2.1. The mode of action of AmtB.

Since the discovery of the Amt proteins many members of this family have been demonstrated to transport the ammonium analog [14C]-methylammonium ([14C]-MA) (Marini et al. 1994; Ninnemann et al. 1994; Tate et al. 1998; Siewe et al. 1996). However, it is likely that most, if not all, of these assays do not directly measure accumulation of [I4C]-MA. This is due to the metabolic conversion of [14C]-MA to impermeable species which, due to a washing step during sampling, are all that remain inside the cell after the assay. In many bacteria, including E. coli, this assimilation is through the action of glutamine synthetase (GS), which converts [14C]-MA to [14C]-methylglutamine (Barnes et al. 1983; Boussiba et al. 1991; Meier-Wagner et al. 2001). Soupene et al. concluded that AmtB does not concentrate NH/ but rather that it facilitates the equilibration of NH3 across the cytoplasmic membrane and that any apparent concentration of [14C]-MA is due to metabolism by GS (Soupene et al. 1998). The problem of the [14C]-MA transport assay was recognized by Jayakumar and Barnes (Jayakumar et al. 1983) who developed a novel filtration method so that unassimilated [14C]-MA as well as assimilated [14C]-MA could be detected. In this way they were able to show that the initial uptake phase for [14C]-MA lasts no more than 5 seconds (Jayakumar et al. 1985). However these crucial aspects of the assay procedure were not taken into account by Soupene et al.

We have revisited this question and compared wild-type and AamtB strains using both assay methods. In the wild-type strain there is a clear difference between the two Figure j Effect of washing filters in [14C]-MA uptake assays. There is rapid phase of uptake in the assays. standard uptake assays were modified by using a absence of assimilation (non-washed assay) double filter arrangement as described by Jayakumar and and after 30 seconds the rates of uptake Barnes (Jayakumar et al 1983). Samples were either become similar in both conditions washed (filled symbols) or left unwashed (unfilled (Figure 1). By contrast the amtB mutant symbols). Data using the wild-type strain ET8000 (□) and shows a negligible rapid uptake in the the amtB mutant GT1001 (A), are shown, absence of washing and, importantly, uptake tails off rapidly as there is no assimilation. The apparent low level of rapid uptake in the amtB mutant is likely to be due to a combination of the small amount of extracellular radiolabel remaining on the filters and non-specific association with the biomass. However, it is clear that in the presence of AmtB there is a much greater initial uptake of methylammonium, indicating that, in contrast to the report of Soupene et al. (Soupene et al. 1998), there is AmtB-dependent concentration of unmodified methylammonium by E. coli.

5 10

Time (mins)

5 10

Time (mins)

Given these results we have used the washed assay method to calculate the apparent K,„ of AmtB for methylammonium. We determined the rates of [14C]-MA assimilation for the wild-type strain over a substrate concentration range of 1 |iVI to 50 mM. The process is saturatable and follows Michaelis-Menton kinetics giving a Km of 200 uM. This compares with two independent estimates in the literature of 78-79 mM for the Km of GS for methylammonium. From these data we conclude that the kinetic measurements we have made reflect the transport step of the reaction. The data also support our thesis that AmtB functions as a secondary transporter that concentrates ammonium. If AmtB functioned in an analogous manner to facilitator proteins such as GlpF, one would expect the two Km values to be very similar.

2.2. Purification of E. coli AmtB

To facilitate the purification of E. coli AmtB we chose to add a C-terminal hexahistidine (His) tag. As there is considerable conservation of primary sequence within the C-terminus we separated the His tag from the native sequence by a linker of ten residues. The intention was that this might improve accessibility of the His tag in the protein and thereby increase the efficiency of Ni2+ affinity chromatography. Addition of the linker sequence and of the His tag did not impair AmtB activity and the engineered amtB gene was expressed from an inducible T7 polymerase-dependent promoter in a derivative of E. coli strain BL21(DE3). After induction the [14C]-MA transport activity was around 40 times greater than that of the host strain alone. The AmtB6H protein was solubilized from isolated cell membranes using dodecyl-p-D-maltoside (DDM) and then subjected to Ni2+ affinity chromatography yielding a preparation of >95% purity (Figure 2).

The estimated monomeric mass of AmtB from primary sequence is 44.5 kDa but the purified protein runs on SDS PAGE with an apparent molecular mass of around 90 kDa, with a minor species at around 33 kDa. However membrane proteins often run with an aberrantly low apparent molecular mass on SDS-PAGE and we assume that the lowest (33 kDa) band is the monomeric species. The fact that the majority of the protein ran at a significantly higher molecular weight suggested that this could be an oligomeric species. Analysis of the purified material by both size exclusion chromatography and dynamic light scattering showed the protein to behave as a homogeneous species. An accurate estimate of the molecular mass of the AmtB:DDM complex was obtained by sedimentation equilibrium analytical ultracentrifugation. This analysis requires an independent estimate of the molar ratio of DDM:AmtB in the complex which we determined using equilibrium column desorption with [14C]-DDM (Friesen et al. 2000).

Figure 2. SDS PAGE showing steps in purification of His-tagged AmtB from E. coli. Lane 1 - mol. wt. markers, Lane 2 - whole cell lysate, Lane 3 - membrane extract, Lane 4 -after Ni2+ affinity purification, Lane 5 - after ion exchange chromatography.

205 116 84

55 45 36

The resultant data gave a molecular weight of 127 ± 17 kDa which implies that the protein in the particles is present as a trimer (predicted as 133.5 kDa). There is evidence that other members of the Amt family, namely the S. cerevisiae Mep proteins and the human Rhesus proteins form oligomeric complexes and it is therefore possible that the trimeric conformation of E. coli AmtB is representative of all members of this protein family.

2.3. The AmtB and GlnK proteins interact.

In E. coli there are two members of the Pn signal transduction protein family, GlnB and GlnK. GlnB facilitates the sensing of the intracellular nitrogen status and regulates the activities of the NtrBC system and adenylyltransferase. While GlnK can partially substitute for some of the functions of GlnB, the primary function of GlnK is unknown (Atkinson, Ninfa 1999). In order to test our hypothesis that AmtB interacts with the signal transduction protein GlnK we have employed cell fractionation together with Western blotting to investigate the cellular localization of the Pn proteins of E. coli in both the wild-type and a range of mutant strains. For all of these experiments cells were grown in a nitrogen-limited minimal medium. The cells were harvested and broken by sonication. Samples of the whole cell lysate were retained and the membrane fraction was then separated by ultracentrifugation at 250,000g. To distinguish between proteins that were weakly or tightly membrane-associated, the membrane fractions were then subjected to extensive washing with 50 mM sodium phosphate buffer supplemented with 600 mM sodium chloride.

In intial experiments we found that in cells grown under nitrogen limitation a significant percentage of the Pn protein in the cell is membrane associated. 'Studies of glnB and glnK mutants suggested that this was the case for both Pn paralogs although it was much more pronounced for GlnK. Most significantly this association is entirely absent in a strain lacking AmtB. These observations fully support our hypothesis that the GlnK and AmtB proteins are capable of forming a complex and raise the question of the likely function of this association. We therefore examined the effects of ammonia-shock, in which nitrogen-limited cells are subjected to a rapid increase in their nitrogen status, typically by the addition of 30 mM ammonium to the culture. Under these conditions glutamine synthetase is known to be rapidly adenylylated and it would make physiological sense to inactivate transport activity of AmtB. Analysis of fractions from cells harvested 15 minutes after ammonia shock showed the expected rapid deuridylylation of Pn and a concomitant dramatic increase in the amount of membrane-associated Pn (Figure 3). Our current hypothesis is that the membrane-sequestration of GlnK by AmtB in response to a rapid increase in the cellular nitrogen status serves to inactivate AmtB. However it is also possible that this change in the localization of GlnK may have important consequences for other functions of GlnK by rapidly changing the intracellular pool of the protein.

Figure 3. Effect of ammonia shock on GlnK association to the membrane, m a AglnB strain. Lane 1, whole cells pre ammonia shock; Lane 2, whole cells post ammonia shock; Lane 3, cytoplasm pre ammonia shock; Lane 4, cytoplasm post ammonia shock; Lane 5, membrane pre ammonia; Lane 6, membrane post ammonia shock.

Figure 3. Effect of ammonia shock on GlnK association to the membrane, m a AglnB strain. Lane 1, whole cells pre ammonia shock; Lane 2, whole cells post ammonia shock; Lane 3, cytoplasm pre ammonia shock; Lane 4, cytoplasm post ammonia shock; Lane 5, membrane pre ammonia; Lane 6, membrane post ammonia shock.

Was this article helpful?

0 0

Post a comment