The pivotal twin histidines and aromatic triad of the Escherichia coli ammonium channel AmtB can be replaced

Abstract

In Escherichia coli, each subunit of the trimeric channel protein AmtB carries a hydrophobic pore for transport of NH₄⁺ across the cytoplasmic membrane. Positioned along this substrate conduction pathway are two conserved elements—a pair of hydrogen-bonded histidines (H168/H318) located within the pore itself and a set of aromatic residues (F107/W148/F215) at its periplasmic entrance—thought to be critical to AmtB function. Using site-directed mutagenesis and suppressor genetics, we examined the requirement for these elements in NH₄⁺ transport. This analysis shows that AmtB can accommodate, by either direct substitution or suppressor generation, acidic residues at one or both positions of the H168/H318 twin-histidine site while retaining near wild-type activity. Similarly, study of the F107/W148/F215 triad indicates that good-to-excellent AmtB function is preserved upon individual and simultaneous replacement of these aromatic amino acids with aliphatic residues. Our findings lead us to conclude that these elements and their component parts are not required for AmtB function, but instead serve to optimize its performance.

Ammonium (used to designate NH₃ and NH₄⁺) functions as both a primary nutrient and waste product and, thus, its transport across biological membranes is of fundamental importance. Because NH₃ readily traverses phospholipid bilayers by simple diffusion the role of protein-catalyzed transport of this compound is unusually interesting. The Amt family, a group of integral membrane proteins having representation across all three domains of life, mediates the transport of NH₄⁺ and is necessary for microbial growth when diffusion of NH₃ becomes limiting for nitrogen uptake (1–4). Current evidence supports the view that these proteins are hybrids between passive channels and active transporters. Consistent with their functioning as channels, members of the Amt family have small temperature coefficients (5, 6), indicating that there are no large conformational changes during the transport process, and contain conduction paths capable of carrying several substrate molecules simultaneously (7–10). However, whereas all other characterized channels facilitate only downhill substrate movement, Amt proteins have been found to transport NH₄⁺ against a concentration gradient (2, 11, 12).

High-resolution structures of Amt family members have been determined (7, 9, 10). These structures indicate that Amt proteins function as homotrimers, with each monomeric unit carrying a pore for substrate conduction (Fig. 1). Each of these pores is lined entirely with hydrophobic residues, save for a pair of histidines, H168 and H318 (numbering for Escherichia coli AmtB), postulated to play a critical role in mediating NH₄⁺ transport (8–10). Situated at the periplasmic entrance to every pore is a collar of residues that includes two aromatic components, F107 and W148. Below this collar the periplasmic opening constricts, and a third aromatic residue, F215, whose elevated temperature (B) factor suggests an increased mobility relative to its surroundings, blocks entrance into the pore. It has been proposed that the recruitment of NH₄⁺ to and its subsequent movement along the transport pathway requires, in part, the cation–π interactions afforded by the presence of these three aromatic residues (9, 15–17). All five residues mentioned here—F107, W148, H168, F215, and H318—are highly conserved within the Amt family, but only a limited number of genetic and biochemical studies have been undertaken to address their importance to Amt protein function (2, 8, 15, 18).

Fig. 1.

Substrate transport path of E. coli AmtB. (A) Ribbon representation of the AmtB monomer viewed from the membrane, with the extracellular surface (Upper) and the periplasmic surface (Lower). Transmembrane spanning segments 7, 8, and half of 9 have been made transparent to reveal the substrate conduction pathway (modeled in dark gray). Components of the F107/W148/F215 aromatic triad (blue) and H168/H318 twin-histidine element (red) are highlighted. The position of S257 is shown in green. The model was created using PyMOL (Schrödinger, LLC; http://www.pymol.org/) from Protein Data Bank ID code 2NS1 (13). The substrate conduction pathway was visualized using the program CAVER (14). (B) Detailed view of the center region of the transport pathway rotated 45° counterclockwise relative to A to reveal the location of I110 (green).

In the present study, we assessed the impact that changes to the aforementioned histidine pair and aromatic residue triad had on the function of the E. coli AmtB protein. Two methods were used to measure AmtB activity: a growth assay at low NH₃ in which the absence of AmtB-mediated conductance of its physiologically relevant substrate results in a pronounced growth defect, and a transport assay that follows the uptake of the nonmetabolizable ammonium analog methylammonium. Our analysis indicates that both the twin-histidine and aromatic triad elements can be replaced, either by engineered substitution or via suppressor accommodation of inactivating mutations, with good-to-excellent retention of AmtB function. This finding not only provides insight into the roles played by these elements in AmtB, but also challenges existing mechanistic models of Amt family function that are based in part on a perceived requirement for their presence.

Results

F107/W148/F215 Aromatic Triad.

To gain insight into the roles played by the F107/W148/F215 aromatic triad in AmtB-mediated NH₄⁺ transport, a set of mutants was generated in which either an alanine or a leucine was introduced at one of these three positions. Growth studies at low external NH₃ (∼100 nM NH₃ and 0.5 mM total ammonium at pH 5.5) showed that each derivative, save for the F215A variant, had good AmtB activity (Table 1). Tests of methylammonium ion (CH₃NH₃⁺) uptake confirmed the functionality of the F107 and W148 mutants. However, the F215L variant accumulated CH₃NH₃⁺ poorly despite having a doubling time on low NH₃ only moderately longer than its wild-type parent (Table 1 and Fig. 2A). The levels of AmtB and GlnK present in these derivatives were also determined—AmtB to assess amounts of this protein during the exponential growth phase and GlnK as a measure of glnK-amtB operon transcription and, thus, AmtB expression (see Materials and Methods for details). Most mutant AmtB proteins were produced in similar amounts (±twofold wild-type strain GlnK levels) and present in quantities (±twofold wild-type strain AmtB levels) comparable to that found in the wild-type strain under both growth conditions used to assay transport function (Table 1). The W148 variants were exceptions. Previous work indicated that W148 restricts substrate flux, and as is shown here, replacing this residue with L increases CH₃NH₃⁺ conductance (2, 20). These studies also provided evidence that the W148L substitution improves the rate of NH₄⁺ transport. That both the W148A and W148L mutants were poorly expressed in cells cultivated on low-NH₃ medium (approximately one-third wild-type strain GlnK levels), presumably to counteract the increased activity of these proteins as a means to maintain near-optimal rates of growth, is consistent with such a finding.

Table 1.

AmtB function and expression in aromatic triad variants

			Protein expression‡
			GlnK		AmtB
Genotype	Doubling time, min*	MA uptake, % wild type†	Low NH3	Glutamine	Low NH3	Glutamine
Wild-type	50	100 ± 17	100	100	100	100
F107L	52	46 ± 2	84	85	100	96
W148L/F215L	53	160 ± 9	180	85	130	48
W148L	53	1,300 ± 130	32	96	52	49
W148A	54	99 ± 9	36	93	21	35
F107A/W148L/F215L	55	25 ± 4	150	120	90	41
F107A	59	290 ± 56	190	84	73	61
F215L	66	≤5	200	86	220	89
F107L/W148L/F215L	73	≤5	280	110	130	55
F215A	240	≤5	160	89	15	53
ΔamtB	250	0	130	89	0	0

*Neidhardt's Mes medium (pH 5.5) containing 0.5 mM NH₄Cl and 0.1% glucose. Strains are ordered by growth rate, fastest to slowest. The doubling times are from a single growth experiment, and are representative of findings made in three independent trials.

^†CH₃NH₃⁺ (MA) transport rates were estimated after a 5-min incubation of cells with labeled substrate. Background uptake values of the ΔamtB strain NCM4590 (44 ± 3 pmol/mL/OD₆₀₀) were subtracted from those of all other strains, and the resulting values were then normalized to those of the wild-type strain (270 ± 47 pmol/mL/OD₆₀₀ after background subtraction). Values are reported as means ± SD for at least three independent experiments. The limit of reliable detection for MA transport was ≤5% of the wild-type strain. Rates of MA uptake were also determined after 1-min and 20-min incubations. With the exception of the W148L variant (720 ± 100% wild-type activity at 20 min), mutant strain transport values relative to their wild-type counterpart remained nearly constant across this time frame.

^‡Protein levels were quantitated by densitometry of digitized Western blot images and then normalized to wild-type values after subtracting background values of the ΔglnK strain NCM4589 or ΔamtB strain NCM4590. Values are means of two independent trials. Growth media used were Neidhardt's Mes medium (pH 5.5) containing 0.5 mM NH₄Cl and 0.1% glucose (low NH₃), and N⁻C⁻ medium (pH 7) containing 3 mM glutamine and 0.4% glucose (glutamine).

Fig. 2.

Growth of selected strains at low NH₃. Strains were grown in low-NH₃ medium [0.5 mM NH₄Cl, 0.1% glucose (pH 5.5)] with aeration at 37 °C. (A) Strain genotypes were wild-type (●), ΔamtB (■), F215L (▲), F107A/W148L/F215L (□), and F107L/W148L/F215L (○). (B) Strain genotypes were wild-type (●), ΔamtB (■), H318E (▲), H318D (□), I110N/H168D/H318D (○), and I110N/H168D/H318E (△). The data in A and B, each from single experiments, are representative of findings made in at least three independent trials. Doubling times on low NH₃ derived from growth curves such as those shown in A and B are reported in Tables 1 and 2. As has been observed previously (19, 20), and for reasons yet to be explored, most active AmtB variants reached a slightly higher final OD₄₂₀ than the wild-type strain when grown on low-NH₃ medium.

Having established that each aromatic residue could be removed individually without total loss of AmtB function, we next asked whether transport activity was retained upon simultaneous replacement of these three amino acids. The W148L/F215L variant of AmtB was constructed as a first step in this analysis. Tests of function and expression indicated that this double mutant behaved much like its wild-type parent (Table 1) and, therefore, could act as a suitable template for introduction of a third lesion at position 107. Two triple variants were generated using this approach and both were expressed and functional. The first, a F107L/W148L/F215L derivative, had a growth rate on low NH₃ slightly longer than the F215L mutant but, as was the case with that strain, did not accumulate CH₃NH₃⁺ above the detection limit. The introduction of a F107A lesion into the W148L/F215L background resulted in a strain that performed considerably better than its triple-leucine counterpart; functional studies found this variant to have a nearly normal (within 10% wild type) growth rate on low NH₃ and one-quarter the CH₃NH₃⁺ transport activity of a congenic wild-type strain (Table 1 and Fig. 2A).

H168/H318 Twin-Histidine Element.

An initial screen of variants of the H168/H318 twin-histidine element showed that introduction of leucine at either site or glutamate at position 318 yielded AmtB proteins that, despite being expressed well (60–130% wild-type strain GlnK expression), could not be readily detected on Western blots and exhibited no activity (Table 2). The H318E mutant strain, for reasons yet to be explored, had a doubling time on low-NH₃ growth medium ∼30% longer than its ΔamtB counterpart (Table 2 and Fig. 2B). That an altered AmtB causes a more pronounced growth defect than the absence of AmtB when cultivated on this medium implies that the H318E substitution affects cellular processes external to AmtB-mediated NH₄⁺ transport. Replacement of H168 with glutamate produced a variant having a normal growth rate on low NH₃ (within 10% wild type) and a readily detectable CH₃NH₃⁺ transport activity (∼15% wild type). These findings are consistent with previously published CH₃NH₃⁺ uptake data for this mutant, and were expected because the equivalent position in some Amt family members is occupied by glutamate instead of histidine (8, 18). Additional experimentation indicated that a derivative carrying a H168D substitution had transport properties roughly equivalent to those of the H168E mutant.

Table 2.

AmtB function and expression in twin-histidine variants

			Protein expression‡
			GlnK		AmtB
Genotype	Doubling time, min*	MA uptake, % wild type†	Low NH3	Glutamine	Low NH3	Glutamine
Wild-type	47	100 ± 17	100	100	100	100
H168E	50	15 ± 2	210	110	82	66
H168D	52	17 ± 4	74	100	33	32
H318D§	53	≤5	250	95	83	37
I110N/H168D/H318D§	57	≤5	510	87	≤5	≤5
I110N/H168D/H318E§	69	≤5	550	78	26	≤5
S257P/H318E§	84	23 ± 5	390	110	28	72
H168D/H318E	220	≤5	160	66	≤5	≤5
H168D/H318D	220	≤5	280	110	≤5	≤5
ΔamtB	240	0	130	89	0	0
H168L	250	≤5	120	89	≤5	≤5
H318L	250	≤5	130	76	≤5	13
H168E/S257P/H318E	260	≤5	160	110	≤5	≤5
H168E/H318E	270	≤5	140	100	≤5	13
H168E/H318D	270	≤5	110	91	≤5	≤5
H318E	320	≤5	60	68	≤5	10

*Neidhardt's Mes medium (pH 5.5) containing 0.5 mM NH₄Cl and 0.1% glucose. Strains are ordered by growth rate, fastest to slowest. The doubling times are from a single growth experiment, and are representative of findings made in three independent trials.

^†CH₃NH₃⁺ (MA) transport rates were estimated after a 5-min incubation of cells with labeled substrate. Background uptake values of the ΔamtB strain NCM4590 (44 ± 3 pmol/mL/OD₆₀₀) were subtracted from those of all other strains, and the resulting values were then normalized to those of the wild-type strain (270 ± 47 pmol/mL/OD₆₀₀ after background subtraction). Values are reported as means ± SD for at least three independent experiments. The limit of reliable detection for MA transport was ≤5% of the wild-type strain. Rates of MA uptake were also determined after 1-min and 20-min incubations. Mutant strain transport values relative to their wild-type counterpart remained nearly constant across this time frame.

^‡Protein levels were quantitated by densitometry of digitized Western blot images and then normalized to wild-type values after subtracting background values of the ΔglnK strain NCM4589 or ΔamtB strain NCM4590. Values are means of two independent trials. Growth media used were Neidhardt's Mes medium (pH 5.5) containing 0.5 mM NH₄Cl and 0.1% glucose (low NH₃), and N⁻C⁻ medium (pH 7) containing 3 mM glutamine and 0.4% glucose (glutamine).

^§Mutants isolated by genetic selection. H318D and S257P/H318E were selected as suppressors of the inactive H318E mutant, whereas I110N/H168D/H318D and I110N/H168D/H318E were selected as suppressors of the H168D/H318D and H168D/H318E variants, respectively. All other mutants were generated by site-directed mutagenesis.

We extended this study by using a genetic selection to isolate suppressors of the low-NH₃ growth defect associated with the H168L, H318L, and H318E mutant strains. Because no suppressor mutations that restore growth on low NH₃ are obtained in an amtB null background, all suppressors isolated in this selection must carry a lesion that restored AmtB function. Application of this method to the H168L and H318L variants yielded no isolates, suggesting that these inactivating substitutions are not easily suppressed by spontaneous mutations. The H318E mutant, however, generated numerous suppressors—all but two of which formed small, near-translucent colonies with irregular margins on the low-NH₃ selection medium. Eight of these unusually shaped/colored isolates were sequenced across their amtB gene, and all were found to carry only the original H318E lesion at this locus. For this reason, these strains were considered not germane to the present work and set aside for future analysis. The two suppressors that exhibited normal colony morphology each contained an intragenic amtB lesion, one resulting in a S257P substitution (see Fig. 1A for location in AmtB), and the other replacing the glutamate at position 318 with aspartate. Examination of the S257P/H318E variant indicated that the S257P lesion only partially compensated for the defects associated with the H318E mutation; growth rate on low NH₃, CH₃NH₃⁺ uptake, and AmtB amounts were all improved but still appreciably worse than those of the wild-type strain (Table 2). The H318E-to-D suppressor restored AmtB function to near-normal levels when assessed using the low-NH₃ growth assay (Table 2 and Fig. 2B). This variant, despite containing similar amounts of AmtB (∼40–80% wild-type strain levels) when cultivated on both mediums used for growth, had no CH₃NH₃⁺ transport activity.

The finding that acidic residue substitutions at each site of the H168/H318 twin-histidine element were tolerated led us to examine what effect replacement of both histidines in this pairing would have on AmtB. Four double mutants (H168D/H318D, H168D/H318E, H168E/H318D, and H168E/H318E) and a single triple-variant (H168E/S257P/H318E) were constructed for such work. Each of these derivatives was tested for activity and, in all cases, found to have doubling times on low NH₃ and CH₃NH₃⁺ uptake values no better than a ΔamtB strain (Table 2). These mutant AmtB proteins, though produced in amounts similar to that of the wild-type strain (±twofold wild-type strain GlnK expression; H168D/H318D ∼threefold wild-type strain GlnK expression), were present at low-to-undetectable levels. Suppression analysis was then used to determine whether activity could be restored to these mutants through the introduction of compensatory lesions. To date, this selection has generated two isolates, one each in the H168D/H318D and H168D/H318E backgrounds, containing the same I110N intragenic missense mutation (see Fig. 1B for location in AmtB). Both of these suppressor strains grew on low NH₃ at rates that were moderately longer than that of the wild-type strain but much improved relative to those of their respective parents (Table 2 and Fig. 2B). Neither suppressor strain displayed CH₃NH₃⁺ transport activity. Protein expression studies showed that the I110N-containing AmtB mutants were present at low levels irrespective of the medium used for growth. Interestingly, when cultivated on low-NH₃ medium—but not on N⁻C⁻ medium containing glutamine as the nitrogen source—GlnK amounts in these two suppressor strains were elevated more than fivefold relative to their wild-type counterpart. Considerable increases in glnK-amtB operon transcription (≥2.5-fold wild-type strain GlnK levels) were also noted in a number of other variants grown under these same conditions (Tables 1 and 2). However, because the same phenomenon is not observed in the amtB null strain, the mechanism by which this occurs remains unclear.

Discussion

The study of membrane transport proteins has oftentimes been hindered by the difficulties involved in obtaining information concerning their structure. This is not the case with the Amt family, where the structures of two members have been solved to resolutions affording detailed molecular analysis (7, 9, 10). Though such structures have proven invaluable to the understanding of this group of proteins, aspects of the various mechanistic models of transport derived from them have yet to be rigorously tested experimentally by either biochemical or genetic means. Here, we have addressed this issue by examining the role of and requirement for two conserved elements situated on the AmtB substrate transport pathway, the F107/W148L/F215L aromatic triad and the H168/H318 twin-histidine pairing.

Based on structural and molecular dynamic simulation studies, it has been suggested that members of the F107/W148/F215 triad are critical to AmtB function because of their ability to recruit and stabilize NH₄⁺ via cation–π interactions (9, 16, 17). Work presented here shows that each component of this element can be replaced with aliphatic amino acids and, as determined by the low-NH₃ growth assay, AmtB retains activity (Table 1). Furthermore, a variant carrying aliphatic substitutions of all three of these aromatic residues (F107A/W148L/F215L) was found to have a doubling time on low-NH₃ medium only 10% longer than that of a congenic wild-type strain. Taken together, our findings indicate that the aromatic triad, and the cation–π interactions it potentially provides, is not required for either NH₄⁺ recruitment or entry into the conduction pore.

Our analysis of the H168/H318 twin-histidine pairing shows that it also is not essential. This conclusion is drawn from suppressor genetic studies and low-NH₃ growth assays that demonstrate the tolerance AmtB has for individual and tandem acidic amino acid substitutions of the twin-histidine element (Table 2). Moreover, such work has identified acidic residue type (glutamate or aspartate) and charge density at positions 168 and 318 as important determinants of AmtB function. For instance, whereas the H318D variant is functional, its H318E counterpart is inactive in the absence of the S257P lesion. S257 is one of a group of residues that make contact with the cytoplasmic C-terminal extension of AmtB (7, 13, 21). Genetic evidence suggests this interaction facilitates an oscillation of transmembrane segment 5 (20)—a motion first predicted from structural studies (7)—that is required for alternately opening the periplasmic entrance and cytoplasmic exit of the conduction pore. The S257P substitution may alter this movement, allowing the pore to accommodate the H318E mutation. A similar scheme based on charge density can explain how the I110N lesion restores function to the H168D/H318D and H168D/H318E variants. These two double-mutant proteins are present at low levels and are nonfunctional despite being well expressed (Table 2, GlnK expression), implying that the presence of negative charge at both positions 168 and 318 destabilize and/or inactivate AmtB. I110 is located at roughly the same depth but on the opposite face of the conduction pore from residues 168 and 318 (Fig. 1B). Although asparagine insertion at this position does not dramatically stabilize the H168D/H318D and H168D/H318E derivatives (Table 2; compare I110N suppressor strain AmtB and GlnK levels in low-NH₃ medium), it likely allows AmtB to tolerate the excess charge associated with them by partially masking it through hydrogen bond interactions.

Finally, this work highlights the problem associated with using CH₃NH₃⁺ uptake as the sole method to evaluate the activity of AmtB derivatives. This ammonium analog exhibits substantially more variability in reporting AmtB function relative to the growth assay on low NH₃. In the present analysis, this results in a vast range of CH₃NH₃⁺ uptake values (≤5–1,300% of wild type) for mutants that all have doubling times on low NH₃ within ∼10% of the control strain. More important, trends in CH₃NH₃⁺ transport and growth on low NH₃ do not always parallel one another. The S257P/H318E variant, which ranks seventh in growth rate on low NH₃ but second in CH₃NH₃⁺ accumulation among members of the twin-histidine mutant set (Table 2), is one example of this phenomenon. Other instances of this behavior found in the literature include the C-terminal delete mutant of AmtB, which is a protein that lacks the last 24 residues of AmtB (19, 22), and the Q57H derivative of Arabidopsis thaliana AMT1;1. The first of these mutant proteins transports CH₃NH₃⁺ at 25–35% of the wild-type rate while allowing E. coli to grow on low NH₃ with a doubling time (∼100 min) longer than all functional mutants described here (19, 20), and the second increases NH₄⁺ conductance but has a reduced CH₃NH₃⁺ uptake activity (23). We show here that the correlation between Amt protein-mediated CH₃NH₃⁺ and NH₄⁺ transport can be completely abolished. Thus, several cases were identified where AmtB substitutions had only a minimal-to-moderate impact on growth rate, even though the CH₃NH₃⁺ uptake assay predicted these lesions to be inactivating. F215 mutagenesis studies illustrate this dichotomy well: previously deemed essential based on assays of CH₃NH₃⁺ transport (15), current work demonstrates that this residue can be replaced without total loss of NH₄⁺ transport. Findings such as this indicate that, for reasons yet to be determined, AmtB handles NH₄⁺, its physiologically relevant substrate, and CH₃NH₃⁺ differently. We therefore suggest that the CH₃NH₃⁺ transport assay be used together with the low-NH₃ growth assay when assessing mutant AmtB function.

Materials and Methods

Strains.

Strain NCM4236, a derivative of the prototrophic E. coli K-12 strain NCM3722 that carries a tesB::kan lesion, served as the wild-type control for tests of AmtB expression and function (2, 24). Strains NCM4589 and NCM4590 are NCM3722 derivatives that carry complete deletions of glnK and amtB, respectively (25). The construction of amtB mutant strains via site-directed mutagenesis was carried out essentially as described by Inwood and coworkers (2, 19). Suppressors of inactive amtB mutants were selected on Neidhardt's medium with 80 mM Mes buffer (pH 5.5), instead of Mops buffer (pH 7.4), containing 0.2% glucose as carbon source and 50 μM NH₄Cl as nitrogen source (19). Selections were done at both room temperature (∼20 °C) and 37 °C, and suppressor strains carrying intragenic lesions were identified by sequencing across the amtB gene. All amtB alleles described in this study were carried on the E. coli chromosome in single copy.

Growth Assays.

Growth at low NH₃ concentrations was performed as previously described (2, 19). Cells, grown overnight in LB medium, were first adapted to minimal medium on Neidhardt's Mops medium (pH 7.4) and subsequently acclimated to growth at low pH by inoculation into Neidhardt's Mes medium (pH 5.5) (2, 26). Glucose (0.1%) and NH₄Cl (5 mM) served as the sole carbon and nitrogen sources, respectively, in these media. Cells were finally diluted into Neidhardt's Mes medium (pH 5.5) containing 0.1% glucose and 0.5 mM NH₄Cl, and growth was monitored by changes in optical density at 420 nm. All culture dilutions were 100-fold, and strains were always grown with aeration at 37 °C.

Transport Assays.

Surveys of CH₃NH₃⁺ uptake were carried out using a glutamine synthetase-coupled assay system in which accumulated CH₃NH₃⁺ is incorporated into methylglutamine (2, 27). Strains were grown first in LB medium and then diluted 100-fold into N⁻C⁻ minimal medium (28, 29) containing 0.4% glucose and 5 mM NH₄Cl. Following overnight incubation, cultures were again diluted 100-fold into the same N⁻C⁻ medium, except with 3 mM glutamine replacing NH₄Cl as the nitrogen source. Glutamine supports good growth as the sole source of nitrogen when cells are cultivated in this medium. However, for reasons not well understood, the use of glutamine as a sole nitrogen source elicits the nitrogen limitation response, resulting in the expression of all genes, including amtB, under control of nitrogen regulatory protein C (27, 30, 31). Cells cultivated under these conditions were grown to exponential phase (optical density at 600 nm of 0.2–0.6), harvested by centrifugation, washed twice with and resuspended in ice-cold assay buffer [50 mM Hepes, 100 mM NaCl, 0.2% glucose (pH 7)] at OD₆₀₀ = 0.5, and then held on ice until use. To initiate tests of CH₃NH₃⁺ transport, cell suspensions were preincubated for 10 min at 37 °C before the addition of a 0.01 volume of [¹⁴C]methylammonium (final concentration, 5 μM; specific activity 5 Ci/mol). After a 5-min incubation, aliquots were removed for filtration on Millipore filters (0.45 μM pore size, type HAWP), rinsed twice with 5 mL ice-cold assay buffer, and counted by liquid scintillation. Strain growth and CH₃NH₃⁺ transport assays were carried out with aeration at 37 °C.

Immunoblot Analysis.

Cell samples were harvested (OD₄₂₀ = 0.25–0.3 and OD₆₀₀ = 0.2–0.6 for growth and transport assays, respectively), resuspended in NuPAGE lithium dodecyl sulfate sample buffer (Invitrogen) containing 50 mM DTT at an OD₆₀₀ = 5.0, and incubated 15 min at room temperature before storage at −80 °C. Cell lysates (10–20 μL) were applied to a SDS/PAGE (4–12%) gel without preheating, and electrophoresed at 150 V at 4 °C. Protein was transferred to nitrocellulose and probed with an anti-AmtB rabbit-derived polyclonal antibody (2) to determine amounts of AmtB at the time of cell harvest. Alternatively, transferred proteins were probed with a rabbit-derived polyclonal antibody reactive to GlnK (32) as an indication of glnK-amtB operon expression. Western blots were developed using an alkaline phosphatase-conjugated anti-rabbit IgG antibody (Invitrogen) and, after scanning the blots, protein expression levels relative to those found in the wild-type strain were quantitated by densitometry of digitized images. Because there is no known translational regulation of AmtB and GlnK, this approach provided a means to determine if the absence or low cellular yields of mutant AmtB proteins was due to a transcriptional or posttranscriptional (e.g., inherent instability, increased protease susceptibility, poor membrane insertion) event. However, these methods do not address whether the lack of transport activity in cases where mutant proteins are present at low levels is caused by such a posttranscriptional event or is the result of AmtB loss of function per se. This issue is presently not significant because we are interested only in those mutants exhibiting good growth at low NH₃ concentrations.