Crystal structures of adenylylated and unadenylylated PII protein GlnK from Corynebacterium glutamicum

Abstract

P_II proteins are ubiquitous signaling proteins that are involved in the regulation of the nitrogen/carbon balance in bacteria, archaea, and some plants and algae. Signal transduction via P_II proteins is modulated by effector molecules and post-translational modifications in the P_II T-loop. Whereas the binding of ADP, ATP and the concomitant binding of ATP and 2-oxoglutarate (2OG) engender two distinct conformations of the T-loop that either favor or disfavor the interaction with partner proteins, the structural consequences of post-translational modifications such as phosphorylation, uridylylation and adenylylation are far less well understood. In the present study, crystal structures of the P_II protein GlnK from Corynebacterium glutamicum have been determined, namely of adenylylated GlnK (adGlnK) and unmodified unadenylylated GlnK (unGlnK). AdGlnK has been proposed to act as an inducer of the transcription repressor AmtR, and the adenylylation of Tyr51 in GlnK has been proposed to be a prerequisite for this function. The structures of unGlnK and adGlnK allow the first atomic insights into the structural implications of the covalent attachment of an AMP moiety to the T-loop. The overall GlnK fold remains unaltered upon adenylylation, and T-loop adenylylation does not appear to interfere with the formation of the two major functionally important T-loop conformations, namely the extended T-loop in the canonical ADP-bound state and the compacted T-loop that is adopted upon the simultaneous binding of Mg-ATP and 2OG. Thus, the P_II-typical conformational switching mechanism appears to be preserved in GlnK from C. glutamicum, while at the same time the functional repertoire becomes expanded through the accommodation of a peculiar post-translational modification.

Keywords: AMPylation; T-loop conformations; adenylylation; bacterial signal transduction; crystal structures; nitrogen starvation; post-translational modifications.

Figure 1

Canonical structure of P_II proteins. (a) Topology plot of a P_II monomer. Structural features such as the B-, C- and T-loops, as well as the post-translational modification site located in the T-loop, are highlighted. (b) P_II protein trimer displaying C ₃ point-group symmetry. Each protomer is depicted in a different color. (c) Hexameric assembly with D ₃ point-group symmetry of two P_II trimers as observed in the crystal structures of unGlnK and adGlnK from C. glutamicum. The symmetry elements present in point groups C ₃ and D ₃ are illustrated as follows: black triangles indicate threefold rotation axes oriented perpendicular to the plane of the illustration and black arrows indicate twofold rotation axes located in the plane of the illustration.

Figure 2

The three-dimensional structure of unGlnK and adGlnK from C. glutamicum. (a) Structure of an unGlnK monomer shown in a cartoon representation. The side chain of Tyr51, which is located in the T-loop (residues 37–55), as well as the phosphate ion bound in the ATP-binding pocket, is shown in a stick representation and labeled accordingly. (b) Structure of the unGlnK trimer. Primes and double primes denote residues from the second and third protomers, respectively. (c) Structure of one adGlnK monomer. The side chain of the adenylylated Tyr51, as well as the AMP bound in the ATP-binding pocket, is highlighted in a stick representation. (d) Structure of the adGlnK trimer.

Figure 3

Detailed view of the adenylylated T-loop region (residues 43–55) in (a) chain A (with C atoms colored gray), (b) chain B (colored green) and (c) chain E (colored orange) of the adGlnK hexamer (chains A–F). The 2mF _o − DF _c electron-density map is shown in blue within a radius of 1.5 Å of any displayed atoms and is contoured at 1σ. (d) Stereo representation of the immediate surroundings of the adenylylated Tyr51 in chain E. Residues from chains B, E and F are colored green, blue and orange, respectively. Potential hydrogen bonds are shown as dotted black lines. The π–π-stacking interaction between the tyrosyl moiety of adTyr51 and Phe11′′ is shown as a dotted red line.

Figure 4

T-loop conformations in unGlnK and adGlnK. Superposition of the T-loops from (a) all three monomers in the unGlnK trimer, (b) all six monomers present in adGlnK and (c) all monomers from both unGlnK and adGlnK. In all panels the T-loops were superimposed using the coordinates of the entire monomers.

Figure 5

Comparison of effector binding and T-loop conformation between adGlnK from C. glutamicum and GlnZ from A. brasilense. (a) Superposition of adGlnK (in green) and ADP-bound GlnZ (red; PDB entry 4co1; Truan et al., 2014 ▸). (b) Superposition of adGlnK (in green) and Mg-ATP and 2OG-bound GlnZ (blue; PDB entry 3mhy; Truan et al., 2010 ▸). For clarity, only a single subunit is shown and hence the contribution of the B- and C-loop residues to effector binding is not shown. (c) Sequence alignment between GlnK and GlnZ. Residues located within 4.5 Å of any bound effector atom are highlighted in red, blue and yellow when involved solely in ADP binding, solely in Mg-ATP/2OG binding or in binding both ADP and Mg-ATP/2OG, respectively