Allosteric regulation of nitrate transporter NRT via the signaling protein PII

Significance

The homeostasis of carbon and nitrogen metabolisms, which is fundamental to all living organisms, needs to be precisely controlled. The uptake of nitrate in cyanobacteria is conducted by an ABC (ATP-binding cassette)-type nitrate transporter NrtABCD, the activity of which is proposed to be regulated by the signaling protein PII. Our structural and biochemical analysis elucidated the fine mechanism on how PII regulates the activity of an ABC transporter via directly binding to the regulatory domain of NrtC, giving an example of an asymmetric regulatory mode of ABC transporter. More importantly, these findings provide structural insights into a different mode of PII interaction with the partners as well as the carbon and nitrogen homeostasis control tuned by PII signaling networks.

Abstract

Coordinated carbon and nitrogen metabolism is crucial for bacteria living in the fluctuating environments. Intracellular carbon and nitrogen homeostasis is maintained by a sophisticated network, in which the widespread signaling protein PII acts as a major regulatory hub. In cyanobacteria, PII was proposed to regulate the nitrate uptake by an ABC (ATP-binding cassette)-type nitrate transporter NrtABCD, in which the nucleotide-binding domain of NrtC is fused with a C-terminal regulatory domain (CRD). Here, we solved three cryoelectron microscopy structures of NrtBCD, bound to nitrate, ATP, and PII, respectively. Structural and biochemical analyses enable us to identify the key residues that form a hydrophobic and a hydrophilic cavity along the substrate translocation channel. The core structure of PII, but not the canonical T-loop, binds to NrtC and stabilizes the CRD, making it visible in the complex structure, narrows the substrate translocation channel in NrtB, and ultimately locks NrtBCD at an inhibited inward-facing conformation. Based on these results and previous reports, we propose a putative transport cycle driven by NrtABCD, which is allosterically inhibited by PII in response to the cellular level of 2-oxoglutarate. Our findings provide a distinct regulatory mechanism of ABC transporter via asymmetrically binding to a signaling protein.

Carbon and nitrogen (C and N) are two most abundant nutrient elements in living organisms, which constitute the major components of biomacromolecules, including carbohydrates, proteins, and nucleic acids. Moreover, besides their importance for building up essential organic molecules, biological assimilations of carbon and nitrogen play an important role in the biogeochemical cycling of elements on Earth and therefore have a key impact on global climate change (1). Autotrophic organisms, such as plants and microalgae, are capable of incorporating inorganic carbon and nitrogen into the biomass, thus driving the main biological production on Earth (2, 3). The homeostasis of carbon and nitrogen metabolisms, which is fundamental to all forms of living organisms to acclimate to fluctuating environments, needs to be precisely controlled. The photosynthetic cyanobacteria, which usually maintain an N/C ratio at ~1:5 in the cell (4), have evolved a sophisticated mechanism to finely tune the N/C balance, including the regulations at the transcriptional and metabolic levels (5).

Cyanobacteria are capable of utilizing various nitrogen sources, including ammonium, nitrate, nitrite, urea, and even atmospheric nitrogen, but ammonium is favorite (6, 7). The presence of ammonium promptly inhibits the transcription of relevant genes and uptake of other nitrogen sources (8). Once the environmental ammonium is deprived or consumed, cyanobacteria can utilize alternative nitrogen sources by activating the corresponding assimilation pathways. The nitrate, which is the most abundant nitrogen source in nature, could be assimilated through the nitrate assimilation pathway encoded by the nirA operon, which is activated by the coordinated action of the global regulator NtcA (9, 10) and the context-specific regulator NtcB (8, 11, 12). The pathway starts with an active nitrate transporter that facilitates nitrate uptake, followed by a nitrate reductase NR and a nitrite reductase NiR that catalyze the two steps of reduction, respectively (7).

Two types of nitrate transporters have been identified in cyanobacteria, the ATP-binding cassette (ABC) transporter NrtABCD (also termed NRT) found mainly in freshwater cyanobacteria (5, 8, 13) and the nitrate/nitrite permease NrtP in the major facilitator superfamily mostly identified in marine cyanobacteria (5, 8, 14). NRT is a bispecific nitrate/nitrite transporter showing high affinity toward both substrates (15). As a type I ABC importer, NRT consists of five subunits: a periplasmic substrate-binding protein (SBP) NrtA that captures the substrates, two identical transmembrane domains (TMDs) NrtB, and two cytosolic nucleotide-binding domains (NBDs) NrtC and NrtD responsible for ATP binding and hydrolysis. Distinct from the canonical ABC transporters (16–18), NrtC possesses an additional C-terminal regulatory domain (NrtC-CRD) fused to its NBD (NrtC-NBD). NrtC-CRD shows a sequence identity of ~30% with NrtA that specifically captures extracellular nitrate/nitrite (19). Deletion of NrtC-CRD has been found to make NRT partly gain the nitrate uptake activity in the presence of ammonium (20, 21).

Previous pull-down assays and bacterial two-hybrid analysis indicated that the signaling protein PII directly interacts with both the NrtC and NrtD subunits of NRT (22). As one of the largest families of signaling proteins in nature (22, 23), PII regulates multiple carbon and nitrogen metabolic processes by sensing the intracellular energy status and N/C level. Canonical PII proteins are homotrimers, each subunit of which is featured with a long extended T-loop that adopts various conformations to recognize different metabolic effectors and/or partner proteins (5, 22). Typically, the ATP- or ADP-bound PII preferentially recognizes partner proteins; however, synergistic binding of 2-oxoglutarate (2-OG) and ATP induces the T-loop bending toward the core structure, leading to the release of partner proteins (5). It was found that nitrate uptake is not subjected to ammonium inhibition in PII-deficient mutants in Synechococcus sp. PCC 7942 and Synechocystis sp. PCC 6803 (21, 24), which possess a phenotype similar to the NrtC-CRD deletion strain (20, 21). These findings strongly indicated that PII might regulate the nitrate uptake activity of NRT, but the fine regulatory mechanism remains elusive.

Using single-particle cryoelectron microscopy (cryo-EM), we solved three structures of NrtBCD complexed with nitrate, ATP, and PII, respectively. Combined with biochemical and genetic analyses, we propose a finely tuned model of NRT transport regulated by interactions between NrtC and the lateral part of PII, representing a mode of PII interaction with the partner proteins. These findings provide a distinct regulatory mechanism of ABC transporters, and also increase our understanding on N/C homeostasis control in cyanobacteria.

Results

Structure of NrtBCD in Complex with the Substrate Nitrate.

To investigate the transport and regulatory mechanisms of NRT, we initially focused on the structural and functional studies of NRT from Synechocystis sp. PCC 6803 (termed Synechocystis for short), as more molecular genetic information available for the NRT in this strain (19, 21, 22). However, we failed in overexpressing the Synechocystis NrtBCD complex in Escherichia coli, which impeded our structural studies. Instead, we screened the NrtBCD homologs from various cyanobacterial strains, and finally, we successfully expressed and purified Anabaena sp. PCC 7120 (termed Anabaena for short) NrtBCD (SI Appendix, Fig. S1A). Notably, the SBP is usually flexibly tethered to the membrane and dynamically interacts with the TMDs; thus, NrtA was not prepared in our complex. Using single-particle cryo-EM, we solved the structure of Anabaena NrtBCD at 3.54 Å resolution (Fig. 1A and SI Appendix, Figs. S2 and S3), with NrtC-CRD invisible in the cryo-EM map.

Fig. 1.

Overall structure of NrtBCD-NO₃⁻ and the substrate translocation channel in TMDs. (A) Cartoon representation of NrtBCD-NO₃⁻. The nitrate is shown in spheres, and the membrane plane is indicated as the gray rectangle. The number of the transmembrane helices of one NrtB subunit is labeled. (B) The substrate translocation channel in TMDs. The nitrate is shown as ball-and-stick models with cryo-EM map density contoured at 15σ shown as a blue mesh. The key residues along the channel are shown as ball-and-stick models. (C) Multiple sequence alignment of residues along the substrate translocation channel of cyanobacterial NrtB homologs, including Ana_7120 (Anabaena sp. PCC 7120), Cyl_CS-505 (Cylindrospermopsis raciborskii CS-505), Syn_6803 (Synechocystis sp. PCC 6803), Cya_8801 (Cyanothece sp. PCC 8801), Cya_7424(Cyanothece sp. PCC 7424), Mic_NIES-843 (Microcystis aeruginosa NIES-843), Syn_JA-3-3Ab (Synechococcus sp. JA-3-3Ab), Art_NIES-39 (Arthrospira platensis NIES-39), Art_CS-328 (Arthrospira maxima CS-328), The_BP-1 (Thermosynechococcus elongatus BP-1), Glo_7421 (Gloeobacter violaceus PCC 7421), and Syn_7942 (Synechococcus sp. PCC 7942). The conserved leucine and arginine residues are marked by gray and blue dots, respectively. (D) Relative ATPase activities of wild-type NrtBCD and mutants upon addition of 8 mM nitrate. Each data point is the average of three independent experiments (n = 3), and error bars represent the means ± standard deviation. One-way analysis of variance is used for the comparison of statistical significance of mutants and wild-type. The P values of <0.0001 are indicated with ****.

The overall structure of NrtBCD adopts an inward-facing conformation (Fig. 1A). It has a canonical ABC-type architecture (17), consisting of two transmembrane subunits of NrtB and two cytosolic NBD subunits, NrtC and NrtD. The two NrtB subunits, each possessing six TM helices (Fig. 1A and SI Appendix, Fig. S1B), pack against each other to form a swapped dimer, with TM1 and TM6 of one subunit extending and attaching to the opposite subunit (Fig. 1A). Besides the six TM helices, NrtB possesses three additional α-helices: One is located at the periplasmic side, whereas the other two, including the conserved coupling helix, are located at the cytosolic side (Fig. 1A). The two NBDs form a pseudosymmetric dimer (SI Appendix, Fig. S1C), each subunit of which adopts a classical NBD fold of ABC transporter composed of two subdomains (RecA-like and α-helical subdomains) harboring the conserved Walker A, Walker B, and ABC signature motifs (SI Appendix, Fig. S1C), in addition to the so-called D- and H-loops (16–18).

The two NBDs in the “semi-open” dimer interact with each other mainly through the hinge helix of NrtC (residues Ser242-Ser269) and the C-terminal α-helix of NrtD (SI Appendix, Fig. S1C), accounting for a total interface area of ~1,000 Ų. The α-helical subdomains of two NBDs are separated from each other, whereas the TMDs in the inner membrane leaflet (IML) are open toward the intracellular side, resulting in an inward-facing conformation of NrtBCD. Meanwhile, the TMDs in the outer membrane leaflet (OML) still contact each other, with a buried interface area of ~3,100 Ų through the interactions of TM3 and TM5 with their counterparts of the opposite TMD (Fig. 1A).

In this structure, we observed an obvious density at a cavity located in the OML, which is formed by TM3 and TM5 of one TMD and their counterparts of the opposite TMD (Fig. 1B). A molecule of the substrate nitrate could be well fitted into this density within the cavity, which is highly hydrophobic and contains residues Pro131 and Leu132 from TM3, as well as Leu216 and Ala217 from TM5 (Fig. 1B). The nitrate is proposed to be transiently trapped within this hydrophobic cavity that constitutes a substrate translocation channel running along the central twofold axis of the TMD dimer, which is formed by TM3 and TM5 in the upper part and TM4 and TM5 in the lower part (Fig. 1B).

In this hydrophobic cavity, two pairs of leucine residues, Leu132–Leu132’ (residues from the opposite subunit are labeled with a prime) and Leu216–Leu216’ from the TMDs, located close to the periplasmic side of the membrane and separated by 4.0 and 3.6 Å, respectively (Fig. 1B), likely act as gatekeepers that govern substrate passage. Moreover, several basic residues in the IML along the channel, including Arg127 in TM3 and Arg208 in TM5, form a hydrophilic cavity (Fig. 1B). These positively charged residues, which are complementary to the negatively charged substrates nitrate/nitrite, might facilitate the substrate translocation. Sequence analysis revealed that all these residues are highly conserved among NRT homologs (Fig. 1C).

To investigate the role of these conserved residues, we mutated each residue to alanine and compared their ATPase activities in the presence or absence of nitrate. The addition of nitrate up to 8 mM significantly increased the ATPase activity of wild-type NrtBCD by ~2.5 folds (Fig. 1D and SI Appendix, Fig. S1D), consistent with the previous reports that substrates usually stimulate the ATPase activity of ABC transporters (25–27). In contrast, addition of nitrate to the mutants failed to stimulate their ATPase activities, suggesting that these mutants no longer respond to the substrate nitrate (Fig. 1D). These findings demonstrated that these residues along the channel, including Leu132, Leu216, Arg127, and Arg208, are indispensable for the substrate recognition of NRT.

NrtBCD Adopts an Outward-Facing Occluded Conformation upon ATP Binding.

It is known that binding of ATP to type I ABC transporters could help reveal how the substrate is translocated and released in the transport cycle (16, 28). Accordingly, we mutated the catalytic glutamate residues (Glu164 in NrtC and Glu179 in NrtD) in two NBDs to glutamine, and pre-incubated ATP and nitrate with the purified NrtBCD. Eventually, we solved the structure of NrtBCD in complex with ATP (NrtBCD-ATP) at 3.10 Å resolution (Fig. 2A and SI Appendix, Figs. S4 and S5). However, the substrate nitrate is absent in this structure.

Fig. 2.

Overall structure of NrtBCD-ATP and conformational changes upon ATP binding. (A) Cartoon representation of NrtBCD-ATP. The Mg²⁺ ions and ATP molecules are shown in green spheres and ball-and-stick models, respectively. (B) Superposition of TMDs in the structures of NrtBCD-NO₃⁻ (gray) and NrtBCD-ATP (green and yellow). All α-helices are displayed as cylinders. The distances between the two coupling helices in the two structures are shown. (C) A zoom-in view of the superimposed substrate translocation channel in TMDs. The key residues along the channel are shown as ball-and-stick models.

In the structure, two ATP-Mg²⁺ molecules symmetrically bind to the cleft between Walker A motif of one NBD and ABC signature motif of the opposite NBD, which triggers the closure of the NBD dimer. This conformational change in the NBDs is further transmitted to TMDs via a pair of coupling helices, the distance of which is reduced from 24 to 13 Å (Fig. 2B). Consequently, the TMDs move closer to each other in the IML, whereas the two TM5 helices slightly separate in the OML, resulting in the TMDs partly open toward the extracellular side (Fig. 2 B and C). These rearrangements make the TMDs exhibit an outward-facing occluded conformation, with the translocation channel closed at both the intra- and extracellular sides (Fig. 2C).

The Structure of NrtBCD Complexed with PII.

Previous studies have demonstrated that the signaling protein PII also participates in the regulation of NRT activity (21, 22, 24), most likely via direct interactions with the intracellular NBDs of NRT (22). To further investigate the binding of PII to NrtBCD in vitro, we performed pull-down assays of NrtBCD against PII in the presence of nitrate, adenosine diphosphate (ADP), adenosine triphosphate (ATP), and/or 2-OG. The results showed that PII binds to NrtBCD upon the addition of ADP or ATP, and is independent of the presence of nitrate (Fig. 3A). In contrast, simultaneous addition of both 2-OG and ATP significantly decreased the yield of PII pulled down from NrtBCD (Fig. 3A). It indicated that 2-OG prevents PII from binding to NrtBCD, in agreement with the general interaction pattern of canonical PII and partner proteins (5).

Fig. 3.

Structure of NrtBCD-PII and the interactions between NrtC and PII. (A) The pull-down assays of NrtBCD against PII in the presence of 10 mM MgCl₂ and various effectors, including ADP (lane 2), ADP and NaNO₃ (lane 3), ATP (lane 4), and ATP and 2-OG (lane 5). The NrtC protein is fused with a FLAG tag. The protein marker is loaded in lane 1. (B) Cartoon representation (Left) and cryo-EM map (Right) of the NrtBCD-PII complex. Different subunits and domains are colored as in Figs. 1A and 2A. NrtC-CRD is colored bright orange, and three PII subunits are colored slate, green, and cyan, respectively. Five ADP molecules are colored gray and shown as ball-and-stick models. (C) The interfaces between NrtC and PII. The NrtC-NBD, NrtC-CRD, and the hinge helix are colored orange, bright orange, and red, respectively. The two distinct interfaces (labeled as I and II) are outlined by blue boxes, and the detailed interactions are shown in the Insets. The interacting residues are shown as ball-and-stick models, and the ADP molecule is colored gray.

To obtain the structure of NrtBCD complexed with PII, we pre-incubated the purified PII and NrtBCD in the presence of ADP and nitrate, which was subjected to cryo-EM analysis (SI Appendix, Figs. S6 and S7). After extensive trials, we solved the 3.53 Å structure of NrtBCD complexed with PII (NrtBCD-PII), which binds to five ADP molecules: two at the NBDs and three at the PII trimer (Fig. 3B). The ADP-bound PII trimer binds to the lateral side of NrtC via interactions with both NBD and CRD domains (Fig. 3 B and C). Similar to the nitrate-bound form, the PII-bound NrtBCD also adopts an inward-facing conformation, sharing a root mean square deviation (RMSD) value of 0.894 Å against 1,022 Cα atoms (SI Appendix, Fig. S8A).

Different from the other two complex structures, NrtC-CRD in the PII-bound structure adopts a stabilized conformation via direct interactions with PII. The CRD and NBD of NrtC, which are connected by a hinge helix, form direct interactions between the C-terminal subdomain of CRD and β2&β11 of NBD (SI Appendix, Fig. S8B). NrtC-CRD is structurally similar to the SBP protein NrtA (19), with an RMSD value of 1.964 Å against 362 Cα atoms (SI Appendix, Fig. S9A). It also has a C-clamp structure of two lobes that harbors an inter-lobe cleft binding to the substrate, in addition to an all-α C-terminal subdomain (SI Appendix, Fig. S9A). Structural comparison showed that NrtC-CRD and NrtA share a similar substrate-binding mode (SI Appendix, Fig. S9 A and B). In particular, the lysine residue, namely Lys480 in Anabaena NrtC-CRD, Lys269 in NrtA and Lys484 in Synechocystis NrtC-CRD that might determine the substrate specificity is highly conserved (SI Appendix, Fig. S9C). Therefore, similar to NrtA, NrtC-CRD might also bind to nitrate, despite which was not observed in our structure. In fact, isothermal titration calorimetry (ITC) assays showed that both Anabaena and Synechocystis NrtC-CRD bind to nitrate, at a dissociation constant (K_d) of 0.22 mM and 1.08 μM, respectively (SI Appendix, Fig. S10 A and B). Notably, Synechocystis NrtC-CRD has a much higher nitrate-binding affinity, which is comparable to that of Synechocystis NrtA (15). Moreover, mutation of this lysine residue to alanine in either Anabaena or Synechocystis NrtC-CRD largely decreased the binding affinity toward nitrate, leading to a K_d value of 1.35 mM and 330 μM, respectively (SI Appendix, Fig. S10 A and B). It suggested that this conserved lysine residue in NrtC-CRD is required for nitrate binding.

In the complex structure, PII also forms a trimer, adopting a canonical ferredoxin-like fold, with each subunit composed of an antiparallel β-sheet with four β-strands and two α-helices at the lateral surface (Fig. 3C and SI Appendix, Fig. S11A). The ADP molecule binds to the nucleotide-binding cleft of PII, in a pattern similar to that in other PII-ADP structures (29, 30). Notably, most residues (Gln42-Thr52) of the T-loop are not visible in the structure, consistent with the previously reported PII structures (29, 31, 32). At the base of the T-loop, the α- and β-phosphate groups of ADP are coordinated by residues Arg38 and Gln39 of one subunit, in addition to Arg101′ and Arg103′ of another subunit, leading to a sharp turn of the T-loop at residue Lys40, the side chain of which protrudes outward and interacts with NrtC-NBD (Fig. 3C).

Notably, two PII subunits contribute interactions with NrtC, forming an interface at the junction between NBD and CRD of NrtC (Fig. 3C). Different from the previous reports that the T-loop is generally stabilized via interacting with various protein partners (33–35), the T-loop in our complex structure remains largely disordered. Instead, the lateral part of PII core structure interacts with the NBD domain of NrtC mainly via polar interactions, forming the interface I with a total buried area of ~650 Ų. Specifically, Lys40 of PII at the base of the T-loop interacts with Arg71 and Glu75 of NrtC-NBD. The main-chain carbonyl oxygen of Gln39 of PII forms a hydrogen bond with Arg74 of NrtC-NBD. Glu85 of B-loop and Arg103′ of C-loop from two subunits of PII form salt bridges with Arg60 and Asp8 of NrtC-NBD, respectively (Fig. 3C). Moreover, the α2 helix and C-loop of PII also forms a smaller interface II of ~560 Ų with the C-terminal subdomain of CRD (Fig. 3C). Besides the hydrophobic and van der Waals interactions at interface II, the residues Gly105′ and Lys107′ of PII form hydrogen bonds with Lys565 and Asn564 of NrtC-CRD, respectively (Fig. 3C).

NrtBCD Is Locked in an Inhibited Inward-Facing Conformation upon PII Binding.

The PII-bound NrtBCD showed pronounced conformational changes compared to NrtBCD-ATP (SI Appendix, Fig. S11B). The NrtC-NBD undergoes a substantial shift relative to NrtD, leading to the coupling helices in TMDs moving away by ~12 Å from each other (SI Appendix, Fig. S11B). In contrast, comparison of the structure of NrtBCD-PII with nitrate-bound NrtBCD (NrtBCD-NO₃⁻) showed that the NrtC-NBD slightly rotates by ~6° toward NrtD around the twofold axis, given the two NrtD subunits superimposed (Fig. 4A). This rotation brings the two NBDs slightly closer to each other, forming a ~100 Ų larger interface area in the PII-bound NrtBCD (Fig. 4A). Consequently, the coupling helix undergoes a slight shift and induces a relative rotation of two TMDs toward each other (Fig. 4A). Notably, the volumes of the two cavities become smaller, shrinking from 65 to 50 ų for the hydrophobic cavity, and from 974 to 708 ų for the hydrophilic cavity (Fig. 4B). Eventually, binding of PII locks NrtBCD in an inhibited inward-facing conformation.

Fig. 4.

Conformational changes and transport inhibition of NrtBCD induced by PII binding. (A) Side view of superimposed NBDs (Left) and TMDs (Right) of NrtBCD-NO₃⁻ (gray) and NrtBCD-PII complex (orange). NrtD and one NrtB subunit are used as the fixed reference for the superposition. All α-helices are displayed as cylinders. The relative torsion angle of NrtC-NBD is labeled. (B) Volume representation of the hydrophobic (orange) and hydrophilic (slate) cavities in the TMDs of NrtBCD-NO₃⁻ and NrtBCD-PII complex. (C) Effects of ammonium on the nitrate uptake of wild-type Synechocystis sp. PCC 6803 and mutants. WT: wild-type Synechocystis sp. PCC 6803; ΔCRD: deletion of NrtC-CRD; Interface I mutation: the K71Q&E72R&E75Q mutation disrupting the interface between NrtC-NBD and PII; Interface II mutation: the Y566A&V567E&K568S&Q569A mutation disrupting the interface between NrtC-CRD and PII; K484A: mutation of the nitrate-binding site. In the experimental groups (solid triangles), nitrate of 200 μM was added to the medium with an OD_{750 nm} value of 1, followed by the immediate addition of 500 µM ammonium. The control groups (open circles) are set in the same condition without the addition of ammonium. The changes of nitrate concentration in the medium are monitored along the time. Each data point represents the mean of three independent measurements, with SD indicated by error bars.

To further investigate the role of PII in the nitrate uptake of NRT in vivo, we utilized the model cyanobacterium Synechocystis sp. PCC 6803, which shares an NRT of 63% sequence identity and highly conserved key residues with that of Anabaena sp. PCC 7120. Similar to previous results (20, 21, 36), the wild-type NRT showed an obvious nitrate uptake activity when nitrate was used as the sole nitrogen source. Once 500 µM ammonium was added to the medium, the nitrate uptake activity was almost completely inhibited (Fig. 4C). Deletion of NrtC-CRD enabled NRT to partly gain the nitrate uptake activity in the presence of ammonium (Fig. 4C). Once the interactions with PII were disrupted via mutating the interface residues on either NrtC-NBD (K71Q&E72R&​E75Q) or NrtC-CRD (Y566A&V567E&K568S&Q569A), we also observed the partly gained nitrate uptake activity in the presence of ammonium (Fig. 4C). Notably, mutation of the nitrate-binding residue Lys484 in NrtC-CRD to alanine showed a profile of nitrate uptake activity similar to the wild-type NRT (Fig. 4C). It suggested that the nitrate-binding site of NrtC-CRD might not be directly involved in the nitrate uptake activity of NRT.

Discussion

A Proposed Transport Cycle of NRT.

Based on the structural and functional data, we propose a putative model for the transport cycle of cyanobacterial NRTs (Fig. 5), representing a unique group of ABC transporters. The transport cycle initiates with the binding of the substrate nitrate to the SBP protein NrtA, which is anchored to the plasma membrane (resting state). Subsequently, the substrate-loaded NrtA is docked to TMDs, which adopt a transient inward-facing conformation (pre-translocation state). Binding of ATP induces the dimerization of the NBDs, making TMDs adopt an outward-facing conformation (outward-facing). Afterward, the nitrate is loaded to a hydrophobic cavity, which most likely functions as a substrate residency station, as shown in our NrtBCD-NO₃⁻ structure (Fig. 1B). Following ATP hydrolysis, the NBDs partially separate, leading to a transition to an inward-facing state (inward-facing). Therefore, the nitrate is attracted by the adjacent hydrophilic cavity located in the IML and finally released to the cytosol. Throughout the active transport cycle, NrtC-CRD remains in a highly flexible state; however, binding of PII stabilizes NrtC-CRD, thereby locks NRT in an inward-facing conformation and effectively inhibits the transport.

Fig. 5.

A proposed model for the transport cycle of NRT and its regulation by PII. The NRT and PII proteins are shown as the schemes, with the subunits and domains colored differently. Under the low N/C condition, the 2-OG-bound PII dissociates from NRT, which activates the transport activity of NRT. The transport cycle starts from the resting state with the TMDs adopting an inward-facing conformation. Upon the docking of substrate-loaded NrtA to TMDs (pre-translocation state), the binding of ATP will trigger the conformational changes, leading to an outward-facing conformation (outward-facing). Following ATP hydrolysis, the NBDs partially separate, leading to a transition to an inward-facing state (inward-facing) to facilitate the substrate release. Under high N/C condition, PII specifically binds to NrtC, which allosterically inhibits the nitrate uptake activity of NRT. The resting and pre-translocation states, which were not observed in this study, are marked by blue dashed boxes. The N/C status is indicated by a blue triangle at the Bottom of the figure.

Regulation of ABC Transporter via Asymmetrically Binding to a Signaling Protein.

Previous studies showed that some type I ABC importers, such as molybdate transporter ModBC (37) or methionine transporter MetNI (38, 39), have additional regulatory domains fused to the NBDs. Once the substrate concentration is high enough, substrate binding triggers the self-dimerization of the regulatory domains and keeps the NBDs separated and retains the transporter in an inward-facing state, therefore trans-inhibiting the substrate translocation. In the case of maltose transporter MalFGK₂ (40), the regulatory protein EIIA^Glc is wedged between the NBD domain of one subunit and the regulatory domain of another subunit, thus locking the transporter in an inward-facing conformation to prevent substrate translocation. NRT differs from the previously reported ABC transporters in the regulatory mode (SI Appendix, Fig. S12). NRT has only one regulatory domain fused to the NBD domain of NrtC, showing an asymmetric architecture. The signaling protein PII specifically binds to NrtC in an asymmetric pattern, which in return inhibits the NRT transport by sensing the intracellular 2-OG level, directly linking the ABC transporter with the cellular N/C metabolism.

The ATPase assays showed that nitrate could stimulate the ATPase activity of NRT, which is distinct from ModBC (37) or MetNI (38), in which the substrates at high concentration significantly inhibit the ATPase activities. Structural analysis provides plausible explanations for why nitrate may not have the trans-inhibition effect. The nitrate/nitrite is rather small in size, and the binding of substrates to the inter-lobe cleft of NrtC-CRD could induce very subtle conformational changes. It is different from the cases in ModBC (37) and MetNI (38) that substrate binding triggers the self-dimerization of two identical CRDs, the conformational changes of which could effectively transmit to NBDs. In contrast, the NrtC-CRD remains highly flexible in the absence of PII. The binding of PII stabilizes NrtC-CRD, and thereby locks NRT at an inhibited inward-facing conformation to prevent substrate translocation.

Coordination of N/C Homeostasis through the PII-NRT Network.

In the present study, we observed that PII specifically binds to the ABC transporter NRT via directly interacting with both CRD and NBD domains of NrtC, representing the first structure of PII-ABC complexes. Notably, the main part of the T-loop does not participate in the NrtBCD-PII interaction, which is distinct from other PII–partner complex structures, such as PII-PipX (35), PII-NAGK (34), and GlnK-AmtB (33), where a symmetric trimeric T-loop is involved in the interaction. As observed in the structure of Azospirillum brasilense PII paralog GlnZ complexed with the dinitrogenase reductase-activating glycohydrolase DraG (41), PII and NrtC adopt a similar interaction mode (SI Appendix, Fig. S13 A and B). In the GlnZ-DraG complex, GlnZ interacts with DraG via the lateral part of GlnZ, involving the B-loop, C-loop, and the base of T-loop (SI Appendix, Fig. S13B). Therefore, the complex structures of NrtBCD-PII and GlnZ-DraG represent a second widely distributed mode of PII interaction with the partners, in which the interactions are mediated by the lateral part of PII, but not the canonical T-loops.

The present structural and functional analyses, together with previous studies (20–22, 36, 42), enabled us to propose a fine model of PII-regulated nitrate uptake in cyanobacteria (Fig. 5). Notably, the coordination of carbon and nitrogen balance is achieved through a network of pathways involving various regulators, such as NtcA (9, 10), PII, and its phosphorylation state. This simplified model only emphasizes the role of PII in regulating nitrate uptake by NRT via sensing different effector molecules. When the environmental ammonium is sufficient, the nitrate assimilation pathway is completely inhibited (20, 21, 36), which is controlled at both the transcriptional and activity levels (8). Being integrated with enriched ammonium via the GS-GOGAT cycle (43), the intracellular 2-OG is maintained at a relatively low level. However, the accumulation of 2-OG is necessary for the activation of the global nitrogen regulator NtcA (10). Meanwhile, in the absence of 2-OG, PII adopts a conformation with the base of T-loop favorite for binding to NrtC (SI Appendix, Fig. S14 A and B). The conformational changes at NBDs upon PII binding are then transmitted to TMDs, thereby locking NRT in an inhibited conformation. This synergistic regulation at both the transcriptional and activity levels, in response to the intracellular 2-OG level, ensures the complete inhibition of nitrate assimilation in the presence of ammonium.

Once ammonium is depleted in the environment, the cellular 2-OG concentration transiently increases due to reduced consumption. In consequence, the elevated 2-OG binds to NtcA, activating the transcription of the nirA operon together with the LysR-type regulator NtcB (11, 12). On the other hand, simultaneous binding of ATP and 2-OG makes the base of T-loop more rigid, with the key residue Lys40 of PII that interacts with NrtBCD in our complex structure pointing to the opposite side (SI Appendix, Fig. S14C). Indeed, the PII binding affinity toward NrtBCD was largely reduced in the presence of 2-OG (Fig. 3A). Upon the release of PII, the activity of NRT could be restored. In consequence, the upregulated nitrate assimilation pathway enables the cells to regain a balanced N/C metabolism.

Notably, previous studies have shown that the phosphorylation state of Ser49 at the tip of PII T-loop might be involved in controlling the nitrate uptake (44, 45). The unphosphorylated PII variant S49A constitutively inhibits nitrate uptake by NRT (44) or has an inhibitory effect only in the presence of ammonium (21), whereas the PII phosphorylation mutants (S49D and S49E) still have a short-term ammonium-inhibited nitrate uptake (21, 44), similar to that of wild-type strains. In fact, the T-loop, which is invisible in our NrtBCD-PII structure, is not involved in the binding of PII to NrtC, thus providing direct evidence that the phosphorylation state of Ser49 at T-loop might not affect the inhibitory effect of PII in the presence of ammonium. However, a mutant of PII phosphatase deficient mutant excretes nitrite when grown in the presence of nitrate, indicating that this mutant has impaired the nitrate assimilation, probably due to the over-phosphorylation of PII (45). More investigations are needed to elucidate the fine mechanism of PII phosphorylation in regulating the physiological functions of various partner proteins.

In summary, a balanced supply of carbon and nitrogen is necessary for the optimal growth of cells under varying environmental conditions. The intracellular 2-OG level serves as a central indicator of the cellular N/C ratio (8). The imbalanced N/C ratio will result in the intracellular accumulation of 2-OG, which is sensed by the global regulators, such as NtcA (10), NdhR (46), and PII (23). Fine coordination of these regulators that switch on or off the downstream genes involved in carbon and nitrogen metabolisms will keep and restore the intracellular N/C balance. Here, we elucidate the fine mechanism of how PII regulates nitrate uptake by NRT to keep N/C homeostasis in cyanobacteria. We provide the first structural insights into PII-ABC complexes, which are widespread among various species, such as the PII-regulated ABC-type urea transporter URT in Synechocystis (22) and the PII-like PotN regulating the ABC-type polyamine transporter PotABCD in lactic acid bacteria (47). More importantly, we propose a distinct regulatory mode of ABC transporters, in which a partner protein asymmetrically binds to one NBD subunit and allosterically inhibits the transport activity.

Materials and Methods

Protein Expression and Purification.

The coding region of nrtBCD was amplified from Anabaena sp. PCC 7120 genomic DNA by PCR and inserted into a modified pET-19b vector by using the One Step Cloning kit (Vazyme), with a 3xFlag-tag at the C-terminus of NrtC. The nrtBCD mutants were also constructed using the One Step Cloning kit according to the manufacturer’s protocol.

The recombinant plasmids harboring wild-type or mutant nrtBCD were transformed into E. coli C43 (DE3) for protein expression. The cells were grown in Luria–Bertani (LB) culture medium supplemented with 100 μg/mL ampicillin at 37 °C. At an optical density at 600 nm of 0.6 to 0.8, 0.2 mM IPTG (isopropyl-β-d-thiogalactopyranoside) was added to induce protein overexpression. After induction of the cultures at 37 °C for 4 h, the cells were collected by centrifugation, resuspended in the lysis buffer (20 mM Tris-HCl pH 7.5, 100 mM NaCl, and 20% glycerol) and stored at −80 °C for further use.

For membrane extraction and protein purification, cells were thawed and lysed by a high-pressure homogenizer at 800 bars. Cell debris was removed by centrifugation at 17,300 × g for 30 min at 4 °C. Membranes were collected from the supernatant by ultracentrifugation at 200,000 × g for 1 h and homogenized in lysis buffer. Proteins were extracted from the membrane suspension using 1% (w/v) dodecyl-β-d-maltopyranoside (Anatrace) and 1% (wt/vol) cocaethylene glycol monododecyl ether (C₁₂E₈; Anatrace) at 4 °C with gentle rotation for 2 h. The resulting supernatant was loaded onto an anti-FLAG M2 affinity gel (Sigma) and incubated on ice for 60 min. The resin was then packed into a column and washed with 5 mL of wash buffer each time. The resin was washed twice with wash buffer A containing 20 mM Tris-HCl pH 7.5, 200 mM NaCl, 20% (v/v) glycerol, 1 mM MgCl₂, 1 mM ATP, and 0.02% (w/v) glycodiosgenin (GDN; Anatrace), followed by two washes with wash buffer B containing 20 mM Tris-HCl pH 7.5, 100 mM NaCl, 10% (v/v) glycerol, and 0.02% (w/v) GDN. The protein was eluted with 6 mL of elution buffer containing 20 mM Tris-HCl pH 7.5, 100 mM NaCl, and 0.02% (w/v) GDN supplemented with 200 μg/mL FLAG peptide. The eluent was concentrated using a 100-kDa cut-off Centricon (Millipore) and further purified using size-exclusion chromatography (Superdex® 200 Increase 10/300, GE Healthcare) in elution buffer. Peak fractions were collected and concentrated for further biochemical studies or cryo-EM experiments.

The gene of PII was cloned from Anabaena sp. PCC 7120 genomic DNA and inserted into the pET-29 expression vector (Novagen) using the One Step Cloning kit with a C-terminal 6xHis tag. The constructed plasmid was transformed into E. coli strain BL21 (DE3) (Novagen) and grown in LB medium supplemented with 50 μg/mL kanamycin. PII expression was induced by IPTG at 37 °C for 4 h. The cells were harvested and resuspended in lysis buffer (20 mM Tris-HCl pH 7.5, 300 mM NaCl, 1 mM ethylenediaminetetraacetic acid) before sonication. After the cell debris was removed by centrifugation, the supernatant was loaded onto nickel affinity resin (Ni-NTA, Qiagen), and the column was washed with wash buffer (20 mM Tris-HCl pH 7.5, 300 mM NaCl). Contaminants were then removed using wash buffer supplemented with 20 mM imidazole. The protein was eluted from the affinity resin with elution buffer (20 mM Tris-HCl pH 7.5, 150 mM NaCl, and 10 mM MgCl₂) containing 400 mM imidazole. The concentrated protein was further purified by size exclusion chromatography (HiLoad 16/60 Superdex 75, GE Healthcare) in elution buffer, and protein purity was assessed by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS–PAGE).

The coding regions of NrtC-CRD from Anabaena sp. PCC 7120 or Synechocystis sp. PCC 6803 were cloned into the pET-28a vector (Novagen) and transformed into E. coli strain Rosetta (DE3) (Novagen). Subsequently, the cells were induced with IPTG and grown at 16 °C for 20 h. Similar to that of PII, NrtC-CRD was purified by nickel-affinity and size-exclusion chromatography, with the buffer of 20 mM HEPES pH 7.5 and 100 mM NaCl.

Cryo-EM Sample Preparation and Data Collection.

To prepare the cryo-EM sample of NrtBCD-NO₃⁻, aliquots of 3.5 µL NrtBCD protein at ~8 mg/mL were applied to glow-discharged QUANTIFOIL (R1.2/1.3, 300 mesh, holey carbon films) Au grids. The grids were blotted with filter paper for 8 s under a blotting force of −2. Then, the grids were plunged into liquid ethane cooled with liquid nitrogen using a Vitrobot Mark IV (Thermo Fisher Scientific) under 100% humidity at 8 °C. A total of 3,824 micrograph stacks were automatically collected with EPU 2 software (48) on a Titan Krios microscope at 300 kV equipped with a K3 detector (Gatan) and a GIF Quantum energy filter (Gatan) at a nominal magnification of ×81,000 with defocus values from −2.3 to −1.5 μm. For these stacks, motion correction and dose weighting were performed using patch motion correction with a Fourier cropping factor of 0.5, resulting in a pixel size of 1.07 Å. Meanwhile, the defocus values were estimated using Patch contrast transfer function (CTF) estimation (49).

To prepare the NrtBCD-ATP sample, the purified mutant NrtBC^E164QD^E179Q at ~8 mg/mL was mixed with 10 mM ATP-Mg²⁺ and 10 mM NaNO₃, followed by incubation on ice for 60 min. Aliquots of 3.5 µL protein mixture were then applied to glow-discharged QUANTIFOIL (R1.2/1.3, 300 mesh, holey carbon films) Au grids. The grids were blotted with filter paper with an 8-s blotting time and −2 blotting force and plunged into liquid ethane cooled with liquid nitrogen using a Vitrobot Mark IV (Thermo Fisher Scientific) under 100% humidity at 8 °C. A dataset with a total of 4,233 micrograph stacks was collected in the same manner as the NrtBCD-NO₃⁻ sample. These stacks were motion-corrected with dose weighting by MotionCor2 (50) with a binning factor of 2, resulting in a pixel size of 1.07 Å. The defocus values were estimated using CTFFIND4 (51).

For the preparation of the NrtBCD-PII sample, the purified wild-type NrtBCD was concentrated to ~8 mg/mL and mixed with purified PII trimer at a molar ratio of 1:2 with the addition of 2 mM ADP and 10 mM NaNO₃. Then, the protein mixture was incubated on ice for 60 min. After that, aliquots of 3.5 µL protein mixture were applied to glow-discharged QUANTIFOIL (R1.2/1.3, 300 mesh, holey carbon films) Au grids. The grids were blotted with filter paper with an 8-s blotting time and −2 blotting force. Then, the grids were plunged into liquid ethane cooled with liquid nitrogen using a Vitrobot Mark IV (Thermo Fisher Scientific) under 100% humidity at 8 °C. Similar to the collection of data for NrtBCD-NO₃⁻, a total of 2,506 movie micrographs for NrtBCD-PII were collected using a Titan Krios microscope at 300 kV equipped with a K3 detector (Gatan). The stacks were motion corrected and dose weighted by patch motion correction with a Fourier cropping factor of 0.5, yielding a pixel size of 1.07 Å, and the defocus values were estimated using Patch CTF estimation.

Image Processing.

For the datasets of NrtBCD-NO₃⁻, a total of 2,854,383 particles were extracted from 3,824 micrographs by reference-based auto-pick from cryoSPARC 3.1 (49). After several rounds of 2D classification, 1,265,340 particles were selected and used for ab initio reconstruction and heterogeneous refinement with C1 symmetry. Then, 319,161 particles from the best class were further subjected to nonuniform refinement and local refinement, yielding a final map at 3.54 Å resolution.

For the datasets of NrtBCD-ATP, a total of 3,366,180 particles were automatically picked from 4,233 micrographs using RELION 3.1 (52) and then subjected to reference-free 2D classification. After 2D classification, 1,827,550 particles merged from good classes were selected and subjected to 3D classifications with a global search. Several rounds of global angular searching 3D classification with C1 symmetry were performed, during which the particles were classified into four classes. A total of 1,222,929 particles from the best classes were merged and subjected to 3D refinement with an adapted mask, yielding a reconstruction with an overall resolution of 3.1 Å, followed by CTF refinement and Bayesian polishing.

The datasets of the NrtBCD-PII were also processed using cryoSPARC 3.1 (49). In total, 1,773,611 particles were automatically picked from 2,506 micrographs and subjected to 2D classification. Then, 815,015 particles were selected for multi-round ab initio reconstruction and heterogeneous refinement with C1 symmetry. Finally, 304,482 particles were selected and applied to nonuniform refinement and local refinement, yielding a reconstruction map at a resolution of 3.53 Å.

All resolutions of the cryo-EM maps were estimated with the gold-standard Fourier shell correlation 0.143 criterion (53, 54).

Model Building and Refinement.

The final sharpened map of NrtBCD-ATP with a B-factor of −168.6 Ų was used for model building in Coot (55). The high quality of electron density in the TMDs allowed unambiguous assignment of residues. Therefore, we manually built the TMD model in Coot. The models of NBDs (NrtC and NrtD) were generated by the SWISS-MODEL server (56), which were docked into the 3.10 Å map of NrtBCD-ATP using Chimera. The ATP-Mg²⁺ molecules were manually built into the corresponding density maps. Structure refinements were carried out by PHENIX (57) in real space with secondary structure and geometry restraints to prevent structure overfitting. The modeling of the NrtBCD-NO₃⁻ and NrtBCD-PII complex was performed by rigid-body fitting of the NrtBCD-ATP structure, followed by manual refinement. The crystal structure of ADP-bound PII (PDB: 4C3K) was manually fitted into the map of the NrtBCD-PII complex and adjusted accordingly. The nitrate and ADP molecules were also manually built into their respective density maps. Similarly, these models were real space refined by PHENIX (57).

All structures were validated by PHENIX and MolProbity (58). The refinement and validation statistics are summarized in SI Appendix, Table S1. The UCSF ChimeraX (59) and PyMOL (https://pymol.org) were used to prepare the structural figures. The final coordinates and cryo-EM maps have been deposited in Protein Data Bank (PDB) under the accession codes listed in SI Appendix, Table S1.

ATPase Activity Assays.

The ATPase activities of wild-type NrtBCD and mutants were measured using the ATPase Colorimetric Assay Kit (Innova Biosciences). The assays were performed in 96-well plates, and the absorbance was measured at 630 nm. To assess the impact of nitrate on the ATPase activity of NrtBCD, a final concentration of 0.1 μM protein was added to a reaction buffer containing 50 mM Tris-HCl pH 8.0, 150 mM NaCl, 0.02% (w/v) GDN, and 2 mM MgCl₂. Nitrate was diluted to different concentrations and added to the reaction mixture to 100 µL as one reaction sample. Reactions were conducted at 37 °C for 30 min, and the amount of released Pi was quantified at OD_{630 nm} using a SpectraMax iD5 Multi-Mode Microplate Reader (Molecular Devices). The control groups in the absence of proteins were subtracted as background for each point of data. The data are presented as the mean ± SD of three independent biological replicates (n = 3).

ITC.

The ITC experiments were performed at 13 °C using a PEAQ-ITC calorimeter (MicroCal) in the buffer consisting of 20 mM Na₂HPO₄ pH 8.0 and 100 mM NaCl. The wild-type NrtC-CRD protein of 56.6 μM was titrated with 10 mM NaNO₃. The ligands of 2 µL were injected 20 times into the 300 µL cell under constant stirring conditions of 750 rpm. The NrtC-CRD mutant (K480A mutant of Anabaena or K484A mutant of Synechocystis) of 97.7 μM was titrated with 20 mM NaNO₃. The ligands of 2 µL were injected 20 times into the 280 µL cell. The binding isotherms were calculated from the received data and fitted to the one set of sites model with MicroCal PEAQ-ITC Analysis Software (Malvern Panalytical).

Pull-Down Assays.

The pull-down assays were performed using ADP (5mM), ADP&NaNO₃ (5 mM ADP and 10 mM NaNO₃), ATP (5 mM), and ATP&2-OG (5 mM ATP and 5 mM 2-OG) as effectors. For each experimental condition, a reaction mixture with a total volume of 1 mL containing 2 μM Flag-tagged NrtBCD and 40 μM His-tagged PII protein was prepared in a binding buffer (20 mM Tris pH 7.5, 150 mM NaCl, 10 mM MgCl₂, 0.02% (w/v) GDN, and the respective effectors). The mixture was then incubated at 13 °C for 60 min with gentle shaking. Next, each mix was incubated with an equilibrated anti-FLAG M2 affinity gel (Sigma) for a further 60 min with gentle shaking. Then, the gel was washed eight times with 1 mL of binding buffer by gently pipetting up and down. Subsequently, the coupled proteins were eluted using 500 µL binding buffer supplemented with 200 μg/mL of FLAG peptide. All eluted products were applied to SDS–PAGE analysis.

Construction of the Mutants.

The nrtCD deletion mutant (ΔnrtCD) in Synechocystis sp. PCC 6803 was constructed by replacing the NrtC and nrtD genes with the cassette encoding a streptomycin resistance gene via homologous recombination. The upstream and downstream sequences of nrtCD were amplified by PCR and further fused with the spectinomycin cassette gene. This fragment was cloned into the pGEM-T easy plasmid, which was mixed with Synechocystis sp. PCC 6803 cells for homologous recombination. The cells were cultured at 30 °C with continuous illumination of 30 μmol photons m⁻² s ⁻¹ for 4 to 6 h. The deletion of nrtCD was validated by PCR.

NrtC mutants containing one or more residue mutations were constructed by homologous recombination based on the ΔnrtCD mutant in Synechocystis sp. PCC 6803. Residue mutations in NrtC were introduced by PCR. The mutated nrtCD fragment was then fused with the kanamycin cassette gene and inserted into the nrtCD deletion site. In addition, the streptomycin resistance gene was excised from the pGEM-T easy plasmid. The recombinant plasmids were mixed with the ΔnrtCD mutant cells of Synechocystis sp. PCC 6803 and incubated at 30 °C under illumination of 30 μmol photons m⁻² s⁻¹ for 4 to 6 h. The mutations of NrtC were confirmed by PCR.

Cultivation of Cyanobacteria.

The wild-type and mutant strains of Synechocystis sp. PCC 6803 were grown photoautotrophically under CO₂-sufficient conditions at 30 °C. The basal medium used was a nitrogen-free medium obtained by modification of BG11 medium. To prepare the ammonium- and nitrate-containing media, 10 mM NH₄Cl and 20 mM NaNO₃ were added to the basal medium, respectively. All media were buffered with 20 mM TES-NaOH pH 8.0. The media were supplemented with 15 μg/mL streptomycin and 30 μg/mL kanamycin as needed. To induce the expression of nrtABCD genes, ammonium-grown cells were collected by centrifugation, washed, and transferred to a nitrate-containing medium for further cultivation.

Nitrate Uptake Assays.

To determine the nitrate uptake of wild-type and mutant strains of Synechocystis sp. PCC 6803, cells in the exponential growth phase (OD_{750 nm} = 0.4 to 0.8) were collected by centrifugation at 3,000 × g for 10 min at room temperature and washed twice with nitrogen-free medium. The culture was then adjusted to an OD_{750 nm} value of approximately 1 using nitrogen-free medium. The assays were initiated by adding 200 μM NaNO₃. The cells were cultivated at 30 °C under constant shaking at 150 rpm and illumination of approximately 40 μmol photons m⁻² s⁻¹. Every 10 min, a 1 mL aliquot of the cell suspension was taken and centrifuged to remove cells. Nitrate uptake was measured by quantifying the nitrate concentration in the medium through the absorbance value at 220 nm of the cell-free medium. When the effects of ammonium on nitrate uptake were examined, 500 μM NH₄Cl was added to the cell suspension immediately after the addition of nitrate, and nitrate uptake was subsequently measured.

Supplementary Material

Appendix 01 (PDF)

Data, Materials, and Software Availability

The cryo-EM maps and coordinates have been deposited at the Electron Microscopy Data Bank (EMD-37645 (60) for NrtBCD-NO₃⁻, EMD-37375 (61) for NrtBCD-ATP, and EMD-37644 (62) for NrtBCD-PII) and PDB (8WM8 (62) for NrtBCD-NO₃⁻, 8W9M (63) for NrtBCD-ATP, and 8WM7 (64) for NrtBCD-PII), respectively.