Phosphorylation and Dephosphorylation Catalyzed in Vitro by Purified Components of the Nitrate Sensing System, NarX and NarL*

The regulation of specific gene expression by nitrate in Escherichia coli is mediated by the NarX/NarQ-NarL system. Based on sequence homologies with a family of two-component regulatory systems in bacteria, NarL has been identified as a putative response regulator while NarX and NarQ were proposed to be alternative membrane-associated sensors that activate NarL in the presence of nitrate. To investigate the interaction of NarX and NarL in vitro, both proteins were purified from overproducing strains. Purified NarX was rapidly labeled when incubated with [yS2P] ATP but not with [a-"P]ATP in a reaction that re- quired M e + but was unaffected by nitrate. Incubation of the labeled NarX with purified NarL resulted in the transient phosphorylation of NarL. Both the phosphorylation and dephosphorylation of NarL required M8+, and neither reaction was affected by the pres- ence of nitrate. NarL-phosphate, stabilized by the addition of EDTA, ran as a monomer on gel filtration. Dephosphorylation of the isolated NarL-phosphate required the addition of both M 8 + and the NarX protein. The relative stabilities of the phosphorylated forms of the two proteins at different pH values were consistent with the proposal that, in analogy to other related two-component regulatory systems, NarX and NarL were phosphorylated on specific histidine and aspartate residues,

The regulation of gene expression by nitrate in Escherichia coli is mediated by the NarX/NarQ-NarL system (1,2). Based on sequence homologies ( 3 4 , this system belongs to a family of two-component regulatory systems in bacteria that are activated by a phosphorylation mechanism and regulate gene expression or other activities in response to specific environmental stimuli (6-8). NarL corresponds to the family of DNA binding response regulators while NarX and NarQ are closely related proteins, corresponding to membrane sensor components. Genetic experiments suggest that in the presence of nitrate NarX and NarQ are alternative activators of NarL (1, 2).
Although specific cis-acting sequences have been identified *This work was supported by Public Health Service Grant GM19511 from the National Institute of General Medical Sciences. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
in NarL-responsive promoters (9, lo), it has not been possible to demonstrate with crude extracts either the phosphorylation of NarL or the specific binding of NarL to these sequences. Therefore, to establish conditions for the activation of NarL in uitro we purified both NarX and NarL from overproducing strains and demonstrate here that phosphorylation and dephosphorylation of NarL are mediated by NarX.

EXPERIMENTAL PROCEDURES
Construction of Plasmids-The host strain for all plasmids was MV1190 (A(lac-pro AB), thi, sup E, A(sr1-rec A) 306::TnlO (tef) [F':tra D36,pro AB, lac Iq ZAM151) obtained from Bio-Rad. Plasmids for overproducing NarX and NarL were constructed by first inserting a 3.6-kilobase Dm1 to HindIII fragment from pSR9 ( l l ) , containing both the nurX and the nurL genes (41, into the polylinker region of pTZ18R (Pharmacia LKB Biotechnology Inc.) between the XbaI site, which had been filled after cutting, and the HindIII site to give pMW74. To overproduce NarX, plasmid pMW747 was constructed by inserting an EcoRI to BamHI fragment containing the tac promoter from the plasmid pDR540 (12) between the EcoRI and BamHI sites in pMW74 and inserting an XbaI linker with stop codons in each frame between the BglII sites in m r L to terminate translation of the nurL gene after the first 25 codons. To overproduce NarL, the Sty1 site in pMW74, 8 base pairs upstream of the NarL start codon, was converted to a BamHI site, and the EcoRI to BamHI tac promoter fragment was inserted as above to give pMW745.
Purification of NarX and NarL-Overproduced NarX was purified from a l-liter culture of MV1190(pMW747) grown aerobically to 100 Klett units in L-broth containing ampicillin (100 pg/ml). 0.5 mM isopropyl-P-D-thiogalactopyranoside (IPTG)' was added, and after 4 h of further aerobic growth the cells were harvested, washed, and frozen. Membranes, prepared as described previously ( l l ) , were suspended in 7 ml of 50 mM Tris-HC1, pH 8.0, 1 mM dithiothreitol (DTT), 0.1 mM phenylmethylsulfonyl fluoride, and 2% Triton X-100. After 1 h on ice the suspension was centrifuged for 15 min at 17,000 X g. The supernatant was added to a 1.5 X 25-cm DEAE-Sephacel column equilibrated with 50 mM Tris-HC1, pH 8.0, 1 mM DTT, and 0.1% Triton X-100 and eluted with a 400-ml gradient from 0.1 to 0.4 M NaCl in the same buffer. Fractions containing NarX were identified by assaying for phosphorylation of the overproduced 67-kDa protein.
Overproduced NarL was purified from the cell pellet obtained from a 2-liter culture of MV1190(pMW745) grown and induced as above. The pellet was suspended and fractionated into membranes and a cytoplasmic fraction (11). The cytoplasmic fraction was added to a 2.5 X 15-cm DEAE-Sephacel column equilibrated with 50 mM Tris-HCI, pH 8.0, and 1 mM DTT and was eluted with a 400-ml gradient from 0 to 0.2 M NaCl in the same buffer. Fractions containing NarL were identified by SDS-polyacrylamide gel electrophoresis, the peak tubes were pooled, and protein was precipitated by 60% ammonium sulfate. The precipitate was dissolved in 1 ml of 50 mM Tris-HC1, pH 8.0, 1 mM DTT, and 5% glycerol, and 0.4 ml was added to a 1 X 55cm Sephadex G-75 column equilibrated and eluted with the same buffer. Fractions containing NarL were identified as above.

RESULTS AND DISCUSSION
In preliminary experiments with a strain that overproduced both NarX and NarL, we were unable to detect the phosphorylation of either NarX or NarL with [-p3'P]ATP in crude extracts. However, with strains that overproduced NarX alone, phosphorylation of a protein with the expected size of NarX was detected in membrane preparations, and this phosphorylation appeared to be suppressed by the addition of a soluble extract containing overproduced NarL.
T o facilitate the purification of NarX and NarL away from possible interfering activities, plasmids were constructed with each of the components under the control of the tac promoter (12), so that each could be induced to high levels by the addition of IPTG to growing cells. NarX, solubilized from the membrane fraction of strain MV1190(pMW747), was purified on DEAE-Sephacel by following the autophosphorylation activity with [y3*P]ATP (Fig. 1, A and B ) . Purified NarX migrated to a position on SDS-polyacrylamide gel electrophoresis corresponding to 67 kDa, and N-terminal sequence analysis yielded a sequence through 10 residues that was consistent with the deduced narX gene product (4). The Molecular weight markers for A , B, and C were prestained phosphorylase b (96,000), glutamate dehydrogenase (55,000), lactate dehydrogenase (36,000), carbonic anhydrase (29,000), @-lactoglobulin (18,400), and cytochrome c (12,400) from Diversified Biotech. (Fig. lB, lane 4 ) . The phosphorylation of the purified NarX, as well as the 36-kDa protein, was dependent on M C ions. Neither protein was labeled with [cY-~'P]ATP, and the addition of nitrate to the incubations had no effect on the rate or extent of phosphorylation of NarX with [y-"PIATP.
NarL was purified from the cytoplasmic fraction of strain MV1190(pMW745) by following the overproduced protein on SDS-polyacrylamide gels (Fig. IC). The purified protein migrated as an apparent 24-kDa monomer on Sephadex G-100, exhibited no autophosphorylating activity, and yielded 20 residues by N-terminal sequence analysis that were identical to residues 2-21 of the sequence deduced for the narL gene product (3, 4).
The phosphorylated form of purified NarX, produced by incubation with [y3'P]ATP and then separated from the ATP by gel filtration, was stable for over 60 min when incubated with M e at 25 "C (Fig. 2, lanes 1-7). When purified NarL was added (lanes 8-14), the 32P rapidly disappeared from the NarX band accompanied by the phosphorylation and then rapid dephosphorylation of NarL. Significantly, the phosphorylated 36-kDa component that contaminated the NarX preparation was stable in both incubations, indicating that it did not contribute to either the phosphorylation or the dephosphorylation of NarX or NarL. Nitrate had The reaction mixture was chromatographed on a 0.7 X 7.5-cm Sephadex G-50 column equilibrated with 50 mM Tris-HCI, pH 8.0, 1 mM DTT, and 0.1% Triton X-100 and eluted with the same buffer. 77 pl of phosphorylated NarX, essentially free of [y"P]ATP, was incubated in 105 pl of TEGD containing 5 mM MgC12 with or without NarL (2 p~) , and 15-pl aliquots were withdrawn at the indicated times, added to 4 p1 of sample buffer, and subjected to SDS-polyacrylamide gel electrophoresis. Autoradiograms show the incubations without NarL (lanes [1][2][3][4][5][6][7] and with NarL (lanes [8][9][10][11][12][13][14].  The reactions were stopped after 10 min by the addition of sample buffer and added to a SDS-polyacrylamide gel.
The gel was dried and autoradiographed.
no effect on the rates of the phosphorylation or dephosphorylation of NarL while EDTA (50 mM) inhibited both the transfer of phosphate from NarX to NarL and the subsequent dephosphorylation of NarL (data not shown).
To compare the relative properties of the phosphorylated forms of NarX and NarL, samples of an incubation mixture that contained the phosphorylated forms of both proteins were incubated a t different pH values after the addition of SDS to terminate the reaction (Fig. 3). The phosphorylated forms of both proteins were stable a t neutral pH but were completely dephosphorylated at pH 1. At pH 11.5 and 13.5 the phosphorylated form of NarX was relatively stable while the NarL-phosphate was completely dephosphorylated.
On the basis of genetic and site-directed mutagenesis studies (2,14) and by analogy to the mechanism involved in the activation of several related two-component regulatory systems (3-5), it has been proposed that NarX acts as an autokinase that phosphorylates itself on a specific histidyl residue and then transfers the phosphoryl group to a specific aspartyl residue on NarL. The relative stability of the phosphorylated forms of NarX and NarL is consistent with this hypothesis (15).
The phosphorylated form of NarL, stabilized by the addition of EDTA after transfer of the phosphate group from NarX, ran as a monomer on a Sephadex G-100 column and was readily separated from the NarX protein. When incubation mixtures containing phosphorylated NarL and no EDTA were run on the same column only a small amount of phosphorylated NarL was recovered, and it all eluted at a position that corresponded to the monomer.
Purified NarL-phosphate was stable when incubated a t 37 "C in the presence of either Mg2+ ions or the purified NarX fraction, but it was rapidly dephosphorylated when both were added (Fig. 4). Longer incubation times (100 min) revealed that in the presence of M$+ ions NarL-phosphate undergoes a slow but significant rate of dephosphorylation that is dramatically enhanced by the addition of NarX. These results demonstrate that NarX catalyzes the dephosphorylation as well as the phosphorylation of NarL and support the conclusion of Collins et al. (16) that mutations in the narX gene that lead to a constitutively induced phenotype likely result in the loss of specific phosphatase activity of the NarX protein.
As predicted by analogy to other two-component regulatory systems studied in bacteria (3-5), these studies demonstrate that NarX acts as an autokinase that transfers the resulting phosphoryl group to NarL and, as found with several other systems, it also enhances the rate of dephosphorylation of NarL-phosphate. NarL appears to be monomeric in both the unphosphorylated and phosphorylated forms. The isolation of stable NarL-phosphate should facilitate DNA binding studies and the determination of the role of NarL in the regulation of specific gene transcription.
Although it is assumed that the activity of NarX in the whole cell is regulated in some way by nitrate, we were unable to detect any effect of nitrate on the activities of purified NarX. A full understanding of how nitrate regulates the phosphorylation will most likely require identification or reconstitution of a nitrate-dependent form of NarX in the cell membrane.