To access the private area of this site, please log in.

Username: 

Password:  

 

Ammonia Transport

NED S. WINGREEN

[SECTION EDITOR: JOHN INGRAHAM]

Posted November 15, 2004

Department of Molecular Biology, Princeton University, Princeton, NJ 08544-1014

Phone: 609-258-8476, Fax: 609-258-8616, E-mail: This e-mail address is being protected from spambots. You need JavaScript enabled to view it

 

INTRODUCTION

This chapter reviews the ammonium/methylammonium transport (Amt) proteins of Escherichia coli and Salmonella enterica serovar Typhimurium. The Amt proteins and their homologs, the methylammonium/ammonium permease proteins of Saccharomyces cerevisiae, constitute a distinct class of membrane-associated ammonia (used in this module to indicate the prevailing mixture of NH₃ and NH₄⁺) transporters. Members of the Amt family are found in archaea, bacteria, fungi, plants, and invertebrate animals. In E. coli and serovar Typhimurium, the Amt proteins are essential to maintain maximal growth at low concentrations of ammonia, the preferred nitrogen source. Transport of other nitrogen sources is reviewed in Chapter The Cold Shock Response.

 

AMMONIA TRANSPORT BY Amt PROTEINS

Members of the Amt family are cytoplasmic membrane proteins that facilitate transport of ammonia into the cell. Kustu and coworkers have argued that this transport is passive (13), allowing ammonia to flow both into and out of the cell. Because dissolved ammonia is present both as neutral NH₃ and as charged NH₄⁺, a relevant question is whether one species, or both, is transported by Amt proteins. Based on voltage-clamp measurements of electrical current, Ludewig et al. (9) have argued that an Amt protein from tomato (Lycopersicon esculentum) is specifically a transporter for NH₄⁺.

The main evidence that the Amt protein in E. coli and serovar Typhimurium (AmtB) facilitates transport of ammonia into the cell comes from studies on the growth of strains of E. coli and serovar Typhimurium containing disruptions of the gene amtB. Decreased rates of growth were observed in amtB mutants of both E. coli (11) and serovar Typhimurium (13) with ammonia as the sole nitrogen source. These growth defects depended on the concentration of extracellular NH₃, not NH₄⁺. Specifically, at pH 5, growth of amtB mutants became slower than wild type for concentrations of the uncharged species (NH₃)below approximately 50 nM. By either increasing to pH 7, thereby increasing NH₃ by ~100-fold without significantly changing NH₄⁺, or by increasing total ammonia (both NH₃ and NH₄⁺) 100-fold at pH 5, wild-type growth rate was achieved. These observations indicate that in the absence of AmtB only the uncharged species (NH₃) enters the cell, presumably by diffusion through the cytoplasmic membrane. This conclusion is consistent with expectations that direct diffusion through the membrane of charged species such as NH₄⁺ is negligible.

An important first question is whether transport of ammonia by AmtB is passive or active. There is some evidence that it is passive, that is, that AmtB simply facilitates equilibration of NH₃ and/or NH₄⁺ across the cytoplasmic membrane. Soupene et al. (13) observed that an amtB mutant of serovar Typhimurium grew up to twofold faster than the isogenic wild-type strain on arginine. In serovar Typhimurium, each arginine is catabolized into two glutamate and two ammonia molecules. The relatively slower growth of the wild-type strain on arginine was attributed to leakage of ammonia via AmtB. To demonstrate that cells do leak ammonia Soupene et al. (13) grew colonies on arginine as a sole nitrogen source and showed that they crossfed nitrogen-starved colonies. The donor and recipient colonies were grown on separate plates separated by an air gap. The crossfeeding molecule must therefore have a gas phase, a property consistent with its being ammonia, but not glutamate. Moreover, recipient colonies lacking the amtB gene were crossfed less well than colonies that were wild type for amtB, again indicating crossfeeding by ammonia. However, donor strains lacking the amtB gene were also able to crossfeed across an air gap. Therefore, it is not definite that leakage of ammonia occurred via AmtB, and it remains an open question whether transport of ammonia by AmtB is passive and bidirectional. Supporting the possibility of passive transport of ammonia by AmtB is an observation that E. coli appears to have no mechanism for concentrating ammonia: Soupene et al. (13) showed that a mutant of E. coli with only the low-affinity glutamate dehydrogenase pathway for assimilation of ammonia, which therefore grows slowly at low ammonia concentrations, is not relieved of its growth defect by overexpression of AmtB.

Kustu and coworkers have suggested that AmtB is a transporter for NH₃ not NH₄⁺. They cite growth studies of serovar Typhimurium with arginine as sole nitrogen source (13). As described above, growth on arginine allows ammonia to leak out of the cell. Wild-type serovar Typhimurium was found to grow slower on arginine at pH 6 than at pH 8. This pH dependence of growth on arginine was attributed to "acidic trapping" of NH₃ in the medium. If AmtB is a channel for NH₃ then low pH in the medium traps ammonia in the form of NH₄⁺ and prevents its return to the cell, resulting in slow growth. A converse situation was suggested to account for accumulation of NH₄⁺in acidic vacuoles in fungi (14).

However, a recent study on an Amt protein (LeAmt1;1) from tomato concluded that it was a specific transporter for NH₄⁺ (9). Voltage-clamp measurements of electrical current performed on Xenopus oocytes expressing LeAmt1;1 indicated transport of NH₄⁺ rather than NH₃ based on the strong voltage dependence of K_m, the ammonia concentration producing half-maximal current, and the observation that K_mwas independent of external pH between pH 5.5 and pH 8.5, and therefore independent of external NH₃ concentration. Argument by homology suggests that AmtB is also a transporter for NH₄⁺, but conclusions on the ammonia species transported by AmtB and whether transport is active or passive await direct biochemical and biophysical studies.

 

STRUCTURE OF Amt PROTEINS

As yet, the structure has not been solved for any member of the Amt protein family. However, the membrane topology of AmtB in E. coli has been carefully studied. AmtB of E. coli is a protein of 428 amino acid residues. Both amino acid sequence analysis (18) and fusions to periplasmic and cytoplasmic reporter proteins support the topology shown in Fig. 1 (Figure 3 in reference 17). The topological model consists of 12 membrane-spanning helices, with both the N terminus and the C terminus in the cytoplasm. Sequence analysis suggests that most other members of the Amt family lack the extreme N-terminal helix, so that the N terminus lies in the periplasm (17). The oligomerization state of AmtB in E. coli has also been studied. Based on purification and analysis of a functional histidine-tagged derivative of AmtB, Blakey et al. (2) argue that the protein forms trimers. The membrane topology and oligomerization state of AmtB in serovar Typhimurium have not been studied in detail. However, AmtB of serovar Typhimurium has 92% identity to AmtB of E. coli, and both have 428 residues, so it is reasonable to assume the same membrane topology and oligomerization state for the AmtB proteins of both organisms.

 

INTERACTION OF AmtB WITH GlnK AND GlnB

A trimeric stoichiometry for AmtB is supported by the observation of a direct interaction between AmtB and the trimeric signal-transduction protein GlnK. In E. coli, GlnK has been observed to associate with the membrane in an AmtB-dependent fashion (5). GlnK and its homolog GlnB are P_II proteins, which are sensors of nitrogen status in prokaryotes, as described in Chapter The Cold Shock Response. The association of GlnK with AmtB was considerably weaker for the uridylylated form of GlnK, which signals nitrogen limitation, than for the unmodified form of GlnK. In the absence of GlnK, its homolog GlnB was similarly observed to associate with AmtB and also preferentially when GlnB was uridylylated (5). Both GlnK (20) and GlnB (4) form trimers in solution, including heterotrimers of the two species (19), suggesting a symmetric interaction between trimers of AmtB and trimers of GlnK/GlnB.

Interaction with GlnK appears to negatively regulate AmtB's ability to transport ammonia (5). Assays using [¹⁴C]methylammonium, an analogue of ammonia, indicate that GlnK suppresses transport via AmtB by a factor of 2. However, these experiments were carried out under nitrogen limitation because transport of methylammonium is competitively inhibited by ammonia. Coutts et al. (5) hypothesize that the interaction with GlnK suppresses AmtB's transport activity after ammonia shock, i.e., a sudden rise in extracellular ammonia concentration. Under these conditions, GlnK is deuridylylated and hence interacts more strongly with AmtB.

Both GlnK and GlnB are sensors of nitrogen status. Their interaction with AmtB suggests a role for AmtB in nitrogen regulation. Indeed, Blauwkamp and Ninfa (3) have shown that AmtB antagonizes GlnK/B signaling through the two-component nitrogen regulatory system composed of NRI (NtrB) and NRII (NtrC) (see  Chapter The Cold Shock Response).

 

REGULATION OF amtBAND glnK

Homologs of amtB and glnK genes form a conserved pair in prokaryotes, and they are almost always cotranscribed (16). In general, expression of the glnKamtB operon is induced specifically by nitrogen limitation. In E. coli, regulation of glnKamtB depends on the nitrogen regulatory system (see Chapter The Cold Shock Response) which increases expression of the glnKamtB operon in the range 200-fold (11) to 2,000-fold (ref:1) under nitrogen-limiting conditions.

 

Rh PROTEINS ARE HOMOLOGS OF Amt PROTEINS

The Rhesus (Rh) proteins, well known for their role in blood-group incompatibility, share significant sequence similarity with the Amt proteins (10). The function of the Rh proteins remains unknown, despite their importance in medicine. Like the Amt family, the Rh proteins are membrane proteins. Their membrane topology is predicted to be largely the same as that of the Amt proteins, with helices IV to XII of Rh corresponding to III to XI of Amt (17).

Three pieces of accumulating evidence suggest that the Rh proteins function as channels for CO₂ (12). First, the homology of Rh proteins to Amt proteins supports a role in transport of a molecule similar to NH₃. Second, the RH1 gene of the eukaryotic microbe Chlamydomonas reinhardtii is regulated by the availability of dissolved CO₂ . Third, the distribution of Rh proteins in organisms, organs, and tissues suggests a role in CO₂ not NH₃ transport; for example, there are ~10⁵ copies of Rh proteins in each human red cell.

 

CONCLUSIONS AND SUMMARY

A central consideration for understanding the transport of ammonia is the fact that the cytoplasmic membranes of bacteria are relatively permeable to small neutral molecules such as NH₃, which therefore limits the ability of cells to concentrate ammonia. Instead, E. coli and serovar Typhimurium "trap" nitrogen derived from ammonia in the form of glutamine and glutamate, the central intermediates of nitrogen metabolism (see chapter 30 in the second edition). Enteric bacteria maintain levels of free glutamine at ~4 mM and free glutamate at ~20 mM (7), levels that are high compared with commonly encountered concentrations of ammonia. Passive rather than active transport of ammonia into the cell appears advantageous to avoid futile cycling of ammonia, although such cycling has been argued to occur in some organisms (8).

Naively, one might expect the nitrogen species transported by AmtB to be NH₄⁺ rather than NH₃ because, below pH 9.25, the former is more abundant. The voltage-clamp measurements on the Amt protein from tomato LeAmt1;1 support this expectation (9). If, however, AmtB in E. coli and serovar Typhimurium provides channels for NH₃ instead of NH₄⁺, as suggested by Kustu and coworkers (13), one reason may be the prokaryotic cell's need to maintain a proton gradient across its cytoplasmic membrane (8, 11). Permeability of the membrane to both NH₃ (by bulk diffusion) and NH₄⁺ (via specific channels) would allow cycling of ammonia, e.g., NH₄⁺ in, NH₃ out, which would lower the internal pH, thereby breaking down the cell's proton gradient. Equilibration between NH₃ and NH₄⁺ is rapid on cellular timescales. The reaction NH₃ + H₂O → NH₄⁺+ OH^- has a time constant t = 1.6 × 10^–6 s (6), so equilibration between the species at the local pH is essentially instantaneous. Note, however, that diffusion times for small molecules through transport channels are much faster; the diffusion time for a water molecule in an aquaporin channel is ~10^–8 s (15), so NH₃ and NH₄⁺can be regarded as distinct species with respect to channel specificity.

In summary, AmtB is a membrane-associated ammonia transporter that is important for growth at external concentrations of the uncharged species (NH ₃) below about 50 nM. The preponderance of evidence suggests that AmtB specifically transports the charged species (NH₄⁺) and that this transport is passive and, hence, bidirectional.

 

ACKNOWLEDGMENTS

I am pleased to acknowledge valuable comments and suggestions from John Ingraham, Sydney Kustu, and Alexander Ninfa