|Home | About | Journals | Submit | Contact Us | Français|
The EntE enzyme involved in the synthesis of the iron siderophore enterobactin, catalyzes the adenylation of 2,3-dihydroxybenzoic acid, followed by its transfer to the phosphopantetheine arm of holo-EntB, an aryl carrier protein. In the absence of EntB, EntE catalyzes the formation of Ap4A, a molecule that is implicated in regulating cell division during oxidative stress. We propose that the expression of EntE during iron starvation produces Ap4A to slow growth until intracellular iron stores can be restored.
Adenylylated bis(5′-nucleosidyl) tetraphosphates (Ap4N, where N is A, G, C, or U) have been proposed to act as signaling molecules in a variety of biological systems in both prokaryotes and eukaryotes (1). The cellular concentration of one such molecule, Ap4A (P1, P3-diadenosine-5′-tetraphosphate), has been shown to rapidly increase when Escherichia coli are exposed to heat shock or oxidative stress (1, 2). Subsequent studies later revealed that Ap4A binds to several E. coli stress-inducible proteins, including DnaK, GroEL, E89, C45, and C40, thus suggesting Ap4A may serve as a modulator of cellular stress (3). Evidence for this role has been further supported by a recent study that investigated the nature of putative Ap4A interactions with E. coli molecular chaperone GroEL (4). At temperatures >37°C, GroEL chaperone activity is known to be decreased in favor of its protein storage function (5). However, Tanner and coworkers showed that the binding of Ap4A to GroEL allosterically modulates the chaperone's function, resulting in the promotion of chaperoning activities over protein storage activity at higher temperatures. This observation suggests that Ap4A may play a role to sustain basic cell physiology and metabolism during stress, then immediately post-stress, aid in the restoration of normal cellular function (4, 6). Although the exact role of Ap4A has yet to be elucidated, the identification of its multiple binding partners, all of which are stress-inducible proteins, suggests that this dinucleoside polyphosphate may be an important modulator of metabolism during cellular stress.
The intracellular concentration of Ap4A increases during heat shock (~10-fold) and oxidative stress (~100-fold) (1, 7). Blanquet and co-workers were the first to identify E. coli aminoacyl-tRNA synthetases (aaRSs) as the enzymes capable of the in vivo synthesis of Ap4N (8). Specifically, lysyl-, methionyl-, phenylalanyl-, and valyl-tRNA synthetases have been shown to divert from amino acyl-tRNA production to that of Ap4A when in the presence of ZnCl2, amino acid, and pyrophosphatase (7, 8). Twenty years since that discovery, aminoacyl-tRNA synthetases remain as the only enzymes in E. coli with known Ap4N synthase activity. Ap4A phosphorylase, an enzyme capable of synthesizing Ap4A when acting in reverse, has been identified in yeast; however this enzyme has yet to be identified in E. coli (9). When considering the changing, and at times quite abundant, concentration of Ap4A in E. coli, it seems plausible that other E. coli enzymes capable of Ap4A synthesis may exist.
When E. coli are starved for the essential nutrient iron, the bacteria respond by synthesizing and secreting enterobactin (tris-(N-(2,3-dihydroxybenzoyl)serine trilactone), which chelates extracellular ferric iron and imports it back into the cell to provide iron for metabolic processes (11). E. coli enterobactin synthetase is a non-ribosomal peptide synthetase composed of six genes (entA-F) of known biochemical function: EntA, -B, and −C divert the central metabolite chorismate to generate 2,3-dihydroxybenzoate (DHB), while EntB, -D, -E, and -F catalyze the ATP-dependent assembly of enterobactin from three molecules each of DHB and L-serine (12). Inhibition of this essential pathway represents a promising strategy for antibacterial drug development and therefore we began a detailed kinetic and mechanistic characterization of EntE, a key component of the synthetase system. EntE ligase catalyzes the ATP-dependent transfer of DHB onto the phosphopantothenylated cofactor that is bound to the aryl carrier protein domain (ArCP) of EntB to yield the covalently arylated EntB. This molecule then serves as the aryl donor for amide bond formation in the final assembly of enterobactin (13). EntE belongs to the family of aryl acid adenylating enzymes (AAAE) that are characterized by a two-step adenylation/ligation reaction (14) (Scheme 1). In the adenylation half-reaction, the enzyme catalyzes the condensation of DHB and ATP to form an adenylate intermediate (15). We sought to use pre-steady state approaches to determine the rate of adenylation for EntE, however the results of this analysis were quite unexpected. Here we describe the EntE catalyzed formation of Ap4A.
As described for other enzymes (e.g., M. smegmatis cysteine ligase, M. tuberculosis pantothenate synthetase), the rate of adenylation in the first half-reaction can be determined by single-turnover experiments using rapid-quench flow techniques with [α-33P]ATP in the absence of the third substrate (16, 17). The expected products of this reaction, radiolabeled adenylate and/or AMP, can be resolved from ATP on a PEI-TLC plate after separation with 0.9 M guanidine as the mobile phase. Reactions mixtures containing either DHB or salicylic acid, ATP, and EntE resulted in a variety of radiolabeled spots on the TLC plate, including one which did not correspond to the location of expected products, adenylate or AMP (Figure S1, Supplementary Materials). In order to determine the identity of the radiolabeled products, we followed the adenylation reaction using 31P NMR (see Supporting Information for experimental details). Comparison of the 31P NMR spectrum from reaction mixtures with and without added EntE revealed that an unexpected product was indeed formed (Figure 1). Analysis of the chemical shift positions and peak integrations allowed us to hypothesize that the identity of this product might be Ap4A. The 31P NMR spectrum is identical to commercially available Ap4A.
In order to confirm that EntE, in the absence of the third substrate (phosphopantothenylated EntB, ArCP), was catalytically producing Ap4A, we sought to isolate the product and confirm its identity via ESI-MS and 31P NMR. The products of the reaction were separated by HPLC using a Mono Q ion exchange column with a gradient of 0.01 to 0.8 M ammonium bicarbonate (see Supporting Information for details). The peak corresponding to Ap4A was collected, lyophilized overnight, and dissolved in water for ESI-MS or 10% D2O for 31P NMR analysis. Using commercial Ap4A as a standard, ESI-MS unambiguously confirmed the identity of the EntE reaction product to be Ap4A (Figure S2, Supplementary Materials) (see Supporting Information for experimental details). Furthermore, 31P NMR analysis was repeated on the purified product and resulted in chemical shifts and splitting patterns corresponding to the resonances of the unique peaks we initially observed (Figure 1C). Validation by both 31P NMR and ESI-MS allowed us to confirm the EntE-catalyzed formation Ap4A.
Considering the rapid increase of intracellular Ap4A concentration during conditions of stress, it was essential to evaluate the kinetics of this reaction as well as the amount of Ap4A formed catalytically (1, 2). The concentration of Ap4A produced, using either 0.1 mM salicylic acid or DHB as the aryl acid substrate, was determined by HPLC separation of Ap4A and ATP and quantitation by peak integration and comparison to a standard curve (see Supporting Information for experimental details). The rate of Ap4A formation is nearly identical when either salicylic acid or DHB is used in the presence of 1 mM ATP (Figure 2). The maximal amount of Ap4A produced appears to be ~80 μM after 16 hours (Figure 2). The initial rate, calculated from 0-30 min, was 7.2 × 10-4 s-1 compared to the kcat for the overall reaction of 2.8 s-1. The maximal amount of Ap4A formed was not equal to the concentration ATP (1 mM) suggesting that Ap4A may act as an inhibitor of the reaction. When Ap4A was tested as an inhibitor versus ATP, a competitive inhibition pattern was observed, yielding a Ki value of 1.2 ± 0.1 μM (Figure S3). The much tighter binding of Ap4A than ATP (Km ~ 400 μM) suggests that this molecule may bind to both the ATP and aromatic acid sites on the enzyme.
In order to determine if the formation of Ap4A occurs while the DHB-adenylate is enzyme-bound or whether the reaction occurs in solution after release of the adenylate, we performed competition studies with the normal arylation substrate EntB. In the presence of 10 μM holo-EntB-ArCP, the rate of formation of Ap4A is significantly inhibited (Figure 2, inset), even after all the holo-EntB-ArCP is arylated at these high EntE concentrations. This is most consistent with the DHB-adenylate being tightly bound, and reacting with ATP while bound in the absence of holo-EntB-ArCP.
We propose that, after aryl acid-adenylate formation in the first half-reaction, the γ-phosphate of ATP acts as nucleophile to attack the α-phosphate group of the enzyme-bound adenylate resulting in production of Ap4A and regeneration of the free aryl acid (Scheme 2). In contrast, the normal reaction occurs between the thiol of the pantothenylated (holo-) EntB-ArCP and the carbonyl group of the mixed carboxyl-phosphoric anhydride to generate the thioester. This differing regioselectivity of reactions between the common adenylate substrate and either ATP or holo-EntB-ArCP, that is, the P-O versus C-O chemistry shift, is also known for the Ap4A formation by aminoacyl tRNA synthetases (7).
Previous studies have shown that the in vivo Ap4A concentration in un-stressed wild-type E. coli cells is ~1-3 μM (3). This concentration of Ap4A quickly increases to ~100-160 μM when the cells experience environmental stress, such as heat shock or oxidative stress (18). Conversely, cellular Ap4A concentrations have not been shown to increase in cells experiencing a variety of metabolic conditions, including deficiencies in nucleic acids, amino acids, fatty acids, carbon, nitrogen, phosphate, or oxygen. Interestingly, iron is absent from this list (7). Our data, as described in this report, suggests that Ap4A may be catalytically produced in E. coli during iron-limiting conditions when the ent genes are expressed. Furthermore, we show that at least 80 μM of Ap4A is catalytically produced by EntE, a concentration that is significantly increased from that of the un-stressed state and in the concentration range proposed to elicit biological responses in cells (1, 2, 19).
Until now, aminoacyl-tRNA synthetases were the only known enzymes capable of Ap4A formation in E. coli (10). However, how these constitutively expressed enzymes would be able to rapidly respond catalytically to oxidative stress and heat shock is unclear. In this report, we demonstrate that the E. coli enzyme EntE catalytically produces Ap4A in the absence of its third substrate. The synthesis of Ap4A catalyzed by EntE provides new implications for the E. coli enterobactin synthetase system. Under iron-limiting conditions, the genes that produce both DHB and the enzymes that generate enterobactin will be expressed. Once produced, and EntB is posttranslationally modified, they will initiate the synthesis and export of enterobactin will begin. If pantothenylation is slow compared to DHB synthesis and EntE production, then the system is perfectly poised to synthesize Ap4A. This product will signal the cell of the iron-limiting condition and slow growth while iron is scavenged from the surroundings. Thus, the synthesis of Ap4A by EntE in concert with enterobactin assembly may provide additional aid to iron-depleted cells such that, while enterobactin is scavenging ferric iron, Ap4A is acting to moderate cellular activities until sustainable iron concentrations are restored (4, 6). The in vivo demonstration that Ap4A is produced under iron limiting conditions is presently under investigation.
The authors wish to thank Edward Nieves for performing mass spectrometry, Dr. Courtney Aldrich for the entB and entE clones, and Dr. Jun Yin for the B. subtilis sfp clone. This work was supported by grants from the NIH: AI60899 (to JSB) and T32 GM08572 (ALS). The instrumentation in the AECOM Structural NMR Resource is supported by the Albert Einstein College of Medicine and in part by grants from the NSF (DBI9601607 and DBI0331934), the NIH (RR017998), and the HHMI Research Resources for Biomedical Sciences.