Active Site Engineering of Copper-Containing Electron Transfer Proteins

• Main Author: Yanagisawa, Sachiko
• Published: University of Newcastle upon Tyne 2009
• Subjects: 572.6
• Online Access: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.484818

Cupredoxins arc electron transfer (ET) proteins which possess type I (Tl) copper sites. A TI copper ion is equatorially coordinated by the thiolate sulfur of a Cys and the imidazole nitrogens of two His residues, along with usually an axially coordinating thioether sulfur of a Met [azurin (Az) possess a second axial interaction with a backbone carbonyl oxygen]. Thr~e of these ligands (Cys, a His and Met) are found on a C-terminal loop which links two of the strands of the cupredoxin p-barrel scaffold. The length and sequence of this Tl copper-binding loop varies. The shOltest known Tl binding loop (that of amicyanin, Ami) has been introduced into three different cupredoxin scaffolds. All of the . loopcontraction variants possess copper centres with authentic TI properties and are redox active. The Cu(lI) and Co(II) sites experience only small structural alterations upon loop contraction with the largest changes in the Az variant (AzAmi) which can be ascribed to the removal of a hydrogen bond to the coordinating thiolate sulfur of the Cys ligand. In all cases loop-contraction leads to an increase in the pKa of the His ligand found on the loop in the reduced proteins, and in the pseudoazurin (Paz) and plastocyanin (Pc) variants the values are almost identical to that of Ami (~6.7). Thus in Paz, Pc and Ami the length of this loop tunes the pKa of the His ligand. In the AzAmi variant the pKa is 5.5 which is considerably higher than the estimated value for Az « 2) and other controlling factors, along with loop length, are involved. The reduction potentials (EmS) of the loop-contraction variants are all lower than those of the wild type (WT) proteins by ~ 30-60 mV and thus this property of a Tl copper site is finetuned by the C-tenninalloop. The electron self-exchange (ESE) rate constant (kESE) of Paz is diminished significantly by the introduction of a shorter loop. However, in PcAmi only a 2-fold decrease is observed and in AzAmi there is no effect, and thus in these two cupredoxins loop contraction does not signi~cantly influence ET reactivity. Loop-contraction provides an active site environment in all of the cup.re~oxins which is preferable for Cu(ll), whereas previous loop elongation experiments always favoure..d the cuprous site. Thus the ligand-containing loop plays an important role in tuning the entatic nature ora TI copper centre. The thiolate sulfur of the Cys ligand in Az is hydrogen bonded from the backbone amides of .Asn47 and Phe114. One of these interactions has been removed in the Phel14Pro variant. A unique . I'~ pectr~s~opj~=..featur~ ,of A.z is-the position of.the S(Cys)~Cu(Il),.ligand-to-metalcharge-transfer . - .. .' ~, . . .' . . (LMCT) band in its UVNis spectrum (~ 630 nm) which is shifted to 599 nm in the Phel14Pro variant, although a site with classic Tl properties is maintained. Shorter CU(Il)-SO(Metl21) and longer Cu(lI}O( Gly45) distances are found at the active site in the crystal structure of the variant compared to WT Az. The copper centre of Phel14Pro Az is more like those of Pc, Ami and Paz than the trigonal bipyramidal arrangement found in Az. The Phel14Pro mutation results in an 80 mV decrease in Em and an order of magnitude smaller kESE value. The influence of this mutation on Em is due to a number of structural effects ofthe mutation, with removal of the hydrogen bond probably most significant. Comparison of the active site structures of Cu(ll) and Cu(I) Phe114Pro· Az indicate larger changes upon redox interconversion than those in the WT protein which increases the reorganization energy and results in slower ET. The axial ligand at Tl copper sites is not conserved. In most cases a weakly coordinated thioether sulfur from a Met [Cu(II)-S5 ~ 2.6-3.3 A] is found in the axial position as in Pc, Paz, Az and Ami. A strong axial bond [Cu(II)-Otl of ~ 2.2 A] is sometimes provided by a Gln.[as in the stellacyanins (STCs)] and the axial ligand can be absent (a Val, Leu or Phe in the axial position) as in ceruloplasmin, FeOp, fungal laccases and some plantacyanins (PLTs). Cucumber basic protein (CBP) is a PLT which has a relatively short Cu(II)-S5(Met89) axial bond (2.6 A). The Met89Gln variant of CBP has a kE?E' a measure of intrinsic ET reactivity, ~ 7 times lower than that of the WT protein. The Met89Vai mutation to CBP results in a 2-fold increase in kESE' As the axial interaction decreases from strong Oel of Gin to relatively w.eak S5 of Met to no ligand (Val), ESE reactivity is enhanced by - 1 order of magnitude whilst Em increases by - 350 mY. The variable coordination position at this ubiquitous ET site provides a mechanism for tuning the driving force to optimize ET with the correct partner without significantly compromising intrinsic reactivity. The enhanced reactivity of a three-coordinate Tl copper site will facilitate intramolecular ET in fungallaccases and Fet3p. The phytocyanins form a sub-family of the cupredoxins and are made up of the STCs, PLTs and uc1acyanins. All of the phytocyanins exhibit an alkaline transition which results in the S(Cys)~Cu(II) LMCT band shifting - 20 nm to higher energy at elevated pH (pKa - 10). The alkaline transition influences all of the coordinating residues with the Cys ligand most affected. The exact cause of alkaline transition is not known, although deprotonation of a group close to the active site must be involved, and the side chain of the axial Gin ligand has been suggested as the trigger for this effect in the STCs. The influe!lce of pH on the spectrosco~ic properties of WT CBP and the Met89Gln and Met89Vai axial ligand variants has been studied. The alkaline transition has a similar influence on . the visible spectrum in all three proteins although the pKa value in Met89Vai CBP is smaller (8.9) than for the other two proteins (- 9.7). Thus the axial ligand is not the cause ofthe alkaline transition. The surface exposed Met16 residue of Paz is situated close to the His81 ligand in the centre of the protein's hydrophobic patch. To study the importance of Met'i6, and to attempt to introduce a 1t-1t. interaction with the imidazole ring of His81, the Met16Phe and Met16Trp variants have been prepared and characterized. NMR studies indicate that the introduced aromatic groups are oriented parallel to the imidazole ring of His8l. UVNis, EPR and paramagnetic IH NMR spectra of the Cu(II) variants highlight very similar active site structures in the two mutants which are less tetragonally distorted than in the WT protein. The pKa value for the His81 ligand in the Cu(I) proteins decreases from 4.9 in WT Paz to 4.5 and 4.1 in Metl6Phe and Metl6Trp Paz respectively, indicating that 1t-1t contact with the introduced aromatic ring stabilizes the Cu-N51 (His81) interaction. The enhanced rigidity at the active site may contribute to decr~ased reorganization energies in the variants resulting in - 2-fold and - 3-fold larger kESE values in Met16Phe alid Metl6Trp Paz respectively. These mutations could also contribute to tl~e increased kESE values by facilitating homo-dimer formation: The Metl6Phe and Metl6Trp mutations give rise to approximately 40-60 mV increases in the Em of Paz. The physiological function of Paz is donation of electrons to nitrite reductace (NiR) and the influence of these mutations on Em result in a decreased driving force for this ET reaction and smaller kC31 are found. The Km for the reaction with NiR is - 2-fold larger for the Met16Phe variant whilst similar values are found for Met 16Trp Paz and the WT protein. Introduction of a 1t-1t interaction at the active site of Paz leads to subtle structural changes but has little effect on the interaction with the physiological ET partner.