake represented by the dashed line. Concentration-dependent inhibition by l–Hcy (black bars) was observed under both conditions. The system L analogue BCH (pale grey bars) significantly inhibited uptake at all concentrations, whereas the system y⁺L substrate l– Arg (dark grey bars) failed to evoke significant inhibition. Data are presented as mean ±s.e.m. (n= 5–7). *P < 0.05, **P < 0.01, ***P < 0.001 vs control (Kruskal–Wallis with Dunn's multiple comparison test).

Effect of amino acids or analogues on [³⁵S]l–Hcy uptake

The effect of exogenously added l–Hcy on [³⁵S]l–Hcy uptake as well as that for substrates of the three candidate systems, BCH (system L), l–Arg (system y⁺L) and MeAIB (system A), are shown in Figs 6 and 7. Under both Na⁺-free (Fig. 6A) and Na⁺-containing (Fig. 6B) conditions, [³⁵S]l–Hcy uptake was significantly inhibited by l–Hcy in a concentration-dependent manner. At 10 mm l–Hcy, ∼80% inhibition of control uptake was observed under both conditions. The inhibitory potential of BCH and l–Arg was only examined under Na⁺-free conditions (Fig. 6A) as these substrates are transported in a Na⁺-independent manner by systems L and y⁺L, respectively. BCH significantly inhibited [³⁵S]l–Hcy uptake at all three concentrations tested (0.5, 10 and 20 mm). In contrast, no significant inhibition of [³⁵S]l–Hcy uptake was observed with 0.5 or 10 mm l–Arg, although a mean reduction of ∼19% was apparent at the higher concentration; this is likely to be accounted for by mechanisms other than system y⁺L, that is system y⁺, mediating l–Arg transport across MVM (Ayuk et al. 2000). Figure 7 shows that 20 mm MeAIB significantly inhibited uptake of [³⁵S]l–Hcy and that this inhibition was dependent of the presence of Na⁺.

Figure 7. MeAIB inhibition of [³⁵S]l–Hcy uptake into MVM vesicles.

Effect of unlabelled MeAIB on uptake of 10 μm[³⁵S]l–Hcy into MVM vesicles at initial rate (30 s) in the presence (black bars) or the absence (K⁺ replacement; grey bars) of an inwardly directed Na⁺ gradient shown as percentage of control uptake. 20 mm MeAIB significantly inhibited uptake and this inhibition was dependent on the presence of Na⁺. Data are presented as mean +s.e.m. (n= 5). **P < 0.01 vs control (Friedman with Dunn's multiple comparison test).

Kinetic analysis of [³⁵S]l–Hcy uptake

Based on these data suggesting the primary involvement of system L amino acid transporter in [³⁵S]l–Hcy transport across MVM, we examined whether kinetic modelling of the individual l–Hcy inhibition curves (Fig. 8) would reveal discrimination of two kinetic components reflecting mediation by both LAT light chains. In all four MVM vesicle isolates examined, and for both assay conditions, the best fit (F test, P < 0.01) was to a two saturable transporter model. Derived estimates of K_m and V_max for the two transport mechanisms were not significantly different between Na⁺-containing and Na⁺-free conditions and have therefore been pooled. The kinetic characteristics for each transport system were K_m1= 72 ± 13 μm, V_max1= 750 ± 161 pmol (mg protein)⁻¹ (30 s)⁻¹ and K_m2= 9748 ± 894 μm, V_max2= 60341 ± 4910 pmol (mg protein)⁻¹ (30 s)⁻¹ (n= 8).

Figure 8. Kinetic modelling of [³⁵S]l–Hcy uptake into MVM vesicles.

Effect of unlabelled l–Hcy (5 μm–10 mm) on uptake of 10 μm[³⁵S]l–Hcy into MVM vesicles at initial rate (30 s) in the presence (▪) or absence (▴; with K⁺ replacement) of an inwardly directed Na⁺ gradient. Kinetic modelling using non-linear regression was performed for each individual experiment under both conditions, and this revealed that a two-site transporter model (dashed lines) was preferred over a one-site model in all cases. There was a significant improvement in the correlation coefficient for a two-site model (r² > 0.98 for all) compared to a one-site model (r² < 0.96 for all). Data are mean ±s.e.m. (n= 4).

Discussion

This study characterises three potential influx transport mechanisms for l–Hcy across MVM of human term placenta, the first plasma membrane barrier to maternofetal nutrient transfer. The three transport systems we elected to study here, systems A, L and y⁺L, were considered prime candidates based on (i) the extensive characterisation of these transport mechanisms in MVM of human placenta (Jansson, 2001; Cleal & Lewis, 2008) and (ii) the ability of all three transport mechanisms to utilise l–Met, the methylated derivative of l–Hcy, as substrate (Devés & Boyd, 1998; Kanai et al. 1998; Meier et al. 2002; Mackenzie & Erickson, 2004; Cleal & Lewis, 2008). Our evidence supports the involvement of each of these neutral amino acid transport mechanisms in mediating l–Hcy transport across MVM of human placenta. Considering the magnitude of maximal inhibition of [³⁵S]l–Hcy uptake into MVM vesicles by substrates of systems A (∼21%), L (∼69%) and y⁺L (∼19%) observed here, it would appear that involvement by other transport systems, such as system X⁻_AG which is localised to the MVM and imports l–Hcy in other cell types (Büdy et al. 2006; Jiang et al. 2007), is minimal. Involvement of system ASC (Büdy et al. 2006; Jiang et al. 2007) can be discounted as its activity is confined to basal plasma membrane in human placenta (Cleal & Lewis, 2008).

We infer that system A transports l–Hcy across MVM based on our observations of (i) concentration-dependent inhibition by l–Hcy of [¹⁴C]MeAIB uptake, a paradigm substrate for system A, (ii) the comparable inhibition of [¹⁴C]MeAIB uptake by Hcy to that of model system A substrates at high substrate concentration, and (iii) the ability of MeAIB to inhibit [³⁵S]l–Hcy uptake. These data demonstrating a system A-mediated component of l–Hcy transport are consistent with the observations of others in some (Naggar et al. 2003; Hultberg, 2004; Büdy et al. 2006) but not all (Ewadh et al. 1990; Jiang et al. 2007) cell types.

SNAT1, 2 and 4 isoforms of system A are all localised to MVM (Desforges et al. 2009), allowing for the possibility that more than one SNAT isoform mediates Hcy transport across MVM. Based on the comparability of IC₅₀ values observed for l–Hcy and l–Met inhibition of [¹⁴C]MeAIB uptake, and the relatively high preference of SNAT1 and SNAT2 for l–Met as substrate (Hatanaka et al. 2000; Wang et al. 2000), we speculate that l–Hcy transport across MVM may involve either, or both, of these two SNAT isoforms. Our observation that the racemic mixture dl–Hcy was equally effective as l–Hcy in inhibiting [¹⁴C]MeAIB uptake into MVM is consistent with the demonstration that both enantiomers of Hcy can be transported (Jiang et al. 2007) and that d–enantiomers of amino acids can be taken up into cells by various transport mechanisms, including systems A and L, although stereoselectivity usually favours the l–enantiomer (Shikano et al. 2007).

Our evidence of (i) concentration-dependent inhibition by l–Hcy and dl–Hcy of Na⁺-independent [³⁵S]l–Met uptake, a preferred substrate for both LAT1 and LAT2 (Kanai et al. 1998; Pineda et al. 1999; Meier et al. 2002), (ii) Hcy inhibition of [³⁵S]l–Met uptake with similar capacity to that of model system L substrates, and (iii) marked inhibition of [³⁵S]l–Hcy uptake by BCH, collectively supports the concept that system L transports Hcy across MVM. Further, our data suggest that system L predominates as the transport mechanism for Hcy across MVM as shown by BCH inhibiting ∼69%[³⁵S]l–Hcy uptake. This agrees well with previous observations in a variety of cell types that system L is a major contributor to cellular Hcy uptake (Ewadh et al. 1990; Limpach et al. 2000; Naggar et al. 2003; Büdy et al. 2006; Jiang et al. 2007).

The notion that system L activity predominates would also be consistent with our kinetic analyses of [³⁵S]l–Hcy uptake into MVM vesicles, which discriminated both a high-, and low-affinity, Na⁺-independent, transport component with a K_m value of ∼72 μm and 9.7 mm, respectively. These K_m values fall within the range of values (12 μm to >10 mm) previously cited for substrate affinities for LAT1 and LAT2, which is governed by LAT isoform substrate preference; this can vary by orders of magnitude between LAT1 and LAT2 for the same amino acid (Rossier et al. 1999). Whilst the CD98/LAT1 variant of system L transports neutral amino acids with large side chains and generally exhibits higher substrate affinity (Rossier et al. 1999), CD98/LAT2 has broader substrate selectivity and accepts smaller neutral amino acids (Pineda et al. 1999). We therefore speculate that the two kinetically distinguishable components of [³⁵S]l–Hcy uptake might represent transport by CD98/LAT1 and CD98/LAT2 heterodimers, respectively. The ability of dl–Hcy to inhibit [³⁵S]l–Met uptake into MVM vesicles with the same efficacy as l–Hcy, and the profoundly greater ability of d-amino acids to inhibit LAT1 over LAT2 (Rajan et al. 2000) strongly implicates the involvement of LAT1. The notion that LAT2 is involved in Hcy transport across MVM would have been greatly bolstered by the positive identification of LAT2 in MVM, but such investigations have been hampered by poor antibody specificity (Lewis et al. 2007).

Our demonstration of a Hcy-sensitive Arg transporter that exhibited a significant reduction in its affinity for Hcy when Na⁺ was replaced by K⁺, and which showed a similar inhibitory capacity to other neutral amino acids at high concentration, led to the designation of this transport mechanism as y⁺L (Eleno et al. 1994; Devés et al. 1998; Devés & Boyd, 1998). The magnitude of the K_i shift for l–Met found here in MVM (5.8-fold), compares favourably to that of other studies with l–Gln (5-fold; Eleno et al. 1994) providing confidence in this attribution. System y⁺L activity demonstrated substrate stereoselectivity, preferring l–Hcy over d-Hcy, as judged by the smaller magnitude of the observed K_i shift for dl–Hcy and the lower inhibitory capacity of dl–Hcy compared to other l–amino acids substrates of system y⁺L over a broad concentration range, entirely consistent with system y⁺L activity (Devés et al. 1998; Devés & Boyd, 1998). Whilst l–Hcy was found to be an effective inhibitor of system y⁺L activity, measured here as neutral amino acid-sensitive [³H]l–Arg uptake into MVM vesicles, the reciprocal experiment revealed that 20 mm l–Arg was only able to inhibit ∼19% of [³⁵S]l–Hcy transport. We attribute this to the overlying action of system y⁺ activity which makes a major contribution to l–Arg transport across MVM, as shown here by the neutral amino acid-resistant component (Fig. 4) and characterised previously (Ayuk et al. 2000). Our demonstration that system y⁺L mediates Hcy transport across MVM agrees with the observation of others in platelets demonstrating Hcy inhibition of a Leu-sensitive, Arg-transporting component, taken to be system y⁺L activity (Brunini et al. 2003; Leoncini et al. 2003). The co-expression of y⁺LAT1 and y⁺LAT2 mRNA in human placenta (Dye et al. 2004) allows for the possibility that either of these light chain variants form a heterodimeric complex with CD98 to mediate system y⁺L activity and transport l–Hcy. Both y⁺LAT1 and y⁺LAT2 exhibit a strong preference for l–Met, l–Gln and l–Leu (Pfeiffer et al. 1999; Bröer et al. 2000), precluding their functional delineation using these amino acid substrates.

The collective evidence presented here reveals that, as for other amino acids, l–Hcy is transported by multiple transport mechanisms across the MVM of human placenta. It is of particular interest that systems L, A and y⁺L all support l–Hcy transport in MVM, as the latter two transport mechanisms may work in parallel to provide system L with amino acid substrates for exchange (Verrey, 2003). However, of these three mechanisms, system L-mediated l–Hcy influx is likely to be of prime importance at maternal physiological concentrations of l–Hcy. This notion is proposed based on our evidence that this transport mechanism makes the major contribution to MVM uptake of 10 μm l–Hcy with the kinetic discrimination a high-affinity (∼72 μm) transport component ascribed to system L. The predominance of a particular transport mechanism for l–Hcy appears to be governed by cell type (Naggar et al. 2003; Jiang et al. 2007), as does the apparent K_m for the same transport mechanism. Hence, the K_m for system l–mediated Hcy transport in human aortic endothelial cells is 19 μm (Büdy et al. 2006) whilst that in human umbilical vein endothelial cell (HUVEC) is 160 μm (Ewadh et al. 1990). It is noteworthy that the K_m value for the high-affinity component observed here falls within this range.

There are important pathophysiological implications of these data. The scenario that Hcy is transported by, or accumulates within, the syncytiotrophoblast raises the possibility that Hcy could impact on syncytiotrophoblast function in several ways: reduction in the provision of amino acids to the developing fetus by competing with other substrates of systems L, A and y⁺L; perturbation of placental metabolism; altered vascular function; and induction of apoptosis. Indeed, the concept that Hcy is transported across the MVM into the syncytiotrophoblast is supported by the observation that Hcy inhibits [¹⁴C]MeAIB uptake into placental villous fragments (Tsitsiou et al. 2008).

System L activity in MVM would be particularly susceptible to perturbations in maternal Hcy status based on the evidence presented here. As system L provides essential amino acids for both fetal development and placental metabolism, with placental supply of essential amino acids only just matching fetal demand for protein synthesis (Chien et al. 1993), such an interaction is likely to impact on fetal growth. This concept is compatible with the pre-disposition to fetal growth restriction and a lower birth weight in mothers with HHcy (Vollset et al. 2000). Furthermore, as systems L and A activities in MVM are reduced in association with fetal growth restriction (Jansson & Powell, 2007), maternal HHcy would serve to further exacerbate the situation by restricting amino acid supply.

In summary, this study has shown that systems L, A and y⁺L each support Hcy as a substrate to mediate transport of this amino acid across MVM of human syncytiotrophoblast. System L activity predominated with kinetic discrimination of a high-affinity and low-affinity transport component that was independent of an inwardly directed Na⁺ gradient across MVM. The involvement of these three amino acid transport mechanisms suggests that in pregnancies compromised by maternal HHcy, Hcy competes with endogenous amino acid substrates for transporter activity, thereby diminishing placental amino acid influx and provision of amino acids to the developing fetus.

Glossary

Abbreviations

Ala

l–alanine

Arg

l–arginine

BCH

2-aminobicyclo[2.2.1]heptane-2-carboxylic acid

EVB

extravesicular buffer

Gln

l–glutamine

Hcy

homocysteine

HHcy

hyperhomocysteinaemia

Leu

l–leucine

Lys

l–lysine

MeAIB

α-(methylamino)isobutyric acid

Met

l–methionine

MVM

microvillous plasma membrane

Ser

l–serine

SNAT

sodium-coupled neutral amino acid transporter

tHcy

total Hcy

Author contributions

E.T., C.P.S., S.W.D'S., D.W.J., J.D.G. designed research. E.T., O.C., J.D.G. performed research. E.T., C.P.S., J.D.G. analysed data. E.T., C.P.S., S.W.D'S., D.W.J., J.G. wrote paper.