Skip to main content
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2012 May 18.
Published in final edited form as: Science. 2011 Nov 18;334(6058):940. doi: 10.1126/science.1214025

Evidence for Interstitial Carbon in Nitrogenase FeMo Cofactor

Thomas Spatzal 1, Müge Aksoyoglu 2, Limei Zhang 4, Susana LA Andrade 1,3, Erik Schleicher 2, Stefan Weber 2, Douglas C Rees 4, Oliver Einsle 1,3,*
PMCID: PMC3268367  NIHMSID: NIHMS349905  PMID: 22096190

The enzyme nitrogenase is the only known biological system able to break the triple bond of dinitrogen to yield bioavailable ammonium (1). Its active site, the FeMo cofactor, is a [Mo:7Fe:9S:X]:homocitrate cluster, the largest and most complex biological metal center known to date. The exact mode and position of N2 binding to the FeMo cofactor is unknown. The identification of a light atom X (C, N, or O) at its center in a 1.16 A resolution crystal structure gave rise to contradictory mechanistic proposals that require clarification (2). Our analysis of the diffraction data indicated that while an unambiguous assignment could not be made, X was most plausibly a nitrogen species. Through further optimization of protein isolation and crystallization we have obtained an improved structural model at dmin = 1.0 Å (Fig. 1A) (3). In the 1.16A resolution structure, the central atom was obscured by the geometry of the FeMo cofactor. Our suggestion of a nitrogen species was based on integrating electron density (ED) at the cofactor center using a probe radius of 1.4 Å, the approximate van-der-Waals radius of a candidate atom. Here we have varied this probe radius on a very fine ED-grid and have performed a statistical analysis for all light atoms in the structure. A plot of ED vs. probe radius for the average C, N, and O atoms shows that these can be clearly distinguished (Fig. 1B), with the distinction clearer for smaller radii. The corresponding curves for the two central atoms in the two copies of the FeMo cofactor in the asymmetric unit overlay perfectly with the curve obtained for all other carbons. A slight deviation to higher values was seen at a radius of 1.4 Å, indicating that the influence of the surrounding Fe atoms is already noticeable, and this in part biased our analysis of the earlier structure towards nitrogen. Note that in fig 1B plotted values do not represent integrated ED, but rather the average ED within the sphere of given radius. Alternatively, we have plotted the average of EDs for each exact atom position (ρ0) vs. its B-factor (Fig. 1C, S2). Again, the different light atoms group into distinct areas of the plot, with the interstitial atoms close to the center of the carbon distribution. While the ρ0 plot indicated the atom to be a carbon both in the 1.16 Å and in the 1.0 Å data, the ED/radius plot was far more ambiguous at the lower resolution and did not allow to distinguish between C and N for the central atom (3). The seemingly modest improvement in resolution obtained with the new data set, corresponds to an increase of the data-parameter ratio (with an anisotropic Uij temperature factor model) of 50%, from 4.3 (1.16 Å) to 6.6 (1.0 Å), underlining that the improved resolution was essential for assigning the identity of the central atom.

Figure 1.

Figure 1

Carbon in the center of the FeMo cofactor. A) The [Mo:7Fe:9S:C]:homocitrate FeMo cofactor. B) Average electron density in a sphere of given radius for all carbon (black), nitrogen (blue) and oxygen (red) atoms in the structure. The two central atoms in the asymmetric unit (green) closely follow the trace for carbon. C) Electron density at the atomic positions (ρ0) vs. residual B-factor shows C, N, and O to occupy distinct areas. The central atoms (green) fall within the carbon area. D) X-band 3-pulse ESEEM frequency domain spectra of WT (green), U-15N (blue) and U-13C (black) labeled nitrogenase. The inset depicts 13C resonances recorded with two different τ values.

To complement the diffraction data, we produced A. vinelandii MoFe protein labeled with 13C or 15N, respectively, for resonance spectroscopy. The isolation procedure was optimized to yield complete incorporation of the isotopes, far exceeding the labeling ratio of ~5 % reported earlier (4). Inspection of wild-type and two isotope-labeled (U-15N; U-13C) nitrogenases (3) by electron spin-echo envelope modulation (ESEEM), a powerful EPR-technique for detection of weak hyperfine couplings (hfcs) in paramagnetic moieties such as the clusters of nitrogenase (3), revealed for the U-13C labeled (I = 1/2) sample an additional spectral pattern, centered at the free 13C Larmor frequency (3.7 MHz, Fig. 1D) with a splitting of 2.5 MHz, that is not detected in the other samples. Two types of resonances can be discriminated, one originating from 13C atoms that are very weakly coupled to the paramagnetic FeMo cofactor (Fig. 1D, [I]), and the other from a more strongly coupled 13C hfc with significant unpaired electron spin density (Fig. 1D, [II]). The latter is expected for a carbon nearby or within the FeMo cofactor. This observation is consistent with X = C because the crystal structure does not reveal any carbons within the first coordination sphere of the cluster that could account for such a large hfc.

Supplementary Material

Supp. Material

Acknowledgments

We thank the staff at Swiss Light Source, Villigen, CH, and S.M. Consiglio for assistance with the calculations. The work was supported by DFG (grants Ei-520/7 to O.E., An-676/1 to S.L.A. and IRTG 1478), Canadian NSERC postdoctoral fellowship to L.Z. and NIH (grant GM45162 to D.C.R.) and HHMI (to D.C.R.). The structural model and structure factors have been deposited with the Protein Data Bank (accession code 3U7Q).

References

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supp. Material

RESOURCES