The EMBO Journal vol.13 no.15 pp.3464-3471, 1994

Identification of a high affinity NH4+ transporterfrom plants

Olaf Ninnemann, Jean-Claude Jauniauxland Wolf B.FrommerInstitut fur Genbiologische Forschung Berlin GmbH, Ihnestrasse 63,D-14195 Berlin and 'Applied Tumor Virology Abt. 0610 andINSERM U375, DKFZ, P. 101949, D-69009 Heidelberg, GermanyCommunicated by J.Schell

Despite the important role of the ammonium ion inmetabolism, i.e. as a form of nitrogen that is taken upfrom the soil by microorganisms and plants, little isknown at the molecular level about its transport acrossbiomembranes. Biphasic uptake kinetics have beenobserved in roots of several plant species. To studysuch transport processes, a mutant yeast strain that isdeficient in two NH4+ uptake systems was used toidentify a plant NH4+ transporter. Expression of anArabidopsis cDNA in the yeast mutant complementedthe uptake deficiency. The cDNA AMTI contains anopen reading frame of 501 amino acids and encodes ahighly hydrophobic protein with 9-12 putative mem-brane spanning regions. Direct uptake measurementsshow that mutant yeast cells expressing the protein areable to take up [14C]methylamine. Methylamine uptakecan be efficiently competed by NH4+ but not by K+.The methylamine uptake is optimal at pH 7 with a Kmof 65 ,uM and a K; for NH4+ of '10 gM, is energy-dependent and can be inhibited by protonophores. Theplant protein is highly related to an NH4+ transporterfrom yeast (Marini et al., accompanying manuscript).Sequence homologies to genes of bacterial and animalorigin indicate that this type of transporter is conservedover a broad range of organisms. Taken together, thedata provide strong evidence that a gene for the planthigh affinity NH4' uptake has been identified.Key words: Arabidopsis thaliana/methylamine uptake/nitrogen transport

IntroductionThe ammonium ion, which is the most abundant nitrogen-ous compound apart from N2, is an important ion involvedin nitrogen metabolism and can be transported acrossthe plasma membrane of many organisms. One possibletransport mechanism is that ammonia, as a small,uncharged, lipophilic molecule, passively diffuses acrossthe membrane. However, the rates do not seem to besufficient to explain the flow required for uptake processes.

Carrier-mediated transport of NH4' could be demon-strated in bacteria, fungi and in the kidney of mammals(Hackette etal., 1970; Dubois and Grenson, 1979; Kleiner,1985; DuBose et al., 1991; Knepper, 1991). In plants,which are strongly dependent on the exogenous nitrogensupply, the uptake of ionic nitrogen from the soil can

occur either in the form of nitrate or NH4+. Plantsliving in symbiosis with nitrogen-fixing organisms areadditionally able to use atmospheric N2; the transfer fromthe symbiotic bacteria into the plant is assumed to occurin the form of ammonia (Glenn and Dilworth, 1984).Ammonium ions are thought to be toxic and therefore arerapidly fixed into amino acids, amides or ureides in theroots. Thus, at least under normal conditions, only verylow concentrations of NH4' have been detected in plantcells (Lee and Ratcliffe, 1991). Both low and high affinityNH4' transport has been described for uptake from thesoil (Fried et al., 1965; Ullrich et al., 1984; Wang et al.,1993). Uptake systems for NH4' seem to be energy-dependent, as carbon skeletons have to be supplied fromphotosynthesis in the form of sucrose to the roots plusenergy for reduction or to create the electrochemicalgradient. In contrast to nitrate uptake, for which a trans-porter has been isolated by analysis of chlorate-resistantArabidopsis mutants (Tsay et al., 1993), little is known atthe molecular level about NH4' transporters in plants orother organisms. It therefore seems important to identifysuch transport systems to characterize nitrogen uptakeprocesses further. Yeast mutants have proven to be aninvaluable tool for identifying genes from homologousand heterologous sources by complementation. By thismeans, multiple amino acid transport systems have beencharacterized from Arabidopsis (Frommer et al., 1993,1994; Hsu et al., 1993; Kwart et al., 1993). The approachwas also effective for identifying and characterizing sev-eral other transport systems from a variety of plants suchas K+ channels and sucrose transporters (Anderson et al.,1992; Riesmeier et al., 1992, 1993; Sentenac et al., 1992).A yeast mutant deficient in two NH4' uptake systems,MEPI and MEP2, provided a good system for isolatingNH4' transporters from plants. These mutations wereidentified by selecting yeast strains on media containingthe toxic NH4' analogue methylamine. A double mutantwas constructed that grows only poorly on media con-taining low concentrations of NH4' as the sole nitrogensource (Dubois and Grenson, 1979). Complementation ofthe mutant with a cDNA library from Arabidopsis seed-lings (Minet et al., 1992) has led to the identification ofa plant protein that mediates NH4+ transport. Characteriza-tion of the cDNA, the derived protein and the biochemicalproperties of this transporter are described. In a parallelstudy a corresponding gene was isolated from yeast (seeaccompanying paper by Marini et al., 1994).

ResultsComplementation of the NH41 uptake yeastmutationsThe Saccharomyces cerevisiae strain 26972c, which carriesmutations in the two NH4' permease genes, MEPI and

36) Oxford University Press3464

Plant NH4' transporter

AIcORI EcORV Not Pstl

11 kb -

5 kb -

;kh-

Fig. 1. Southern analysis of the AMTI gene. Ten micrograms ofgenomic DNA isolated from Arabidopsis digested with differentrestriction enzymes were separated by gel electrophoresis, transferredto a nylon membrane and hybridized to the 1.75 kb AMTJ cDNA.

100 -

V o-S3 90-

5 80-

X 70-

C6

60-

I I I I I

4 5 6 7 8 9[pH]

B

c

oe

c2-

0.

cE

-)1800 [it

0 5 10 15[min]

Fig. 2. Expression of the AMTJ gene in different organs ofArabidopsis. Twenty micrograms of total RNA isolated from leaves,stems, roots and seedlings were hybridized to the 1.75 kb AMTJcDNA after gel electrophoresis and subsequent transfer of the RNA tonylon membranes.

MEP2, grows very poorly on media containing 1 mMNH4' as the sole nitrogen source. To identify plantNH4' transporter genes, 26972c was transformed with an

episomal plasmid containing a cDNA library derivedfrom Arabidopsis seedlings under the control of thephosphoglycerate kinase promoter (Minet et al., 1992).Transformants were selected on uracil-free medium, thecells were washed from the plates and an aliquot was

plated on medium containing 1 mM NH4' as the solenitrogen source. A cDNA clone with a size of 1.75 kb(pAMT1) was able to complement the yeast mutation. Toconfirm that no reversion or second site mutation hadoccurred in the transformants, the recombinant plasmidswere isolated and reintroduced into the mutant strain.Mutant cells transformed with pAMT1 regained the abilityto grow on 1 mM NH4' containing minimal mediumwhereas those transformed with the control plasmid pFL61did not. This shows that the 1.75 kb cDNA insert fromArabidopsis mediates NH4' uptake in yeast. High strin-gency Southern blot analysis indicates that the AMTJ geneis present in one or a few copies in the Arabidopsisgenome (Figure 1). According to Northern analysis theAMTJ gene is highly expressed as a transcript with a

C

c

-

g3

A

.S

1s

c

.9

'---* 180 x5 -

4

3

21 x

1 8x11

0 10 20 30 40 50

[min]Fig. 3. Methylamine uptake properties of AMT1. (A) pH activityprofile for methylamine transport. Uptake was measured understandard conditions in 20 mM sodium phosphate buffer at the pHvalue indicated. The uptake rate at pH 7 was set to 100% andcorresponds to 6 nmol [14C]methylamine/mg protein/min. (B)Dependence of the [14C]methylamine uptake by AMT1 in the presence(+ glu) or absence (- glu) of 100 mM glucose that was added 4 minafter the start of the uptake experiment as indicated by the arrow.

(C) Accumulation of methylamine in yeast cells expressing AMT1 inthe absence of protonophores (*), with the addition of 2,4-DNP priorto (LI) or 5 min after (0) the start of the uptake experiment. Theaccumulation was determined at pH 5 and otherwise as described inthe legend to Table II. The calculated accumulation factors are givento the right.

3465

'l_l_VWF

10

,.-

O.Ninnomann, J.-C.Jauniaux and W.B.Frommer

length of -1800 nucleotides in roots and leaves of maturesoil-grown plants and to lower levels in stems and inseedlings (Figure 2).

Biochemical properties ofAMTIA frequently used method for determining transportkinetics is the measurement of the accumulation of aradioactive substrate. Since no suitable nitrogen isotopeis commercially available, [14C]methylamine was used asan NH4' analogue to characterize the AMT1 plant NH4'uptake system in yeast.To determine optimal conditions for methylamine trans-

port, uptake assays were performed at different pH values.The maximal rate of methylamine uptake by transformedyeast cells was between pH 6 and pH 8, with only halfthe activity remaining at pH 5 and pH 9 (Figure 3A). Toachieve maximal uptake rates, cells were grown in mediumcontaining a high concentration of glucose (2%) andsubsequently were incubated with 100 mM glucose beforethe start of the uptake measurements. Without glucoseaddition the uptake rate reached only background levels(Figure 3B). This requirement for glucose suggests thatthe AMT1-mediated uptake is an active process that isdependent on ATP synthesis. This conclusion is supportedby the fact that inhibition of de novo protein synthesiswith cycloheximide 5 min prior to the start of the uptakeassay did not strongly inhibit the uptake rate for methyl-amine (Table I), ruling out the possibility that the carbonsource is utilized for the synthesis of proteins directlyinvolved in methylamine transport. Table I shows theinfluence of protonophores and several inhibitors of ATP

Table I. Effect of inhibitors on methylamine (MA) uptake

Inhibitor Relative uptake of MA (%)

None 10010 gg/ml cycloheximide 85100 lpM diethylstilbestrol 820 IM DCCD 34200 pM DCCD 15100 gM 2, 4-DNPa 1810 FM CCCPa 40

Standard conditions were used. Methylamine was present at a finalconcentration of 100 gM. The inhibitors were added at a 5-fold molarexcess 5 min prior to the reaction start. For each inhibitor the 100%reference value was measured by adding the appropriate volume ofsolvent without the inhibitor substance. (DCCD, diethylstilbestrol and2,4-DNP were dissolved in ethanol, and CCCP was dissolved indimethylformamide).aAssayed at pH 5.

synthesis on AMT1 activity. Inhibitors of plasma mem-brane H+-ATPase activity such as diethylstilbestrol andDCCD (N,N'-dicyclohexylcarbodiimide) strongly reducemethylamine uptake. Protonophores such as CCCP(carbonylcyanide m-chlorophenylhydrazone) and DNP(2,4-dinitrophenol) that cause the collapse of the electro-chemical gradient also block the transport of methylamine.

Since methylamine does not serve as a nitrogen sourcefor yeast (Roon et al., 1975), it was possible to measurethe accumulation of this substance simply by determiningthe intra- and extracellular concentrations of the substrateby measuring the distribution of 14C. In an uptake assaycontaining an initial methylamine concentration of100 ,uM, 26972c yeast mutants expressing AMT1 wereable to concentrate methylamine -700-fold within a10 min period (Table II). The same strain transformedwith the control plasmid pFL61 was capable of con-centrating methylamine only 16-fold. This backgroundactivity might be due to the presence of a third uptakesystem for NH4' besides MEPI and MEP2 (Dubois andGrenson, 1979). The capability to accumulate methylamineagainst a concentration gradient is strongly reduced whenprotonophores are present (Figure 3C).

26972c mutants expressing AMT1 take up [14C]methyl-amine in a saturable, concentration-dependent manner thatis consistent with a carrier-mediated uptake. The apparentKm of this uptake system is -65 ,uM when measured in20 mM sodium phosphate buffer at pH 7. Inhibitionstudies with NH4Cl show competitive inhibition with a K,of 5-10 jiM. These values suggest that NH4' is a strongcompetitive inhibitor of methylamine uptake and thatAMT1 represents a high affinity NH4' uptake system(data not shown).

Specificity of the AMT1 transport systemThe results of inhibition studies with a variety of potentiallycompeting substances of methylamine uptake are summar-ized in Table III. With the exception of NH4', whichinhibited methylamine uptake by 90%, none of the mono-valent cations used in 5-fold excess showed a significanteffect. In this context it is of special interest that K+obviously does not compete for methylamine uptake evenat 5-fold excess. This clearly distinguishes the AMT1transport system from other plant cation transporters likethe K+ channel KATI, which has a significant conductivityfor NH4' ions (up to 30% of the K+ conductance;Schachtman et al., 1992). However, in methylamine uptakeassays in which the 20 mM sodium phosphate buffer wasreplaced by 20 mM potassium phosphate buffer, the uptakerate was lowered to 70%. A further increase of the

Table H. Accumulation of [14C]methylamine (MA)

X26972c transformed Initial MA in the Final MA concentration Internal/externalwith: medium (,uM)

External (jiM) Internal (jiM)

pFL61 100 66 1000 15pAMTI 100 4.6 3100 670

Methylamine transport was assayed by adding [14C]methylamine to a cell suspension in 20 mM NaPO4 pH 7 at the given final concentration of MA.Aliquots were withdrawn at several time points, centrifuged and assayed for relative amounts of [14C]methylamine. Reactions were terminated whenthe methylamine levels remained constant for at least 20 min. For estimation of the internal concentrations, the cell volume was assumed to be fourtimes the dry weight.

3466

Plant NH4+ transporter

potassium buffer concentration up to 100 mM led to adecrease of uptake down to 30% of the value measuredin 20 mM sodium phosphate buffer. These data indicatethat K+ only competes for methylamine uptake whenpresent in very high excess (>1000-fold in the case of100 mM potassium phosphate buffer). The fact thatamines with multiple substituents such as dimethylamine,trimethylamine and ethylamine showed no significantinhibitory influence on methylamine uptake rules out thepossibility that the AMT1 uptake system is a generaltransporter of amines rather than a specific NH4+ carrier(Table III). Taken together these data confirm that AMT1represents an NH4' transporter with high specificity.

Sequence analysis ofAMTIDNA sequence analysis of AMTJ reveals an open readingframe of 501 amino acids which is preceded by a stopcodon and has a coding capacity for a protein of 53 kDa.Analysis of the hydrophilicity shows that the predictedprotein is highly hydrophobic and contains 9-12 putativemembrane spanning regions (Figure 4; Kyte and Doolittle,1982). No homology to the putative nucleotide bindingsite (P-loop) of ATP binding proteins involved in primaryactive transport could be found (Higgins et al., 1990). Incontrast to other transporter proteins like the yeast andArabidopsis amino acid permeases (Weber et al., 1988;Kwart et al., 1993) and the homologous protein MEPIfrom yeast (Marini et al., 1994), the N-terminus of

Table Ill. Effect of potential competitors on methylamine uptake

Substance Relative uptake of MA (%)

None 100Methylamine 40Dimethylamine 89Trimethylene 89Ethylamine 93KCI 98RbCl 96CsCl 98NH4Cl 10

Standard conditions were used. Methylamine was present at a finalconcentration of 100 g±M. 100% corresponds to 4 nmol[14C]methylamine/mg protein/min. Competitors were present at a 5-fold molar excess.

k~~~~~~~

Fig. 4. Hydropathy plot calculated from the amino acid sequence ofthe NH4' transporter AMT1. The calculation was performed accordingto Kyte and Doolittle (1982) with a window of 11 amino acids.Hydrophobic regions are given a positive hydropathy index.

the AMT1 protein is hydrophobic. Whether this featureindicates the presence of a signal sequence remains tobe shown.A database search for related protein sequences revealed

significant homology to three open reading frames withso far unknown functions (Figure 5). One of these putativegenes, nrgA, is part of the Bacillus subtilis nrg29 operon(Atkinson and Fisher, 1991, Wray et al., unpublished:accession no. L03216). The second open reading framewith homology to AMT1 is the 5' part of a gene ofprobably bacterial origin named HI. This DNA has beenfound as a contamination in a rat liver cDNA library(Zheng and Weiner, unpublished: accession no. M98350).The third sequence CEL3 represents a fragment of animalorigin and is a translation of a Caenorhabditis eleganscDNA clone (Waterston et al., 1992; accession no.M89248). The high homology of these sequences supportsthe notion that they might also represent NH4+ transporters.This conclusion is supported by analyses of the yeastMEPI gene (see accompanying paper by Marini et al.,1994). This gene encodes a high affinity NH4+ uptakesystem from S.cerevisiae that is also homologous toAMT1 (Figure 5). No homology was found to other iontransporters such as K+ channels or nitrate transporters.

DiscussionPlants can take up nitrogen from the soil as either nitrateor NH4+. Under most conditions, nitrate is the predominantform of nitrogen in the soil since NH4+ produced bymineralization processes or supplied as fertilizer is rapidlyconverted into nitrate by nitrifying microorganisms. Underfield conditions, however, a mixed assimilation of nitrateand NH4+ is probably the common situation due to thepartial inhibition of nitrification in acidic soils and inagricultural fields. The simultaneous uptake of both nitro-gen sources also reduces the problem of pH stress for theplant, since the by-products (H+ and OH-) of NH4+ andnitrate assimilation can neutralize each other. After theuptake of NH4+ into the root, it is either transported inthe form of ammonium malate or is readily fixed intoamino acids which are translocated through the vascularsystem to the shoot (Raven and Smith, 1976). In bothcases the charge balance would be maintained, since theinflux of positively charged NH4+ ions could readily beneutralized by exporting an equivalent amount of protonsdirectly to the soil.

Although numerous NH4+ carriers have recently beendescribed genetically or physiologically, the molecularbasis for such a transport remains unclear and is a matterof debate. For many years it was accepted that NH3 asa small uncharged lipophilic molecule can pass theplasma membrane without the necessity for specific uptakesystems. The identification of mutants for NH3 uptake inmultiple species has challenged this model (Kleiner, 1985).This and the fact that at neutral and acidic pH mostammonia is protonated and in this form not able to passthe membrane has led to the proposition of specific carriersfor NH4+. In unicellular organisms like bacteria and yeastan important role for such transport systems might be tocounteract the passive efflux of NH3 as retrieval systems(Kleiner, 1985; Penia et al., 1987).

3467

O.Ninnomann, J.-C.Jauniaux and W.B.Frommer

Molecular approaches for isolating NH41transporter genesRecently transporters for nitrate and K+ could be identified(Anderson et al., 1992; Sentenac et al., 1992; Tsay et al.,1993). In contrast, very little is known about the molecularbasis of NH4' transport. Attempts to isolate NH4' trans-port systems by complementation of NH4+ uptake mutantsfrom a variety of species have not been successful. Theseattempts allowed the identification of several genes whichmost likely do not code for carrier proteins. TheEscherichia coli amtA gene seems to encode a cytoplasmiccomponent of the NH4+ transport system (Fabiny et al.,1991). The Rhodobacter adgA gene which complementsa mutation of a starvation-inducible NH4+ uptake systemseems to represent a regulatory protein (Zinchenko et al.,

1990). The same Rhodobacter adgA mutation can becomplemented by the E.coli gene efg, whose function isunknown (Willison, 1993).

Expression ofAMT1 in yeastKinetic studies indicate the presence of at least threeNH4+ transport systems in S.cerevisiae. Two of theseuptake systems, MEPI and MEP2, are kinetically wellcharacterized, whereas a third uptake system has beenpostulated based on the remaining uptake activity in ameplmep2 double mutant (Dubois and Grenson, 1979).The MEPI uptake system is a high affinity and highcapacity transporter (Km: 2 mM), whereas MEP2 has aneven higher affinity but lower capacity to transport theNH4' analogue methylamine (Km: 250 ,uM). These trans-

Amti N S C S A T D L A V L L G P N A T A A A N Y C G Q L G D V N N F IDjTjA F A IjEJN TjY L F SNep - - - - - - - - - - - - - - - - - - - - -M-ESRTTGPL TTETYg -GP-- TV AFMILGA

Nrga - r Q V F M F F CAHi - - - -,---------L NIT GI A WFV Lu A

Cel3 - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - E C Qu VUiP L FHG

Amtl Y --LjQl I A vS| A K N T XNiI T PviA g Y L F G Y A F A FMep L V F FM MG L FL Y SG L BR RE L AL I T VL G I |Y F W G Y S L A F

Nrga L V W L M T P G L A L F Y G G M S|tSRNV L S T HSIS S I A I VSIVWl L F G Y T L A FHi EL V L M T P G L A L F_Y G G L ER XNLLH T 1 F IPI1I A THV F T VIG Y S L A F

Cel3A N HL S[H A M RgV WRY LE A S CNTU I C T , A I GwAIL A Y

Amt l - - G FF--I HFZGLF7 I| P T A[DAt - - - - - - - F L 1L r ACI] A A(DGM{ep S ISA PSNIIGNLD FIRNV|G K K L F Q M M F|S C V N L SNrga - - - A P lS I I GF2 L E W[A G L K G V F D P GI D T M[IS IF F Q*IF| A V LlWUT -AlHi -- - GGG D I|I G N FiKXHLSG FTHS TIPDGAF A L FO GMM IITAACel3 G D[j]G E V NUF L - - - - - - - ULV T LISFFS(DP V2J - II PDSX X N T X SUJQULJP L L - - -

Amtl I AE T AY I Y |SSL T G YP v isl iS V DG WAS P R T D G D LMep |IaGgA R| GR L|L3 N V F F I LNT Y V W I W|S P G G WAY W|- - - - - -

Nrga gI I S G A E RMR FG L S SL V Y T P V A G -G G W GI L- - - - - -Hi I I G rJI GR|VV1K T TJW LIFIY G - G GL A L- - - - - -

Cel3 MlIV S G A V A E R C EJ - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -

5026131311

10076636361

140125110110100

190169153153112

Anti L F S T GAIDV GIG L G R F DrNiR AIAilG 240Nep - - - G L D W A G G GjI VSVG F V FWJL G X R - - - - -LNELKL LlIlIF R P NN 210

Nrga - - - - G A L D F AGGNV V H I S S G V A G L V LA|I L G K R - - - - - j- - TIAJS SIP _H 192Hi - - - G A I D FA V V H I S S G v LG_L V L A 1-- - - - HND--KH-V V R 191

Cel3 ---- - - - ---- - --- - - --- - - - - 112

Amtl ASLVVLGTF_ WFGWYGF FNKILVTYETGT NGQWSAVGRTA V Tl TMep V L VTL G T S I LW F G W LiF N S|AS - _-- S L S PP L R A FHNTC

Nrga L I Y FL G G ALW F G W F G F N S - L T L_jG V A Y F I N T NHi PoY G G G S- - -_ A A|@ A - A L V N T W

Cel3 - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -----------

Amtl1 T A A TEL F G 17 L[oS GfilN V T DV L L GGIAA-T GGC S V V[F-w-lAMep T | M T Wl C L D Y R S E 1tK_JTK mW mOm Ai P S Gm I tYL L

Nrga ,rAAA A G A AUIwI 'IINRS S L A G L V A I T P A A G F V T P F A SHi L A G V L T \I G L A E yK(3H G p G I X - - - - - _

Cel3 - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -

Amtl I C Cl N;| ELK L Y D DP TLT|A Q1A GI CZAW G L I F T A IMEMp IQGI-- KAXVViXVLaY D D A N DDI 1 iAGV] LIF gL

Nrga I J Al WC nX VFS KKFG GALA1 H GI G[T T THi - -

Cel3 - - - - _

Amti K Y L N Q I Y G N P G R PHNVG TDGVG L LG [Q I TIG S G T L FiFAIMep WVI G - - - M D G T T E HEAWTVT H N Y Y It Y S IITAA C

Nrga SAUN S - - - A G A - - - - -JNF YPGD AX I AI A AITY-F-I- F IHi - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -

Cel3 - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -

Aetl mIKKMKM| E G11KDT. RGfFAD YHMAFMDDDESHKA I Q L GARVEPRMep UL G Y I P G M R L R I ESEE|E2G M DI D Q_J Y D YYDJV EV R R D Y Y L W G V D E D S Q R

Nrga V S L F - - LP [ S LIG L D L Yl Q D S M - - - - - - - - - - - - - - - -Hi - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -

Cel3 - - - - - - - - -

Amtl FSIP S P S G A iN T T P T P V - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -

NrgaJAT

Nep US DV N G-R-V N TNAHLAAE R GVVGT NNYjS D GYKGQE YIIQ IE I L PYI HAQ PAIT FVRNrga - GA--D-- FY-- I-

Hi ----- - - - - - - - - - - - - - -- - - -- - -

Cel3 - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -_

290248230229112

340298280252112

389347330252112

439394372252112

487444404252112

501492404252112

Fig. 5. Amino acid sequence alignment of the plant NH4+ transporter AMT1 (this work), the yeast NH4+ transporter MEP1 (Marini et al., 1994)and sequences from the database with unknown functions from Bacillus subtilis (NrgA), unknown origin (Hi) and Caenorhabditis elegans (Cel3).The analysis was performed using PILEUP/PRETTYBOX programs set to standard parameters. Identical and similar residues are boxed.

3468

Plant NH4' transporter

Metabolism

- W K+/NH4+

NHme transporter

NH+NH4

Ndiffusion

NH3 NH3

Fig. 6. Possible mechanisms for NH4+ uptake. The H+-ATPase, linkedto metabolism, generates the membrane potential that drives NH4+inside the cell. Besides carrier molecules with high affinity andspecificity like AMT1, K+ channels can mediate NH4+ uptake. Atneutral or acidic pH, passive diffusion of ammonia contributes onlymarginally to the transport.

port functions are inhibited by NH4+ with Ki values of 20jM (MEPI) and 1 ,uM (MEP2). The double mutant growsonly poorly on 1 mM NH4+ as the sole source of nitrogen.Complementation of this mutant has proven to be aneffective system for isolating and characterizing bothhomologous and heterologous NH4+ transporter genes.The data summarized below demonstrate the existenceof an Arabidopsis NH4+ transport system which withregard to its biochemical properties and protein sequenceresembles high affinity NH4+ uptake systems fromS.cerevisiae.

Both methylamine transport in yeast (Roon et al., 1975)and AMTI-mediated uptake show a pH optimum in theneutral range. This activity profile suggests that methyl-amine (pKa 10.6) as well as NH4+ (pKa 9.2) are transportedor initially bound by the transporter in their protonatedform. Furthermore, this pH optimum distinguishes it fromother Arabidopsis transporters such as amino acid andsucrose carrier proteins that show highest activities at lowexternal pH, a property which agrees with their predictedrole as proton cotransporters.

Regarding its high affinity for methylamine (Km: 65,M)and the low Ki value for inhibition by NH4+ (<10 ,uM),AMT1 seems to be equivalent to the high affinity NH4+uptake systems of yeast. This finding is supported by thefact that AMT1 and MEPI show strong homology on theprotein level.

Both wild type yeast cells and the plant transporterexpressed in the meplmep2 double mutant are capable ofconcentrating methylamine several hundred-fold against aconcentration gradient, a process that is indicative ofan active transport mechanism. For efficient uptake a

fermentable energy source is required. More detailedstudies with S.cerevisiae support the view that the energyfor NH4' transport is provided by ATP through theoperation of a plasma membrane H+-ATPase (Pefiaet al., 1987).

Mechanism of NH41 uptakeThe sensitivity ofAMT1 activity in yeast towards protono-phores and its dependence on an energy source and ATPsynthesis can be taken as indications that this systemtransports along the electrochemical gradient and isdependent on the plasma membrane ATPase activity. TheH+-ATPase activity would allow the cell to compensatefor the depolarization of the membrane potential causedby the influx of the positively charged NH4' ions. Sincedepolarization of the membrane will stimulate the H+-ATPase and thus lead to a high turnover of ATP theenergy dependence of the uptake system can be explained(Figure 6). Such a model of a secondary active uptakesystem has also been proposed for yeast NH4+ transport(Penia et al., 1987). Since the characteristics of NH4+uptake resemble those of K+ uptake in yeast, it has beenspeculated that NH4' and K+ are transported by the samegeneral mechanism but different carriers (Pefia et al.,1987). For the transport of K+ into plant cells severalauthors (e.g. Behl and Raschke, 1987) propose thatplasmalemma K+ influx is coupled to the active extrusionof protons. Recently Arabidopsis K+ channels (KATI andAKT1) have been isolated by complementation of K+-transport-deficient yeast mutants (Anderson et al., 1992;Sentenac et al., 1992). Electrophysiological analysisallowed characterization of KAT1 as an inwardly-rectify-ing K+ channel (Schachtman et al., 1992), suggesting thatK+ influx might be driven by the negative membranepotential of the cells. However, the precise mechanism ofK+ uptake into plant cells remains a matter of debate.Calculations of the electrochemical potential for K+ inthe cytoplasm versus the solution at the root surfacesuggest that the electrochemical potential gradient for K+could not facilitate K+ uptake from a solution containingK+ in the micromolar range under physiological conditions(Kochian and Lucas, 1993). A similar conclusion wasdrawn from a thermodynamic analysis of K+ absorptioninto Arabidopsis roots (Maathuis and Sanders, 1993). Thisanalysis suggests that, due to the high cytosolic K+concentration found in the plant cell, the driving force forK+ will be outwardly directed and thus K+ uptake cannotbe solely energized by the proton motive force. In caseof NH4' uptake the situation seems to be much simpler.The fixation of NH4+ into amino acids by metabolicprocesses results in the intracellular levels of NH4' beinglow enough to allow channel-mediated uptake drivenby the internal negative membrane electrical potential.However, the present data on AMT1 cannot rule outalternative transport mechanisms such as symport orantiport with H+ or other ions and it may even be possiblethat NH4+ is actively accumulated into the vacuole byAMT1 and that transport into the cell simply follows aconcentration gradient. Electrophysiological studies arerequired to unambiguously demonstrate the transportmechanism.

ADP + Pi _>

ATP

3469

I

O.Ninnemann, J.-C.Jauniaux and W.B.Frommer

Specificity of the NH41 transporterThe AMTl uptake system is strongly inhibited by NH41but only slightly by K+ and other monovalent cations(Table III). This high specificity for NH4' clearly dis-tinguishes AMT1 from plant K+ channels like KATIwhich has been shown to transport NH41 with about one-third of its capacity for K+ (Schachtman et al., 1992).Electrophysiological studies have provided evidence forcross-competition of K+ and NH4' for both the low andhigh affinity transport systems (Vale etal., 1988; Henriksenet al., 1992). In many cases, however, the analysis at thewhole plant and plasma membrane level is complicatedby the presence of multiple transport systems that cannotbe resolved by kinetic studies. The characteristics ofAMT1 indicate the existence of a new class of transportersthat are highly specific for the transport of NH4'.

Sequence analysis and homologies to otherproteinsAnalysis of the hydrophobicity predicts that 9-12 hydro-phobic segments are present in AMT1 (Figure 4). Such ahydrophobicity pattern is characteristic of a variety ofintegral membrane transporter proteins belonging to theclass of single transport proteins that mediate both substraterecognition and translocation. The absence of a putativeATP binding site suggests that the energy required foractive transport of NH4' is not acquired by direct hydro-lysis of ATP and indicates that AMT1 is a secondarytransport protein. AMT1 is homologous to the yeastcounterpart (53% similarity). An interesting differencebetween AMT1 and the yeast MEPI protein is the presenceof a 21 amino acid long hydrophobic domain at the N-terminus of the plant transporter protein which might beinvolved in plasma membrane targeting. No such leadersequences were found in other transporters such as thesucrose or amino acid transporters (Riesmeier et al., 1992,1993; Kwart et al., 1993). A search for other homologoussequences led to three yet uncharacterized putative genes.Alignment of the four protein sequences from differentkingdoms (Figure 5) shows significant homologies andmay indicate that the AMT1 type NH4' carrier is evolu-tionarily conserved in bacteria, fungi, plants and animals.

Regulation of NH4' transportIn yeast both MEPI and MEP2 transport activities arestrongly repressed in cells grown in the presence of NH4'and thus are subject to nitrogen catabolite repression.There is evidence that this repression mechanism acts atthe level of permease synthesis and not on the carrierprotein itself (Dubois and Grenson, 1979). Preliminaryresults show that AMT1 uptake activity indeed is notlowered in cells grown in the presence of NH4+ (data notshown). To obtain further information about the regulationof AMTJ it will be interesting either to analyse the genewith its own transcriptional control region in yeast or toassay AMTJ expression directly in plants grown undervarious nutritional conditions.

Function in the plant and outlookThe expression of AMTJ in roots and the similarity of itsproperties to the high affinity root uptake system fromrice (Wang et al., 1993) strongly support a function in the

uptake of NH4' from the soil. Such a function would beanalogous to that of the related yeast MEP system whichis responsible for uptake of NH4' from the environment.However, currently we cannot exclude the possibility thatthe protein is involved in retrieval of NH3 that has leakedout of the cells by diffusion-a function suggested for thebacterial transporters (Kleiner, 1985). The expression ofAMTJ in leaves and in the stem might be an indicationfor such a role of the transporter besides uptake of NH41from the soil. Under stringent conditions one gene couldbe detected in Southern analysis. However, this does notrule out the possibility of the presence of further moredistantly related genes. Analysis of regulation and in situlocalization of AMTJ, e.g. by analysing differences inexpression in the presence of different nitrogen sources,may help to unravel the actual function of this protein.These studies will be combined with use of the AMTJgene to identify related genes. Furthermore, expression ofthe gene in the antisense orientation in the plant willpossibly be a useful tool to analyse role and function inthe plant as has been shown for the sucrose transporter(Riesmeier et al., 1993, 1994). The availability of suchgenes in the future might enable the manipulation of thetransport characteristics of different organs and the uptakecharacteristics for nitrogen in transgenic plants.

Materials and methodsStrains and plant materialThe yeast strain 26972c (mepi mep2 ura3) was constructed by introducinga ura3 deficiency into strain 01220b (Dubois and Grenson, 1979;B.Andre, Bruxelles, personal communication). The plant material wasArabidopsis thaliana L. Heynh., ecotype C24.

Yeast growth, transformation and selectionThe yeast strain 26972c was transformed with a cDNA expression libraryderived from Arabidopsis seedlings in the yeast expression vector pFL61(Dohmen et al., 1991; Minet et al., 1992). Transformants were selectedon SD medium without uracil, washed from the plates in nitrogen-free medium (Difco, Augsburg) and selected on nitrogen-free mediumsupplemented with 10 mg/ml glucose and 1 mM ammonium sulfate asthe sole nitrogen source. Colonies able to grow were selected in liquidmedium, and plasmid DNA was isolated and reintroduced into themutant 26972c to demonstrate that the ability to grow was dependenton the presence of the plant cDNA.

DNA manipulations and sequence analysisThe AMTJ cDNA was excised from the yeast expression vector andsubcloned. Both strands were sequenced with T7 polymerase (Pharmacia,Freiburg) from a set of deletions and by using synthetic oligonucleotides.The DNA sequence analysis was performed using the UWGCG package(Devereux et al., 1984) and has been submitted to the EMBL DataBankunder the accession number X75879. The search for related proteinsequences was done using the BLAST program (Altschul et al., 1990)for screening of the NCBI databank. Isolation of genomic DNA fromleaf material was performed according to Rogers and Bendich (1985)and RNA was isolated from different organs of Arabidopsis as describedby Logemann et al. (1988). Southern and Northern hybridizations wereperformed at 68°C in a buffer containing 7% SDS, 1 mM EDTA, 1%BSA and 250 mM sodium phosphate, pH 7.2. The 1.75 kb NotI fragmentfrom pAMTl was used as a probe. Filters were washed at 68°C threetimes in 2 x SSC, 0.1% SDS.

Transport measurementsFor standard uptake studies, yeast cells were grown to the logarithmicphase in NAAG medium [2% glucose, 1.7 g/l nitrogen base withoutNH4+ and amino acids (Difco)] supplemented with 500 ,ug/ml L-proline.Cells were harvested at an OD6W of 0.5-0.7, washed and resuspendedin 20 mM phosphate buffer, pH 7, to a final OD6W of 8. Prior to the

3470

Plant NH4+ transporter

uptake measurements, the cells were supplemented with 100 mM glucoseand incubated for 5 min at 300C. To start the reaction, 100 lt of thiscell suspension were added to 100 gl of the same buffer containing18.5 kBq of [14Clmethylamine (NEN) and different concentrationsof unlabelled methylamine (Sigma) and inhibitors depending on theexperiment. The standard assay contained 100 gM methylamine. Samples(50 1l) were taken after 10, 60, 120 and 180 s, transferred to 4 ml ofice-cold water, filtered on glass fibre filters, and washed with 8 mlof water. To reduce non-specific binding, the washing solution wassupplemented with 5 mM methylamine. The uptake of 14C was deter-mined by liquid scintillation spectrometry (Beckman).

AcknowledgementsWe are very grateful to M.Minet (CNRS, Gif-sur-Yvette, France) forproviding the excellent Arabidopsis cDNA library and to B.Andre forproviding the NH4+ transport mutant. We would like to dedicate thisarticle to Marcelle Grenson who laid the basis for a lot of our work.

122, 502-509.Schachtman,D.P., Schroeder,J.I., Lucas,W.J., Anderson,J.A. and

Gaber,R.F. (1992) Science, 258, 1654-1658.Sentenac,H., Bonneaud,N., Minet,M., Lacroute,F., Salmon,J.M.,

Gaynard,A.R.D. and Grignon,C. (1992) Science, 256, 663-665.Tsay,Y., Schroeder,J.I., Feldmann,K.A. and Crawford,N.M. (1993) Cell,

72, 705-713.Ullrich,W.R., Larsson,M., Larsson,C.M., Lesch,S. and Novacky,A.

(1984) Physiol. Plant., 61, 369-376.Vale,F.R., Volk,R.J. and Jackson,W.A. (1988) Planta, 173, 424-431.Wang,M.Y., Siddiqi,M.Y., Ruth,T.J. and Glass,A.D.M. (1993) Plant

Physiol., 103, 1259-1267Waterston,R. et al. (1992) Nature Genet., 1, 114-123.Weber,E., Chevallier,M. and Jund,R. (1988) J. Mo. Evol., 27, 341-350.Willison,J.C. (1993) FEMS Microbiol. Rev., 104, 1-38.Zinchenko,V.V., Babakin,M.M., Shetstakov,S., Allibert,P., Vignais,P.M.

and Willison,J.C. (1990) J. Gen. Microbiol., 136, 2385-2393.

Received on February 21, 1994; revised on May 20, 1994

ReferencesAltschul,S.F., Gish,W., Miller,W. and Lipman,D.J. (1990) J. Mol. Biol.,

215, 403-410.Anderson,J.A., Huprikar,S.S., Kochian,L.V., Lucas,W.J. and Gaber,R.F.

(1992) Proc. Natl Acad. Sci. USA, 89, 3736-3740.Atkinson,M.R. and Fisher,S.H. (1991) J. Bacteriol., 173, 23-27.Behl,R. and Raschke,K. (1987) Planta, 172, 531-538.Devereux,J., Haeberli,P. and Smithies,O. (1984) Nucleic Acids Res., 12,

387-395.Dohmen,R.J., Strasser,A.W.M., Honer,C.B. and Hollenberg,C.P. (1991)

Yeast, 7, 691-692.Dubois,E. and Grenson,M. (1979) Mol. Gen. Genet., 175, 67-76.DuBose,T.D.,Jr, Good,D.W., Hamm,L.L. and Wall,S.M. (1991) J. Am.

Soc. Nephrol., 11, 1193-1203.Fabiny,J.M., Jayakumar,A., Chinault,A.C. and Barnes,E.M. (1991)

J. Gen. Microbiol., 137, 983-989.Fried,M.F., Zsoldos,F., Vose,P.B. and Shatokaim,I.L. (1965) Physiol.

Plant., 18, 313-320.Frommer,W.B., Hummel,S. and Riesmeier,J.W. (1993) Proc. Natl Acad.

Sci. USA, 90, 5944-5948.Frommer,W.B., Hummel,S. and Reutsch,D. (1994) FEBS Lett., in press.

Glenn,A.R. and Dilworth,M.J. (1984) J. Gen. Microbiol., 130, 1961-1968.

Hackette,S.L., Skye,G.E., Burton,C. and Segal,I.H. (1970) J. Biol. Chem.,245, 4241-4250.

Henriksen,G.H., Raman,D.R., Walker,L.P. and Spanswick,R.M. (1992)Plant Physiol., 99, 734-747.

Higgins,C.F., Hyde,S.C., Mimmack,M.M., Gileadi,U., Gill,D.R. andGallagher,M.P. (1990) J. Bioenerg. Biomembr, 22, 571-592.

Hsu,L., Chiou,T., Chen,L. and Bush,D.R. (1993) Proc. Natl Acad. Sci.USA, 90, 7441-7445.

Kleiner,D. (1985) FEMS Microbiol. Rev., 32, 87-100.Knepper,M.A. (1991) Kidney Int. Suppl., 33, 95-102.Kochian,L.V. and Lucas,W.J. (1993) Plant Cell, 5, 720-721.Kwart,M., Hirner,B., Hummel,S. and Frommer,W.B. (1993) Plant J., 4,

993-1002.Kyte,J. and Doolittle,R.F. (1982) J. Mol. Biol., 157, 105-132.Lee,R.B. and Ratcliffe,R.G. (1991) Planta, 183, 359-367.Logemann,J., Schell,J. and Willmitzer,L. (1987) Anal. Biochem., 163,

21-26.Maathuis,F.J.M. and Sanders,D. (1993) Planta, 191, 302-307.Marini,A., Vissers,S., Urrestarazu,A. and Andre,B. (1994) EMBO J., 13,

3456-3463.Minet,M., Dufour,M.E. and Lacroute,F. (1992) Plant J., 2, 417-422.Pefia,A., Pardo,J.P. and Ramirez,J. (1987) Arch. Biochem. Biophys., 253,

431-438.Raven,J.A. and Smith,F.A. (1976) New Phytol., 76, 415-431.Riesmeier,J.W., Willmitzer,L. and Frommer,W.B. (1992) EMBO J., 11,

4705-4713.Riesmeier,J.W., Hirner,B. and Frommer,W.B. (1993) Plant Cell, 5,

1591-1598.Riesmeier,J.W., Willmitzer,L. and Frommer,W.B. (1994) EMBO J., 13,

1-7.Rogers,S.O. and Bendich,AJ. (1985) Plant Physiol., 95, 669-674.Roon,RJ., Even,H.L., Dunlop,P. and Larimore,F.L. (1975) J. Bacteriol.,

3471