Metalloproteins: Methods and

Protocols Yilin Hu
Visit to download the full and correct content document:
https://textbookfull.com/product/metalloproteins-methods-and-protocols-yilin-hu/
More products digital (pdf, epub, mobi) instant
download maybe you interests ...

Zebrafish Methods and Protocols Koichi Kawakami

https://textbookfull.com/product/zebrafish-methods-and-protocols-
koichi-kawakami/

SNAREs: Methods and Protocols Rutilio Fratti

https://textbookfull.com/product/snares-methods-and-protocols-
rutilio-fratti/

Epitranscriptomics: Methods and Protocols Narendra

Wajapeyee

https://textbookfull.com/product/epitranscriptomics-methods-and-
protocols-narendra-wajapeyee/

Phytoplasmas: Methods and Protocols Rita Musetti

https://textbookfull.com/product/phytoplasmas-methods-and-
protocols-rita-musetti/
Nanotoxicity: Methods and Protocols Qunwei Zhang

https://textbookfull.com/product/nanotoxicity-methods-and-
protocols-qunwei-zhang/

Autoantibodies: Methods and Protocols Gunnar Houen

https://textbookfull.com/product/autoantibodies-methods-and-
protocols-gunnar-houen/

Oligodendrocytes: Methods and Protocols David Lyons

https://textbookfull.com/product/oligodendrocytes-methods-and-
protocols-david-lyons/

Bacteriophages Methods and Protocols Ebenezer Tumban

https://textbookfull.com/product/bacteriophages-methods-and-
protocols-ebenezer-tumban/

Antibody Engineering Methods and Protocols Damien

Nevoltris

https://textbookfull.com/product/antibody-engineering-methods-
and-protocols-damien-nevoltris/
Methods in
Molecular Biology 1876

Yilin Hu Editor

Metallo-
proteins
Methods and Protocols
METHODS IN MOLECULAR BIOLOGY

Series Editor
John M. Walker
School of Life and Medical Sciences
University of Hertfordshire
Hatfield, Hertfordshire, AL10 9AB, UK

For further volumes:

http://www.springer.com/series/7651
 Metalloproteins

Methods and Protocols

Edited by

Yilin Hu
Department of Molecular Biology and Biochemistry, University of California-Irvine, Irvine, CA, USA
Editor
Yilin Hu
Department of Molecular Biology and Biochemistry
University of California-Irvine
Irvine, CA, USA

ISSN 1064-3745 ISSN 1940-6029 (electronic)

Methods in Molecular Biology
ISBN 978-1-4939-8863-1 ISBN 978-1-4939-8864-8 (eBook)
https://doi.org/10.1007/978-1-4939-8864-8
Library of Congress Control Number: 2018957140

© Springer Science+Business Media, LLC, part of Springer Nature 2019

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the material is
concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction
on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation,
computer software, or by similar or dissimilar methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply,
even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations
and therefore free for general use.
The publisher, the authors, and the editors are safe to assume that the advice and information in this book are believed to
be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty,
express or implied, with respect to the material contained herein or for any errors or omissions that may have been made.
The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

This Humana Press imprint is published by the registered company Springer Science+Business Media, LLC part of
Springer Nature.
The registered company address is: 233 Spring Street, New York, NY 10013, U.S.A.
Preface

Complex metalloproteins catalyze some of the most remarkable chemical transformations in

biological systems. Many of the reactions catalyzed by these enzymes involve small mole-
cules, such as N2, CO, and H2, which are used to generate chemical building blocks and
energy for metabolic purposes. Despite intense efforts in this research area, the mechanisms
and biosynthesis of many of these complex metalloproteins are still poorly defined and
represent substantial and continuing challenges to biochemists, biophysicists, and synthetic
chemists. This volume attempts to provide an up-to-date, in-depth overview of the methods
that have been applied to studying the complex metalloproteins at a molecular level. A large
ensemble of approaches is covered in this volume, ranging from genetic, biochemical,
spectroscopic, and chemical methods to theoretical calculations. A project of this scope
requires the timely cooperation of many participants and I greatly appreciate the willingness
of all authors to face and meet such a challenge. I hope that this volume, written by
recognized experts in this research area, will be useful for anyone who is interested in
metalloprotein research and who is willing to take charge of addressing the unanswered
mechanistic and biosynthetic questions of these fascinating enzyme systems.

Irvine, CA, USA Yilin Hu

v
Contents

Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v
Contributors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix

PART I METALLOPROTEINS
1 Nitrogenases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
Nathaniel S. Sickerman, Yilin Hu, and Markus W. Ribbe
2 Enzymatic Systems with Homology to Nitrogenase: Biosynthesis
of Bacteriochlorophyll and Coenzyme F430. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
Jürgen Moser and Gunhild Layer
3 Carbon Monoxide Dehydrogenases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
Jae-Hun Jeoung, Berta M. Martins, and Holger Dobbek
4 Molybdenum-Containing Enzymes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
Dimitri Niks and Russ Hille
5 Hydrogenases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
Nathaniel S. Sickerman and Yilin Hu

PART II GENETIC AND BIOCHEMICAL METHODS

6 Genomic Manipulations of the Diazotroph Azotobacter vinelandii . . . . . . . . . . . . 91

Patricia C. Dos Santos
7 Purification of Nitrogenase Proteins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
Chi-Chung Lee, Markus W. Ribbe, and Yilin Hu
8 Expression, Purification, and Activity Analysis of Chlorophyllide
Oxidoreductase and Ni2+-Sirohydrochlorin a,c-Diamide Reductase . . . . . . . . . . . 125
Jürgen Moser, Jan Jasper, José Vazquez Ramos, Sven T. Sowa,
and Gunhild Layer
9 Reconstitution of Molybdoenzymes with Bis-Molybdopterin
Guanine Dinucleotide Cofactors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
Paul Kaufmann, Chantal Iobbi-Nivol, and Silke Leimkühler

PART III STRUCTURAL AND SPECTROSCOPIC METHODS

10 Crystallization of Nitrogenase Proteins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
Belinda B. Wenke, Renee J. Arias, and Thomas Spatzal
11 X-Ray Crystallography of Carbon Monoxide Dehydrogenases . . . . . . . . . . . . . . . . 167
Jae-Hun Jeoung, Berta M. Martins, and Holger Dobbek
12 X-Ray Absorption Spectroscopy of Metalloproteins . . . . . . . . . . . . . . . . . . . . . . . . . 179
Limei Zhang
13 Electron Paramagnetic Resonance Spectroscopy of Metalloproteins . . . . . . . . . . . 197
Andrew Jasniewski, Yilin Hu, and Markus W. Ribbe

vii
viii Contents

14 Magnetic Circular Dichroism Spectroscopy of Metalloproteins . . . . . . . . . . . . . . . 213

Brian J. Hales

PART IV CHEMICAL METHODS AND THEORETICAL CALCULATIONS

15 Chemical Synthesis of an Asymmetric Mimic of the Nitrogenase

Active Site . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229
Kazuki Tanifuji and Yasuhiro Ohki
16 Computational Methods for Modeling Metalloproteins . . . . . . . . . . . . . . . . . . . . . 245
Martin T. Stiebritz and Yilin Hu

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 267
Contributors

RENEE J. ARIAS  Division of Chemistry and Chemical Engineering, California Institute of

Technology, Pasadena, CA, USA
HOLGER DOBBEK  Institute of Biology, Structural Biology and Biochemistry, Humboldt-
Universit€at zu Berlin, Berlin, Germany
PATRICIA C. DOS SANTOS  Department of Chemistry, Wake Forest University, Winston-
Salem, NC, USA
BRIAN J. HALES  Department of Chemistry, Louisiana State University, Baton Rouge, LA,
USA
RUSS HILLE  Department of Biochemistry, University of California, Riverside, Riverside,
CA, USA
YILIN HU  Department of Molecular Biology and Biochemistry, University of California,
Irvine, Irvine, CA, USA
CHANTAL IOBBI-NIVOL  Aix-Marseille Université, CNRS, Marseille, France
ANDREW JASNIEWSKI  Department of Molecular Biology and Biochemistry, University of
California, Irvine, Irvine, CA, USA
JAN JASPER  Institut für Mikrobiologie, Technische Universit€
a t Braunschweig, Braunschweig,
Germany
JAE-HUN JEOUNG  Institute of Biology, Structural Biology and Biochemistry, Humboldt-
Universit€at zu Berlin, Berlin, Germany
PAUL KAUFMANN  Department of Molecular Enzymology, Institute of Biochemistry and
Biology, University of Potsdam, Potsdam, Germany
GUNHILD LAYER  Institut für Pharmazeutische Wissenschaften, Albert-Ludwigs-Universit€ at
Freiburg, Freiburg, Germany
CHI-CHUNG LEE  Department of Molecular Biology and Biochemistry, University of
California, Irvine, Irvine, CA, USA
SILKE LEIMKÜHLER  Department of Molecular Enzymology, Institute of Biochemistry and
Biology, University of Potsdam, Potsdam, Germany
BERTA M. MARTINS  Institute of Biology, Structural Biology and Biochemistry, Humboldt-
Universit€at zu Berlin, Berlin, Germany
JÜRGEN MOSER  Institut für Mikrobiologie, Technische Universit€ a t Braunschweig,
Braunschweig, Germany
DIMITRI NIKS  Department of Biochemistry, University of California, Riverside, Riverside,
CA, USA
YASUHIRO OHKI  Department of Chemistry, Graduate School of Science, Nagoya University,
Nagoya, Japan
MARKUS W. RIBBE  Department of Molecular Biology and Biochemistry, University of
California, Irvine, Irvine, CA, USA; Department of Chemistry, University of California,
Irvine, Irvine, CA, USA
NATHANIEL S. SICKERMAN  Department of Molecular Biology and Biochemistry, University of
California, Irvine, Irvine, CA, USA
SVEN T. SOWA  Institut für Biochemie, Universit€ at Leipzig, Leipzig, Germany

ix
x Contributors

THOMAS SPATZAL  Division of Chemistry and Chemical Engineering, California Institute of

Technology, Pasadena, CA, USA
MARTIN T. STIEBRITZ  Department of Molecular Biology and Biochemistry, University of
California, Irvine, Irvine, CA, USA
KAZUKI TANIFUJI  Department of Molecular Biology and Biochemistry, University of
California, Irvine, Irvine, CA, USA
JOSÉ VAZQUEZ RAMOS  Institut für Biochemie, Universit€
a t Leipzig, Leipzig, Germany
BELINDA B. WENKE  Division of Chemistry and Chemical Engineering, California Institute
of Technology, Pasadena, CA, USA
LIMEI ZHANG  Department of Biochemistry, University of Nebraska-Lincoln, Lincoln, NE,
USA; Redox Biology Center, University of Nebraska-Lincoln, Lincoln, NE, USA; Nebraska
Center for Integrated Biomolecular Communication, University of Nebraska-Lincoln,
Lincoln, NE, USA
 Part I

Metalloproteins
 Chapter 1

Nitrogenases
Nathaniel S. Sickerman, Yilin Hu, and Markus W. Ribbe

Abstract
Biological nitrogen fixation, the conversion of dinitrogen (N2) into ammonia (NH3), stands as a particu-
larly challenging chemical process. As the entry point into a bioavailable form of nitrogen, biological
nitrogen fixation is a critical step in the global nitrogen cycle. In Nature, only one enzyme, nitrogenase,
is competent in performing this reaction. Study of this complex metalloenzyme has revealed a potent
substrate reduction system that utilizes some of the most sophisticated metalloclusters known. This chapter
discusses the structure and function of nitrogenase, covers methods that have proven useful in the elucida-
tion of enzyme properties, and provides an overview of the three known nitrogenase variants.

Key words Biological nitrogen fixation, Nitrogenase, MoFe protein, Fe protein, P-cluster, M-cluster

1 Introduction

Molecular nitrogen (N2, dinitrogen) is an extremely stable mole-

cule that makes up 78% of the Earth’s atmosphere. The stability and
relative inertness of N2 are due to the strength of its diatomic triple
bond, whose scission represents one of the most challenging chem-
ical reactions in Nature. Although atomic nitrogen (N) is an essen-
tial component for life, its incorporation from N2 into a bioavailable
form, a process called biological nitrogen fixation, is a major bot-
tleneck of the global nitrogen cycle. The cleavage of N2 and its
conversion into bioavailable ammonia (NH3) is performed solely by
diazotrophic bacteria and archaea that express the enzyme nitroge-
nase. This enzyme is a set of complex metalloproteins that catalyzes
the nucleotide-dependent reduction of N2 into NH3, and at mini-
mum, each type of nitrogenase contains an α2β2 catalytic compo-
nent and a γ2 reductase component [1, 2]. Catalysis with
nitrogenase is enabled by complex Fe-S metalloclusters contained
within the protein components, and the clusters work in concert to
shuttle electrons and reduce bound substrates [3]. Due to the
multielectron reducing capabilities of nitrogenase, the system can

Yilin Hu (ed.), Metalloproteins: Methods and Protocols, Methods in Molecular Biology, vol. 1876,
https://doi.org/10.1007/978-1-4939-8864-8_1, © Springer Science+Business Media, LLC, part of Springer Nature 2019

3
4 Nathaniel S. Sickerman et al.

also facilitate the reduction of a number of other small molecule

substrates, including H+, N3
, C2H2, CO, and CO2 [4–7].
Three nitrogenase variants have been identified in diazotrophs:
an Fe-only system (Fe-nitrogenase), a variant that contains vana-
dium (V-nitrogenase), and a molybdenum-containing system
(Mo-nitrogenase) [8, 9]. Of these three variants, Mo-nitrogenase
is the most well-studied and characterized, while the other two
types are considered “alternative” nitrogenases. This chapter covers
many of the important aspects of Mo-nitrogenase and concludes
with a discussion of V-nitrogenase and the Fe-only system.

2 Mo-nitrogenase

All extant diazotrophs currently known contain Mo-nitrogenase.

This variant possesses the highest competency for fixing nitrogen
and accomplishes the feat according to the following reaction [3]:
N2 þ 8e
þ 8Hþ þ 16MgATP ! 2NH3 þ H2 þ 16MgADP þ 16Pi ð1Þ
Substrate reduction in Mo-nitrogenase is facilitated by the
products of nitrogen fixation (nif) genes [10], and two protein
components in particular comprise the enzyme (Fig. 1) [1]. The
reductase component, Fe protein (NifH), serves as the canonical

Fig. 1 Mo-nitrogenase complex between the MoFe protein (NifDK) and the Fe protein (NifH) from A. vinelandii,
with labeled components (a). Visualization of the complex clusters (b) indicating the ATP-coupled transfer of
electrons from the Fe4S4 cluster of NifH to the P-cluster at the αβ-subunit interface of NifDK and on to the
M-cluster within the α-subunit. The pseudo-twofold symmetry axis of the complex is indicated by a dashed
line. PDB ID: 1N2C
 Nitrogenases 5

electron source for substrate reduction in Mo-nitrogenase. This

protein contains a Fe4S4 cluster and two binding sites for MgATP,
and the hydrolysis of MgATP is coupled with electron transfer from
the Fe4S4 cluster of NifH to the Mo-nitrogenase catalytic compo-
nent [2, 3]. This component, referred to as the molybdenum–iron
(MoFe) protein or NifDK, is an α2β2 heterotetramer that contains
two unique metalloclusters, designated the [Fe8S7] P-cluster and
the [R-homocitrate(MoFe7S9C)] cofactor (called M-cluster or
FeMo-co). The structure and function of the proteins and metal-
loclusters are outlined in more detail below. Unless otherwise
noted, this section will discuss Mo-nitrogenase protein compo-
nents and amino acid compositions as derived from the soil bacte-
rium Azotobacter vinelandii.

2.1 Fe Protein (NifH) The Fe protein of Mo-nitrogenase is a ~60 kDa homodimer that is
encoded by the gene nifH. Deletion of this gene gives rise to
bacterial strains that are unable to fix nitrogen [11]. In addition,
these nifH deletion strains express a NifDK protein that contains a
pair of Fe4S4-like clusters (P*-cluster) instead of the P-cluster
[12]. Furthermore, nifH deletion variants of NifDK completely
lack the catalytic M-cluster [11]. These observations point to
three important roles played by NifH. First, NifH acts as the
reductase component during catalytic substrate reduction, transfer-
ring electrons to NifDK [3]. Second, the P-cluster is matured on
NifDK through the actions of NifH, whereby the precursor
P*-cluster is reductively coupled to form the unique [Fe8S7] species
[13]. Finally, NifH delivers Mo and R-homocitrate to a precursor
of the M-cluster to facilitate a critical step in cofactor assembly [14].
The reductive capabilities of NifH are enabled by a solvent-
exposed Fe4S4 cluster that sits at the homodimer interface, with
each subunit providing two Cys residues for cluster ligation
[15]. Under reducing conditions, the cluster of NifH resides in
the [Fe4S4]+ state, which is maintained in vivo by physiological
reductants such as flavodoxin and in vitro by the reductant sodium
dithionite (Na2S2O4). Perpendicular-mode electron paramagnetic
resonance (EPR) spectra of the [Fe4S4]+ species indicate that the
system exists as a mixture of S ¼ 1/2 and S ¼ 3/2 signals, the ratio
of which is sensitive to solution conditions [16]. Potentiometric
titration with NifH has established that the midpoint potential
(Em) for the one-electron oxidation to the [Fe4S4]2+ state occurs
at approximately
300 mV vs SHE [17]. The oxidized [Fe4S4]2+
species is diamagnetic, displays a featureless EPR spectrum, and can
be reversibly reduced back to the [Fe4S4]+ state. An additional “all-
ferrous” or “super-reduced” oxidation state for the NifH Fe4S4
cluster can be accessed as well through the use of strong reductants
such as TiIII- and EuII-containing compounds [18–21]. This
[Fe4S4]0 species exhibits a reddish-pink hue with a new electronic
absorbance peak at 540 nm, an unusual absorbance for [Fe4S4]-
6 Nathaniel S. Sickerman et al.

containing proteins, which are normally a dark brown color

[20]. Mössbauer spectroscopy confirms the all-ferrous assignment
of the [Fe4S4]0 system, and parallel-mode EPR spectroscopy reveals
a g ¼ 16 signal that is indicative of a species with an S ¼ 3 or 4 spin
state [19]. The physiological relevance of the [Fe4S4]0 state of NifH
is unclear, but the species has been proposed to facilitate the trans-
fer of two electrons with the hydrolysis of just two MgATP mole-
cules, thereby presenting a more economical means to transfer
electrons for substrate reduction [22].
During normal catalytic operation, the cluster of NifH is pro-
posed to cycle between the [Fe4S4]2+ and [Fe4S4]+ states [3, 23],
the properties of which are critically influenced by the ability of the
protein to bind and hydrolyze MgATP. Each protein subunit half
features one Walker A motif to which the nucleotide can bind [15],
and the effects of MgATP versus MgADP binding to the protein
lead to observable changes in the magnetic circular dichroism
(MCD), EPR, and small angle X-ray scattering (SAXS) spectro-
scopic signals [16, 24, 25]. Importantly, nucleotide binding
also influences the reduction potential of the Fe4S4 cluster,
which is located ~15 Å from the binding sites. The Em for the
[Fe4S4]2þ/1þ couple as observed in the absence of nucleotides
(~
300 mV vs SHE) decreases by over 120 mV (to ~
430 mV)
when MgATP is bound within the protein [17]. In the MgADP-
bound state of NifH, the Em of the [Fe4S4]2þ/1þ couple is shifted
roughly 10 mV lower than its MgATP-containing counterpart.
Thus, binding and hydrolysis of ATP within NifH yields a more
reducing Fe4S4 cluster.
Illustrative of the connection between the Fe4S4 cluster and
nucleotide binding, the dissociation constants (Kd) for nucleotide
binding affinity increase upon cluster oxidation. More specifically,
the affinity of MgATP binding to NifH at the [Fe4S4]+ state
(Kd ¼ ~ 500 μM) increases by an order of magnitude (Kd ¼ ~45 μM)
when the cluster is oxidized to the [Fe4S4]2+ form, and a similar
phenomenon is observed with MgADP [26]. Both X-ray crystal-
lography and SAXS studies on the protein indicate that structural
rearrangements occur in the different nucleotide-bound and
nucleotide-free states [15, 27, 28]. These changes in protein con-
formation from the binding of nucleotides binding appear to mod-
ulate the [Fe4S4] cluster reduction potential, which influences both
intraprotein electron transfer and association/dissociation events
during catalytic complex formation (see Subheading 3).

2.2 MoFe Protein The MoFe protein in Mo-nitrogenase is a ~220 kDa heterotetra-
(NifDK) mer that is encoded by the genes nifDK. The separate NifD (α) and
NifK (β) subunits each consist of three domains that feature alter-
nating α-helix / parallel β-sheet folds [29]. Overall, folding of the
αβ-subunit pairs generates an open, water-filled channel with ~8 Å
diameter that passes through the heterotetramer center, and the
 Nitrogenases 7

Fig. 2 Molecular structures of the A. vinelandii NifDK metalloclusters. The reduced (PN) and two-electron-
oxidized (POX) states of the [Fe8S7] P-cluster (a) and the [(R-homocitrate)MoFe7S9C] M-cluster (b), with
selected amino acid residues labeled. Atoms are colored as follows: Fe, orange; S, yellow; Mo, cyan; C,
white; N, blue; O, red. PBD IDs: 3U7Q, 2MIN

protein exhibits a pseudo-twofold axis of symmetry. Amino acid

regions on the protein surface allow NifDK to interface with a
number of other relevant proteins, including its redox partner
NifH [30], the cofactor assembly scaffold NifEN [14, 31], and
the O2-protection protein FeSII [32]. Additionally, each αβ half
of this protein contains two metalloclusters that are essential for
catalytic function [2]. The capacitor-like P-cluster sits at the α-
β-subunit interface (Fig. 2a) and transfers electrons to the catalytic
active site located in the α-subunit. Buried within the α-subunit
active site of NifDK, the catalytic M-cluster (Fig. 2b) serves as the
site of substrate binding and reduction. The properties of these two
metalloclusters are discussed below.

2.2.1 P-Cluster The [Fe8S7] P-cluster bridges between the α and β subunits of
NifDK at a location ~11 Å from the protein surface
[29, 33]. This cluster serves as a conduit for electrons as they are
shuttled between the Fe4S4 cluster of NifH and the α-subunit
M-cluster. Three reversible, physiologically relevant oxidation
states have been identified for the P-cluster: the reduced PN state,
one-electron-oxidized P1+ state, and the two-electron-oxidized
POX state. Under conventional dithionite-reduced conditions, the
PN “resting state” predominates; Mössbauer and EPR spectrosco-
pies indicate that this species is diamagnetic, consisting of eight
ferrous Fe atoms [34, 35]. Crystal structures of NifDK in the PN
state present a P-cluster structure that can be viewed as the fusion of
8 Nathaniel S. Sickerman et al.

two Fe4S4 clusters that share a corner S atom at the cluster center
(Fig. 2a, top). This central μ6-S atom is a unique feature of P-cluster
PN state—one that has never been observed in any other biological
cluster. Additionally, for the all-ferrous PN state, the Fe atoms are
ligated by a total of six Cys residues, with three provided by each
subunit. In this capacity, the cysteinyl ligands provide four terminal
and two bridging linkages.
The important electron-gating role played by P-cluster is facili-
tated by its ability to reversibly reorganize its structure, thereby
maintaining its ability to cycle through oxidation states and deliver
electrons to the M-cluster. Oxidation of PN by two electrons to the
POX state (Em ¼ ~
290 mV) dramatically alters the linkages
between the amino acid residues and the Fe atoms (Fig. 2b, bot-
tom) [24, 36]. More specifically, the POX state features a configura-
tion in which one cubanoid half of the cluster core opens up,
breaking two Fe-SCys bonds and forming two new Fe-X bonds
with a α-Ser188 O atom and the backbone N atom of α-Cys88.
The binding of harder N and O donor atoms to the Fe atoms of the
oxidized P-cluster stabilizes the open cluster structure and appears
to enable the reversibility of this state. Unlike the S ¼ 0 PN state,
which produces featureless EPR spectra, parallel-mode EPR spec-
troscopy of the POX state reveals a diagnostic signal at g ¼ 11.8
[37]. This POX EPR signal indicates the presence of an integer-spin
system that likely originates from an S ¼ 3 or 4 spin state. An EPR
spectrum for the intermediate P1+ state can be observed as well
and contains signals that suggest a mixture of S ¼ 1/2 and S ¼ 5/2
states [34, 38].

2.2.2 M-Cluster The M-cluster is housed within the α subunit of NifDK

[29, 39]. The initial crystal structure of NifDK revealed the
M-cluster to be an unusual dicubanoid cofactor containing a
[MoFe7S9] core [29], and an improved structure later located elec-
tron density consistent with a light atom within the cluster architec-
ture [39]. Subsequent X-ray emission spectroscopy (XES) analyses
have unequivocally identified the light atom as C4
(carbide), coor-
dinated to the six core Fe atoms [40]. Furthermore, biochemical
studies have pinpointed the C atom source to be from the S-methyl
group of S-adenosylmethionine, which enters the cofactor biosyn-
thetic pathway via the assembly protein NifB [41, 42]. The overall
structure of the [MoFe7S9C] core consists of the following features:
a [MoFe3S3] cubanoid half; a [Fe4S3] cubanoid half; three
Fe-bridging μ2-S atoms occupying the “belt” region of the cluster;
and the interstitial μ6-C at the cluster center. An additional ligand,
R-homocitrate, is also part of the M-cluster and coordinates to the
Mo atom through a hydroxide and one carboxylate moiety.
Only two amino acid residues within the NifDK active site
provide covalent contacts to the M-cluster: the α-Cys275 residue
 Nitrogenases 9

coordinates to the peripheral Fe atom opposite Mo, and the

α-His442 side chain is bound to the remaining coordination site
on the opposing Mo center. These amino acid linkages can be
severed through protein-denaturing techniques to labilize and
extract intact M-cluster into organic solvents [43, 44]. The
M-cluster extracts can be combined with cofactor-deficient
NifDK (apo-NifDK), which is inactive for substrate reduction
[43, 45]; apo-NifDK reconstituted with M-cluster in this way
completely regains its catalytic substrate reduction activity, clearly
illustrating the essentiality of the cofactor. Additionally, cluster
extracts can be used in the presence of a proton source, strong
reductant, and C1-substrate source (i.e., CN
, CO, or CO2) to
form short-chain hydrocarbon products [46–48].
Although the M-cluster displays catalytic activity when
extracted from NifDK, reduction of N2 has only been demon-
strated for the protein-bound cofactor. Therefore, the active-site
environment of the protein is critical for facilitating the proper
movement of substrates, protons, and electrons. The roles of the
amino acid residues comprising the M-cluster secondary sphere
have been discussed in previous articles [2, 49–52].
Consistent with its multielectron processing capabilities, the
M-cluster can be observed in several different oxidation states.
Under dithionite-reduced conditions, the M-cluster resides in the
so-called resting state (MN). Perpendicular-mode EPR spectros-
copy of the M-cluster MN state reveals a diagnostic rhombic
S ¼ 3/2 signal with features at g ¼ 4.7, 3.7, and 2.0 [53]. Chemical
oxidation of the MN species (~
40 mV vs NHE) yields the
one-electron oxidized MOx state, which exhibits an S ¼ 0 spin
state and displays a featureless EPR spectrum [54]. More reduced
forms of the MN state are also accessible: reduction of resting-state
NifDK by reduced, MgATP-bound NifH leads to generation of a
new MR state, with a midpoint reduction potential of approxi-
mately
495 mV determined for the MN/R couple [55]. This
reduced MR state of the M-cluster possesses an integer spin state
(S > 1) and produces a featureless EPR spectrum, even when
probed in parallel mode [56].
A combination of Mössbauer, X-ray absorption spectroscopy
(XAS), and EPR spectroscopic techniques has improved our under-
standing of the M-cluster MN state electronic structure [35, 53,
56]. Particularly, high-energy resolution fluorescence detected
(HERFD) Mo K-edge XAS of the MN state indicates that the
heterometal exists as an S ¼ 1/2 MoIII center [57]. This result
suggests that the M-cluster possesses the first physiological example
of Mo in the 3þ oxidation state, and its low-spin state is attributed
to a spin-coupled “quasi non-Hund” ("##) configuration.
Recently, the oxidation states of the MN state Fe atoms have been
evaluated using Fe K-edge XAS-derived spatially resolved
10 Nathaniel S. Sickerman et al.

anomalous dispersion (SpReAD) methods [58]. According to the

SpReAD analysis, which has enabled the refined determination of
electron density from the individual, spatially separated Fe
atoms, the formal MN state configuration is best represented as
[3FeII:4FeIII:MoIII]. Taken together with the formal oxidation
states of closed-shell carbide (C4
) and sulfide (S2
) ions, the
overall charge of the [MoFe7S9C] core in its resting state is most
likely
1 [58]. This conclusion tracks well with experimental and
theoretical treatments of the M-cluster. Furthermore, the presence
of the polyanionic ligand (R)-homocitrate ensures that even the
MOX state, whose core is a neutral charge, would be anionic overall.

2.3 Mechanism The mechanism underlying the multielectron, multiproton reduc-

of Mo-Nitrogenase tion of N2 into NH3 by NifDK has been elucidated over decades
through a combination of kinetic, spectroscopic, and structural
2.3.1 The NifH:NifDK
biology approaches. Electron equivalents in Mo-nitrogenase are
Complex
transferred when reduced, MgATP-bound NifH binds to NifDK,
and the cycle followed by the reductase component during catalysis
is shown in Fig. 3. Visualizations of this dynamic process have been
made possible through the crystallization of the NifH:NifDK com-
plex in various forms: NifH MgADP adducts [59], NifH contain-
ing nonhydrolyzable ATP analogs [30], NifH mutants incapable of
hydrolyzing ATP [60], and NifH covalently linked to NifDK [61]
have all yielded solid-state snapshots of the encounter complex.
The study of mismatched NifH:NifDK complexes using protein
components that originate from different organisms has proven
useful as well [62–65]. Coupled with biochemical and

Fig. 3 Proposed Mo-nitrogenase Fe protein (NifH) cycle and its association with the MoFe protein (NifDK), with
indicated points of ATP hydrolysis. States are labeled as follows: NifHR, reduced NifH; NifHOX, oxidized NifH;
NifDK, state of NifDK prior to reduction event; NifDKR, state of NifDK following reduction event
 Nitrogenases 11

spectroscopic studies, a more comprehensive view of electron

transfer between the two Mo-nitrogenase components has
emerged. Kinetic experiments suggest that formation of the tran-
sient NifH:NifDK complex initially induces electron transfer in
NifDK, resulting in the intramolecular shuttling of an electron
from the resting-state P-cluster to the M-cluster [66]. According
to this “deficit spending” model, the one-electron-oxidized P1+
state is then re-reduced by rapid “backfilling” with an electron
from the [Fe4S4]+ cluster of NifH. The conformational changes
associated with electron transfer in NifH [67] lead to ATP hydro-
lysis, which in turn results in a release of Pi and dissociation of the
complex [68]. The details concerning the re-reduction of NifH and
replacement of MgADP with MgATP are still unclear. Regardless,
the formation of the transient complex between reduced, MgATP-
bound NifH and NifDK must repeat multiple times until the
catalytic component is capable of performing substrate reduction.

2.3.2 The The initial kinetic treatment of the Mo-nitrogenase mechanism

Lowe–Thorneley Model involved the work of many groups and culminated with the Low-
e–Thorneley (LT) kinetic model (Fig. 4a) [69–71]. This model was

Fig. 4 Modified Lowe–Thorneley kinetic scheme (a) for the reduction of N2 by Mo-nitrogenase. The En notation
represents one αβ half of the MoFe protein that has accumulated n electrons. The resting state is denoted as
E0, and the general points of NH3 release are noted. The proposed intermediates of the distal and alternating
pathways for N2 hydrogenation (b) are shown, with the corresponding En states indicated
12 Nathaniel S. Sickerman et al.

constructed from the results of steady-state, stopped-flow, and

freeze-quench kinetic studies and outlines the delivery of protons
and electrons to NifDK. The convention for naming the various
reduced states during catalysis uses the En notation, which accounts
for the number (n) of electrons added to one αβ-half of NifDK
along the catalytic pathway. Importantly, the En notation applies to
the overall number of reducing equivalents added to the system and
does not distinguish whether the electrons reside on the P-cluster
or the M-cluster. The designation for the resting state, E0, repre-
sents the start of the catalytic cycle. The subsequent E1–E3 states
precede the binding of N2, which is proposed to occur at the E4
state, and further addition of electrons and protons to form the
E5–E8 states leads to the release of two NH3 equivalents and a
return to the E0 state.
Two substrates that are isoelectronic to N2, acetylene (C2H2)
and carbon monoxide (CO), have aided in the construction and
understanding of the kinetic model [72]. Both molecules act as
inhibitors of N2, partly due to their ability to bind the M-cluster at
the E2 level, whereas N2 must bind at the E4 level to be reduced.
Many other pathways are accessible within the framework of the LT
model beyond what are discussed in this chapter [3, 73].
Spectroscopic studies of Mo-nitrogenase proteins have been
invaluable in validating the assignment of many of the intermedi-
ates originally invoked in the LT mechanistic scheme. In particular,
pulsed-wave EPR techniques such as electron nuclear double reso-
nance (ENDOR) and electron spin-echo envelope modulation
(ESEEM) spectroscopy have provided the means to illuminate key
steps of N2 binding, activation, and reduction [74–78]. Using
these spectroscopic methods, two complementary approaches
have been employed: the use of NifDK variants containing key
amino acid point mutations within the active site, and the use of
substrate analogs that can trap certain intermediate states. Certain
amino acid substitutions made within the active site of NifDK can
slow or alter the reactivity, trapping intermediate species that may
have relevance within the catalytic cycle [77]. Moreover, pulsed-
EPR analyses of NifDK mutants can be used to provide additional
information about the mechanism by studying the reactions with
substrate analogs such as methyldiazine (CH3N¼NH) and partially
reduced substrates such as hydrazine (H2NNH2). Isotopically
labeled analogs of these molecules can provide additional informa-
tion as well. The contribution of point mutants and select sub-
strates toward evaluating the mechanism of Mo-nitrogenase has
been covered in detail elsewhere [73].

2.3.3 The E4 State: One perplexing aspect of the nitrogenase catalysis has been the
Binding of N2 cause underlying the release of H2 as a byproduct during N2
reduction. The loss of H2 is an unavoidable process to which a
 Nitrogenases 13

quarter of the ATP consumed by Mo-nitrogenase is diverted, sug-

gesting that the process may be critically tied to N2 binding or
reduction. Methods to uncover the answer to this long-standing
nitrogenase mystery have included freeze-quench techniques with
NifDK point mutants, the use of isotopically labeled substrates, and
a combination of EPR and ENDOR spectroscopic analyses
[79]. The results of these studies point to the E4 state of the LT
scheme as a critical juncture for N2 binding and activation that is
coupled to concomitant H2 release.
An effective strategy for trapping intermediate species has
involved the substitution of key amino acid residues found within
the active site of NifDK. One such mutant, α-Val(70)Ile, has a
congested active site due to the increased steric bulk of the Ile
side chain, and this mutation abolishes the binding, activation,
and reduction of all substrates except protons. Since only electrons
and protons can access the active site of the NifDK α-Val(70)Ile
mutant, freeze-quench methods can be used under turnover con-
ditions to accumulate an intermediate species that corresponds to
the E4 state in the LT scheme. Analyzing the putative E4 state with
ENDOR spectroscopy indicates that this state features two
Fe-bridging hydride (H
) units; this species, which is also proposed
to be loaded with two sulfur-bound protons, has been termed the
“Janus intermediate” [79]. End-on binding of N2 to a core Fe
atom in the E4 “Janus” state putatively triggers the formal “reduc-
tive elimination” of H2 to render an E4(2N2H) state. Observation
of this species provides a mechanism that accounts for both the
binding/activation of N2 and the loss of H2.

2.3.4 N2 Hydrogenation Following the E4-state binding of N2 and release of H2 within the
Pathways NifDK active site, successive electrons and protons are added to the
system to afford two equivalents of NH3. The exact mechanism of
N2 hydrogenation is still a debated subject, and two main pathways
have been proposed to account for the process: the distal and
alternating mechanisms (Fig. 4b). Both pathways begin with the
N2 molecule bound in an end-on fashion to a metal center (M-N2)
and converge at the formation of a terminal amido (M-NH2)
species. The overall catalytic cycle ends with the addition of a
proton and electron to the amido unit, which leads to a liberation
of NH3 and a return to the E0 state. The major differences between
the two proposed pathways concern the sites of hydrogenation,
intermediates formed, and the steps at which NH3 is released.
The distal pathway for N2 hydrogenation is supported by the
landmark studies of inorganic Mo complexes by Chatt [80–83] and
Schrock [84, 85]. In this mechanism, the distal N atom of the
metal-bound N2 molecule is hydrogenated by successive H
atoms, first yielding a hydrazido (M¼NNH2) intermediate. An
additional hydrogenation results in the release of one equivalent
14 Nathaniel S. Sickerman et al.

of NH3 and a terminal nitrido species. Further reduction and

protonation of the nitrido unit leads to the formation of the termi-
nal M-NH2 species. Conversely, the proposed alternating pathway
first involves the single hydrogenation of both the proximal and
distal N atoms of N2 to result in a bound diazene (HN¼NH)
species. The hydrogenation events continue to alternate, producing
a hydrazine (H2NNH2)-bound intermediate. Subsequent addition
of protons and electrons releases NH3 and furnishes the terminal
M-NH2 species.
Although a distal hydrogenation pathway has been demon-
strated in model Mo systems, the N2-reducing ability of
Fe-nitrogenase [86, 87] suggests that a heterometal is unnecessary
for this type of chemical transformation. However, the Peters
group has demonstrated that single-Fe sites in synthetic complexes
can facilitate the reduction of N2, and some of the intermediates
formed comport with the proposed distal mechanism [88, 89]; in
fact, these complexes also form H2NNH2, an alternating pathway
intermediate, suggesting that a hybrid distal/alternating mecha-
nism could be operative. Hydrazine can be used as a substrate for
Mo-nitrogenase to generate two equivalents of NH3, and the com-
pound can also be detected when quenching the enzyme with
strong acid or base under turnover conditions. Furthermore,
V-nitrogenase produces trace amounts of H2NNH2 during cata-
lytic reduction of N2. Assuming that the mechanism for N2 reduc-
tion is conserved among the nitrogenase variants, the results are
consistent with H2NNH2 as an intermediate, which supports the
alternating mechanism. Regardless of the exact pathway of N2
hydrogenation, consensus seems to growing that N2 reduction
occurs at M-cluster Fe atoms [73, 90]. While the alternating hydro-
genation pathway appears to have the most experimental support,
the debate over the mechanism of Mo-nitrogenase is far from
settled.

3 Alternative Nitrogenases

The alternative nitrogenases act as important backup systems for

diazotrophs when the trace metal Mo is not available. These sys-
tems lack the efficiency of Mo-nitrogenase in terms of catalytic N2
reduction but have interesting properties in their own right. For
instance, the reduction of C2H2 by Mo-nitrogenase almost exclu-
sively yields ethylene (C2H4), whereas a mixture of C2H4 and fully
reduced ethane (C2H6) is observed in the case of V- and
Fe-nitrogenase [91, 92]. During C2H2 reduction with the alterna-
tive nitrogenases, the electron flux is substantially more biased
toward proton reduction, leading to a lower total amount of
reduced C2H2 compared to Mo-nitrogenase. Furthermore, wild-
 Nitrogenases 15

type V-nitrogenase from A. vinelandii has been shown to reduce

CO and CO2 into short-chain hydrocarbon products [4, 93]. In
contrast, CO acts as a strong noncompetitive inhibitor of Mo-ni-
trogenase at all concentrations [94]. The ability of Fe-nitrogenase
to reduce CO into hydrocarbons has not yet been evaluated, but
the enzyme shows unusual behavior in C2H2/CO competition
experiments [92].
The V-containing and Fe-only variants are encoded by vnf and
anf genes, respectively [8, 9, 95]. However, some nif gene pro-
ducts are still essential for the functioning of the alternative nitro-
genases. For example, in organisms such as A. vinelandii and
Rhodobacter capsulatus, both of which express all three nitrogenase
variants, the alternative nitrogenases require expression of nifM,
nifU, nifS, nifV, and nifB [96–99]. The gene product NifM has
been identified as a proline isomerase that facilitates the proper
folding of the nitrogenase Fe proteins [98, 99]. The other gene
products are involved in the assembly of the catalytic cofactors
[100]: NifS, a cysteine desulfurase protein, mobilizes S atoms for
the construction of Fe-S clusters [101]; NifU assembles Fe4S4 units
from Fe and S atoms [102, 103]; NifV synthesizes R-homocitrate
for incorporation into the cofactor [104]; and NifB couples two
Fe4S4 clusters with inclusion of S and C atoms to form a [Fe8S9C]
cofactor precursor called the L-cluster [42, 105, 106]. The
L-cluster formed on NifB is proposed to serve as a common pro-
genital cluster to the catalytic cofactors of all three nitrogenase
variants [100]. Accordingly, deletion of the nifB genes from a
given diazotrophic strain provides a reliable method to obtain
cofactor-deficient (apo) catalytic components for any variant.

3.1 V-Nitrogenase The vnf genes are repressed by the presence of molybdate
(MoO42
) and NH3; in their absence, along with the presence of
a suitable vanadium source, expression of V-nitrogenase can occur
[95, 107]. This enzyme consists of the reductase component VnfH
and catalytic component VnfDGK. The structure of VnfH is not yet
known; however, this reductase component is highly homologous
to the Fe protein of Mo-nitrogenase, NifH, with the two proteins
sharing ~91% amino acid sequence identity. Based on biochemical
and spectroscopic analyses, VnfH also contains an Fe4S4 cluster that
differs only somewhat from NifH. The Em of the VnfH [Fe4S4]2þ/
1þ
couple is about 40 mV lower than that of NifH [108], and the
cluster structure also exhibits subtle differences compared to its
Mo-nitrogenase counterpart [109]. However, the VnfH Fe4S4
cluster can still access all three oxidation states known for NifH,
including the all-ferrous [Fe4S4]0 state [108].
The catalytic component of V-nitrogenase, which is encoded by
the genes vnfDGK, is referred to as VnfDKG or the VFe protein. In
A. vinelandii, the gene products VnfD and VnfK share approxi-
mately 32% sequence identity compared to NifD and NifK of
16 Nathaniel S. Sickerman et al.

Mo-nitrogenase. An additional subunit, VnfG, is also present in

V-nitrogenase. This small ~26 kDa protein is homologous to other
nitrogenase and nitrogenase-associated accessory proteins such as
NifY [10], but its exact function is not known [110, 111]. Depend-
ing on the protein purification protocols for VnfDGK, the catalytic
component as-isolated contains a variable number of the VnfG
subunits per α2β2 heterotetramer. For example, purified protein
with α2β2δ2 (heterohexameric) [112] and α2β2δ4 (heterooctameric)
[113] compositions have both been isolated. Despite differing
subunit compositions, in vitro studies with VnfDGK indicate that
the overall catalytic activity does not seem to be affected by the ratio
of VnfG to the catalytic component [9].
A variety of spectroscopic and biochemical analyses have been
applied to investigate VnfDGK. In the dithionite-reduced state, the
EPR spectrum of VnfDGK displays both S ¼ 1/2 and S ¼ 3/2
signals, and the S ¼ 1/2 signal has been assigned to the P-cluster
within VnfDGK. While the P-cluster of reduced NifDK (PN state)
does not give rise to an observable EPR signal, Mössbauer analysis
of dithionite-reduced VnfDGK suggests that 5–10% of the
P-cluster may exist in the oxidized P1+ state, which is consistent
with the spectral data [114]. The rhombic S ¼ 3/2 signal contains
g-values at 5.5, 2.05, and 1.94 and is attributed to the active-site
VFe-cofactor, also termed V-cluster [115, 116]. X-ray methods
such as XAS, XES, and extended X-ray absorption fine structure
(EXAFS) spectroscopy indicate that the V-cluster structure is very
similar to the M-cluster in NifDK [113, 117–119]. Both clusters
possess an interstitial C4
atom [120, 121], and the overall arrange-
ment of the Fe atoms appears to be analogous. The most notable
spectroscopic difference is that based on the measured Fe-V dis-
tances versus the Fe-Mo distances, the V-cluster exhibits a more
elongated structure compared to the M-cluster [118, 122]. The
observed elongation of the V-cluster is also apparent even when the
cluster is extracted into organic solvent [123].
To date, only one structure of VnfDGK has been reported,
isolated from A. vinelandii Fig. 5a [124]. This 1.35 Å structure
features a symmetrical, 240 kDa α2β2δ2 species consisting of an
overall α2β2 core arranged similarly to NifDK, with the addition of
two small VnfG subunits. Each VnfG polypeptide folds into a
bundle of four α-helices, and the subunit binds to the surface of
VnfD. The VnfG:VnfD docking site appears to be sufficiently dis-
tant from the VnfDK αβ-subunit interface to still allow for com-
plexation by VnfH. Similar methods used to obtain diffraction-
quality crystals of NifH:NifDK complexes, e.g., the formation of
non-dissociable complex systems, will likely be useful in future
experiments for obtaining crystal structures to visualize the VnfH:
VnfDGK complex.
The crystal structure of VnfDGK has provided a detailed
glimpse into the architecture of its metalloclusters [124]. The
 Nitrogenases 17

Fig. 5 Structural representation of the VFe protein (VnfDGK) from A. vinelandii (a) with clusters for one
αβ-subunit half indicated. The α-subunit [(R-homocitrate)VFe7S8C(CO3)] V-cluster is shown (b), with selected
amino acid residues labeled. Atoms are colored as follows: Fe, orange; S, yellow; V, magenta; C, white; N,
blue; O, red. PDB ID: 5N6Y

P-cluster structure within VnfDGK has been modeled to be nearly

identical to that of NifDK, with the exception of the binding of one
Fe atom whose bond with the central S atom is broken to form a
bond with the harder β-Ser153 O atom. The authors note the
fractional occupancy of oxidized P-cluster (P1+ state), which is
corroborated by spectroscopic studies and may account for the
partially opened P-cluster that is observed [114]. Also consistent
with previous spectroscopic analyses, the V-cluster within VnfDGK
possesses a more elongated structure compared to the M-cluster. As
with the cofactor in NifDK, the V-cluster is coordinated by Cys at
the peripheral Fe atom and by a His residue at the V center, which is
coordinated by a homocitrate ligand as well. In the most striking
difference between the two cofactors, the V-cluster electron density
reveals that one of the belt S2
atoms is replaced by a species that
refines best as a bridging μ-1,3-carbonate (CO32
) ligand (Fig. 5b).
This revelation results in a V-cluster composition of [(R-homoci-
trate)VFe7S9C(CO3)]. The CO32
unit fits into a binding cleft that
is not present within NifDK; in VnfDGK, the swapping of the
sequence order of two key active-site residues, leucine and proline,
compared to NifDK produces the binding pocket, which provides
hydrogen-bonding stabilization to the CO32
O atoms.
The unexpected composition of the V-cluster prompts many
questions concerning the origin of the CO32
ligand, its point of
addition along the cluster biosynthetic pathway, and its possible
effect on the catalytic activity of V-nitrogenase. Regarding the first
18 Nathaniel S. Sickerman et al.

two points, the putative vnf gene products VnfP1, VnfP2, and
VnfP3 have been identified and appear to be homologs of molyb-
dopterin biosynthesis proteins [124]. These proteins may play a
role in replacing S2
with CO32
, but more genetic, biochemical,
and spectroscopic studies will be required to evaluate this hypothe-
sis. The presence of CO32
may also explain the unique reactivity of
V-nitrogenase toward CO, but more examples of VnfDGK struc-
tures are needed to verify that the bridging ion is truly a represen-
tative feature of the V-cluster.

3.2 Fe-Nitrogenase The Fe-only nitrogenase system is encoded by anf genes and repre-
sents the least characterized of the three variants. Methods to
achieve the selective expression of this alternative nitrogenase sys-
tem include genetic manipulation to delete the main Mo- and
V-nitrogenase gene clusters (nifHDK and vnfHDGK, respectively)
and the use of tungstate (WO42
) in cell cultures, which represses
expression of Mo-nitrogenase [8, 87, 92]. Isolation of the catalytic
component has only been reported for a few organisms, including
A. vinelandii, R. capsulatus, and Rhodospirillum rubrum [8, 9].
The fact that this enzyme is capable of catalyzing nitrogen fixation
without the need for a heterometal such as Mo or V makes the
Fe-nitrogenase a compelling system for studying substrate reduc-
tion. Hydrogenation of N2 by an all-Fe system resembles the indus-
trial Haber–Bosch process, which serves as a high-temperature,
high-pressure analog for nitrogen fixation [125, 126]. Thus, the
insights that may be derived from Fe-nitrogenase have particular
relevance for understanding how Nature can convert N2 to NH3
under ambient conditions, using only the Earth-abundant transi-
tion metal Fe.
The reductase component of Fe-nitrogenase, AnfH, bears 63%
and 64% primary sequence identity with NifH and VnfH, respec-
tively. The three Fe protein variants as-isolated from A. vinelandii
have been investigated side-by-side using EPR and XAS spectro-
scopic analyses [109], and while all have similar properties, the
Fe4S4 clusters from the alternative Fe proteins VnfH and AnfH
appear to possess more closely aligned geometric features. Notably,
all three Fe protein homologs are capable of supporting substrate
reduction with A. vinleandii NifDK as the catalytic component.
The catalytic component of Fe-nitrogenase, AnfDGK, appears to
be best represented as a α2β2δ2 heterohexamer, and perpendicular-
mode EPR spectra of AnfDGK reveal a rhombic S ¼ 1/2 signal
centered at g ¼ 2 and lack the S ¼ 3/2 features exhibited by its
heterometal-containing counterparts [87]. Similar to VnfDGK, the
S ¼ 1/2 signal observed for AnfDGK is believed to arise from a
small percentage P-cluster in the P1+ state. The catalytic cofactor of
Fe-nitrogenase (referred to as the FeFe-cofactor) is hypothesized to
be a [(R-homocitrate)Fe8S9C] cluster. This assignment would
make the structure of the FeFe-cofactor similar to the [Fe8S9C]
 Nitrogenases 19

cofactor precursor (L-cluster) that is constructed on NifB [100],

albeit with the addition of the ligand homocitrate. Dithionite-
reduced L-cluster is EPR-silent [127], consistent with the lack of
a cofactor-associated signal from reduced AnfDGK as compared to
its Mo- and V-containing counterparts. Whether or not the cata-
lytic cofactor of Fe-nitrogenase contains a bridging CO32
ligand,
similar to the reported structure of VnfDGK [124], remains to be
seen. Clearly, there is plenty of room for improving our under-
standing of Fe-nitrogenase structure and function.

4 Summary and Outlook

The undeniably complex nitrogenase metalloproteins and their

associated metalloclusters remain captivating subjects of study due
to their unique structural and reductive catalytic properties. An
interdisciplinary combination of molecular biology, biochemistry,
inorganic chemistry, spectroscopy, and computational theory has
been employed to elucidate the features of this elaborate enzyme
system. While significant advances in our understanding have
brought the structure–function relationships of Mo-nitrogenase
proteins into greater focus, considerable work remains to elucidate
the complete mechanisms and functions of related gene products
responsible for modulating enzymatic regulation, cluster biosyn-
thesis, and substrate reduction. Furthermore, researchers have only
begun to scratch the surface regarding the alternative nitrogenases,
with the recent structure of VnfDGK representing the first in an
inevitable series probing deeper into V-nitrogenase. Using similar
methods and techniques, a more intimate understanding of
Fe-nitrogenase will likely follow as well. These steps are necessary
to fully appreciate and take advantage of the nature’s premier
system for nitrogen fixation.

Acknowledgments

The authors are supported by the National Institutes of Health

grant GM67626 (to M.W.R. and Y.H.).

References
1. Brill WJ (1980) Biochemical genetics of nitro- 4. Lee CC, Hu Y, Ribbe MW (2010) Vanadium
gen fixation. Microbiol Rev 44:449–467 nitrogenase reduces CO. Science
2. Howard JB, Rees DC (1996) Structural basis 329:642–642
of biological nitrogen fixation. Chem Rev 5. Hardy RWF, Knight E (1967)
96:2965–2982 ATP-dependent reduction of azide and
3. Burgess BK, Lowe DJ (1996) Mechanism of HCN by N2-fixing enzymes of Azotobacter
molybdenum nitrogenase. Chem Rev vinelandii and Clostridium pasteurianum.
96:2983–3011 Biochim Biophys Acta 139:69–90
20 Nathaniel S. Sickerman et al.
6. Dilworth MJ (1966) Acetylene reduction by
nitrogen-fixing preparations from Clostrid-
ium pasteurianum. Biochim Biophys Acta
127:285–294
7. Seefeldt LC, Rasche ME, Ensign SA (1995)
Carbonyl sulfide and carbon dioxide as new
substrates, and carbon disulfide as a new
inhibitor, of nitrogenase. Biochemistry
34:5382–5389
8. Joerger RD, Bishop PE (1988) Bacterial alter-
native nitrogen fixation systems. Crit Rev
Microbiol 16:1–14
9. Eady RR (1996) Structure-function relation-
ships of alternative nitrogenases. Chem Rev
96:3013–3030
10. Jacobson MR, Brigle KE, Bennett LT et al
(1989) Physical and genetic map of the
major nif gene cluster from Azotobacter vine-
landii. J Bacteriol 171:1017–1027
11. Robinson AC, Dean DR, Burgess BK (1987)
Iron-molybdenum cofactor biosynthesis in
Azotobacter vinelandii requires the iron pro-
tein of nitrogenase. J Biol Chem
262:14327–14332
12. Ribbe MW, Hu Y, Guo M et al (2002) The
FeMoco-deficient MoFe protein produced by
a nifH deletion strain of Azotobacter vinelan-
dii shows unusual P-cluster features. J Biol
Chem 277:23469–23476
13. Hu Y, Fay AW, Lee CC et al (2007) P-cluster
maturation on nitrogenase MoFe protein.
Proc Natl Acad Sci U S A 104:10424–10429
14. Hu Y, Corbett MC, Fay AW et al (2006)
Nitrogenase Fe protein: a molybdate/homo-
citrate insertase. Proc Natl Acad Sci U S A
103:17125–17130
15. Georgiadis MM, Komiya H, Chakrabarti P
et al (1992) Crystallographic structure of the
nitrogenase iron protein from Azotobacter
vinelandii. Science 257:1653–1659
16. Lindahl PA, Day EP, Kent TA et al (1985)
Moessbauer, EPR, and magnetization studies
of the Azotobacter vinelandii iron protein.
Evidence for a [4Fe4S]1+ cluster with spin
S ¼ 3/2. J Biol Chem 260:11160–11173
17. Lanzilotta WN, Ryle MJ, Seefeldt LC (1995)
Nucleotide hydrolysis and protein conforma-
tional changes in Azotobacter vinelandii
nitrogenase iron protein: defining the func-
tion of aspartate 129. Biochemistry
34:10713–10723
18. Watt GD, Reddy KRN (1994) Formation of
an all ferrous Fe4S4 cluster in the iron protein
component of Azotobacter vinelandii nitroge-
nase. J Inorg Biochem 53:281–294
19. Angove HC, Yoo SJ, Burgess BK et al (1997)
Mössbauer and EPR Evidence for an
all-ferrous Fe4S4 cluster with S ¼ 4 in the Fe
protein of nitrogenase. J Am Chem Soc
119:8730–8731
20. Angove HC, Yoo SJ, Munck E et al (1998) An
all-ferrous state of the Fe protein of nitroge-
nase: interaction with nucleotides and elec-
tron transfer to the MoFe protein. J Biol
Chem 273:26330–26337
21. Vincent KA, Tilley GJ, Quammie NC et al
(2003) Instantaneous, stoichiometric genera-
tion of powerfully reducing states of protein
active sites using Eu(II) and polyaminocar-
boxylate ligands. Chem Commun
20:2590–2591
22. Lowery TJ, Wilson PE, Zhang B et al (2006)
Flavodoxin hydroquinone reduces Azotobac-
ter vinelandii Fe protein to the all-ferrous
redox state with a S ¼ 0 spin state. Proc Natl
Acad Sci U S A 103:17131–17136
23. Johnson JL, Tolley AM, Erickson JA et al
(1996) Steady-state kinetic studies of dithio-
nite utilization, component protein interac-
tion, and the formation of an oxidized iron
protein intermediate during Azotobacter vine-
landii nitrogenase catalysis. Biochemistry
35:11336–11342
24. Morgan TV, Mortenson LE, McDonald JW
et al (1988) Comparison of redox and EPR
properties of the molybdenum iron proteins
of Clostridium pasteurianum and Azotobacter
vinelandii nitrogenases. J Inorg Biochem
33:111–120
25. Corbett MC, Hu Y, Fay AW et al (2007)
Conformational differences between Azoto-
bacter vinelandii nitrogenase MoFe proteins
as studied by small-angle X-ray scattering.
Biochemistry 46:8066–8074
26. Lanzilotta WN, Parker VD, Seefeldt LC
(1999) Thermodynamics of nucleotide inter-
actions with the Azotobacter vinelandii nitro-
genase iron protein. Biochim Biophys Acta
1429:411–421
27. Sarma R, Mulder DW, Brecht E et al (2007)
Probing the MgATP-bound conformation of
the nitrogenase Fe protein by solution small-
angle X-ray scattering. Biochemistry
46:14058–14066
28. Schlessman JL, Woo D, Joshua-Tor L et al
(1998) Conformational variability in struc-
tures of the nitrogenase iron proteins from
Azotobacter vinelandii and Clostridium pas-
teurianum. J Mol Biol 280:669–685
29. Kim J, Rees DC (1992) Crystallographic
structure and functional implications of
the nitrogenase molybdenum-iron protein
from Azotobacter vinelandii. Nature
360:553–560

30. Schindelin H, Kisker C, Schlessman JL et al
(1997) Structure of ADP·AlF4
-stabilized
nitrogenase complex and its implications for
signal transduction. Nature 387:370–376
31. Kaiser JT, Hu Y, Wiig JA et al (2011) Struc-
ture of precursor-bound NifEN: a nitrogenase
FeMo cofactor maturase/insertase. Science
331:91–94
32. Maier RJ, Moshiri F (2000) Role of the Azo-
tobacter vinelandii nitrogenase-protective
Shethna protein in preventing oxygen-
mediated cell death. J Bacteriol
182:3854–3857
33. Chan MK, Kim J, Rees DC (1993) The nitro-
genase iron-molybdenum-cofactor and
P-cluster pair: 2.2 Å resolution structures.
Science 260:792–794
34. Pierik AJ, Wassink H, Haaker H et al (1993)
Redox properties and EPR spectroscopy of
the P clusters of Azotobacter vinelandii
molybdenum-iron protein. Eur J Biochem
212:51–61
35. Huynh BH, Henzl MT, Christner JA et al
(1980) Nitrogenase XII. Mössbauer studies
of the MoFe protein from Clostridium pas-
teurianum W5. Biochim Biophys Acta
623:124–138
36. Peters JW, Stowell MHB, Soltis SM et al
(1997) Redox-dependent structural changes
in the nitrogenase P-cluster. Biochemistry
36:1181–1187
37. Surerus KK, Hendrich MP, Christie PD et al
(1992) Moessbauer and integer-spin EPR of
the oxidized P-clusters of nitrogenase: POX is
a non-Kramers system with a nearly degener-
ate ground doublet. J Am Chem Soc
114:8579–8590
38. Rupnik K, Hu Y, Lee CC et al (2012) P+ State
of nitrogenase P-cluster exhibits electronic
structure of a [Fe4S4]+ cluster. J Am Chem
Soc 134:13749–13754
39. Einsle O, Tezcan FA, Andrade SLA et al
(2002) Nitrogenase MoFe-protein at 1.16 Å
resolution: a central ligand in the FeMo-
cofactor. Science 297:1696–1700
40. Spatzal T, Aksoyoglu M, Zhang L et al (2011)
Evidence for interstitial carbon in nitrogenase
FeMo cofactor. Science 334:940–940
41. Wiig JA, Hu Y, Lee CC et al (2012) Radical
SAM-dependent carbon insertion into the
nitrogenase M-cluster. Science 337:1672–1675
42. Wiig JA, Hu Y, Ribbe MW (2015) Refining
the pathway of carbide insertion into the
nitrogenase M-cluster. Nat Commun 6:8034
43. Shah VK, Brill WJ (1977) Isolation of an iron-
molybdenum cofactor from nitrogenase. Proc
Natl Acad Sci U S A 74:3249–3253
Nitrogenases 21
44. McLean PA, Wink DA, Chapman SK et al
(1989) A new method for extraction of iron-
molybdenum cofactor (FeMoco) from nitro-
genase adsorbed to DEAE-cellulose. 1. Effects
of anions, cations, and preextraction treat-
ments. Biochemistry 28:9402–9406
45. Christiansen J, Goodwin PJ, Lanzilotta WN
et al (1998) Catalytic and biophysical proper-
ties of a nitrogenase apo-MoFe protein pro-
duced by a nifB-deletion mutant of
Azotobacter vinelandii. Biochemistry
37:12611–12623
46. Lee CC, Hu Y, Ribbe MW (2012)
ATP-independent formation of hydrocarbons
catalyzed by isolated nitrogenase cofactors.
Angew Chem Int Ed 51:1947–1949
47. Lee CC, Hu Y, Ribbe MW (2015) Catalytic
reduction of CN
, CO, and CO2 by nitroge-
nase cofactors in lanthanide-driven reactions.
Angew Chem Int Ed 54:1219–1222
48. Tanifuji K, Sickerman N, Lee CC et al (2016)
Structure and reactivity of an asymmetric syn-
thetic mimic of nitrogenase cofactor. Angew
Chem Int Ed 55:15633–15636
49. Igarashi RY, Dos Santos PC, Niehaus WG
et al (2004) Localization of a catalytic inter-
mediate bound to the FeMo-cofactor of
nitrogenase. J Biol Chem 279:34770–34775
50. Brigle KE, Setterquist RA, Dean DR et al
(1987) Site-directed mutagenesis of the nitro-
genase MoFe protein of Azotobacter vinelan-
dii. Proc Natl Acad Sci U S A 84:7066–7069
51. Dilworth MJ, Fisher K, Kim C-H et al (1998)
Effects on substrate reduction of substitution
of histidine-195 by glutamine in the
α-subunit of the MoFe protein of Azotobacter
vinelandii nitrogenase. Biochemistry
37:17495–17505
52. Fisher K, Dilworth MJ, Kim C-H et al (2000)
Azotobacter vinelandii nitrogenases contain-
ing altered MoFe proteins with substitutions
in the FeMo-cofactor environment: effects on
the catalyzed reduction of acetylene and eth-
ylene. Biochemistry 39:2970–2979
53. Orme-Johnson WH, Hamilton WD, Jones
TL et al (1972) Electron paramagnetic reso-
nance of nitrogenase and nitrogenase compo-
nents from Clostridium pasteurianum W5
and Azotobacter vinelandii OP. Proc Natl
Acad Sci U S A 69:3142–3145
54. O’Donnell MJ, Smith BE (1978) Electron-
paramagnetic-resonance studies on the redox
properties of the molybdenum-iron protein of
nitrogenase between þ50 and
450
mV. Biochem J 173:831–838
55. Watt GD, Burns A, Lough S et al (1980)
Redox and spectroscopic properties of
Another random document with
no related content on Scribd:
 It certainly was a poser. If, he said, he described a man as having
a very cowardly dog, what should I think he meant? I suggested
various possible solutions: that he was a brutal man who thrashed
his dog unmercifully; that he was a very poor man who could not
afford to buy a good one; or a very mean man who nearly starved
the beast to death. As none of them was correct, I asked him to
explain. But he preferred to keep me on tenterhooks and declined to
do so, chuckling with delight at the way in which he was making me
“work my brain.”
Sheykh Senussi.—Clerk to the Qadi in Haggi Qwaytin.—My guide during my
Mut and the village poet. (p. 44). last year in the desert. (p. 199)
 Sheykh Ibn ed Dris.—One of the Left, Haggi Qwaytin. Right, Haggi
Senussi Sheykhs in the Zawia in Qway.—Qway was my guide during my
Farafra. (p. 228). first two years in the desert. (p. 26).

It was not until the train began to move, that he condescended to

solve the problem. When he said that a man had a cowardly dog, he
meant that he was so very hospitable that his dog got tired of barking
at the innumerable guests who came to his house! As a literary
language Arabic must be very hard to beat.
Near Nazali Genub, where I camped while buying my camels, I
found an acquaintance whose job consisted in surveying and
drawing the hieroglyphics on the tombs, where, to relapse into
metaphor, in mere English:

“The long phantasmal line

Of Pharaohs crowned divine
Are dust among the dust that once obeyed them.”
 I shared a very comfortable tomb with him during the time I spent
there.
After about a week, during which Qwaytin with Abd er Rahman
visited the surrounding villages and markets in search of the camels
that he required to complete the caravan, we moved over to
Qwaytin’s house, some seven miles away, and on the following
morning, 24th March, started off for our journey into the desert;
Ibrahim at the start, as usual, banging off his gun after the Arab
custom to scare away evil spirits during our journey.
I decided first to make for an uninhabited oasis called Bu Gerara,
which had not previously been reported, that Qwaytin said he knew
of, and that lay some little distance off the Derb et Tawil, and not very
far to the north-east of Dakhla. This oasis he said contained palms,
wells and some old buildings, but had been deserted for many years.
The Derb ed Deri—the monastery road—that we followed starts
from near an old der, or monastery, called the Der Muharug, from
which it takes its name, and is a branch of the great Derb et Tawil
—“the long road”—that, starting from the Nile Valley near Manfalut,
runs right across the desert plateau to Dakhla Oasis.
On getting up on to the plateau, Qwaytin pointed out in the
distance to the west a low hill, which he said was called the Jebel
Jebaïl, where, he stated, there are many tombs.
The plateau was level and as featureless as that between the Nile
Valley and Kharga Oasis, to which it bore a strong general
resemblance. There were the same patches of sand and pebbles,
interspaced with areas of limestone, showing all the same types of
sand erosion—kharashef, kharafish, battikh and rusuf.
In many places the limestone appeared as marble, sometimes
polished by the action of the sand blast. White, black, grey, yellow
and beautiful rose pink, in various combinations, were seen in the
stone. Much of it showed large cracks on the surface, but there were
considerable areas of stone, especially of the grey marble with
darker grey lines, that seemed to be quite solid. The rock in places
was translucent and appeared to be alabaster, but of very inferior
quality. Some of the pink marble looked to be of a fine colour and
texture; but it is doubtful, in such an inaccessible position, if it would
ever repay working.
Early on our second day in the desert we joined the main road—
the Derb et Tawil, or “long road.” Close to the east of the point,
where the two roads met, was one of the low rocky hills with which
the plateau was studded. From the foot of this ran a little used road
leading to ’Ain Amur, via ’Ain Embares, an undiscovered well that I
had tried to reach by way of the small depressions leading out of the
’Ain Amur wady. It was reported to be almost sanded up and to give
very little water.
On the following day we passed the point where a road branches
off to the west from the Derb et Tawil to go direct to Qasr Dakhl. This
road, which does not appear to be much used, is known as the “Derb
el Khashabi,” or “wooded road,” owing to the fact that about two
days’ journey from the point where it leaves the Derb et Tawil, there
is a patch of dead trees about ten feet high. It is said to be an easy
one to traverse.
The next day we reached the Abu Moharik dune belt that took us
three hours and twenty minutes to cross, which at the rate of two and
a half miles an hour, was equivalent to a distance of eight and a half
miles. The dunes of which it was composed were all, so far as we
saw, of the crescentic type, and were probably all considerably under
fifteen feet in height. Where the road crossed the belt, was a sand-
free gap, the dunes in that part being rather thinly distributed, though
farther to the north they appeared to lie much closer together. The
whole of the road, where it ran through the belt, was entirely free
from drift sand. We camped that night in the middle of the dunes.
On leaving the dunes for Bu Gerara, I sent Qwaytin and Abd er
Rahman off to look for another oasis that the former had heard of,
that was said to lie some distance to the west of our road, which,
however, he failed to find. In the evening before his departure, he
came into my tent and announced that “his book said” that on the
following day we should reach the Gara bu Gerara. There, he said,
the road forked, and one branch, leaving the usual road followed by
caravans going to Dakhla, and keeping more to the west, led to Bu
Gerara—the oasis we were in search of.
This was the first mention he had made of any “book,” so I
enquired what the book was to which he referred. Qwaytin seemed
rather surprised that I had not heard of it before, and said that it was
his “book of treasure!”
Cautious questioning elicited the fact that he had never been to
Bu Gerara before, but that he was relying entirely on the directions
given in this precious volume to take me there, and evidently
expected, when we reached the place that we should all fall to
digging in search of the buried riches that the book said were to be
found there, instead of getting on and mapping the desert.
He was clearly under the impression that he was conferring a
great favour on me by taking me into the secret of the vast wealth
that he expected to find. To have thrown any doubt on the reliability
of his wonderful book would have mortally offended him, as natives
are very sensitive on these subjects. But as following out his
instructions did not seem likely to take me far from the part to which I
wanted to go, and would lead me into new ground, I thought it best
to humour him, trusting that, when he failed to find the place, he
would be willing to come down to the more mundane occupation of
mapping the desert. So, much against my inclination, I found myself
at last fairly launched on a treasure-hunting expedition!

PINNACLE ROCK ON DESCENT TO BU GERARA VALLEY.

 Soon after our start on the following morning one of the camels
fell ill. What the particular disease was I was unable to discover; but
the remedy that Mohanny applied was to bleed her from the tip of
her tail—an operation that appeared to afford some relief. The
bedawin veterinary methods are simple, but, on the whole, effective.
They may be summed up in three words—“bleed, butter, or burn.”
Eleven miles’ march from our camp brought us to the Gara bu
Gerara—a long, low, flat-topped hill with a small peak at its eastern
extremity, where the road to Bu Gerara branched off from the Derb et
Tawil, according to Qwaytin’s book of treasure.
Qwaytin had given minute instructions to Mohanny as to finding
the place, so, on reaching the hill, he came up to me and announced
that it was now necessary to leave the Derb et Tawil, which turned
off towards the east, and to follow a road that led straight on.
Our departure for Gara bu Gerara though had to be postponed for
a short time owing to the camel developing another attack of
whatever the complaint was that she was suffering from,
necessitating that she should be again bled—this time from the
nose. The operation having been successfully performed, we started
off to look for our treasure.
Much to my surprise, we found a very well marked road branched
off from the Derb et Tawil, though, judging from its appearance, it
had not been used for a very long time. Away to the east of our route
—by the side of the Derb et Tawil—was a small, but very
conspicuous mound of bright yellow earth—probably ochreous—
which I was told was the Garet ed Dahab (golden hillock).
Shortly after we passed through a tract of desert thickly studded
with stones. Through this stony area ran a made road. The stones
had all been cleared off its surface, which had then been smoothed
over with a thoroughness that made it extremely unlikely that the
work had ever been done by the bedawin, whose contempt for all
forms of manual labour always induces them to put up with a bad
part in a road, if they cannot circumvent it by a slight detour.
 After traversing this stony part of the desert, we reached the top of
a steep descent, where again it was evident that some more civilised
race than the desert bedawin had been at work, for the road down
on to the lower level had been notched out of the side of the scarp in
a way that would not have done discredit to a modern engineer. After
that I felt prepared for any developments.
After negotiating the Negeb Shushina, as the descent from the
plateau to the level of the depression is called, we came on to a level
sandy plain, where for a time Mohanny, who had been acting as
guide by following the directions Qwaytin had given him from his
wonderful book, succeeded in getting completely lost. After
wandering about for a time, seeking the marble palaces and gilded
domes of Bu Gerara, we at length caught sight of two figures in the
distance, who, when examined through the glass, proved to be
Qwaytin and Abd er Rahman.
 CHAPTER XXII

W E found Abd er Rahman and Qwaytin diligently engaged in

grubbing about in the ground. In reply to my question as to
whether they had seen anything of Bu Gerara, I was told that we
were standing on it. Qwaytin pointed out the foundations of several
walls that could just be seen showing above the sandy surface of the
ground and a lot of broken pottery lying about on the desert. He then
led me a few yards away to where a circular patch of unusually
sandy soil, a few feet in diameter, was to be seen, which he said was
the mouth of a well, and produced as the first instalment of the
“treasure” to be found, a piece of broken purple glass, that had
apparently once formed part of a cup or bowl, and a copper coin of
the Ptolemaic period, which he had dug up.
The sight of that coin was too much for my men. It was all I could
do to get them to unload the camels and pitch my tent, before they
were all off digging away into the ground for dear life, expecting
every moment to find the untold wealth that the book had described.
They continued until it was too dark for them to see. They then set to
work to cook their evening meal.
Qwaytin’s men were even more primitive in their culinary
arrangements than Abd er Rahman and Ibrahim. Their food supply
consisted of the usual leathern sack of flour, an earthenware jar,
covered with raw hide, which contained clarified butter, that they
slung on a camel, and several tin canisters containing a very
anæmic-looking cheese. They mixed the flour, water, salt and butter
together into a dough, which they rolled out into thick slabs with a
stick about three-quarters of an inch in diameter on the top of one of
my provision boxes. They then lighted a huge bonfire from the
surrounding scrub, and when the sand was sufficiently heated and
the wood was reduced to glowing embers, scraped the fire away, laid
the slab of dough down on the heated sand and covered it over
again with the cinders. After about a quarter of an hour’s baking the
bread was considered to be ready to eat. My men cooked their
dough on a slightly dished iron plate called a saj.
Qwaytin came into my tent in the evening, highly elated at having
found Bu Gerara. To my disgust I discovered that he was not thinking
of going on into Farafra at all, but was fairly off on a treasure hunt,
and seemed to imagine that he was going to drag me all over the
desert with him searching for buried riches. His book, he explained,
not only described the road as far as Bu Gerara, but said that close
by there was a hill to the west, standing in the same wady—Qwaytin
pointed out a hill standing by itself to the west as a conclusive proof
that his book was correct—and that a road ran past the foot of the
hill that, if followed, led to a big hill, on the top of which was a well in
which the treasures of three Sultans were buried. He said he had
seen the road at the foot of the hill in the morning. He had never
followed it to see where it led to; but he had seen a hill some years
before in the direction of which the road was going, and had noticed
a lot of pigeons alight on the top of it, and thought that perhaps it
was the one the book referred to, and that they had gone there to
drink from the well.
His mind was fairly obsessed with the idea of treasure, and I could
get him to talk of nothing else.
I tried to get some information from him about Farafra and Iddaila
—but it was of no use; he got round again to treasure at once. The
last time he was in Farafra, he said, he was on his way back from
Tibesti, where he said he had found a seam of diamond that stood
up two feet above the ground like a wall, and ran for a long distance.
He had chopped some lumps of diamond off and they cut glass. But
the Senussi sheykhs had apropriated them for the benefit of the
Senussia. He was clearly very sore on the subject.
I tried to switch him off on to the Bedayat country—but it was
almost equally useless. He said he knew of a number of places there
where there were ruins that probably contained treasure and asked
to see my map.
 He studied it for some time, asking me to read out the names and
then declared that it was all wrong, and asked for a pencil and piece
of paper, saying that he would draw me a better one. He laid the
paper on the ground, sucked the end of the pencil and began to
draw the roads joining the places he was going to tell about, now
and then consulting a cheap bar compass to make sure that he had
got them running in the right direction. This framework, when
completed, looked more like a broken spider’s web than anything
else.
He next proceeded to place on his “map” a number of enormous
dots, to represent the positions of the places he wished to insert in it,
and began to reel out a lot of names that are not found on any
printed map. The conversation commenced to become interesting,
and a good deal more in my line than fables relating to buried
treasure.
But the names given to places by Arabs, Tibbus and Sudanese
are not easily remembered, or even grasped, the first time they are
heard. I allowed him to run on until he had completed his list, and
then took the pencil from him and began to write opposite his dots
the names of the places they stood for, and their distances and
bearings from each other. Qwaytin at once became impatient, as I
found he always did whenever I questioned him as to his statements,
or he saw me writing them down, and I was not able to get more
than two or three of the names before he took himself off in a huff.
He was a most exasperating man to get information from. If I
questioned him about some part of the world that he had visited, he
would very often give me no information at all. Sometimes a
suggestion on my part that perhaps he did not know anything about
it, would put him on his mettle and cause him to divulge some of his
knowledge; but as often as not it only had the effect of making him
reel out a lot of lying statements that probably appealed to his
yokelish mind as being facetious, but which subsequent questioning
showed to be incorrect—every statement that he made I had to
check in order to guard against this peculiar sense of humour that he
possessed.
 I usually verified his information by getting him to repeat what he
had said after an interval of a week or more, by which time I
calculated that, if he had lied, he would have forgotten what he had
said.
The first time I caught him misleading me, I pulled him up. But this
proved to be quite the wrong line to take, and nearly had the effect of
making him withhold his information altogether. He was greatly
incensed, and for many days he would tell me nothing at all. Then
late one night, while I was sitting up, waiting for a star to come on to
the meridian in order to get a latitude, just when it had almost got
there and I was going to take the observation, Qwaytin came quietly
up and said that he would tell me about the country of the Bedayat. I
had to give up the observation and go and sit down in the tent,
where under a coat thrown over my knees, under pretence of feeling
cold, so that he should not see me doing so, I took down all he said
in writing.
The information that he volunteered in this way I found to be
nearly always reliable—but he generally chose some time when I
was going to bed, or in the middle of some other work, to come and
impart his knowledge, and broke off at once if he saw me writing it
down. During the day’s march, too, I would sometimes get him in a
communicative mood, and he would describe to me the route that
joined two places. The difficulty of writing all he said down before I
forgot it was easily solved in this case, by stopping to take a
compass bearing, as he had often seen me do before, and then by
writing in my route book the information he had volunteered under
pretence of keeping up the survey.
Collecting data from natives on which to base a map is not quite
so simple as it sounds. The habit that so many bedawin have of
deliberately misleading one, makes it necessary to check the routes
described most carefully. But even bona-fide information collected in
this way will make a most inaccurate map unless some means for
adjusting it can be found.
The plan I adopted was to get the data whenever possible, in the
form of through routes, joining two places whose positions had been
previously fixed. Roads so described can then be plotted on the
map, their accuracy tested and the route as a whole adjusted, so as
to fit the positions of the places that are already known. In this way
the errors can be minimised as far as possible.
But collecting this information from Qwaytin was anything but
easy. Like nearly all the bedawin he was entirely illiterate, and so
could not give me the spelling of the names of the places that he
gave me in the Libyan Desert, the Sudan, Tibesti and Endi, and
these could not be found on any map. Many of them were almost
unpronounceable, and in some cases introduced sounds that could
not be reproduced by even the Arabic alphabet. They were
presumably those of the Tibbu or Bedayat languages—the latter
being a tongue that Qway had described to me as sounding like “the
chattering of monkeys.”
It will easily be imagined that to take down a long string of new
names such as these, when rapidly reeled off, was a matter of some
difficulty.
But it seemed worth taking some trouble to get them. I had been
asked to get any information I could about the unknown parts of the
desert, for the Senussi question was in the air—the Government
were by no means so fast asleep as people were led to suppose—
and at that time, moreover, a rod was being laid in pickle for ’Ali
Dinar, the Sultan of Darfur, whose goings on did not always meet
with the approval of the authorities; so information on these unknown
parts was likely to be of some practical use, beyond spoiling the
virginal whiteness of this part of the map of Africa.
Qwaytin’s knowledge of the least known parts of North and
Central Africa was profound, and he had the great virtue, from my
point of view, of being so densely stupid that he was unable to
realise the value of all that he let out. Before I had done with him, he
gave me enough data to form a fairly complete map of the unknown
portions of the Libyan Desert, with a great deal of the Bedayat
country and Endi. In addition I learnt from him much information of
the orography of the desert and the distribution of the sand dunes.
The map when completed contained the names of some seventy
places that, I believe, had not previously been recorded; many of
them have been found since, approximately in the position in which
they were shown. Maps of this kind do not, of course, err on the side
of accuracy, but they have their uses—principally in giving future
travellers that definite objective to look for, that I had so greatly
needed when starting work in this desert. If I had been able to collect
during my first year the information I eventually obtained, mainly
from Qwaytin, in my last, I should certainly have tackled the job in an
entirely different way.
The morning after our arrival at Bu Gerara, the men fell seriously
to work to dig about in the site, with the result that by midday a few
pieces of broken pottery and glass and two or three more small
copper coins had been unearthed. As I wished to ascertain whether
any water was to be found in the well, in the afternoon, much to their
disgust, I set the men to clear it out.
Qwaytin’s men, after they had been working for a short time,
downed tools, declaring that that sort of work was a job that was only
fit for the fellahin, and beneath the dignity of the Arabs. It was not
until I pointed out that the well was the most likely place in which to
look for treasure, and reminded them that the three Sultans were
said to have buried theirs in the well on the top of the hill described
in Qwaytin’s book, that they could be induced to resume their work.
Then they fell to with a will and soon had the well cleared out to the
bottom.
It was about nine feet in diameter, and eight feet deep. On the
side towards the site of Bu Gerara, a ramp, cut in its side, led down
to the bottom. The part of the desert in which it had been sunk was
covered with a thin layer of rock, below which lay a bed of clay,
extending down to another layer of rock, which formed the bottom of
the well. Before we commenced to dig, the well was completely filled
with sand that had drifted into it. About half-way down our
expectations were raised by the sand becoming damp; but though
the well was cleared out to the very bottom, and the sand got
considerably damper as we descended, no water was to be seen.
This was a considerable disappointment, as a well at this point on
the long Derb et Tawil would have been of the greatest value to the
travellers using the road.
The nut of a dom palm that I dug up, and the trunks of some palm
trees that had been built into the walls, showed that, at one time,
there must have been a plentiful supply of water in the
neighbourhood. The place was probably a small fortified station built,
or at all events occupied, during Ptolemaic times, to protect the well,
which from its position on “the long road” must once have been of
considerable importance.
The existence of fossilised tree-trunks and old river-beds in the
Libyan Desert shows conclusively that, in the remote past, this
portion of the world must have been a well-watered country. But
whether this desiccation reached its limit before historical times, or
whether it is still going on is one of the most disputed points in
connection with this district. The failure of the well at Bu Gerara may,
of course, have been due to some purely local cause, which was not
apparent. But in the absence of some explanation as to its nature,
this little abandoned settlement affords a very strong argument in
favour of the view that the water supply—apart from that derived
from the artesian wells—has failed appreciably in comparatively
recent times, owing probably to a decrease in the rainfall. From this
point of view, our discovery was of some importance, though the
place itself was of no consequence at all.
But it was found by following instructions in Qwaytin’s book of
treasure. Works of this description, to put it mildly, are regarded in
Egypt with a considerable amount of incredulity. This scepticism, I
own, I fully shared—until my discovery of Bu Gerara.
Since then, however, I have taken a different view of the case,
and believe that the almost universal suspicion with which these
books are regarded, may not, after all, be entirely justified, and that
part of it at least is due to the strong prejudice that so often exists
towards any native beliefs or customs that do not admit of a ready
explanation, or that savour in any way of the occult, or of buried
riches.
 These books of treasure, it is true, are mostly written by natives of
the astrologer class, who clearly expect their readers to rely largely
upon charms and various occult means to discover the hidden riches
to which they profess to give the key. Many of them lived in the
Middle Ages, but the race has not died out. There are hundreds of
the same class of men still to be found practising their arts on the
credulous natives of Egypt, and one of the principal subjects upon
which they are even now consulted is the recovery of buried
treasure. There is a sheykh el afrit (ruler of spirits) in almost every
village, to whom the inhabitants resort to induce him by means of the
pool of ink, or some similar method, to foretell the future or to guide
them to where treasure has been buried. Some of them perhaps
believe in their own powers, but the majority are probably little more
than impostors. So far, so bad.
But there appears to be a very fair amount of grain among all the
chaff contained in these books, for in many cases they not only refer
to places, such as Esna, which are perfectly well known, but they
describe the roads that lead to them, when these roads are still in
use.
For this reason I think they are worth careful—or perhaps it would
be better to say cautious—consideration. It is true that in many
cases they mention places and describe roads which perhaps were
perfectly well known at the time when the books were written, but
that cannot now be identified; but this proves nothing. There are
many old sanded-up wells, little deserted oases and small outpost
stations of the Roman and other periods, such, for instance, as Bu
Gerara, scattered about in the desert, and the vestiges of many old
roads are still to be seen, whose ultimate destination is now
unknown, but which, I believe, lead to these abandoned oases,
which very probably, in the fifteenth century, when the book of
Johnson Pasha’s that has already been referred to was written, were
populous villages and oases; but which, owing to failure of the water
supply, the encroachment of the sand, or to some other cause, have
long since become deserted.
On these grounds I believe these books contain—among a great
deal that is useless except as a curiosity—some valuable information
as to old places in the desert that have long since been lost to sight,
and whose very names may now be forgotten, information that is of
a geographical character.
Why not? Is it the age of the book, or the fact that the descriptions
in it are associated with magic and hidden treasure, that presents the
difficulty? If it be the former, ask any archæologist whether he would
hesitate to look for the site of some ancient city, because the only
references to it were to be found in some old papyrus or temple
hieroglyphics; besides, did not the Royal Geographical Society have
a paper read before them on the identity of the Garden of Eden with
Mesopotamia? The description that led to the identification being
taken from the Book of Genesis, which was written long before any
of these books were thought about.
If it be the treasure that presents the difficulty, has there not been
endless discussion among geographers as to the identity of the
Wakwak Islands and other places, mentioned in the “Arabian
Nights,” a large proportion of the stories in which have no
pretensions to be anything more than fairy-tales—and certainly there
is enough buried treasure mentioned in them to satisfy the most
ardent fortune hunter in Egypt.
Are not educated Europeans, even now, continually setting out to
look for the fabulous riches hidden, a hundred or two years ago, by
some old pirate or buccaneer, usually on an island—say, in the West
Indies? Of the identity of the island there is generally no doubt at all
—but the treasure does not seem to be often found!
We stayed for a day or two more at Bu Gerara, during which time
the men found a small earthenware pot, some broken fragments of
glass and pottery and one or two more copper coins—and that was
all! Then as we had drawn a blank, so far as treasure was concerned
at Bu Gerara, the men all wanted to be taken off to the hill where the
riches of the three Sultans was buried with the least possible delay.
Qwaytin was the most insistent of them all, evidently assuming that I
had given up my plan of going to Farafra and had committed myself
to a whole season’s treasure hunting instead.
 The hill where the mystical Sultans had buried their riches was not
far off, though it did not lie in the direction in which I had intended to
go; but it was in a part of the desert that had never been mapped, so
I thought it best to humour him once more and let him take me there.
We got off early the next morning. Qwaytin led us straight towards
the hill in the wady, near the foot of which we found the promised
road.
As we increased our distance from the cliff lying to the north of Bu
Gerara, the surroundings of the place could be better seen. The view
to the north was, of course, cut off by the cliff, which as soon as we
had got some distance from it, could be seen stretching away for
many miles to the east, forming the continuation of the escarpment
that bounded the Kharga depression on its northern side.
To the south-east was a considerable expanse of elevated
ground, evidently the plateau in which lay some small depressions I
had found to the north of ’Ain Amur. So far as I could see, there was
no cliff on the northern side of this tableland, the ground only sloping
up to it from the lower level. A well—’Ain Embares—that I had tried
to reach by way of the chain of small depressions, with little doubt
was situated between the foot of the scarp of the main plateau and
this high ground that lay to its south.
On the west, the scarp of the plateau was visible for a long way.
Qwaytin’s old road led us in a southerly direction, roughly parallel to
the cliff of a detached plateau. It was chiefly noticeable for the large
number of small patches of bushes, known as roadhs, that were
scattered along it. These seemed to be a favourite feeding ground
for gazelle, to judge from the number of tracks we saw, most of
which, however, were fairly old ones. In one place, instead of the
usual small bushes, a couple of small acacia (sunt) trees were seen.
We sighted the hill we were in search of in the afternoon, and, an
hour before camping, reached the top of a steep descent on to lower
ground, about two hundred and fifty feet below us, that Qwaytin said
was called in his book the “Negeb er Rumi” (descent of the
European). The road down to the valley below was obviously to
some extent an artificial one, and, though extremely steep, was
negotiated without difficulty. We pitched our camp below.
This lower ground was so covered with sand and pebbles that I
was unable to see whether we were still on the limestone. But the
ground level rose again considerably as we neared the hill, and for
the last part of the way the limestone was showing again on the
surface. Possibly a fault exists in the neighbourhood.
We reached the hill itself at noon, and camped on its southern
side. It was a small limestone-capped hill, chiefly remarkable for the
extent to which the limestone was honeycombed by the wind-driven
sand. At the foot of the hill, near the camp, was a boulder that had
evidently rolled down from the top. It was almost four feet in
diameter, and literally riddled with holes like a sponge.
As soon as the camp was pitched, the men rushed up the hill and
began minutely searching every nook and cranny for the reported
well, while Qwaytin wandered disconsolately along its base, vainly
searching for the broken glass that his book had foretold would be
found there. He thought that perhaps he had mistaken the hill, and
said, if we could not find the well or any glass, that we had better
follow the road farther, to see if there was not another hill upon it that
might be the one referred to in his book; but to waste any more time
in looking for that hill was the last thing I intended to do.
By sunset the well had not been found, though every inch of that
hill must have been most carefully examined several times.
This was a serious blow to Qwaytin’s hopes, and a distinct wet
blanket to the whole caravan with the exception of Ibrahim, who, as
he explained to me, would not have minded a little bakhshish, but
would not in the least know what to do with sacks of gold, or
diamonds, even if he found them.
In the evening Qwaytin, Abd er Rahman, Dahab and I held a
serious consultation. The position of the hill tallied so well with the
description of it in Qwaytin’s book that he felt sure that it was the
right one; but he was terribly worried over the failure to find the well.
 Dahab said that he thought that probably it was there all right, but
that it was hidden by enchantment, and that it would be necessary
for us to burn some incense before it would become visible.
Qwaytin cheered up rather at this idea; but said that we had no
incense with us, and added it was awkward stuff to play with, as it
was most important that we should have the right kind, and should
be quite sure that we knew how to use it.
Abd er Rahman agreed with this, and was very emphatic in saying
that we ought to be quite sure that we had enough of it, as he had
heard a story of a Maghrabi Arab, who had joined with two fellahin in
a search for treasure that was buried in some tombs in the side of a
hill that had a spell over them, and so could not be opened without
proper formalities. They found the place where the tombs were
hidden, and then had gone through the necessary incantations and
burnt some incense and the tombs immediately opened. The two
fellahin had then gone in to collect the treasure while the Maghrabi
had remained outside to look after their camels and to keep the
incense burning. Unfortunately the incense ran short, and, as soon
as the last of it had been burnt, the tombs closed again with a bang,
burying the two unfortunate fellahin alive. The Maghrabi had then
gone home with their camels, and Abd er Rahman was clearly of
opinion that that Arab had done something that was quite
exceptionally clever.
He suggested that, to be on the safe side, we had better go and
fetch Sheykh Ibrahim, the Sheykh el Afrit from Dakhla, to come out
and do the necessary incantations. But this did not meet with
Qwaytin’s approval at all. Sheykh Ibrahim, he said, was a member of
the Senussia and he knew all about him. He had the right books and
the proper incense and was very clever at his work. But he was such
a bad man that sometimes the spirits would not obey him; and he
pointed out that if an afrit went on strike in the middle of the
performance, we might find ourselves rather badly in the soup.
After much serious discussion, we came to the conclusion that, in
the circumstances, it was no use for us to waste any more time in
examining the hill, but that at the end of the trip we would go and get