New ways to look at enzyme structures

Metal-catalysed activation of oxygen in biology is achieved using either copper or iron. In the case of iron, this includes the heme-containing enzymes which activate oxygen.

The heme peroxidases use H2O2, forming two enzymes intermediates – the first is Compound I and the second is Compound II (Fig. 1). Compound I contains an oxidized ferryl heme (Fe(IV)) plus either a porphyrin π-cation radical or a protein radical. Reduction of Compound I by one electron (in the peroxidases) or by hydrogen atom abstraction from the substrate (in the P450s) yields the closely related Compound II intermediate, which contains only the ferryl heme (and no radical). Nature uses these Compound I and II intermediates for oxidations, in a diverse group of enzymes that include the cytochrome P450s and peroxidases, as well as the nitric oxide synthases, cytochrome c oxidases, and heme dioxygenases.

Establishing the true nature of these intermediates to establish whether the ferryl heme is protonated or not (Fe(IV)=O) or (Fe(IV)-OH) has been the major objective of a collaboration with Prof. Emma Raven (now University of Bristol). Previous approaches using cryo-trapping and X-ray crystallography to determine bond length as an indirect reporter have been informative, but have several disadvantages. The sensitivity of the high valent ferryl centre to direct or indirect photoreduction means that there will always be some doubt about the results (even if followed spectroscopically, does a proportion become reduced?). The measurement of bond lengths also posed problems, there is a possibility of Fourier ripples from the relatively electron dense iron atom affecting the apparent position of the oxygen. Other compounding factors reduce positional confidence. This is especially important as a only small differences (<0.2 Å) lead to different interpretations.

The low electron density of hydrogens means reliable observation of these is not possible, even at exceptionally high resolution. We overcame the problem with the first structures of cryo-trapped enzyme intermediates to be determined with neutron crystallography (Casadei et al., (2014) Science 345; 193-197, Kwon et al., (2016) Nature Communications 7, Art. 13445) showing that Compound I of CcP is the Fe=O species and Compound II in Ascorbate Peroxidase is Fe-OH.. The neutron structures also showed that the distal histidine was protonated/deuterated in these intermediates, which had not been considered before. This, with other details revealed, provoked re-assessment and further questions about the enzymes’ mechanism. It also becomes important to know if the structures of Compound I of cytochrome c peroxidase and Compound II of ascorbate peroxidase are typical of this class of enzymes or are exceptions.

Despite the information gained from the neutron crystallography experiments, they are only suitable for intermediates that can be isolated in the large crystals required and are stable for long periods (weeks), this inevitably means that the experiments are conducted at ca. 100K. X-ray free electron lasers will not allow us to see hydrogen atoms but offer the prospect of using the femtosecond bursts and pump-probe techniques to determine the structures of very short lived states at room temperature. We have conducted preliminary serial fs experiments using the SCALA XFEL in Japan. These give us confidence we can use fs exposures to out-run photoreduction. To establish that good data could be obtained from microcrystals we tried the “chip” technique at Diamond Light Sources, this showed the iron centre is photoreduced in milliseconds and the microcrystals then bind dioxygen from the surrounding air, displacing the water normally seen.

The obvious approach to determine the structures of short-lived states is to use pump-probe techniques. The use of micro- or nano-crystals in a stream orthogonal to the X-ray pulse from a FEL permits mixing with reactant in a manner analogous to a stopped-flow enzyme kinetics experiment. The time between mixing and recording diffraction data can be altered by varying the distance between “pump” (adding reactant) and “probe” (fs X-ray diffraction). Shorter and precise time intervals can be accessed by the use of caged compounds, whereby a laser is used to photolyse and release active reactants at intervals before X-ray data is recorded. The obvious first objective will be “Compound 0” (Figure 1). This can be formed by either reacting ferric enzyme with H2O2 and catching it before it becomes Compound I, or by oxidizing the ferrous- oxy Compound III. Although this seems trivial, finding the exact temporal and chemical conditions will require extensive research.

Project details at https://www2.le.ac.uk/research-degrees/doctoral-training-partnerships/bbsrc/MOODYPeter_MIBTPProjectProposal201819.pdf