The main active-site loop of the copper-binding protein azurin (a cupredoxin) has been shortened from C 112 TFPGH 117 SALM 121 to C 112 TPH 115 PFM 118 (the native loop from the cupredoxin amicyanin) and also to C 112 TPH 115 PM 117 . The Cu(II) site structure is almost unaffected by shortening, as is that of the Cu(I) center at alkaline pH in the variant with the C 112 TPH 115 PM 117 loop sequence. Subtle spectroscopic differences due to alterations in the spin density distribution at the Cu(II) site can be attributed mainly to changes in the hydrogen-bonding pattern. Electron transfer is almost unaffected by the introduction of the C 112 TPH 115 PFM 118 loop, but removal of the Phe residue has a sizable effect on reactivity, probably because of diminished homodimer formation. At mildly acidic pH values, the His-115 ligand protonates and dissociates from the cuprous ion, an effect that has a dramatic influence on the reactivity of cupredoxins. These studies demonstrate that the amicyanin loop adopts a conformation identical to that found in the native protein when introduced into azurin, that a shorter than naturally occurring C-terminal active-site loop can support a functional T1 copper site, that CTPHPM is the minimal loop length required for binding this ubiquitous electron transfer center, and that the length and sequence of a metal-binding loop regulates a range of structural and functional features of the active site of a metalloprotein.copper proteins ͉ electron transfer ͉ metalloproteins ͉ protein engineering N umerous approaches are being used to design metal-binding sites in proteins (1), with many of these studies informed by an understanding of the basic structural requirements for biological metal centers. Metal-binding sites in proteins are commonly formed from loops, because these regions are reasonably tolerant to sequence modifications outside of the coordinating residues (1). Cupredoxins are copper-containing electron transfer (ET) proteins that provide a significant challenge for protein-design experiments (2-4) because their scaffold is thought to constrain the metal site structure (5). In the type 1 (T1) copper sites of cupredoxins (see Fig. 1), three of the four canonical ligands Cys, His, and, usually, Met are present on a loop linking the C-terminal strands of a rigid -barrel (7, 8). The fourth ligand, a His, is donated from a -strand more in the core of the fold (see Fig. 1). The lengths of the metal-binding loops in known cupredoxins range from 7 to 16 residues and have a variety of primary structures. These proteins, therefore, provide a suitable system for investigating the importance of loop length and structure for the active-site integrity of a metalloprotein. Loop-directed mutagenesis has been used to swap loops between different cupredoxins, giving sites with authentic T1 properties (8)(9)(10)(11)(12). In this work, we present studies that have been aimed at assessing the structural consequences of shortening the active-site loop of a cupredoxin and have determined the short...