Time‐resolved fluorescence and computational studies of adenylylated glutamine synthetase: Analysis of intersubunit interactions

Abstract

Adenylylation of Tyr‐397 of each subunit of Escherichia coli glutamine synthetase (GS) down‐regulates enzymatic activity in vivo. The overall structure of the enzyme consists of 12 subunits arranged as two hexamers, face to face. Research reported in this paper addresses the question of whether the covalently attached adenylyl group interacts with neighboring amino acid residues to produce the regulatory phenomenon. Wild‐type GS has two Trp residues (positions 57 and 158) and the adenylylation site lies within 7–8 Å of the Trp‐57 loop in the adjacent subunit of the same hexameric ring; Trp‐158 is about 35 Å from the site of adenylylation. Fluorescence lifetimes and quantum yields have been determined for two fluorophores with wild‐type and mutant GS. One fluorophore is ε‐AMP adenylylated GS (at Tyr‐397), and the other fluorophore is the intrinsic protein residue Trp‐57. These experiments were conducted in order to detect possible intersubunit interactions between adenylyl groups and the neighboring Trp‐57 to search for a role for the Trp‐57 loop in the regulation of GS. The fluorescence due to ε‐AMP of two adenylylated enzymes, wild‐type GS and the W158F mutant, exhibits heterogeneous decay kinetics; the data adequately fit to a double exponential decay model with recovered average lifetime values of 18.2 and 2.1 ns, respectively. The pre‐exponential factors range from 0.66 to 0.73 for the long lifetime component, at five emission wavelengths. The W57L‐ε‐AMP enzyme yields longer average lifetime values of 19.5 and 2.4 ns, and the pre‐exponential factors range from 0.82 to 0.85 for the long lifetime component. An additional residue in the Trp‐57 loop, Lys‐58, has been altered and the K58C mutant enzyme has been adenylylated with ε‐AMP on Tyr‐397. Lys‐58 is near the ATP binding site and may represent a link by which the adenylyl group controls the activity of GS. The fluorescence of ε‐AMP‐adenylylated K58C mutant GS is best described by a triple exponential decay with average recovered lifetime values of 19.9, 4.6, and 0.58 ns, with the largest fraction being the median lifetime component. Relative quantum yields of ε‐AMP–Tyr‐397 were measured in order to determine if static quenching occurs from adenine–indole stacking in the wild‐type GS. The relative quantum yield of the ε‐AMP‐adenylylated W57L mutant is larger than the wild‐type protein by the amount predicted from the difference in lifetime values: thus, no static quenching is evident. Intrinsic tryptophan fluorescence was also studied in the presence and absence of covalently attached adenylyl groups. The fluorescence decay parameters of Trp‐57 are not significantly affected by the presence of ε‐AMP or AMP attached to Tyr‐397. Enzymatic activity of the mutant proteins was also studied. The Mg‐activated enzymes are nearly completely inhibited upon adenylylation, and regulation is not affected by mutation at Trp‐57 or Lys‐58. These results indicate that the Trp‐57 loop dynamically interacts with the adenylyl group, but a “ring stacked” complex is not formed and is not a structural feature of the regulatory mechanism. This conclusion was further corroborated by computational minimization and molecular dynamics studies of GS with an AMP moiety built onto Tyr‐397. Relative energies were sampled at various points as the Trp‐57 ring and purine ring of AMP were targeted toward one another, the endpoint being a 3‐Å parallel and stacked conformation. Dynamics simulations were performed with the parallel, stacked conformation, as well as an extended conformation, as the starting point. All studies indicate that the stacked indole–purine conformation is not favorable; however, a potential hydrogen bonding interaction between the amine nitrogen of the tryptophan and the ring oxygen of ribose is implied by dynamics simulations for certain conformations of the loop. Copyright © 1993 The Protein Society