Identifying the structure of the active sites of human recombinant prolidase

Abstract

In this paper we provide a detailed biochemical and structural characterization of the active site of recombinant human prolidase, a dimeric metalloenzyme, whose misfunctioning causes a recessive connective tissue disorder (prolidase deficiency) characterized by severe skin lesions, mental retardation and respiratory tract infections. It is known that the protein can host two metal ions in the active site of each constituent monomer. We prove that two different kinds of metals (Mn and Zn) can be simultaneously present in the protein active sites with the protein partially maintaining its enzymatic activity. Structural information extracted from X-ray absorption spectroscopy measurements have been used to yield a full reconstruction of the atomic environment around each one of the two monomeric active sites. In particular, as for the metal ion occupation configuration of the recombinant human prolidase, we have found that one of the two active sites is occupied by two Zn ions and the second one by one Zn and one Mn ion. In both dinuclear units a histidine residue is bound to a Zn ion.

Appendix

The EXAFS signal,$\chi (k ) ,$ is defined by the formula

$$ {\chi (k )= \frac{{\mu (k )- \mu_{ 0} (k )}}{{\mu_{ 0} (k )}}}, $$

(4)

where $\mu (k )$ is the measured total absorption coefficient and $\mu_{0} (k )$ is the absorption coefficient of the isolated absorber. k is the wave vector of the extracted electron which is related to the incident photon energy, $E,$ and the ionization energy, $E_{0} ,$ by the obvious relation

$$ {k = \sqrt {\frac{{2m\left( {E - E_{0} } \right)}}{{\hbar^{2} }}} }, $$

(5)

with m the electron mass and ħ the (reduced) Planck constant.

For the reader’s convenience we recall the theoretical formula describing the EXAFS signal. We give it for simplicity in the single scattering approximation which is valid for photon energies sufficiently higher than the threshold energy. Following the standard notation of Boland et al. (1982), it reads

$$ \chi (k) = S_{0}^{2} \sum\limits_{l} {\frac{{N_{l} }}{{kr_{l}^{2} }}\left| {f_{l} (k,{{\uppi}})} \right|\sin \left( {2kr_{l} + \varphi_{l} (k)} \right){\text{e}}^{{ - 2\sigma_{l}^{2} k^{2} }} {\text{e}}^{{ - {{2r_{l} } \mathord{\left/ {\vphantom {{2r_{l} } \lambda }} \right. \kern-\nulldelimiterspace} \lambda }(k)}} } , $$

(6)

where the sum runs over the different coordination shells around the absorber. N _l is the number of scatterers of the l-th shell, located at a distance r _l from the absorber and $\sigma_{l}^{2}$ is the Debye–Waller factor. $\left| {f_{l} (k,{{\uppi}})} \right|$ is the modulus of the back-scattering amplitude and $\phi_{l} (k)$ the total scattering phase. $S_{0}^{2}$ is an empirical quantity that accounts for all the many-body losses in photo-absorption processes and $\lambda (k)$ is the photo-electron mean free path. In cases where also MS processes become important a formally similar expression holds in which, however, r _l represents the length of the full MS path. Modulus and phase functions are in this case rather complicated expressions which depend on all the scattering events occurring along the MS path (see Lee and Pendry 1975; Benfatto et al. 1986; Gurman et al. 1986; Koningsberger and Prins 1988; Rehr and Albers 1990).

The quantitative analysis of the structural EXAFS signals is performed with the help of the FITEXA code, developed by one of the authors (C. Meneghini).⁴ FITEXA exploits the FEFF8.2 package (Zabinsky et al. 1995) for the calculation of backscattering amplitudes, total phase shifts and photoelectron mean free path. Best fit is achieved by minimizing the quantity

$$ \xi^{2} = \frac{1}{{N_{\text{p}} }}\sum\limits_{i = 1}^{{N_{\text{p}} }} {\left[ {k_{i} (\chi_{\exp } (k_{i} ) - \chi_{\text{th}} (k_{i} )} \right]^{2} } , $$

(7)

where $\chi_{ \exp }$ and $\chi_{\text{th}}$ are the experimental and theoretical data, respectively, and the sum is over the number, N _p, of the values of k at which data were collected. The MINUIT (James 1994) subroutine of CERN library is used for data refinement and statistical error analysis.

The fit quality is measured by the associated R ²-factor defined by the formula⁵

$$ R^{2} = 100\frac{{N_{\text{p}} \xi^{2} }}{{W_{0}^{2} }}\% \;{\text{with}}\;W_{0}^{2} = \sum\limits_{i} {\left[ {k_{i} \chi_{\exp } (k_{i} )} \right]^{2} } $$

(8)

A value of R ² of about 10% is to be considered satisfactory for complex biological molecules.