Regulation of High-Affinity Nitrate Transporter Genes and High-Affinity Nitrate Influx by Nitrogen Pools in Roots of Barley

Abstract

To investigate the regulation of HvNRT2, genes that encode high-affinity NO₃⁻ transporters in barley (Hordeum vulgare) roots, seedlings were treated with 10 mm NO₃⁻ in the presence or absence of amino acids (aspartate, asparagine, glutamate [Glu], and glutamine [Gln]), NH₄⁺, and/or inhibitors of N assimilation. Although all amino acids decreased high-affinity ¹³NO₃⁻ influx and HvNRT2 transcript abundance, there was substantial interconversion of administered amino acids, making it impossible to determine which amino acid(s) were responsible for the observed effects. To clarify the role of individual amino acids, plants were separately treated with tungstate, methionine sulfoximine, or azaserine (inhibitors of nitrate reductase, Gln synthetase, and Glu synthase, respectively). Tungstate increased the HvNRT2 transcript by 20% to 30% and decreased NO₃⁻ influx by 50%, indicating that NO₃⁻ itself does not regulate transcript abundance, but may exert post-transcriptional effects. Experiments with methionine sulfoximine suggested that NH₄⁺ may down-regulate HvNRT2 gene expression and high-affinity NO₃⁻ influx by effects operating at the transcriptional and post-transcriptional levels. Azaserine decreased HvNRT2 transcript levels and NO₃⁻ influx by 97% and 95%, respectively, while decreasing Glu and increasing Gln levels. This suggests that Gln (and not Glu) is responsible for down-regulating HvNRT2 expression, although it does not preclude a contributory effect of other amino acids.

The uptake of NO₃⁻ in terrestrial plants is mediated by at least three transport systems that coexist in the plasma membranes of root cells (for review, see Glass and Siddiqi, 1995; Crawford and Glass, 1998). These fall into two classes, referred to as low- and high-affinity transport systems (LATS and HATS, respectively). The HATS are further subdivided into constitutive (CHATS) and inducible (IHATS) systems (Siddiqi et al., 1990; Aslam et al., 1993). The LATS are involved in NO₃⁻ uptake at high external concentrations of NO₃⁻ (>0.2 mm), while the CHATS and IHATS are saturated at low external NO₃⁻concentrations (approximately 100 μm). In barley (Hordeum vulgare) roots, LATS activity is expressed without prior exposure to NO₃⁻ (Siddiqi et al., 1990), and this transport system appears to be subject to down-regulation by accumulated N (Clement et al., 1978; Siddiqi et al., 1990). A gene considered to encode the LATS (AtNRT1, originally CHL1) was the first higher plant NO₃⁻ transporter gene to be cloned from Arabidopsis (Tsay et al., 1993). However, in contrast to the apparent constitutive nature of this transport in barley roots, the AtNRT1 gene transcript is undetectable prior to exposure to NO₃⁻ in Arabidopsis (Tsay et al., 1993; Huang et al., 1996). This apparent anomaly is unexplained and warrants exploration. Recently, it has been proposed that the AtNRT1 protein may participate in both high- and low-affinity NO₃⁻ uptake (Wang et al., 1998; Liu et al., 1999).

In higher plants, genes considered to encode IHATS have been cloned from barley (Trueman et al., 1996; Vidmar et al., 2000), Arabidopsis (Filleur and Daniel-Vedele, 1999; Zhuo et al., 1999), Nicotiana plumbaginifolia (Quesada et al., 1997), and soybean (Amarashinghe et al., 1998). In barley, putative IHATS are encoded by a multigene family of seven to 10 members (Trueman et al., 1996). To date, four members of this family, originally named the BCH family and renamed HvNRT2, have been isolated from barley (Trueman et al., 1996; Vidmar et al., 2000). The HvNRT2 genes encode proteins composed of 507 to 509 amino acids, with molecular masses of 54 to 55 kD, including 12 hydrophobic (transmembrane) regions that belong to the major facilitator superfamily (as do the other plant IHATS). It has been shown that the mRNA levels of these IHATS genes increase rapidly following the provision of NO₃⁻ (a process referred to as induction) to NO₃⁻-deprived plants (Trueman et al., 1996; Quesada et al., 1997; Amarashinghe et al., 1998; Filleur and Daniel-Vedele, 1999; Zhuo et al., 1999). This increase in transcript levels is correlated at the physiological level with increased NO₃⁻ influx when NO₃⁻ is first supplied (Siddiqi et al., 1989; Glass et al., 1990; Hole et al., 1990; Zhuo et al., 1999).

In barley, all four of the HvNRT2 transcripts investigated are coordinately up-regulated following the provision of NO₃⁻ (Vidmar et al., 2000). Under quasi-steady-state conditions of NO₃⁻ supply, the highest levels of IHATS mRNA and ¹³NO₃⁻ influx were obtained when the external NO₃⁻ concentration was maintained at 50 μm (Vidmar et al., 2000). In N. plumbaginifolia, genes involved in N acquisition and assimilation, NpNRT2.1, NIA (nitrate reductase [NR]), and NII (NR), were coordinately expressed under conditions of NO₃⁻ induction and N repression (Krapp et al., 1998). Furthermore, NR mutants of N. plumbaginifolia and Arabidopsis showed elevated levels of NRT2.1 transcript (Krapp et al., 1998; Filleur and Daniel-Vedele, 1999; Lejay et al., 1999), which is consistent with the proposal that NRT2.1 transcript abundance is regulated by feedback from reduced forms of N rather than from NO₃⁻ itself. Nevertheless, at the physiological level, the down-regulation of IHATS to a lower steady-state level following peak induction has been argued to result from effects of accumulated NO₃⁻ and/or a product(s) of its assimilation (Siddiqi et al., 1989; King et al., 1993). This conclusion was based on correlations between root [NO₃⁻] and ¹³NO₃⁻ influx (Siddiqi et al., 1989), as well as data gained from NR double mutants of barley (King et al., 1993). Additional physiological evidence for a role of tissue NO₃⁻ in down-regulating NO₃⁻ influx came from studies by Ingemarsson et al. (1987) based upon the use of tungstate (an inhibitor of NR) and from Doddema and Otten (1979), based on kinetic studies of NO₃⁻ uptake in Arabidopsis. Evidence consistent with the down-regulation of NO₃⁻ influx by NH₄⁺ (Aslam et al., 1996) and/or by amino acids (Doddema and Otten, 1979; Breteler and Siegerist, 1984; Muller and Touraine, 1992; Muller et al., 1995) has also been advanced.

The effects of NH₄⁺ on NO₃⁻ uptake are more complex, due to the possibility of affecting NO₃⁻ uptake at a number of levels (transcript abundance, protein level, or direct effects of NH₄⁺ on the NO₃⁻ transporter). This has resulted in a lack of consensus concerning the mechanism(s) of the NH₄⁺ effects on NO₃⁻ fluxes. Using NO₂⁻ as a tracer of NO₃⁻, Aslam et al. (1994) suggested that NH₄⁺ increased NO₃⁻ efflux rather than diminished influx, while the use of ¹³NO₃⁻ demonstrated that influx was strongly reduced (Glass et al., 1985; Lee and Drew, 1989; Kronzucker et al., 1999). A recent paper by Kronzucker et al. (1999) established that in barley roots, the provision of NH₄⁺ in the external medium simultaneously decreased NO₃⁻ influx and increased efflux, the absolute effect upon influx being more significant. Moreover, this effect occurred within minutes of supplying NH₄⁺, suggesting that NH₄⁺ itself was acting directly upon the IHATS.

This direct effect of NH₄⁺ on NO₃⁻ influx does not preclude long-term effects. For example, in longer term experiments, Breteler and Siegerist (1984) showed that Met sulfoximine (MSO), an inhibitor of Gln synthetase, relieved the inhibitory effect of NH₄⁺ on NO₃⁻ uptake in dwarf bean. These authors concluded that the NH₄⁺ effect arose from products of the assimilation of NH₄⁺ rather than from NH₄⁺ itself. However, King et al. (1993) observed no relief of NH₄⁺ inhibition of NO₃⁻ influx by MSO in barley roots. Likewise, de la Haba et al. (1990) suggested that NH₄⁺, and not its assimilation products, was responsible for inhibiting NO₃⁻ uptake into sunflower roots. Clearly, part of the confusion in the literature has resulted from the aforementioned multiple levels at which NH₄⁺ is capable of inhibiting NO₃⁻ uptake. There is every reason to expect that NH₄⁺ might have direct effects on the transport system as well as effects at the level of transcription via products of NH₄⁺ assimilation.

Feeding plants with amino acids to mimic putative shoot signals and to investigate the effects of down-stream metabolites of NO₃⁻ that might control NO₃⁻ uptake by roots has frequently been undertaken (Lee et al., 1992; Imsande and Touraine, 1994). Amino acids, provided to roots exogenously in the nutrient solution decreased NO₃⁻ uptake in Arabidopsis (Doddema and Otten, 1979), common bean (Breteler and Arnozis, 1985), soybean (Muller and Touraine, 1992), and wheat (Rodgers and Barneix, 1993). This down-regulation of NO₃⁻ uptake was usually preceded by a lag period of 3 h or more, indicating that the effects of amino acids on NO₃⁻ uptake were not direct or allosteric (Muller and Touraine, 1992). However, it is necessary to interpret the results of such experiments with caution, since the effects of particular amino acids may be influenced by the extent of their uptake, the extent of their biochemical interconversion and assimilation, and their effects upon the expression of the NO₃⁻ transport system. In some cases (e.g. Muller and Touraine, 1992), amino acids were applied to cotyledons or to skin flaps to simulate the “normal” pathway of amino acid delivery from shoot to root. The results of such experiments confirm those derived from exogenous application of amino acids, but it is still necessary to consider the extent of amino acid transport to the roots and the potential for interconversion of applied amino acids when attempting to identify critical regulatory compounds.

In this report we investigate the regulation of NO₃⁻ influx, HvNRT2 transcript abundance, and changes in NO₃⁻, NH₄⁺, and amino acid concentrations in roots during the down-regulation of NO₃⁻ influx. For this purpose, barley seedlings were exposed to NO₃⁻ with or without NH₄⁺, amino acids (Asn, Asp, Gln, and Glu), and inhibitors of key enzymes of the N-assimilation pathway.

RESULTS

Effects of Treatment with 10 mm NO₃⁻ on ¹³NO₃⁻ Influx, Tissue N Concentrations, and HvNRT2 Transcript Accumulation in NO₃⁻-Deprived Plants

When seedlings that had previously been grown on N-free medium were fed 10 mm NO₃⁻, ¹³NO₃⁻ influx (measured at 50 μm) displayed a typical time course response, increasing from 0.42 to 2.9 μmol g⁻¹ fresh weight h⁻¹, equivalent to a 7-fold increase of ¹³NO₃⁻ influx, within the first 12 h. This confirms earlier reports to this effect (Siddiqi et al., 1989; Vidmar et al., 2000). Thereafter, influx decreased to 2.4 μmol g⁻¹ fresh weight h⁻¹ by the end of the experiment (at 48 h). Parallel northern-blot analysis showed that HvNRT2 transcript was initially undetectable in root tissue, but increased dramatically within the first 6 h after exposure to NO₃⁻, and then steadily decreased to undetectable levels by 24 h. Analysis of inorganic N (Fig. 1A) showed that the root NO₃⁻ concentration increased from 5.1 ± 2.7 μmol g⁻¹ fresh weight at time 0 to 46.2 ± 11.2 μmol g⁻¹ fresh weight by 6 h of NO₃⁻ provision. By 48 h, root NO₃⁻ had increased to 89.2 ± 8.5 μmol g⁻¹ fresh weight. In contrast, NH₄⁺ levels remained basically unchanged, varying between 3.5 and 4.5 μmol g⁻¹ fresh weight (Fig. 1A) during the 48 h of exposure to 10 mm NO₃⁻. Figure 1B shows that during the course of the experiment the concentrations of Gln, Glu, Asp, and Asn increased to maximum levels that were 10-, 9-, 13-, and 4-fold higher, respectively, than initial values. Gln and Glu concentrations increased rapidly during the first 12 h after the onset of NO₃⁻ supply, then decreased to slightly lower concentrations by 48 h.

Figure 1.

Plant inorganic and organic N concentrations as a function of feeding 10 mm NO₃⁻ to N-starved barley plants. Seven-day-old-seedlings were grown in one-tenth-strength N-free modified Johnson's solution, and then supplied with 10 mm NO₃⁻ for 0, 2, 6, 12, 24, and 48 h. A, Root concentrations of inorganic N: NO₃⁻ (●) and NH₄⁺ (▪). B, Root concentrations of amino acids as percentage of N-starved control Asn (▪), Asp (●), Gln (▴), and Glu (▾).

Effects of Exogenously Applied Amino Acids on ¹³NO₃⁻ Influx and HvNRT2 Transcript Accumulation

¹³NO₃⁻ influx was reduced by all amino acids tested (Fig. 2A) in plants pretreated with 1 mm Gln, Glu, Asn, or Asp for 6 h during induction of NO₃⁻ influx with 10 mm NO₃⁻. The strongest effect was due to Asp (89% inhibition), followed by Glu (79% inhibition), Asn (45% inhibition), and Gln (29% inhibition). Northern-blot analysis of RNA isolated from the same amino acid-treated roots showed a similar pattern for the abundance of HvNRT2 transcript (Fig. 2B), except that Glu was more inhibitory than Asp. The exogenous application of each of the amino acids (Glu, Gln, Asp, and Asn) generally increased concentrations of all of the four amino acids in root tissue (Table I). For example, the application of Asp increased root Asp 1.6-fold, while at the same time increasing Asn 2-fold, Glu 1.8-fold, and Gln 1.6- fold. By reference to the NO₃⁻ treatment, root NO₃⁻ concentrations were either increased by 40% by the Gln treatment, unaffected by the Glu treatment, or reduced by 43% and 32% respectively, after treatment with Asp and Asn (Table I). The changes of HvNRT2 transcript levels were most strongly correlated with increases of Glu (r² = 0.92) and Gln (r² = 0.68) concentrations, while Asp and Asn concentrations were poorly correlated (r² values of 0.4 and 0.22, respectively).

Figure 2.

Effects of exogenous treatment with amino acids on ¹³NO₃⁻ influx and HvNRT2 transcript accumulation in roots. Seven-day-old-seedlings were grown in one-tenth-strength N-free modified Johnson's solution. A, ¹³NO₃⁻ influx was measured at 50 μm. Each of the treatments consisted of four replicates. B, Northern-blot analysis of HvNRT2 transcript from RNA isolated from barley roots. Treatments: lane 1, N starved; lanes 2 through 6, treated with 10 mm NO₃⁻ alone (lane 2) or with 10 mm NO₃⁻ plus amino acids (lanes 3–6) for 6 h. Lane 3, 1 mm Asn; lane 4, 1 mm Asp; lane 5, 1 mm Gln; and lane 6, 1 mm Glu. Quantification of transcript levels was by phosphor imager average of two experiments (standardized by 25S transcript). Transcript abundance was calculated in proportion to the 6 h-NO₃⁻ treatment (as 1 relative unit).

Table I.

Effects of exogenously supplied amino acids on inorganic and organic N levels in plant roots

Treatment	Amino Acid				Inorganic N
	Asn	Asp	Gln	Glu	NO3−
	nmol g−1 fresh wt				μmol g−1 fresh wt
−N	0.82 ± 0.41	0.24 ± 0.02	0.71 ± 0.06	2.11 ± 0.81	5.1 ± 2.7
6 h-NO3−	1.32 ± 0.25	0.59 ± 0.14	3.41 ± 0.21	5.29 ± 1.07	46.2 ± 11.2
6 h-NO3− + Asn	4.02 ± 2.01	1.32 ± 0.22	4.62 ± 1.01	8.50 ± 2.13	26.2 ± 12.7
6 h-NO3− + Asp	2.58 ± 1.28	0.98 ± 0.16	5.52 ± 0.42	9.42 ± 2.61	31.5 ± 2.1
6 h-NO3− + Gln	2.51 ± 1.02	0.62 ± 0.21	5.02 ± 0.58	8.51 ± 2.61	64.8 ± 9.4
6 h-NO3− + Glu	1.46 ± 0.68	0.92 ± 0.11	6.82 ± 0.51	10.21 ± 1.94	47.5 ± 4.1

Values shown are the means of four independent replicates ± sd of the mean.

Effects of Inhibitors of N Assimilation on ¹³NO₃⁻ Influx and HvNRT2 Transcript Accumulation

The effects of 0.5 mm tungstate, 1 mm MSO, and 0.25 mm aza-Ser (AZA), inhibitors of the enzymes NR, Gln synthetase, and Glu synthase, respectively, were evaluated in separate experiments by supplying these inhibitors together with 10 mm NO₃⁻ for 6 h to N-starved plants. By examining the effects of these inhibitors on transcript abundance and NO₃⁻ influx, the role of these assimilates in regulating IHATS influx may be assessed. All three inhibitors decreased ¹³NO₃⁻ influx compared with the control treatment, which was the 6 h of exposure to 10 mm NO₃⁻ (Fig. 3A). The AZA effect was the most dramatic, decreasing influx by 97%. Tungstate and MSO reduced NO₃⁻ influx by 44%. In contrast to its effect upon influx, tungstate increased HvNRT2 transcript by 20% to 30% (Figs. 3B [lane 4] and 4 [lane 3]). This treatment failed to significantly affect NO₃⁻, NH₄⁺, Asp, or Glu concentrations, but decreased both Asn and Gln concentrations (Table II).

Figure 3.

The effect of N assimilation inhibitors on ¹³NO₃⁻ influx and HvNRT2 transcript accumulation. Seven-day-old barley seedlings were grown in one-tenth-strength N-free modified Johnson's solution. A, ¹³NO₃⁻ influx measured at 50 μm. Each treatment consists of four replicates. B, Northern-blot analysis of HvNRT2 transcript accumulation in RNA isolated from barley roots. Treatments: lane 1, N starved; lanes 2 through 6, treated with 10 mm NO₃⁻ for 6 h; inhibitors were added to treatments in lanes 3 through 6. Lane 3, 1 mm MSO; lane 4, 0.5 mm tungstate; lane 5, 0.25 mm AZA. Quantification of transcript levels was by phosphor imager average of two experiments (standardized by 25S transcript). Transcript abundance was calculated in proportion to the 6 h-NO₃⁻ treatment (as 1 relative unit).

Table II.

Effects of inhibitors of N assimilation on inorganic and organic N levels in plant roots

Treatment	Amino Acid				Inorganic N
	Asn	Asp	Gln	Glu	NO3−	NH4+
	nmol g−1 fresh wt				μmol g−1 fresh wt
−N	0.82 ± 0.41	0.24 ± 0.02	0.71 ± 0.06	2.11 ± 0.81	5.1 ± 2.7	4.2 ± 0.2
6 h-NO3−	1.32 ± 0.25	0.59 ± 0.14	3.42 ± 0.21	5.29 ± 1.07	46.2 ± 11.2	4.5 ± 0.34
6 h-NO3− + Tungstate	0.61 ± 0.11	0.57 ± 0.21	0.65 ± 0.28	4.80 ± 0.72	42.5 ± 15.6	3.6 ± 0.47
6 h-NO3− + MSO	0.71 ± 0.41	0.54 ± 0.12	2.12 ± 0.81	2.90 ± 0.62	34.2 ± 7.4	18.5 ± 1.4
6 h-NO3− + AZA	1.78 ± 0.87	0.27 ± 0.12	4.88 ± 1.47	0.54 ± 0.24	8.4 ± 2.1	10.4 ± 2.1

Values shown are the means of four independent replicates ± sd of the mean.

When exogenous Glu was added to the tungstate treatment solution, the HvNRT2 transcript level was reduced by 59% compared with the tungstate-plus-NO₃⁻ treatment (Fig. 4). MSO had only a small effect upon HvNRT2 transcript abundance when applied in the presence of NO₃⁻ as the sole source of N. Thus, transcript abundance was reduced by 15%, 7%, or increased by 22%, respectively, in separate experiments (Table III; Fig. 3B, lane 3, and Fig. 4, lane 4). When 1 mm Glu was added to the MSO treatment, a 62% decline in HvNRT2 transcript abundance was observed (Fig. 4). AZA (Fig. 3B, lane 5) decreased transcript abundance by 95%. This was the largest observed effect on transcript abundance of any of the inhibitors used. As would be expected of inhibitors of Gln synthetase and Glu synthase activities, the MSO and AZA treatments increased root NH₄⁺ levels by 4.5- and 2-fold, respectively (Table II), compared with the control treatment (plants treated with 10 mm NO₃⁻ without inhibitor). MSO treatment also decreased Gln, Asn, and Glu concentrations, but had little effect on the Asp concentration (Table II). Likewise, blocking Glu synthase with AZA increased Gln and Asn concentrations 1.4- and 1.3-fold, respectively, and reduced Asp and Glu to 0.45 and 0.1, respectively, of their control values (Table II). AZA and MSO treatments decreased concentrations of NO₃⁻ in root tissue by 80% and 25%, respectively (Table II).

Figure 4.

Effects of combinations of N-assimilation inhibitors, tungstate, and MSO in the presence of Glu and NO₃⁻ on HvNRT2 transcript accumulation in barley roots. Lane 1, Northern-blot analysis of RNA isolated from roots of N starved plants; lanes 2 through 6 contain RNA isolated from roots treated with 10 mm NO₃⁻. Transcript abundance was calculated in proportion to the 6 h-NO₃⁻ treatment (as 1 relative unit). Lane 2, 10 mm NO₃⁻; lane 3, 0.5 mm tungstate plus 10 mm NO₃⁻; lane 4, 1 mm MSO plus 10 mm NO₃⁻; lane 5, 0.5 mm tungstate plus 10 mm NO₃⁻ and 1 mm Glu; lane 6, 1 mm MSO plus 10 mm NO₃⁻ and 1 mm Glu. Quantification of transcript levels was by phosphor imager average of two experiments (standardized by 25S transcript).

Table III.

Effects of NH₄⁺ and MSO on HvNRT2 transcript abundance, NH₄⁺ concentrations, and Gln concentrations in plant roots

Treatment	Relative Transcript Level	NO3− Influx	[NH4+]	[Gln]
		μmol g−1 fresh wt h−1	μmol g−1 fresh wt	nmol g−1 fresh wt
−N	0.12	0.48 ± 0.12	4.2 ± 0.21	0.71 ± 0.06
6h NO3−	1.00	3.20 ± 0.25	4.5 ± 0.34	3.41 ± 0.21
6h NO3− + MSO	0.85	1.74 ± 0.21	18.9 ± 1.4	2.12 ± 0.81
6h NO3− + NH4+	0.14	0.97 ± 0.13	47.2 ± 3.1	6.70 ± 0.54
6h NO3− + NH4+ + MSO	0.09	0.84 ± 0.08	55.2 ± 2.8	4.54 ± 0.52

Values shown are the means of four independent replicates ± sd of the mean. Plants were grown on one-tenth modified (−N) Johnson's solution, then treated for 6 h with the treatments shown below, using 10 mm NO₃⁻ ± 1 mm MSO, or 10 mm NO₃⁻ + 10 mm NH₄⁺ ± 1 mm MSO. Transcript abundance was calculated using the ratio of HvNRT2/25s hybridizing signal, quantified on a phosphor imager.

Effects of Exogenous NH₄⁺ on ¹³NO₃⁻ Influx and HvNRT2 Transcript Accumulation

It was reported previously (Vidmar et al., 2000) that HvNRT2 transcript abundance was not reduced during the first 2 h of exposure to 10 mm NH₄⁺ supplied together with 10 mm NO₃⁻ to roots of barley plants. However, treatments of longer duration, e.g. 4 or 6 h, resulted in a dramatic decrease in the abundance of HvNRT2. This decrease was accompanied by a marked decrease of ¹³NO₃⁻ influx, and it was concluded that either NH₄⁺ itself or a product(s) of its assimilation was responsible for down-regulating HvNRT2 mRNA levels. To explore this question further, we determined ¹³NO₃⁻ influx, transcript abundance of HvNRT2, and NH₄⁺ and amino acid concentrations in roots grown on NO₃⁻ alone or on NO₃⁻ together with NH₄⁺ with or without MSO. These treatments were designed to investigate the effects of elevated root NH₄⁺ concentrations and diminished Gln levels (plus MSO treatments) on ¹³NO₃⁻ influx and HvNRT2 transcript abundance (Table III). ¹³NO₃⁻ influx under control conditions (corresponding to 6 h of NO₃⁻) was reduced from 3.2 to 1.74 μmol g⁻¹ fresh weight h⁻¹, with a concomitant 4.2-fold increase of root NH₄⁺ concentration (4.5–18.9 μmol g⁻¹ fresh weight) as a result of exposure to NO₃⁻ plus MSO (Table III). Exposure to NO₃⁻ plus NH₄⁺ reduced ¹³NO₃⁻ influx to 0.97 μmol g⁻¹ fresh weight h⁻¹, and increased root NH₄⁺ concentration 10.5-fold (4.5–47.2 μmol g⁻¹ fresh weight), while NO₃⁻ plus both NH₄⁺ and MSO reduced ¹³NO₃⁻ influx to 0.84 μmol g⁻¹ fresh weight h⁻¹ and increased root NH₄⁺ concentration 12.3-fold (4.5–55.2 μmol g⁻¹ fresh weight). Additions of NO₃⁻ or NH₄⁺ increased root Gln concentration compared with the minus-N treatment (Table III), while MSO had the opposite effect. Thus, root Gln increased from 0.71 to 3.41 nmol g⁻¹ fresh weight after 6 h of NO₃⁻ treatment, while 6 h of NO₃⁻ plus MSO treatment reduced this value to 2.12 nmol g⁻¹ fresh weight. Exposure to NO₃⁻ plus NH₄⁺ for 6 h produced the highest value for root Gln concentration, 6.7 nmol g⁻¹ fresh weight, while the addition of MSO reduced this value to 4.54 nmol g⁻¹ fresh weight. Using the 6-h NO₃⁻ treatment as a control, these treatments reduced HvNRT2 transcript levels to 0.85 (NO₃⁻ plus MSO treatment), 0.14 (NO₃⁻ plus NH₄⁺ treatment), and 0.09 (NO₃⁻ plus NH₄⁺ plus MSO treatment) relative units.

DISCUSSION

NO₃⁻ uptake is subject to regulation by both positive (induction) and negative (down-regulation) effects. The latter appear to depend upon the N status of the whole plant (for reviews, see Glass and Siddiqi, 1995; Crawford and Glass, 1998). It has been suggested that the cycling of amino acids between shoots and roots serves to provide the necessary information regarding whole-plant N status, enabling roots to regulate N uptake accordingly (Cooper and Clarkson, 1989; Muller and Touraine, 1992). Furthermore, when plants simultaneously absorb different N forms (e.g. NO₃⁻, NH₄⁺, and amino acids), there is a need to integrate information from several putative feedback signals. While the N-cycling model is based on physiological data, there is presently little information available regarding the molecular mechanism(s) responsible for translating these proposed signals into processes that modulate NO₃⁻ transport at root plasma membranes.

At the physiological level, it has long been known that NO₃⁻ is the signal for the induction of the IHATS (Jackson et al., 1973; Goyal and Huffaker, 1986; Aslam et al., 1996), although NO₂⁻ can also serve as an inducer for this process (Siddiqi et al., 1992; Aslam et al., 1996). The cloning of genes that encode this transport system has made it possible to show that this is equally the case at the transcriptional level (Trueman et al., 1996; Quesada et al., 1997; Amarashinghe et al., 1998; Zhuo et al., 1999; Filleur and Daniel-Vedele, 1999). Thus, transcript levels for these genes increase dramatically in response to the provision of NO₃⁻ in the media, and the expression patterns have been correlated with NO₃⁻ influx, giving indirect evidence that NRT2 encode high-affinity transporters (Quesada et al., 1997; Lejay et al., 1999; Zhuo et al., 1999). Heterologous expression of the NRT2 genes in an Hansenula polymorpha NO₃⁻ transport mutant defective in the YNT1 gene (a member of the CRNA family) demonstrated that these genes encode functional NO₃⁻ transporters (Zhang et al., 1998).

By contrast, the identity of the N intermediate(s) responsible for the characteristic down-regulation of NO₃⁻ influx following peak induction has remained in controversy (see introduction). To investigate the factors responsible for this negative feedback control of IHATS and abundance of HvNRT2 transcript, we designed a series of experiments according to three criteria. First, plants were maintained on minus-N media. These plants contained low concentrations of all N derivatives, allowing changes of N pools to be readily detected following provision of NO₃⁻. Second, minus-N plants were exposed to high levels of NO₃⁻ (10 mm), so that NO₃⁻ might enter root cells via the LATS for NO₃⁻, even if inhibitor treatments reduced high-affinity transport. Third, the treatments with amino acids and inhibitors of N-assimilatory enzymes were short-term (6 h) to minimize the possibility of secondary effects of these treatments.

Several possible signals for the down-regulation of IHATS have been proposed. These include root NO₃⁻, root NH₄⁺, and/or amino acids (Siddiqi et al., 1989; Lee et al., 1992; Muller and Touraine, 1992; King et al., 1993). At the molecular level, Quesada et al. (1997) demonstrated that NpNRT2.1 transcript levels in NO₃⁻-grown plants decreased due to the supply of NH₄⁺ or Gln. Unfortunately, these preliminary results failed to identify whether NH₄⁺ itself or products of its assimilation were responsible for the observed effects. A study by Zhuo et al. (1999) using inhibitors of NO₃⁻ assimilation concluded that NH₄⁺ and Gln were both active in the down-regulation of the Arabidopsis AtNRT2.1 gene.

Time Course of the Down-Regulation of NO₃⁻ Influx

To investigate the down-regulation of NO₃⁻ influx in barley roots, we monitored five parameters: NO₃⁻ influx, HvNRT2 transcript levels, and NO₃⁻, NH₄⁺, and amino acid concentrations of root tissue. The typical time profile of ¹³NO₃⁻ influx was observed upon providing NO₃⁻ to plants previously starved of N (Siddiqi et al., 1989; Vidmar et al., 2000). This pattern correlated well with HvNRT2 transcript accumulation during the first 6 h of NO₃⁻ provision (Vidmar et al., 2000), but thereafter transcript levels decreased to levels that were undetectable, while influx remained relatively high. Two possibilities might account for this anomaly. First, turnover rates of the IHATS protein may be relatively slow compared with those of the corresponding mRNA, and therefore the abundance of IHATS mRNA need not correlate with influx capacity. To evaluate this it will be necessary to make use of antibodies to the NRT2 protein. Second, other transporter types (CHATS or LATS) may contribute to the observed fluxes; e.g. CHATS activity has been demonstrated to increase 2- to 3-fold following exposure to NO₃⁻ (Aslam et al., 1993; Kronzucker et al., 1995). During this prolonged exposure to NO₃⁻, NH₄⁺ levels increased only very slightly (10%), whereas root NO₃⁻ concentrations increased 17-fold (Fig. 1A), following the same pattern as reported by Siddiqi et al. (1989). Gln, Glu, Asn, and Asp levels increased from 4- to 13-fold (Fig. 1B). The low levels of accumulated NH₄⁺ suggest that, under normal conditions, NH₄⁺ itself does not participate in the down-regulation of HvNRT2 transcript abundance. However, the changes in root concentrations of both NO₃⁻ and amino acids are consistent with their involvement in this process.

Effects of NO₃⁻ on HvNRT2 Transcript Accumulation and NO₃⁻ Influx

To investigate the role of NO₃⁻ in the down-regulation of HvNRT2 transcript abundance, we made use of tungstate, a well-known inhibitor of the enzyme NR (Deng et al., 1989). The response of ¹³NO₃⁻ influx to tungstate treatment has been described in Lemna by Ingemarsson et al. (1987). Growth at high external [NO₃⁻] in the presence of tungstate increased cell NO₃⁻ concentration and reduced ¹³NO₃⁻ influx. By contrast, ¹³NO₃⁻ influx remained high when plants grown at low external [NO₃⁻] were treated with tungstate, and the authors proposed that this provided evidence for the regulation of NO₃⁻ influx by NO₃⁻ itself. Any potential for indirect inhibitory effects of tungstate was controlled for by the high values of ¹³NO₃⁻ influx reported in the low-NO₃⁻ plants. In our experiments using high-N plants, ¹³NO₃⁻ influx was also substantially reduced by exposure to tungstate during induction (Fig. 3A, treatment 4). This observation suggests that NO₃⁻ may regulate HvNRT2 expression in barley roots. However, compared with control plants, HvNRT2 transcript levels actually increased by 20% to 30% in response to the same treatment (Fig. 3B, lane 4, and Fig. 4, lane 3), a finding that makes it unlikely that NO₃⁻ inhibits influx at HvNRT2 transcript levels. The increased HvNRT2 transcript level associated with tungstate treatment is consistent with a similar effect of tungstate on levels of NR mRNA in tobacco (Deng et al., 1989), and with elevated levels of NRT2.1 transcripts observed in Arabidopsis and N. plumbaginifolia mutants lacking NR activity (Krapp et al., 1998; Filleur and Daniel-Vedele, 1999; Lejay et al., 1999).

These observations suggest that the down-regulation of both NR and NRT2 transcript abundance depends upon reduced forms of N rather than upon NO₃⁻. By adding Glu to the tungstate treatment (Fig. 4), the block of N assimilation was bypassed and HvNRT2 transcript levels were reduced. This confirms the importance of reduced N in regulating levels of HvNRT2 transcript, although the experiment failed to distinguish between Glu and products of Glu metabolism; the data presented in Table I demonstrate that the addition of Glu doubled the concentrations of Gln, Asn, and Glu. Table I reveals that the root NO₃⁻ concentration was virtually unaffected by tungstate treatment, although cytoplasmic NO₃⁻ may have increased under these conditions. Asp and Glu concentrations also remained constant following tungstate treatment, whereas Asn and Gln concentrations were reduced by 53% and 81%, respectively (Table II). The evidence therefore suggests that the elevated transcript abundance observed under these conditions may result from relief of negative feedback associated with lowered concentrations of Asn or Gln. If this is the case, the observed decrease of ¹³NO₃⁻ influx associated with tungstate treatment can only be accounted for by effects at a post-transcriptional level. This hypothesis is in agreement with physiological experiments using the NAR1/NAR7 NR double mutants of barley that have 1% to 5% of wild-type NR activity (Warner and Huffaker, 1989; King et al., 1993). These experiments revealed that mutant plants expressed the typical pattern of IHATS induction on supplying exogenous NO₃⁻, as well as the typical down-regulation of NO₃⁻ influx that follows peak induction. The hypothesis of post-translational effects of NO₃⁻ may also account for the findings of Ingemarsson et al. (1987), who also reported a reduction of ¹³NO₃⁻ influx in Lemna following exposure to tungstate.

Effects of NH₄⁺ on NO₃⁻ Uptake and HvNRT2 Transcript Abundance

NH₄⁺ has been demonstrated to inhibit NO₃⁻ influx into barley roots within minutes of NH₄⁺ application (Lee and Drew, 1989). Likewise, Kronzucker et al. (1999) demonstrated that ¹³NO₃⁻ influx decreased and ¹³NO₃⁻ efflux increased immediately following the addition of NH₄⁺ in barley roots. These observations suggest that inhibitory effects of NH₄⁺ result from direct effects of NH₄⁺ at the plasma membrane at least. In addition, Quesada et al. (1997), Krapp et al. (1998), and Filleur and Daniel-Vedele (1999) reported that transcript abundance of the NRT2 family of genes was reduced by NH₄⁺ treatment. However, the experimental design of the latter studies failed to distinguish between effects of NH₄⁺ itself and/or products of NH₄⁺ assimilation. This distinction can be achieved by using MSO (an inhibitor of the enzyme Gln synthetase), which typically increases root [NH₄⁺] and reduces [Gln]. For example, in the study by Lee et al. (1992), MSO treatment elevated cytoplasmic [NH₄⁺] from 8 to 80 mm. In the present experiments, MSO increased root [NH₄⁺] 4-fold without significantly reducing transcript abundance (Table III) in plants supplied with NO₃⁻ as the sole source of N. When plants were supplied with NO₃⁻ and NH₄⁺ in the absence or presence of MSO, transcript abundance declined to 14% and 9%, respectively, of the NO₃⁻ treatment alone. The strong reduction of transcript abundance in the absence of MSO might be attributed to either an effect of NH₄⁺ or a product(s) of NH₄⁺ assimilation (e.g. Gln). However, in the presence of MSO, the reduction in transcript level indicates that NH₄⁺ itself may participate in regulating transcript abundance. These results demonstrate that HvNRT2 transcript abundance is strongly reduced when tissue [NH₄⁺] is elevated by inhibitors or by provision of a concentrated exogenous source of NH₄⁺.

Nevertheless, since ¹³NO₃⁻ influx decreased by 47% in the presence of NO₃⁻ plus MSO, without significant effects on transcript levels (Table III), it is possible that NH₄⁺ may also act at a post-transcriptional level. Indeed, accumulated NH₄⁺ has been suggested to act directly upon the high-affinity NH₄⁺ transporter (Rawat et al., 2000).

Regulation of HvNRT2 Transcript Abundance and NO₃⁻ Influx by Amino Acids

Both NO₃⁻ influx and HvNRT2 transcript abundance declined in root tissue in response to exogenous amino acid treatments. The inhibitory effects of the four amino acids tested on ¹³NO₃⁻ influx were as follows: Asp > Glu > Asn > Gln. HvNRT2 transcript abundance was reduced according to a similar pattern: Glu > Asp > Asn > Gln, with the decrease being greater than 50% in all cases. Earlier studies have demonstrated that exogenous application of amino acids can inhibit NO₃⁻ uptake/influx, but there has been a lack of agreement as to which amino acids are most active (Breteler and Arnozis, 1985; Glass, 1988; Muller et al., 1995). Breteler and Arnozis (1985) found that exogenously supplied Arg, Asp, Cys, and Glu inhibited NO₃⁻ uptake in NO₃⁻-induced plants. Rodgers and Barneix (1993) demonstrated the same pattern of inhibition, noting that Gln had only a minor effect. In the report by Muller and Touraine (1992), amino acids were fed to root tissue either via the cotyledons, to mimic shoot signals (via the phloem), or by exogenous feeding. Regardless of the method used, several amino acids were inhibitory and, of these, Gln was not the most effective. Tillard et al. (1998) reported that transferring castor bean plants grown on NO₃⁻ to N-free medium resulted in a transient increase in NO₃⁻ influx, and that this increase was correlated with a 40% decrease in the amino acid concentration of the phloem sap (predominantly Gln and Ser), which is consistent with a role for Gln in down-regulating NO₃⁻ influx.

However, without examining the fate of exogenously applied amino acids with respect to their differential uptake and/or metabolism, it is impossible to identify the putative regulators of NO₃⁻ influx. For example, feeding maize plants either Asn or Gln increased endogenous levels of both of these amides (Sivasankar et al., 1997). Likewise, exogenous application of Gln, Glu, or Asn significantly increased root NH₄⁺ concentrations greater than 4-fold and altered concentrations of other amino acids in roots of rice plants (Wang, 1994; A. Kumar, personal communication). In our amino acid-treated barley plants, exogenous application of Gln, Glu, Asp, or Asn increased root concentrations of the applied amino acid and also those of the other amines and amides (Table I). Thus, as a strategy to identify putative regulators of HvNRT2 transcript abundance, this method is inadequate.

Inhibiting Glu synthase (using AZA) in plants supplied with NO₃⁻ reduced ¹³NO₃⁻ influx and HvNRT2 transcript levels by 97% and 95%, respectively. (Fig. 3, A and B). This was the most potent inhibitor employed in the present study. Compared with the NO₃⁻ control, AZA treatment also affected other measured parameters: root [NO₃⁻] was reduced by 82%; root [NH₄⁺] increased by 130%; root [Glu] decreased to approximately 10% of control; and [Gln] increased by 43% (Table II). The observed effects of AZA on HvNRT2 transcript abundance might result from either decreasing the tissue concentration of NO₃⁻, thus affecting induction, or from increasing concentrations of a feedback inhibitor (Gln or NH₄⁺). The first hypothesis is unlikely, because the data in Table II reveal that significant quantities of NO₃⁻ were absorbed during AZA treatment. This is evident in the increase of root [NO₃⁻] and in the larger increases of root [NH₄⁺]. Moreover, concentrations of Asn and Gln were actually higher in the AZA treatments than in the NO₃⁻ controls. Both the elevated [NH₄⁺] and elevated [amide] indicate that NR activity was at least equivalent to that of the controls, which is consistent with significant NO₃⁻ entry, induction of NR (and hence HvNRT2), and significant assimilation of NO₃⁻.

The second hypothesis is supported, with respect to Gln, by the elevated Gln concentrations (and by the virtual elimination of Glu) associated with the AZA treatment (Table II). This treatment reduced the HvNRT2 transcript level by 97% and ¹³NO₃⁻ influx by 95%. If Glu were an important negative feedback regulator, its very low concentration (equivalent to 10% of controls) should have relieved the negative feedback effects on transcript and ¹³NO₃⁻ influx. Support for a major role of Glu therefore relies upon the effects of exogenous application of Glu (Table I; Fig. 2). However, the data in Table I established that exogenous application of Glu increased root Gln 2-fold, compared with a 1.47-fold increase when Gln itself was supplied. The same was true for the exogenous application of Asp, which increased root Gln by 1.62-fold.

A critical role for Gln is also suggested by the results of the tungstate treatment. This was the only treatment that resulted in elevated levels of HvNRT2 transcript, and also the only treatment that substantially reduced concentrations of root Gln (Glu levels were unchanged by this treatment). The argument for a role of root NH₄⁺ is less convincing. During the time course experiment, root [NH₄⁺] remained essentially constant despite a gradual diminution of HvNRT2 mRNA (Vidmar et al., 2000). Likewise, the strong effects of NH₄⁺ on HvNRT2 abundance and ¹³NO₃⁻ influx (Table III) can be explained as arising from a conversion of NH₄⁺ to Gln. However, when [NH₄⁺] becomes particularly high, as was evident when both NO₃⁻ and NH₄⁺ were provided or when MSO was supplied together with NO₃⁻ and NH₄⁺ (Table III), a case may be made for direct effects of NH₄⁺ on transcript levels. A similar conclusion was reached in studies of Arabidopsis (Zhuo et al., 1999).

A model summarizing the main findings of the current experiments is provided in Figure 5, and the main hypotheses and supporting data are presented in Table IV.

Figure 5.

Proposed model for the regulation of IHATS NO₃⁻ transport and HvNRT2 expression. Positive regulatory effects (induction) are shown by a solid line; negative effects are represented by dotted lines.

Table IV.

Hypotheses concerning the regulation of IHATS expression and supporting observations

Hypothesis	Observation
NO3− is not responsible for down-regulating HvNRT2 transcript levels.	Tungstate treatment increased HvNRT2 transcript abundance (Figs. 3B and 4).
NO3− may act post-transcriptionally.	Tungstate treatment decreased NO3− influx (Fig. 3A).
Elevated concentrations of NH4+ down-regulate HvNRT2 transcript levels.	MSO decreased HvNRT2 transcript levels, when applied together with NH4+ (Table III).
NH4+ exerts post-transcriptional effects.	MSO was capable of reducing NO3− influx substantially without significantly affecting HvNRT2 transcript levels (Figs. 3, A and B, and 4).
Gln is the main down-regulator of HvNRT2 transcript levels.	NH4+ treatment dramatically decreased HvNRT2 transcript level when its conversion to Gln was not blocked by MSO (Table III).
	Addition of all amino acids increased root Gln levels, and decreased HvNRT2 transcript levels (Table I; Fig. 2B).
	When root Glu and Gln concentrations were increased by exogenous application, HvNRT2 transcript levels were inversely correlated with Glu and Gln concentrations.
	Tungstate treatment reduced root Gln by 81%, without affecting root Glu. This treatment produced a 20%–30% increase of HvNRT2 mRNA.
	Treatments with AZA decreased root [Glu] by 90%, increased root [Gln] by 43%, and decreased HvNRT2 transcript levels and 13NO3− by 97% and 95%, respectively (Table II; Fig. 3).

MATERIALS AND METHODS

Plant Material

Seven-day-old barley (Hordeum vulgare cv Klondike) seedlings were used in all experiments. Seeds were surface-sterilized with 1% (v/v) hypochlorite solution and rinsed with de-ionized water. The seeds were placed on nylon mesh (pore size, 4 mm) which was fixed on to 20-mm (eight seeds) or 60-mm (25 seeds) plexiglass discs, depending on the experiment. The discs were placed on moist sand and seeds were covered with 10 mm of moist sand in the dark. After 3 d, seedlings were transferred to 40-L hydroponic tanks and grown in N-free 1/10-strength modified Johnson's solution for 4 d (Siddiqi et al., 1990). Prior to influx measurements, plants were treated with 10 mm KNO₃, with or without 5 mm (NH₄)₂SO₄ or 1mm amino acids, and with or without inhibitors of N assimilation. K concentrations of the growth media were monitored daily and a concentrated nutrient solution was supplied to the tanks in the same ratio as in the original modified Johnson's solution to prevent depletion. The pH of the solution was maintained at 6.2 ± 0.3 by the addition of excess CaCO₃ powder. Plants were grown in a controlled-environment chamber with a 16-h/8-h light/dark cycle at 20°C ± 2°C and 70% relative humidity. Light with a photon flux density at plant level equal to 300 μmol m⁻² s⁻¹ was provided by fluorescent tubes with a spectral composition similar to sunlight.

NO₃⁻ Influx

NO₃⁻ influx experiments were carried out essentially as described by Siddiqi et al. (1989). Seven-day-old barley plants, grown in hydroponic tanks and treated according to the particular experimental design, were transferred to 0.5-L of unlabeled influx solution containing 50 μM NO₃ for 5 min to equilibrate roots to the conditions to be employed for influx determination. They were then transferred to 0.5 L of influx solution containing 50 μm NO₃⁻ labeled with ¹³NO₃⁻. After a 5-min influx period, plants were transferred back into a 0.5-L vessel of unlabeled solution for 3 min to remove unabsorbed tracer residing in the cell wall space. Roots and shoots were harvested separately and placed into 20-mL scintillation vials for counting in a Packard gamma counter (Minaxi δ, Auto-γ 5000 series, Packard, Downers Grove, IL). Production of ¹³NO₃⁻ was as described by Kronzucker et al. (1995).

RNA Isolation and Northern-Blot Analysis

Total RNA was isolated using Trizol reagent (Life Technologies, Ontario, Canada), with two modifications. First, after the tissue was ground in a mortar and Trizol reagent was added at a ratio of 0.2 g of tissue to 1 mL of Trizol, the homogenate was centrifuged at 8,000g for 30 min to remove cellular debris. Second, after the total RNA was isolated, it was re-extracted with phenol:chloroform:iso-amyl alcohol (25:24:1), and precipitated with 0.3 m sodium acetate (final concentration) and two volumes of ethanol. Total RNA was separated on a 1.2% (w/v) agarose gel containing 1× 3-(N-morpholino)-propanesulfonic acid (MOPS) with 2.2 m formaldehyde, at 60 V for 3.5 h, then washed twice in water, and RNA transferred by capillary action to N⁺ nylon membranes (Amersham, Quebec, Canada). The membranes were baked for 2 h at 80°C to fix the RNA, and then placed in prehybridization solution for 1 h. Following this procedure, membranes were transferred to hybridization solution with ³²P-labeled probe for 12 to 16 h. Prehybridization and hybridization solutions contained 6× SSC, 5× Denhardt's solution, 0.5% (w/v) SDS, and 20 μg/mL sonicated herring sperm DNA. Randomly labeled probes were made with the Prime-A-Gene kit (Promega, Madison, WI) using an internal fragment (AflIII–EcoRV) from HvNRT2.3, selected for its ability to recognize all known members of the HvNRT2 family of cDNAs. Control levels of total RNA were probed using a fragment of the 25S gene on plasmid pV25S by digestion with XhoI. Membranes were washed as recommended by the manufacturer's instructions with 0.25× SSC buffer and 0.1% (w/v) SDS at 42°C for 15 min for the final wash.

NO₃⁻ Analysis

NO₃⁻ concentrations were determined from fresh tissues, extracted with boiling water at a ratio of 1 g of roots to 10 mL of water. The extracts were centrifuged at 8,000g, and the supernatant was filtered through a 0.45-μm filter. NO₃⁻ was analyzed using the cadmium-copper reduction method on a Technicon Autoanalyzer (Henricksen and Selmer-Olsen, 1970).

Amino Acid and NH₄⁺ Measurements

Amino acids and NH₄⁺ were extracted from root material in a buffer containing 58% (v/v) ethanol, 0.2 m formic acid, and 0.25 mm α-amino butyric acid as an internal standard by use of mortar and pestle at 4°C (Finnemann and Schjoerring, 1999). After centrifugation at 21,000g for 5 min and filtration through a 0.45-mm PVDF microcentrifuge tube filter (Whatman, Maidstone, UK), amino acids were measured with AccQ·Tag on HPLC (two 626 HPLC pumps; 4-mm Nova-pak C₁₈ column, 3.9 × 150 mm, thermostatted at 39°C; 474 scanning fluorescence detector; 717_plus autosampler; 600S controller; all components Waters, Milford, MA). Mobile phase A consisted of 100 mm NaAc (Sigma-Aldrich, St. Louis), 5.4 mm triethylamine (Fluka, Buchs, Switzerland), and 3.5 mm EDTA (Sigma-Aldrich) adjusted to pH 5.7 with phosphoric acid. Mobile phase B had a composition similar to that of A except for the pH, which was 6.7. Mobile phase C was acetonitrile (J.T. Baker, Amsterdam), and mobile phase D was ultrapure water (Milli-Q, Millipore, Bedford, MA; resistance 18.2 MΩ). All solutions were degassed before use. Gradient conditions were (v/v): 0.5 min with 90% A and 10% B; 16.5 min with 89% A, 10% B, and 1% C; 9 min with 80% A, 18% B, and 2% C; 6 min with 68% A, 27% B, and 5% C; 1.5 min with 63% A, 27% B, and 10% C; 3.5 min with 87.5% B and 12.5% C; 11 min with 87% B and 13% C; 0.1 min with 85% B and 15% C; 2.90 min with 60% C and 40% D; and 9 min with 90% A and 10% B. The initial flow rate was 1.0 mL min⁻¹, changing to 1.3 mL min⁻¹ after 33.8 min. Standard curves were made using the appropriate concentrations of authentic amino acid standards (Sigma-Aldrich).