Chapter 86

Regulation of Nitrogen Utilization

BORIS MAGASANIK

OVERVIEW

Enteric bacteria can use one of a variety of nitrogen-containing compounds as the sole source of nitrogen, but their preferred nitrogen source is ammonia. This fact is demonstrated by the observation that in minimal medium these organisms achieve their highest growth rate when ammonia serves as the source of nitrogen.

The utilization of ammonia is described in chapter 24, where it is shown that the most important, and in some conditions the only, reaction in which ammonia serves as a substrate is the synthesis of glutamine from glutamate and ammonia with the concomitant hydrolysis of ATP to ADP and P_i, a reaction catalyzed by glutamine synthetase (GS). Glutamine can subsequently react with α-ketoglutarate and NADPH in a reaction catalyzed by glutamate synthase, which, together with that catalyzed by GS, results in the net synthesis of glutamate from α-ketoglutarate and NH₃. In cells growing with ammonia limitation, these reactions are solely responsible for the synthesis of glutamine and glutamate, which, in turn, provide the nitrogen for all other cell constituents.

Cells growing in an excess of ammonia can also produce glutamate directly from α-ketoglutarate and ammonia, using NADPH as the reducing agent, in a reaction catalyzed by the NADP-linked glutamate dehydrogenase. These cells still require GS for the synthesis of glutamine to be incorporated into protein and to be used for the synthesis of a number of compounds whose nitrogen is derived from the amide group of glutamine. Nevertheless, since 85% of the cell nitrogen has its origin in the amino group of glutamate and only 15% in the amide group of glutamine, it is apparent that in cells growing under ammonia limitation, when the amino group of glutamate is derived from the amide group of glutamine, GS must increase its activity at least sevenfold to meet the demand for cellular nitrogen. It is therefore not surprising that both the activity of GS and the expression of glnA, the structural gene for GS, are tightly regulated in response to the nitrogen source in the growth medium.

The potential sources of nitrogen other than ammonia are amino acids, organic nitrogen compounds such as urea, and the inorganic nitrogen compounds nitrate and dinitrogen. Klebsiella strains are the most versatile, in that they can utilize histidine, urea, nitrate, and dinitrogen, as well as the amino acids which can serve Escherichia coli as sources of nitrogen, such as proline and arginine (51). The utilization of these compounds as sources of nitrogen, which results in the formation of ammonia and/or glutamate, is not required when the growth medium contains ammonia. The synthesis of many of the enzymes and permeases necessary for the utilization of these compounds as sources of nitrogen is appropriately under nitrogen control in that it is activated by nitrogen deficiency.

It has been shown that it is advantageous for the cell to use the NADP-linked glutamate dehydrogenase, rather than GS and glutamate synthase, for the synthesis of glutamate, because this reaction does not require the hydrolysis of ATP (34). However, because of the unfavorable equilibrium of the reaction catalyzed by glutamate dehydrogenase, a drop in the extracellular concentration of ammonia below 1 mM precludes the use of this reaction for the synthesis of glutamate. This reduction in the extracellular ammonia concentration results in the drop in the intracellular concentration of glutamine which triggers the activation of partly inactive GS and the increased transcription of glnA, which together provide a level of GS capable of harvesting the remaining ammonia for the synthesis of glutamine and glutamate.

The glnA gene is a member of an operon that contains in addition the structural genes for nitrogen regulators NR_I and NR_II, which autogenously control the expression of this operon, as well as that of other genes whose products are either themselves permeases and enzymes for the utilization of nitrogen compounds or the activators of the transcription of genes and operons whose products play this role. The activation of the expression of most of these genes depends on the rise in the intracellular level of NR_I brought about by the increased expression of its structural gene, which is the immediate response to ammonia deficiency (reviewed in reference 52).

An increase in the extracellular concentration of ammonia results in the rise in the intracellular concentration of glutamine, which is responsible for the conversion of NR_I from an activator to a repressor of the glnA-containing operon, and causes the inactivation of GS. The decrease in the level of GS resulting from growth of the cells in this medium eventually restores the composition of the cells to that characteristic of the ammonia-replete environment.

The subsequent sections of this chapter describe the mechanisms responsible for the function of this control network.

σ 54-DEPENDENT PROMOTERS

The possibility that the initiation of transcription of glnA and of other nitrogen-regulated genes required a σ subunit other than σ ⁷⁰ was suggested by the observations that (i) a mutation in a gene called glnF (or ntrA) prevented the increased expression of glnA in response to the lack of ammonia as well as that of the nif genes of Klebsiella pneumoniae, whose products are responsible for the fixation of dinitrogen, and (ii) ca. 10 bp from the start of transcription of these genes, a consensus sequence was located with no similarity to that of known bacterial promoters (reviewed in reference 33). The subsequent demonstration that the initiation of transcription of glnA could be achieved with core polymerase combined with the product of the glnF gene rather than with σ ⁷⁰ identified the product of glnF, renamed rpoN, as a previously unknown σ factor, σ ⁵⁴ (36, 38). Subsequently, it was discovered that the transcription not only of nitrogen-regulated promoters but also of some others for genes whose products are not essential under all conditions of growth requires σ ⁵⁴-RNA polymerase (43).

This polymerase differs from RNA polymerases associated with σ subunits of the σ ⁷⁰ family in several fundamental ways. Although it can bind to its promoters to form a closed complex, transition to the open complex requires an activator with binding sites located generally 100 to 150 bp upstream from the transcriptional start site. The activators use the hydrolysis of ATP to obtain the energy required to catalyze the isomerization of the closed σ ⁵⁴-RNA polymerase promoter complex to the open complex. On the other hand, closed σ ⁷⁰-RNA polymerase complexes can generally make the transition to the open complex without the aid of an activator, and the process does not require the hydrolysis of ATP (24, 43).

The affinity of the promoters for σ ⁵⁴-RNA polymerase is determined by their nucleotide sequence. Table 1 lists promoters of E. coli, Salmonella typhimurium (official designation, Salmonella enterica serovar Typhimurium), and K. pneumoniae whose transcription is activated by NR_I phosphate and those of K. pneumoniae whose transcription is activated by NifA.

Table 1?54-dependent nitrogen-regulated promoters

The promoters are located ca. 10 bp upstream of the transcriptional start site, and for comparison of the promoters, an invariant C residue is assigned position –12. In all cases, a GG dinucleotide is located exactly 11 bp upstream from the C in position –24. Furthermore, all promoters with the exception of nifLp have a T in position –26 and all promoters with the exception of glnHp2 have G at position –13. Other highly conserved features are the presence of either C or T in position –14, of A or T in position –11, and the sequence PyPuPyPu in positions –23 to –20.

A comparison of the ability of some of these promoters to bind σ ⁵⁴-RNA polymerase in a closed complex has shown that glnAp2 and glnHp2 have much greater affinity for this polymerase than nifLp, nifHp, and nifUp do; replacement of the G in position –26 of nifLp by T and of C in position –15 or –17 of nifHp by T increases their affinity, and a further increase results from the placement of T at both –15 and –17 of nifHp (14, 20, 90). The already high affinity of glnHp2 for the RNA polymerase can be further increased by replacing the T in position –13 by the canonical G (23). This promoter has greater affinity for RNA polymerase than does the strong promoter glnAp2, suggesting the possible role of T in position –14 and of A in position –11 as determinants of the affinity (F. Claverie-Martin and B. Magasanik, unpublished data). Promoters with T in positions –17 to –14 and A in position –11 are able to bind σ ⁵⁴ in the absence of core polymerase but not as well as in its presence (19; Claverie-Martin and Magasanik, unpublished data).

THE glnALG (glnA ntrBC) OPERON

Structure and Operation

The striking feature of the glnALG operon (illustrated in Fig. 1) is the presence of three promoters. Transcription can be initiated at glnAp1 (187 bp upstream of glnA) and at glnLp (32 bp upstream of glnL) by σ ⁷⁰-dependent RNA polymerase and at glnAp2 (73 bp upstream of glnA) by σ ⁵⁴-dependent RNA polymerase. In addition, the operon contains a rho-independent terminator located between the end of glnA and the start of glnL, which is responsible for preventing three of four transcripts initiated at glnAp1 or glnAp2 from progressing into glnL and glnG. Other features of the operon are a binding site for cyclic AMP (cAMP)- catabolite activator protein (CAP) located 73 bp upstream from the transcriptional start site at glnAp1 and two binding sites for NR_I with their centers located 100 and 140 bp, respectively, upstream from the start site at glnAp2, overlapping the start site at glnAp1. A third binding site for NR_I overlaps the transcriptional start site at glnLp (52).

Fig. 1Transcription of the glnALG operon in cells grown with an excess of nitrogen (High N) or limiting nitrogen (Low N). , promoters; ?, NRI-binding sites; , terminator. Heavy line, strong expression.

Source: Reprinted from reference 52 with permission.

The regulation of the expression of this operon in response to the availability of ammonia depends on the three proteins encoded by this operon (GS, NR_I, and NR_II) and on three additional proteins (σ ⁵⁴, P_II, and uridylyltransferase [UTase]), encoded by the unlinked genes rpoN, glnB, and glnD, whose expression is not subject to this regulation (16, 84). Finally, the expression of glnA in media with an abundant nitrogen supply, such as a protein digest fortified with glutamine, requires CAP and adenyl cyclase, the products of crp and cya (Table 2) (70). The role of these proteins was elucidated initially by the study of mutations and subsequently by the study of the transcription of the glnALG operon with purified components.

Table 2Regulators of glnALG expression

Considering first cells growing in a medium with an abundant nitrogen supply, we find that transcription of the genes of the operon is initiated by σ ⁷⁰-RNA polymerase at glnAp1 and glnLp (70). The product of glnG, NR_I, regulates its own synthesis by partially blocking initiation of transcription at glnLp; in this manner, the intracellular concentration of NR_I is maintained at a level of ca. five molecules of dimeric NR_I per cell (69). In addition, the initiation of transcription at glnAp1 requires cAMP-CAP and is reduced fivefold by the presence of NR_I (70, 74). It is not obvious why the synthesis of GS in the extremely nitrogen-rich medium should require cAMP-CAP and should consequently be reduced by the presence of glucose in the growth medium.

The replacement of the nitrogen-rich medium by one containing a single nitrogen source other than ammonia results in the immediate activation of transcription initiation at the σ ⁵⁴-dependent promoter glnAp2. This transition is favored by the fact that σ ⁵⁴-RNA polymerase is already bound in a closed complex at the glnAp2 promoter in the cells growing with an excess of nitrogen (68, 78). The trigger for the initiation of transcription at glnAp2 is NR_II, which, by phosphorylating NR_I, endows it with the ability to activate the initiation of transcription at glnAp2 (60).

The failure of NR_II to phosphorylate NR_I in the nitrogen-rich medium is the result of its association with the small protein P_II (60). When combined with P_II, NR_II becomes a phosphatase responsible for the dephosphorylation of NR_I phosphate (52). The lack of ammonia in the culture medium results in a drop in the intracellular concentration of glutamine, which enables the enzyme UTase to use UTP to uridylylate P_II (see chapter 24). The resulting P_II-UMP is released from its association with NR_II, which enables the latter to bring about the phosphorylation of NR_I.

NR_I phosphate bound to its sites located upstream of glnAp2 catalyzes the isomerization of the closed σ ⁵⁴-RNA polymerase-glnAp2 complex to the open complex, causing the rapid initiation of transcription at glnAp2 (59, 61, 66). As a consequence, the level of GS in nitrogen-deprived cells is approximately 30-fold higher than in cells grown in the nitrogen-rich medium, making GS one of the major proteins of these cells, capable of assuming its role as the sole agent of ammonia assimilation. The rapid initiation of transcription at glnAp2 also increases the intracellular level of NR_I from 5 to approximately 70 molecules per cell, causing a total block in the initiation of transcription at glnAp1 and glnLp (64, 69). The increase in the intracellular concentration of NR_I occurs because not all transcripts initiated at glnAp2 are terminated at the end of glnA and because the presence of NR_I at the sites overlapping glnLp does not interfere significantly with the elongation of the transcripts. It is therefore apparent that in cells grown with an excess of nitrogen, the glnALG genes are organized as separate glnA and glnLG units of transcription, but in cells deficient in nitrogen, they function as a single glnALG operon (64). The importance of the increase in the level of NR_I for the activation of other nitrogen-regulated genes and operons is discussed below.

An increase in the intracellular level of glutamine which results from the addition of ammonia to the nitrogen-deprived medium causes the removal of the uridylyl group from P_II-UMP by UTase followed by the dephosphorylation of NR_I phosphate by NR_II combined with P_II. In consequence, there is no further initiation of transcription at glnAp2 and the high intracellular concentration also prevents the initiation of transcription at glnAp1 and glnLp. Eventually, growth in the nitrogen-rich medium dilutes the intracellular NR_I to a level that again permits the initiation of transcription at glnAp1 and glnLp.

Molecular Mechanism

The activation of transcription at glnAp2, like that at all other σ ⁵⁴-dependent promoters, requires an activator (43). The critical event in the initiation of transcription at glnAp2 is therefore the conversion of NR_I to NR_I phosphate. Normally, the phosphate donor in this reaction is NR_II, which has the ability to transfer the γ-phosphate group of ATP to a histidine residue located at position 139 in its conserved carboxy-terminal region (58, 89). NR_I is able to catalyze the transfer of this phosphate group from NR_II phosphate to the aspartate residue in its position 54 (75, 89). The recognition that NR_I rather than NR_II is the agent responsible for the phosphorylation of NR_I comes from experiments demonstrating that incubation of NR_I with acylphosphates, such as acetylphosphate, carbamyl phosphate, or phosphoramidate, but not with ATP results in phosphorylation at the aspartate in position 54 of NR_I; nevertheless, NR_II phosphate is the most effective phosphate donor (48).

NR_I phosphate has an autophosphatase activity. In the presence of Mg²⁺ at 37°C, half of its phosphate is liberated as P_i in 5 to 10 min; however, denaturation of NR_I phosphate extends its half-life to 5.5 h. The rate of dephosphorylation of NR_I phosphate is greatly increased by its incubation with NR_II and P_II: under these conditions, ca. half of its phosphate is released in 1 to 2 min (40, 47, 89).

The two activities of NR_II, phosphate donor and phosphatase, can be separated by mutations in its structural gene, glnL. Replacement of histidine 139 by asparagine resulted in the loss of the ability of NR_II to serve as phosphate donor but not of its ability to catalyze the dephosphorylation of NR_I phosphate; moreover, this mutant form of NR_II could catalyze the dephosphorylation of NR_I phosphate in the absence of P_II. On the other hand, alterations in the amino-terminal portion of NR_II had no effect on its ability to serve as the phosphate donor but eliminated its ability to dephosphorylate NR_I phosphate in the presence of P_II (2). These observations suggest that NR_II can exist in two conformations and that P_II stimulates the transformation of its phosphate-donating to its phosphate-removing conformation. Accordingly, P_II is not directly involved in the dephosphorylation of NR_II phosphate. Furthermore, it is not known whether NR_II, in its phosphate-removing conformation, directly causes the dephosphorylation of NR_I phosphate or stimulates its autophosphatase activity.

P_II-UMP is not involved in the dephosphorylation of NR_I. This is in contrast to it other role as regulator of the deadenylylation of GS by adenylyltransferase (ATase). In this case, the ability of ATase to adenylylate and thus inactivate GS in the presence of glutamine is stimulated by P_II and the ability of ATase to deadenylylate GS-AMP requires P_II-UMP (see chapter 24).

The determination of the nucleotide sequence of glnG and comparison with that of other structural genes for response regulators of two-component systems have revealed the presence of three domains which together account for the ability of NR_I phosphate to activate transcription of glnAp2 (25, 82). The target of phosphorylation, aspartate 54, is located in the highly conserved amino-terminal domain of these response regulators. The central domain is common to activators at σ ⁵⁴- but not at σ ⁷⁰-dependent promoters. The carboxy-terminal domain contains the helix-turn-helix motif found in many proteins capable of binding specifically to DNA and could be shown to be responsible for the ability of NR_I to bind specifically to DNA. In addition, the carboxy-terminal domain contains the site which allows NR_I to assume its dimeric configuration (41).

As described in the preceding section, growth in a nitrogen-deprived medium increases the intracellular level of NR_I phosphate by the transcription of glnG initiated at glnAp2. When present in this high intracellular concentration, NR_I phosphate can activate the transcription at a glnAp2 promoter lacking the binding sites for NR_I; on the other hand, when present in low concentration in cells grown with an excess of nitrogen when the transcription of glnG is initiated at glnLp, NR_I-binding sites are essential for the initiation of transcription at glnAp2 (71). A study of transcription initiation at glnAp2 by using purified components led to the same conclusion: with templates present at 5 nM concentration, 100 to 200 nM NR_I, together with NR_II and ATP to bring about its phosphorylation, was required for full activation of transcription at glnAp2 without binding sites for NR_I, but only 10 nM NR_I was required when these binding sites were present (61). Apparently, NR_I phosphate need not be bound to DNA to bring about the activation of transcription. Additional support for this view is provided by the observation that NTRC-3ala, a mutant form of NR_I with three altered amino acids in the carboxy-terminal domain, fails to bind DNA but can activate transcription at glnAp2 when present in high concentration (67). These observations indicate that NR_I phosphate can act directly on the closed σ ⁵⁴-RNA polymerase-glnAp2 complex. Such complexes have been shown to be present in cells grown in nitrogen-rich medium and on a template containing glnAp2 incubated with σ ⁵⁴-RNA polymerase in the absence of NR_I (68, 78).

The central domain of the NR_I polypeptide contains a characteristic ATP-binding motif, and phosphorylation of NR_I endows it with the ability to hydrolyze ATP to ADP and P_i (87). This ability is apparently essential for its role as activator of transcription at glnAp2: ATP is required for the isomerization of the closed to the open σ ⁵⁴-RNA polymerase-glnAp2 complex and cannot be replaced by analogs of ATP not subject to hydrolysis by NR_I phosphate. The need for ATP hydrolysis for the initiation of transcription is a characteristic that distinguishes σ ⁵⁴-dependent promoters from σ ⁷⁰-dependent promoters.

The apparent role of the binding sites is to increase the concentration of NR_I phosphate in the vicinity of the promoter. This view receives strong support from the results of an experiment in which the promoter glnAp2 and the two strong NR_I-binding sites were located on separate small rings of DNA. NR_I phosphate in low concentration stimulated the initiation of transcription at glnAp2 when the two rings formed a singly linked catenane but not when they were decatenated (85).

The region of DNA upstream of glnAp2 contains not only the two strong NR_I-binding sites 100 and 130 bp from the transcriptional start but also three weak binding sites between the strong sites and the promoter (36, 71). However, as long as the two strong sites are present, the weak sites seem to play no role in the expression of the glnALG operon. The strong sites can be placed as far as 1,000 bp upstream or downstream from the promoter without a significant decline in the ability of NR_I phosphate bound to these sites to activate transcription at glnAp2, irrespective of the presence or absence of the weak sites (61). Apparently, NR_I phosphate bound to the strong binding sites interacts directly with the σ ⁵⁴-RNA polymerase bound to the promoter to catalyze the isomerization of the closed to the open complex, an interaction made possible by the flexibility of the intervening DNA. The resulting looping of the DNA has been visualized in electron micrographs (83). The binding sites are therefore the prokaryotic equivalent of eukaryotic enhancers.

In addition to increasing the local concentration of the activator, the two strong binding sites play another important role in the activation of transcription when the concentration of NR_I is low. Each of the two binding sites consists of an inverted repeat capable of binding the dimeric NR_I. Although the phosphorylation of NR_I does not alter its affinity for a single site, it greatly increases its affinity for the two sites, indicating a cooperative interaction between NR_I phosphate molecules (67, 88). Apparently, it is the phosphorylation of the NR_I phosphate dimers that enables them to form tetramers or higher oligomers, and the presence of the two binding sites, located at the same face of the DNA helix, stimulates their interaction. The fact that moving one of the binding sites to the opposite face of the helix diminishes the expression of glnA is in good accord with this view (72). The oligomerization of NR_I phosphate endows it with ATPase activity, as shown by the observation that in the absence of DNA NR_I acquires ATPase activity upon phosphorylation in a cooperative manner and that DNA containing binding sites for NR_I greatly stimulates the ATPase activity when the concentration of NR_I is low (3, 87).

These observations indicate that the actual activator of transcription initiation at glnAp2 is not the NR_I phosphate dimer but a tetramer or higher oligomer which has the required ATPase activity. Accordingly, the role of the phosphorylation is to stimulate the oligomerization. The fact that there are mutant forms of NR_I capable of serving as transcriptional activators in the unphosphorylated state supports this view. The change in one of the mutant forms is the substitution of serine in position 160, not far from the ATP-binding motif in the central domain, by phenylalanine (66, 86). The other change is the substitution of glutamate for aspartate in position 54 (42, 54). The unphosphorylated forms of these mutant proteins possess ATPase activity. The activity of the S160F protein is greatly increased by phosphorylation, but the D54E protein cannot be phosphorylated. A doubly changed S160F, D54E protein, which cannot be phosphorylated, appears to be as good an activator of transcription at glnAp2 as is the NR_I phosphate of wild-type cells (42). The mutations resulting in these changes at positions 54 and 170 apparently accomplish the goal of phosphorylation, i.e., a change in the conformation of NR_I required for the oligomerization of the dimers. The phosphate attached to aspartate 54 is therefore not directly involved in the activation of transcription.

The evidence that the dimerization or oligomerization of NR_I phosphate bound to the strong binding sites is responsible for the activation of transcription at glnAp2 is the comparison of transcription initiation at templates with the usual two binding sites, a single binding site, or no binding sites illustrated in Fig. 2 (88). In Fig. 2, it can be seen that the increased occupation of the two binding sites by NR_I phosphate results in a corresponding increase in the activation of transcription at glnAp2 whereas full occupation of a single site is without effect: in this case, a further increase in NR_I phosphate concentration is required for the activation of transcription. Apparently, an NR_I phosphate dimer bound to a single site can attract a second dimer and form the required tetramer, held together by cooperative interaction. This view is strongly supported by the demonstration that the second molecule of NR_I need not bind to DNA to exert its effect: the addition of NTRC-3ala, a mutant form of NR_I unable to bind to DNA, allows NR_I phosphate in low concentration to activate transcription on a template with a single binding site for NR_I (67).

Fig. 2Analysis of NRI binding to two adjacent binding sites or to a single site and the resulting activation of transcription initiation at glnAp2. Plasmid pVW7 carries two binding sites for NRI, plasmid pVW9 carries a single strong site, and plasmid pAN6 carries no sites. (A) DNase I protection by NRI. Lanes 1 and 7, none; 2 and 8, 2.5 nM; 3 and 9, 5 nM; 4 and 10, 12.5 nM; 6 and 12, 12.5 nM. (B) Open-complex formation at glnAp2.

Source: Reprinted from reference 88 with permission.

Examination of the binding of NR_I to DNA by electron microscopy showed that NR_I, like other DNA-binding proteins, bends the DNA (73, 83). This feature may account for the unexpected observation that replacement of the specific NR_I-binding sites by a spiral DNA segment permits NR_I phosphate in low concentration to activate transcription at glnAp2 (12). The spiral insert consists of many repeats of five or six units of A, separated by five other nucleotides. This stretch of DNA, without any homology to the specific NR_I-binding sites, is as effective as the normal enhancer: it could be shown to compete effectively for NR_I phosphate with DNA carrying the two normal strong binding sites, and it stimulated the ATPase activity of NR_I phosphate. Direct binding of an oligomeric NR_I complex to this sequence-induced supercoiled spiral DNA segment could be demonstrated by electron microscopy (73). It is likely that NR_I, as a DNA-bending protein, has affinity for this bent DNA and that the tight packing of NR_I phosphate in the DNA spiral greatly increases the cooperative interaction, resulting in the formation of oligomeric NR_I phosphate. Apparently, in this case the tertiary structure of the DNA is an effective substitute for a specific recognition sequence.

"Fallback Positions" and "Fossils"

The multicomponent nature of the system required for the proper regulation of the glnALG operon suggests that failure in any single component should result in the crash of the system. It is therefore surprising that almost any one of the components can be singly eliminated without serious consequences.

The most serious effect is the loss of σ ⁵⁴. In such a mutant, transcription cannot be initiated at glnAp2, and because cAMP-CAP is required for the initiation of transcription at glnAp1 and this initiation is blocked by NR_I, the mutant is unable to grow in a medium containing glucose without the addition of glutamine (31). Mutations in glnG, resulting in the lack of NR_I, suppress the need of the mutant lacking σ ⁵⁴ for glutamine (30, 74). In wild-type cells growing on glucose with ammonia as the source of nitrogen, transcription is initiated mostly at glnAp2 (70). Apparently, in these cells the intracellular concentration of glutamine is not sufficiently high for full suppression of the formation of NR_I phosphate, but replacement of ammonia by a poorer source of nitrogen further decreases the level of glutamine and results in a 10-fold increase in the rate of transcription initiation at glnAp2. Elimination of NR_I prevents initiation of transcription at glnAp2 but does not result in a requirement for glutamine, since it simultaneously removes the transcriptional block at glnAp1 (74). In that sense, the existence of glnAp1 constitutes a backup system, permitting growth of the cells in a minimal medium when NR_I is not available. Nevertheless, the loss of NR_I has the consequence that other nitrogen-regulated promoters cannot be activated.

Since the role of NR_II is to bring about the phosphorylation of NR_I, which is essential for its role as activator of transcription, the deletion of glnL, the structural gene for NR_II, should have serious consequences for the ability of the cell to activate transcription at NR_I phosphate-dependent promoters. Actually, however, the elimination of NR_II has very little effect on the response of glnALG or other nitrogen-regulated systems to changes in the availability of a source of nitrogen (16, 21). In the case of glnAp2, the only difference between wild-type cells and the glnL mutant is a slower response to the presence or absence of ammonia in the growth medium (70). The fallback position in this case is the ability of NR_I to catalyze its own phosphorylation by acetylphosphate and of NR_I phosphate to catalyze its own dephosphorylation. Acetylphosphate is a normal metabolite of glucose, and its intracellular concentration increases when lack of ammonia decreases the rate of utilization of glucose metabolites for protein synthesis. Consequently, ammonia deprivation increases the rate of the phosphorylation of NR_I and, as a result, transcription initiation at nitrogen-regulated promoters. In cells with functional NR_II, the level of acetylphosphate does not affect the phosphorylation of NR_I, since NR_II catalyzes both the phosphorylation of NR_I and the dephosphorylation of NR_I phosphate through the activities of P_II and UTase (26, 48).

As may be expected, the loss of P_II results in a high level of GS, even in cells grown with an excess of nitrogen. Nevertheless, the activity of GS is still regulated in response to nitrogen availability by adenylylation and deadenylylation. This is because even in the absence ofP_II, adenylylation of GS by ATase is activated by glutamine; lack of glutamine results in slow deadenylylation of GS-AMP, perhaps because of the presence of other phosphodiesterases (16, 28).

Finally, loss of UTase due to a mutation in glnD, which results in the inability to convert P_II to P_II-UMP, does not completely eliminate the rise in the intracellular concentration of GS in response to nitrogen deprivation (16, 27). In cells with functional UTase, the interconversion of P_II and P_II-UMP is regulated exclusively by the intracellular concentration of glutamine. The observation that in the absence of UTase and α-ketoglutarate, the ability of P_II in combination with NR_II to dephosphorylate NR_I phosphate depends on the presence of glutamate suggests that in cells lacking UTase, the intracellular concentration of glutamate, rather than of glutamine, may regulate the phosphorylation and dephosphorylation of NR_I (47).

How can one account for the existence of these backup mechanisms that enable the cell to function in the absence of the complete regulatory system? It is possible that they are the fossils of earlier stages in the evolution of this complex and elegant control system. For example, the activity of GS may have been regulated exclusively by ATase in response to glutamine before P_II arrived on the scene. The interconversion of NR_I and NR_I phosphate may have been autogenously regulated in response to the intracellular concentration of acetylphosphate before the evolution of NR_II. Subsequently, the ability of NR_II to catalyze the interconversion of the NR_I and NR_I phosphate may have been determined by P_II in response to the intracellular level of glutamate before the evolution of UTase.

OTHER NITROGEN-REGULATED SYSTEMS

General Properties

A common characteristic of nitrogen regulated systems is their dependence on NR_I phosphate. The low intracellular concentration of NR_I in cells grown with an excess of nitrogen is adequate for full activation of transcription at glnAp2 but inadequate for initiation of transcription at other nitrogen-regulated promoters. The activation of transcription at these promoters is delayed until transcription initiation at glnAp2 has increased the intracellular concentration of NR_I phosphate (64, 70).

In the case of some nitrogen-regulated operons, a further delay results from the indirect role of NR_I phosphate in the activation of their transcription. Activation of transcription at the σ ⁵⁴-dependent promoters of the nif operons, coding for components of the nitrogen fixation system, requires NifA, a product of the nifLA operon. The expression at the σ ⁵⁴-dependent nifLp is activated by NR_I phosphate and results in an increase in the intracellular concentration of NifA required for activation at the other nif promoters (13). Similarly, activation of transcription of a number of operons which have σ ⁷⁰-dependent rather than σ ⁵⁴-dependent promoters in response to nitrogen deficiency requires the product of the nac gene. The role of NR_I phosphate in this case is to activate transcription at the σ ⁵⁴-dependent nac promoter (6).

Activation by NRI Phosphate

The sequences of the NR_I phosphate-activated, σ ⁵⁴-dependent promoters are presented in Table 1, and the sequences of the corresponding binding sites for NR_I are presented in Table 3. In every case, with the exception of argTp, in which no binding sites have been found, two binding sites for NR_I are located between 100 and 150 bp upstream of the promoter, and in several cases (glnHp2, nifLp, and hisYp) it has been shown that deletion of these sites greatly decreases the initiation of transcription. The initiation of transcription at these promoters depends on the affinity of the promoters for σ ⁵⁴-RNA polymerase, discussed in an earlier section of this chapter, and on the affinity of the binding sites of NR_I. Both determine the probability of a successful interaction of NR_I phosphate bound to its sites with σ ⁵⁴-RNA polymerase bound to the promoter in a closed complex.

Table 3NRI-binding sites

The consensus sequence for an NR_I-binding site, a perfect inverted repeat of 7 bp separated by 3 bp, has the highest affinity for NR_I (Table 3). A binding site with equally high affinity is that overlapping glnLp, which differs from the consensus by the presence of T rather than C in position 5. The binding sites located upstream of glnAp2, which differ from the consensus in one and two positions, respectively, have somewhat lower affinity; nevertheless, the cooperative interaction of the NR_I phosphate molecules bound to these two sites ensures their full occupancy at the low intracellular concentration of NR_I in cells grown with an excess of nitrogen (88).

The NR_I-binding sites located upstream of glnHp2 are also close to the consensus and have been shown to have almost the same affinity for NR_I as the binding sites at glnAp2 (22). On the other hand, the binding sites upstream of nifLp, which differ much more from the consensus, have considerably lower affinity for NR_I than do those associated with glnAp2 (4). Although the affinity of the remaining binding sites for NR_I has not been reported, their divergence from the consensus indicates a low affinity for NR_I.

In summary, the presence of two binding sites for NR_I upstream of these promoters enhances the dimerization of NR_I phosphate, which is essential for the activation of transcription initiation at these promoters. The fact that these binding sites have less affinity for NR_I than those at glnAp2 ensures that the initiation of transcription at these promoters depends on the earlier initiation of transcription at glnAp2 to provide the necessary intracellular concentration of NR_I phosphate. The affinity of these binding sites for NR_I may determine the temporal order in which transcription at the corresponding promoters is initiated.

As mentioned in the discussion of transcription initiation at glnAp2, the position of the binding sites for NR_I relative to that of the promoter can be altered without affecting the ability of NR_I phosphate bound to these sites to activate the initiation of transcription. In the case of glnAp2, NR_I phosphate bound to its sites remains effective whether it is located on the same face of the DNA as the RNA polymerase or on the opposite face. On the other hand, in the case of nifLp, in which the binding sites have much less affinity for NR_I than those at glnAp2, NR_I phosphate must be bound to a site on the same face of the DNA as the RNA polymerase to exert its effect (53).

The signal for the activation of transcription of nitrogen-regulated genes is a drop in the intracellular concentration of glutamine. It is of some interest that in enteric organisms, GS plays an essential role in maintaining glutamine at a sufficiently high level to prevent activation at nitrogen-regulated promoters. Growth with glutamine as the sole source of nitrogen does not provide a sufficiently high intracellular concentration of glutamine to block transcription initiation at glnAp2, presumably because its transport is too slow. It is therefore not surprising that in certain glnA mutants which lack GS, the addition of ammonia to the glutamine-containing medium does not prevent expression of nitrogen-regulated genes. However, another class of glnA mutants has the opposite phenotype: inability to activate the expression of the nitrogen-regulated genes irrespective of the composition of the growth medium. The nature of the mutations in glnA accounts for the difference in the phenotypes: the first class consists of missense mutants, and the second class consists of nonsense mutants. In the latter case, polarity prevents the elongation of the transcripts initiated at glnAp2 and consequently the increase in the level of NR_I required for the activation at promoters with low-affinity binding sites for NR_I (51).

The products of the operons whose expression can be activated by NR_I phosphate are the activators NifA and Nac and uptake systems for glutamine, histidine, and arginine, the products, respectively, of the glnHPQ operon of E. coli (62) and of the hisJQMP operon and the argT gene of S. typhimurium (79). It is likely that corresponding systems exist in the other enteric organisms. In addition to the σ ⁵⁴-dependent promoters, these operons are equipped with σ ⁷⁰-dependent promoters, and in the case of the hisJQMP operon and argT, the initiation of transcription at the σ ⁷⁰ promoters is activated by CAP-cAMP.

In addition, the nasFEDCBA operon, whose products are responsible for the uptake and the utilization of nitrate as source of nitrogen by K. pneumoniae, has a σ ⁵⁴-dependent promoter. This operon, not found in E. coli and S. typhimurium, contains the structural genes for nitrate reductase (nasA) and for nitrite reductase (nasB), which together catalyze the conversion of nitrate to ammonia. In addition to NR_I phosphate, nitrate and nitrite play an essential role in activating the expression of this operon; they counteract the attenuation of transcription by the product of the nasR gene at a site downstream from the promoter (32, 46; J. T. Lin, Ph.D. thesis, Cornell University, Ithaca, N.Y., 1995).

Activation by NifA

In contrast to other enteric organisms, K. pneumoniae is able to use atmospheric dinitrogen as its only source of nitrogen. This utilization requires a complex set of enzymes and auxiliary factors and is expensive in terms of ATP and of reducing power. It is therefore not surprising that these proteins are produced only when no other source of nitrogen is available and when the oxygen tension is very low (33).

The 19 known nif genes are organized as a continuous set of seven operons, comprising 24,206 bp of DNA (1). The general aspects of the regulation of this system are well understood. Transcription is initiated at all nif promoters by σ ⁵⁴-RNA polymerase, and these promoters, listed in Table 1, have the characteristic nucleotide sequence of promoters of this class. The initiation of transcription at nifLA is activated by NR_I phosphate, and the initiation of transcription at all other nif promoters is activated by NifA. The other product of the nifLA operon, NifL, inactivates NifA in response to the presence of oxygen and ammonia. Mutants defective in NifL can express the nif genes during aerobic growth as long as a low intracellular concentration of glutamine raises NR_I phosphate to a level adequate for the activation of transcription at nifLp (33).

There is therefore an interesting difference in the mechanism of the response of NR_I- and of NifA-regulated systems to the presence or absence of exogenous ammonia. In the case of NR_I, the same system, comprising NR_II, P_II, and UTase, is responsible for its phosphorylation, endowing it with the ability to activate transcription in ammonia deprived cells and for its dephosphorylation, which disables it in ammonia-replete cells. In the case of NifA, the critical event in ammonia-deprived cells is the activation of transcription of the nifLA operon by NR_I phosphate, raising the intracellular concentration of NifA which, without modification, activates the transcription of the nif genes. The initiation of transcription at nifLp also increases the intracellular level of NifL, and it is the responsibility of this protein to disable NifA in response to an increase in exogenous ammonia.

The product of the nifA gene is highly homologous to NR_I in its central domain and, like NR_I, has a helix-turn-helix motif in its carboxy-terminal domain. However, its amino-terminal domain is totally different from that of NR_I and does not contain the sequences characteristic of the response regulators of two-component systems (25). Apparently, phosphorylation is not required for the ability of NifA to activate the initiation of transcription at nif promoters. The binding sites for NifA are located generally between 100 and 150 bp upstream from the transcriptional start sites, except for the one at nifEp at –74 and the one at nifFp at –263. These binding sites consist of the inverted repeat TGT-ACA separated by exactly 10 bp. In contrast to NR_I phosphate, a single binding site is adequate for the activation of transcription by NifA (1, 15, 18).

The study of transcription activation by NifA has been greatly hampered by the instability of NifA and by its great tendency to aggregate, so that overproduction of NifA results in the formation of an insoluble material. As a consequence, much of the information regarding the role of NifA has been derived from studies on intact cells. The methods used included measurement of β-galactosidase synthesis in fusions of lacZ to the different nif genes, measurements of the occupancy of the binding sites for NifA and for the σ ⁵⁴-RNA polymerase, and examination of open-complex formation by strand opening in the promoter region, using wild-type cells as well as nifA mutants with defects in the central and carboxy-terminal domains. In this manner, it was shown that the carboxy-terminal domain is responsible for the binding of NifA to its site but not for its ability to catalyze the conversion of the closed to the open σ ⁵⁴-RNA polymerase complex and that mutations in the central domain, affecting the ATP-binding motif shared by NR_I and NifA, greatly diminish the ability of NifA to activate transcription (18, 55, 56).

Similar methods were used to identify the NifA region responsible for its inactivation by NifL. It was found that removal of the amino-terminal domain diminished but did not abolish the ability of NifA to activate transcription at nifHp. The deleted product had greatly increased sensitivity to NifL: this protein inactivated the defective NifA under all conditions but inactivated the normal NifA only in the presence of ammonia or oxygen. These observations suggested that the site of interaction with NifL is in the central domain of NifA and that the amino-terminal domain serves to protect NifA from NifL in the absence of both ammonia and oxygen (17).

Recent experiments with cell extracts, which are hampered by the insolubility of overproduced NifA, have confirmed these results. In some experiments, the small amount of unpurified soluble NifA present in extracts of cells overproducing NifA was used. Other experiments were done with a maltose-binding protein–NifA fusion protein which was soluble and could be purified; nevertheless, these preparations were still aggregated, making it impossible to assess quantitatively the activity of NifA. These experiments confirmed the ability of NifA and of the carboxy-terminal domain of NifA to bind to its specific sites on the DNA and demonstrated that the ability of NifA to activate the initiation of transcription at nifHp resides in the central domain and requires a nucleoside triphosphate subject to hydrolysis (5, 9, 44, 76). It is therefore apparent that NifA activates the isomerization of a closed σ ⁵⁴-RNA polymerase promoter complex to the open complex in the same manner as NR_I phosphate, but, as expected from the lack of the appropriate domain, NifA need not be phosphorylated to exert its effect on transcription. Finally, it could be shown that purified NifL inhibited the activity of NifA, irrespective of the presence or absence of ammonia and oxygen, but had no effect on the ability of NR_I phosphate to activate transcription. The NifL used in these experiments was isolated from cells grown aerobically with an excess of ammonia and may therefore have been present in its inhibitory conformation. The mechanism by which NifL or NifA senses the presence of ammonia and oxygen remains unknown (8, 45).

Another important discovery resulting from the use of a cell- free transcription system was that the activation of transcription of nifHp depends on the presence of an additional protein capable of bending DNA, the integration host factor (IHF). The role of IHF in the activation of transcription at σ ⁵⁴-dependent promoters is discussed in the next section.

Role of IHF

The possibility that a protein in addition to σ ⁵⁴-RNA polymerase and NifA plays a role in the regulation of transcription at nif promoters was suggested by the observation that a protein present in cell extracts, irrespective of how the cells had been grown, could bind to an AT-rich region upstream from the promoters (11). A study of the activation of transcription at nifHp identified this protein as IHF, known for its ability to bend DNA (76). Subsequent experiments showed that IHF could bind to specific sites located between the promoters and the NifA-binding sites in all NifA-activated promoters with the exception of nifFp (37). On the other hand, so far only a single NR_I phosphate-activated promoter, glnHp2, is associated with a binding site for IHF, which in this case is also located between the promoter and the binding sites for the activator (22).

The importance of IHF for the effective activation of transcription was demonstrated in the case of nifHp. Neither NifA bound to its site nor NR_I phosphate bound to sites replacing accurately the binding site for NifA was able to activate the initiation of transcription at nifHp to any significant extent in the absence of IHF, but the addition of IHF caused a 20-fold increase in transcription initiation (77). Similar results were obtained in the study of the activation of transcription at glnHp2, except that in that case transcription could be activated by NR_I phosphate in the absence of IHF and was increased only twofold by the presence of IHF (22).

The role of IHF is to bend the DNA and in this manner to facilitate the contact of the activator, NifA or NR_I phosphate bound to its site(s), with the σ ⁵⁴-RNA polymerase bound to the promoter. In support of this view, it could be shown that IHF affects neither the binding of the activator nor that of the σ ⁵⁴-RNA polymerase to their respective sites on the DNA (22). It exerts its effect by increasing the probability of a successful encounter between the activator bound to the enhancer and the polymerase bound to the promoter that results in the isomerization of the closed to the open polymerase-promoter complex (37).

IHF can play this role only when the three macromolecules, σ ⁵⁴-RNA polymerase, IHF, and the activator, are bound in this order to the same face of the DNA. For glnHp2, it could be shown that changing the distance between two of the three binding sites by 10 or 20 bp, which keeps them on the same face of the DNA, does not interfere with the ability of IHF to stimulate the activation of transcription; on the other hand, placing 5 bp between two of the binding sites, which places them on opposite faces of the DNA helix, does not prevent the activation of transcription initiation in the absence of IHF but results in converting IHF from an activator to a strong inhibitor of transcription. As illustrated in Fig. 3, location of IHF or the activator NR_I phosphate or both on the face of the helix opposite the one to which the σ ⁵⁴-RNA polymerase is bound results in IHF bending the DNA site to which the activator is bound away from rather than toward the RNA polymerase at the promoter. It is apparent that a regulator, which, like IHF, exerts its effect by bending the DNA, can, depending on its location, serve to enhance or diminish the initiation of transcription (23).

Fig. 3Schematic model for the stimulation or inhibition of glnHp2 transcription activation by IHF. The indentation in the oval that designates ?54-RNA polymerase indicates a receptor site for NRI. By bending the DNA, IHF stimulates transcription activation when the binding sites for NRI, IHF, and ?54-RNA polymerase (shaded) are placed correctly on the DNA helix (A and B). When the binding sites for NRI (C) or IHF (D) or both (E) are placed on the other side of the DNA helix with respect to the RNA polymerase, IHF becomes an inhibitor of transcription.

Source: Reprinted from reference 23 with permission.

For both nifHp and glnHp2, it could be shown that increasing the affinity of the promoters for σ ⁵⁴-RNA polymerase decreased the need for IHF. In the case of nifHp, that was accomplished by replacing the C at position –15 or the Cs at positions –15 and –17 by T (Fig. 1); in the latter case, NifA was almost as potent an activator of transcription in the absence of IHF as in its presence (37); in the case of glnHp2, replacement of the T at position –13 by the canonical G allowed as good activation of transcription by NR_I phosphate in the absence of IHF as in its presence when supercoiled DNA served as the template (23). As mentioned in a previous section of this chapter, this mutant promoter, as well as a mutant promoter of nifH containing T in positions –17 to –14, has the highest affinity for σ ⁵⁴-RNA polymerase and, in contrast to glnAp2, can bind σ ⁵⁴ in the absence of core polymerase. It is therefore surprising that on a linear template in the case of glnAp2, transcription can be readily activated by NR_I phosphate, but in the case of the mutant glnHp2, it can be activated only when IHF is provided in addition to NR_I phosphate (23). Since the binding sites for NR_I phosphate upstream of these promoters do not differ greatly in their affinity for NR_I, it appears that the region of DNA between the promoters and the enhancers plays a role in the activation of transcription on linear DNA.

The fact that, at least on supercoiled DNA, the improved affinity of the promoter for σ ⁵⁴-RNA polymerase obviates the need for IHF raises the question why the cell employs a weaker promoter with an IHF site in preference to a stronger promoter. A very reasonable explanation was offered by Hoover et al. (37): the weaker nif promoter is less likely to be activated by NR_I phosphate, which, as mentioned above, can activate σ ⁵⁴-dependent promoters lacking binding sites for NR_I when present in high concentration. This interpretation is strongly supported by the way in which cells with high-affinity nifHp promoters were isolated. This was accomplished by selecting mutants of cells lacking NifA which were capable of activating the expression of nifH and discovering that this expression depended on the increase in the level of NR_I during growth in a nitrogen-limited medium (63).

In addition, IHF may play a direct role in preventing the inappropriate activation of transcription. This view derives from experiments on the activation of transcription at the Pu promoter for a gene whose product is required for the catabolism of xylene, located on the TOL plasmid of Pseudomonas putida (65). The σ ⁵⁴-dependent Pu promoter is associated with binding sites for IHF and for the xylene-activated response regulator. In intact wild-type cells, there is almost no initiation of transcription at Pu unless xylene is added to the growth medium; however, in mutants lacking IHF, there is appreciable initiation of transcription in the absence of xylene and a smaller response to the presence of xylene. The initiation of transcription in the absence of xylene is presumably due to the presence in the cell of other activators of transcription at σ ⁵⁴-dependent promoters without affinity for the xylR-binding site. Apparently, the bend in the DNA caused by IHF interferes with this illegitimate activation. This view is supported by the observation that introduction of a plasmid expressing the nifA gene of K. pneumoniae at a high level stimulates transcription initiation at Pu only in cells lacking IHF and that introduction of a static bend in the DNA in the region of the binding site for IHF prevents the inappropriate activation of transcription in cells lacking IHF. The authors therefore propose that IHF plays the role of "an active suppressor (restrictor) of promiscuous activation by heterologous regulators" and thus directly increases promoter specificity (65). This view is in good accord with the observation that the initiation of transcription by NR_I phosphate at a glnHp2 promoter lacking binding sites for NR_I is partially inhibited by IHF (22).

Regulation by Nac

The original observation which eventually led to our present understanding of nitrogen regulation concerned the effect of ammonia on the formation of the enzymes responsible for histidine degradation in Klebsiella (then called Aerobacter) aerogenes. It was found that these enzymes were not produced during growth of the cells in a medium containing glucose, ammonia, and histidine, but that omission of either glucose, resulting in the use of histidine as the source of carbon, or ammonia, resulting in the use of histidine as the source of nitrogen, permitted the cells to produce these enzymes. The formation of these enzymes is thus under triple control in that it requires the presence of histidine and the absence of either glucose or ammonia (57). The enzymes responsible for histidine degradation are not found in E. coli but are present in certain strains of S. typhimurium. In this organism, as in K. aerogenes, their formation depends on induction by histidine and the absence of glucose; however, the lack of ammonia in the glucose-containing medium fails to stimulate their synthesis (50).

The initiation of transcription at the promoter of the hutUH operon is blocked by the product of hutC, a closely linked gene. This operon comprises hutH, coding for histidase, the enzyme responsible for the deamination of histidine yielding urocanate and ammonia, and hutU, the enzyme responsible for the further metabolism of urocanate. HutC is inactivated by urocanate; consequently, induction depends on the formation of urocanate from the exogenous inducer histidine by histidase. Urocanate is in turn degraded by urocanase, preventing its buildup, which otherwise results in endogenous induction of histidase. The repressor binds to a site overlapping the promoter, hutUp (50).

The initiation of transcription at hutUp of K. aerogenes requires either CAP-cAMP, which binds to sites on the DNA centered at –42.5 and –82.5, or Nac, which binds to a site centered at –64. These sites appear also to be present at hutUp of S. typhimurium; however, S. typhimurium lacks Nac, since the transfer of the nac gene of K. aerogenes to S. typhimurium allows the activation of expression of the resident hut genes. As may be expected, a mutation in the nac gene prevents the activation of the hut genes of K. aerogenes in response to ammonia deprivation (6, 7, 10).

The availability of mutants lacking Nac made it possible to determine which systems in addition to hut are subject to regulation by Nac. So far, it has been shown that the expression of the put operon, whose products are responsible for the degradation of proline, and of the ure operon, whose products catalyze the conversion of urea to ammonia and CO₂, can be activated by Nac. The put operon but not the ure operon is also found in E. coli, and the fact that the expression of this put operon, as well as that of hutUH transferred from K. aerogenes to E. coli, is activated in response to ammonia deprivation indicates that E. coli possesses the nac gene. In addition, Nac represses the NADP-linked glutamate dehydrogenase of K. aerogenes but does not affect the corresponding enzyme of E. coli (49).

Nac, a member of the LysR family of regulators of transcription, binds to a site on the DNA with the inverted repeat ATA-TAT separated by 9 bp. It resembles NifA in that it can activate transcription without having to be modified, but in contrast to NifA, which in response to the presence of ammonia is inactivated by NifL, Nac is unaffected by ammonia. Consequently, only the formation of Nac, not its activity, is regulated in response to nitrogen availability. This was demonstrated by replacing the nac promoter by the isopropyl-β-d-thiogalactopyranoside (IPTG)-inducible tac promoter and finding that the activation of Nac-responsive promoters was affected only by the presence of IPTG and not by the nitrogen source of the growth medium (6). For this reason, the Nac-responsive promoters are not regulated as accurately as those responsive directly to NR_I phosphate or to NifA. Addition of ammonia to cells growing on a poor source of nitrogen does not inactivate Nac but merely blocks its continued synthesis. Consequently, synthesis of histidase and urease does not halt but, rather, diminishes as the intracellular Nac is diluted by growth in the new ammonia-containing medium (29).

Unexplored Systems

The formation of a number of additional enzymes and permeases is activated by ammonia deprivation. The availability of mutants of K. aerogenes lacking Nac made it possible to show that this protein is not required for the synthesis of asparaginase and of tryptophan permease in response to an increase in the intracellular level of NR_I phosphate (49). Similarly, the formation of enzymes responsible for the degradation of arginine by this organism requires NR_I phosphate but not Nac. So far, no nac mutant of E. coli has been isolated. It has therefore not been possible to determine whether the NR_I phosphate-dependent activation of the expression of the operons coding for the enzymes required for the degradation of γ-aminobutyrate and for the uptake of ammonia requires Nac (39, 81). It will be of interest to discover whether transcription at the as yet unidentified promoters for these genes and operons in K. aerogenes and E. coli is initiated by σ ⁵⁴-RNA polymerase.