全 文 ：Received 1 Jul. 2003 Accepted 24 Sept. 2003
Supported by the National Natural Science Foundation of China (30270296) and the State Key Basic Research and Development Plan of China
(“973” Program 001CB1089-06).
* Author for correspondence. E-mail: .
http://www.chineseplantscience.com
Acta Botanica Sinica
植 物 学 报 2004, 46 (4): 392-400
Advances in Studies on Nitrogenase Crystallography
ZHAO Jian-Feng, ZHOU Hui-Na, ZHAO Ying, BIAN Shao-Min, HUANG Ju-Fu*
(Key Laboratory of Photosynthesis and Environmental Molecular Physiology, Institute of Botany,
The Chinese Academy of Sciences, Beijing 100093, China)
Abstract : The function of biomacromolecule is dependent on its space structure. X-ray diffraction
analysis is generally an important way to obtain structural information of biomacromolecules. Here, the
main advances in the growth and X-ray diffraction analysis of nitrogenase crystals are briefly introduced
and reviewed. At last, the challenge and prospect of nitrogenase crystallography are discussed.
Key words: nitrogenase; crystal growth; crystallographic structural analysis
Biological nitrogen fixation is performed by nitrogenase,
which catalyzes the reduction of atmospheric N2 to NH3
(Milagros et al., 1999). Mo-containing nitrogenase is com-
posed of two s eparable components designated the com-
ponen t Ⅰ(MoFe protein) and the componen t Ⅱ(Fe
protein). Fe protein is a 60 000 Da dimer of identical g sub-
units connected through a single [4Fe-4S] cluster. It serves
as an electron donor for MoFe protein in a reaction some-
how couple to MgATP hydrolysis. MoFe pro tein is a
220 000-240 000 Da tetramer of a2b2. It is often viewed as
being composed of two identical halves that do not com-
municate with each other. Each half has one α subunit and
b subunit, one FeMoco, one P-cluster and one binding site
fo r Fe p rotein (Kim and Rees, 1992; Peters et al . 1997;
Schindelin et al. 1997).
Biosyn thes is of MoFe protein (Av1) and Fe p rotein
(Av2) in Azotobacter vinelandii Lipmann requires at least
15 different nif gene products (Nif). NifD, NifK and NifH
have been implicated in the in vivo synthesis of a, b sub-
units of MoFe protein and g subunit of Fe protein,
respectively. NifZ has been implicated in the in vivo syn-
thesis of P-cluster. If nifZ is deleted , Av1 s ynthesized in
vivo is one of FeMoco-containing, P-cluster-changed form
of Av1 (Zhong et al., 1996; Huang et al., 1999). And NifN,
NifE, NifB, NifH, NifQ and NifV have been s trongly impli-
cated in the in vivo synthes is o f FeMoco (Brig le e t al .,
1987; Paustian et al., 1989; Roll et a l., 1995). If any of the
later six genes is deleted or mutated, Av1 synthesized in
vivo is one of P-cluster-containing, FeMoco-deficient form
of Av1. The conditions for their full activation in vitro are
not the same (Shah et al., 1977; Paustian et al., 1989; Tal et
al., 1991; Zhao et al ., 2003). It indicates that there is a
st ructural d ifference between FeMoco-deficien t Av1.
However, the structural differences between the FeMoco-
deficient or P-cluster-changed Av1 from mutants and Av1
from wild-type strain have not well been known.
Besides Mo-containing nitrogenase, there are other two
genetically dis tinct nitrogenease systems in bacteria: a V-
containing nitrogenase and an “iron only” nitrogenase lack-
ing both Mo and V (Hales et al., 1990; Müller et al., 1992).
Biosynthesis of their nitrogenases requires vnf and anf gene
products , respect ively . They are als o composed of two
separable components designated the component Ⅰ(VFe
protein and FeFe protein, respectively) and the component
Ⅱ(Fe protein). The three component Ⅰ proteins have simi-
lar P-clusters and different FeMco in which M is Mo, V and
Fe, respectively. But their st ructural differences have not
well been known.
It was shown by a measurement using the perturbed
angular correlation’s that the 99Mo had gone to the define
place of Mo in FeMoco after an incubation of a partial
metallocluster-deficient Av1 with a reconst ituent solution
containing 99Mo (Dong et a l., 1996). And the partial
metalloclus ter-deficien t Av1 was act ivated by a
reconstituent solution containing either Mn or Cr, leading
to a suggestion that there could be nitrogenases MnFe
protein and CrFe protein (Huang et a l., 1994; 1995). The
suggestion was supported by the later results that the MnFe
protein and CrFe protein have been partially purified from
a mutan t UW3 of A. Vinelandi i g rown on the Mn-, Cr-
containing medium, respectively (Huang et al., 2001; 2002).
Of course, more structural and functional evidence should
be obtained to demonstrate their existence.
The function of biomacromolecule is dependent on its
.Review.
ZHAO Jian-Feng et al.: Advances in Studies on Nitrogenase Crystallography 393
space structure. A final elucidation of nitrogenase function
is also dependent on its space structure. X-ray diffraction
analysis is generally an important way to obtain structural
information of biomacromolecules. However, the growth of
crystals suitable for the analysis usually is very difficult,
and often became a main hindrance fo r crystallography
(Drenth et al., 1987). Fortunately, many crystals of nitroge-
nase component proteins have been obtained and analyzed
by X-ray d iffraction after a lot of efforts were made in the
past more than 20 years.
1 Crystallization of Nitrogenase
The growth of protein crystal is indeed a complex physi-
cal and chemical process. There are many factors affecting
crystallization of proteins, such as temperature, purity and
concentration of the protein, kind and concentration of pre-
cipitants and stabilizer, concentration and pH value of dif-
ferent buffer, gravitation , method for crys tallization and
techn ical b ias , etc. (Mcphers on , 1976). Like some
metalloenzymes, n itrogenase pro teins are sus ceptible to
O2 and this makes their crystallization more complicated
and difficult.
The first method for crystallization of Av1 was put for-
ward by Burns et al. (1970), then was modified by Shah et
al. (1973) and was simplified by the 7th Laboratory, Insti-
tute of Botany, The Chinese Academy of Sciences (1973).
But the crystals obtained in these laborato ries were small
needle-shape. The crystal is not suitable for X-ray diffrac-
tion analysis. Weininger et al. (1982) attempted to do X-ray
diffraction analysis of MoFe protein crystals g rown by
using microdialysis method. But they did not succeed in
analyzing the crystallographic structure of MoFe protein.
In 1992，both Kim and Rees, and Georgiadis et a l.
succeeded in g rowing the big crystals of Av1 and Av2 by
the microcapillary batch method and in analyzing of these
crystals by X-ray diffraction, respectively. Since then, the
liquid/liquid diffusion method has been us ually used for
crystal growth of nitrogenase proteins (Table 1).
In the past 10 years, many factors affecting crystalliza-
tion of nitrogenase proteins have been well studied in our
laboratory, leading to formatiom of larger crystals of Av1,
DnifZ Av1, NifB- Av1, DnifE Av1, DnifH Av1, MnFe protein,
CrFe p rotein and bacterio ferritin from wild -type A.
vinelandii. Some of them are being analyzed by X-ray dif-
fraction (Table 1).
The p recipitan t solutions used in our laboratory con-
tain PEG, Na2S2O4 (DT) and glycerin. The solutions could
help to protect n itrogenase protein from denaturing. The
protection of the solutions could come from the following
factors: (1) excess DT which is able to reduce O2; (2) glyc-
erin which is used to be a stabilizer for many proteins; and
(3) viscosity o f PEG and g lycerin which obs tructs O2 in
diffusing to protein molecules (Zhao et al., 2003). PEG could
abs orb water from the environment around p ro tein
molecules, resulting in decreasing protein dissolution, and
cou ld obvious ly decreas e the dielectric constan t of the
medium, resulting in decreasing an effectively electrostatic
shield between protein molecules (Mcpherson et al., 1976).
So , PEG was us ually us ed as a p recip itate fo r pro tein
crystallization. In fact, most of the big crystals with good
quality in Table 1 were obtained only when PEG was used
as one of the precipitants for crystallization of nitrogenase
proteins. Like salts (MgCl2, NaCl, etc.), PEG has an opti-
mum concentration for the crystallization. Only a few crys-
tal nuclei are formed and subsequently grown to large crys-
tals of good quality when the protein loses water at a suit-
able rate. The optimum concentration of a precipitant is
changed with the protein kind, the other chemical concen-
trations and the method for crystallization.
Hepes or Tris is one kind of salt, but its basic role is to
stabilize the pH value o f the protein solution, s ince pH
value is very important for the electric charge on protein
and the stability of protein conformation (McPherson et
al., 1976; Huang et al., 2001). Hence, pH value is an impor-
tant factor affecting the size, number, shape and quality of
crystals. The pH value of 8.1-8.4 is the optimum value for
crystallization of DnifE Av1, DnifH Av1, DnifZ Av1, NifB-
Av1, MnFe protein and CrFe protein. Perhaps the optimum
pH value is also dependen t on pro tein kind and other
conditions. For example, the crystals of DnifB Av1 used for
X-ray diffraction analysis were formed at pH value of 9.5
(Schmid et al., 2002).
Excellent protein crystals are usually obtained under
such a condition that the solution convection and wall ef-
fect are minimized. The microgravity on the spacecraft could
decrease the convection and wall effect (Drenth et al., 1991).
Up to date, the effect of the microgravity on the crystalliza-
tion of nitrogenase has not been reported. Fortunately, both
MnFe protein by the vapor diffusion in sitting drop method
and CrFe protein by the liquid/liquid diffusion method were
crys tallized on the spacecraft (“Shenzhou” No. 3) in the
space (Zhao et al., 2003). All formed crystals were single on
the spacecraft, while twin crystals appeared on the ground
under the same conditions. The size of the largest crystal
grown in s pace from CrFe pro tein was 2-fold larger than
that on the ground. But the size of the largest crystal grown
in space from MnFe protein was not larger than that on the
ground. The results show that the crystallization in space
Acta Botanica Sinica 植物学报 Vol.46 No.4 2004394
either by the vapor diffusion method or by the liquid/liquid
diffusion method might benefit the avoidance of twin crys-
tal formation of the proteins. And the crystallization in space
benefited the great diminution of crystal nuclei number and
significant growth of CrFe protein crystals. The differences
in the effect of microgravity on crystal number and s ize
between the two proteins could be resulted from the diffu-
sion method, or the protein kind, or other factors including
the precipitant solutions, etc. It is reported that the liquid/
liquid diffusion method benefits a gradually growth of both
CrFe protein crystals and MnFe protein crystals (Zhang et
al., 2002; Lü et al., 2003). It is indicated that the main factor
affecting crystal growth is the diffusion method, rather than
the kind of nitrogenase.
Salemme (1972) pointed out that the liquid/liquid method
utilizes free diffusion between protein solution and precipi-
tant solution to attain the conditions of protein supersatu-
ration requisite for the nucleation and subsequent growth
of large single crystals. Protein and precipitant solutions
are layered over each o ther, and allowed to diffuse to
equilibrium. The sharp interface is helpful to decrease the
diffusion rate, leading to the slow formation of nuclei and
growth of a few large crystals. Perhaps the microgravity in
space cou ld be helpful to further decrease the diffus ion
rate. However, using the sitting drop method or hanging
drop method, the liquid/liquid diffusion always exists after
Table 1 Crystal growth and diffraction of nitrogenase proteins
Component
Name(1)
Crystalline
Main precipitant (pH) Å (3) Author
(strain) method(2)
Ⅰ Cp1 - PEG6K, MgCl2 (7.5) 2.4 Weininger et al. (1982)
(Wild type) Cp1 MD PEG8K, MgCl2 5.0 Bolin et al. (1990)
Cp1 MB PEG4K, MgCl2/CsCl (8.0) 3.0 Kim et al. (1992)
Cp1 MB PEG4K, MgCl2/CsCl (8.0) 3.0 Kim et al. (1993)
Kp1 MB PEG6K, MgCl2 (8.0) 1.6 Mayer et al. (1999)
Av1 - PEG6K, MgCl2 (7.5) 3.0 Weininger et al. (1982)
Av1 MB PEG4K, NaCl, Na2MoO4 (8.0) 2.7 Kim and Rees (1992)
Av1 MB PEG8K, Na2MoO4 (8.5) 2.0 Peters et al. (1997)
Av1 VS PEG6K, MgCl2, NaCl (8.2) - Huang et al. (1998)
Av1 MB PEG8K, NaCl (8.0) 1.16 Einsle et al. (2002)
Ⅰ Δ nifZ Av1 VH PEG8K, MgCl2, NaCl (8.2) - Huang et al. (2000)
(Mutant) α-Gln195-Av1 MB PEG400, Na2Mo4 (8.0) 2.5 Sorlie et al. (2001)
Δ nifE Av1 VH,VS & MB PEG8K, MgCl2, NaCl (8.2) - Zhao et al. (2003)
Δ nifH Av1 VH,VS & MB PEG8K, MgCl2, NaCl (8.2) - Bian et al. (200?)
Δ nifB Av1 MB PEG8K, CHES (9.5) 2.3 Schmid et al. (2002)
NifB- Av1 VH, VS & MB PEG8K, MgCl2, NaCl (8.2) - Zhao et al. (2003)
NifV- Kp1 MB PEG6K, MgCl2 (7.4) 1.9 Mayer et al (2002)
MnFe pro. VH & MB PEG8K, MgCl2, NaCl (8.2) - Huang et al. (2001)
CrFe pro. VH & MB PEG8K, MgCl2, NaCl (8.2) - Lü et al. (2003)
Ⅱ Cp2 MB PEG4K, CaCl2 (7.5) 2.8 Georgiadis et al. (1992)
(Wild type) Cp2 MB PEG4K, Na2MoO4, (8.0) 1.93 Schlessman et al. (1998)
Av2 Dialysis 2-methyl-2,4-pentanediol (7.8) 3.5 Rees et al. (1983)
Av2 - - 3.0 Georgiadis et al. (1990)
Av2 MB PEG4K, Na2MoO4,NaCl (8.3) 4.0 Georgiadis et al. (1992)
Av2 MB PEG4K, Na2MoO4 (8.0) 2.2 Schlessman et al. (1998)
Av2 MB PEG4K,NaAc , Glycerol (8.5) 2.15 Jang et al. (2000)
Av2 MB PEG4K, NaCl, Glycerol (8.0) 2.25 Strop et al. (2001)
Ⅱ (Mutant) ΔAv2 MB PEG4K, NaAc (8.5) 2.4 Jang et al. (2000)
Complex Av 2-ADP·ALF4--Av 1 MB PEG8K, CaCodylate, MgCl2 (6.5) 3.0 Schindelin et al. (1997)
ΔAv2-Av1 MB PEG4K, Tris, NaoAc 2.2 Chiu et al. (2001)
Av2-Av1 MB PEG6K, MPD, NaCl (8.0) 3.2 Schmid et al. (2002)
Av 2-ADP·ALF4--Av 1 MB PEG8K, MgCl2, NaCl, AlCl3,NaF (6.5) 2.3 Schmid et al. (2002)
(Wild type) Bacterioferritin VS PEG6K, MgCl2, NaCl (8.2) - Huang et al. (1998)
(1), Av1, Kp1 and Cp1 are MoFe protein from Azotobacter vinelandii, Klebsiella pneumoniae and Clostridum pasteurianum , respectively, and
Av2, Kp2 and Cp2 are Fe protein from the strains, respectively; MnFe protein and CrFe protein are component 1 purified from a mutant UW3
strain of A. vinelandii that grew on a Mo-free nitrogen-fixation m edium containing Mn and Cr, respectively. (2), MD, micro dialysis; MB,
microcapillary batch method; VH, the vapor diffusion using hanging drop method; VS, the vapor diffusion using sitt ing drop method. (3), X-
ray diffaction at Å resolution, the data published was in the corresponding paper of references.
ZHAO Jian-Feng et al.: Advances in Studies on Nitrogenase Crystallography 395
add ition of the sample and the sharp interface is hardly
formed, leading to increase of the diffusion rate. It is easy
to form a larger amount of small crystals.
Mos t of biomacromolecule crys tallization proces ses
belong not only indeed to studies in scien tific field, but
also to technological studies with a half of experience. It
needs a given techniques and relative knowledge, and some-
times needs to use unique technique. Therefore, the growth
of crystals suitable for X-ray d iffraction is still a main hin-
drance for protein crystallographic research although the
knowledge and techniques of crystallography have been
greatly improved.
2 Crystalline Structural Analysis of Nitroge-
nase
Georgiadis et al. (1992), Kim and Rees (1992) reported
the first crystal structures of Av2 and Av1. It is viewed as a
fieldstone in the stud ies on nitrogenase structure. From
then on, 10 MoFe proteins, 7 Fe proteins and 5 nitrogenase
complexes have been crystallographically determined and
put into PDB (Protein Data Bank).
2.1 Nitrogenase Fe protein
Fe protein (NifH) has three different roles in the nitroge-
nase enzyme system. Apart from serving as the physiologi-
cal electron donor to Av1, NifH is involved in FeMoco bio-
synthesis and in maturation of the FeMoco-deficient Av1
(Rangaraj et al., 1999). Binding of Fe protein with ATP plays
a key role in electron transfer. The crystallographic struc-
ture of Av2 demonstrated that a single [4Fe-4S]-Cys4 clus-
ter is bridged symmetrically between two identical subunits.
Although Av2 and Cp2 share only 69% amino acid sequence
identity, their subunit folds and dimer arrangements are
very similar to one another (Schlessman et al., 1998). Av2
and Cp2 possess the common core elements of nucleotide-
binding proteins. The elements consist of : (1) a predomi-
nantly parallel b-sheet flanked by a-helices; (2) a phos-
phate-binding loop (P-loop), or Walker A motif; and (3) two
switch regions，switch Ⅰ and switch Ⅱ.
In order to provide the driving force required fo r elec-
tron transfer, the redox potentials of Fe protein have to be
changed. The [4Fe-4S] cluster cont ributes largely in the
process. Mutagenesis studies have revealed that the con-
served Phe. at position 135 in Av2 plays an important role
in defining several biochemical and biophysical properties
of the [4Fe-4S] cluster (Ryle et a l., 1996). The crystallo-
graph ic structure of the Trp-substituted Av2 shows that
the s ubstitut ion d id no t obviously change Fe protein
conformation, indicating that the changes in the properties
of the cluster could be on ly resulted from change in its
local environment (Jang et al., 2000). The cluster is gener-
ally thought to undergo a one-electron redox cycle between
the [4Fe-4S]2+ and the [4Fe-4S]1+. Watt et a l. (1994) p re-
sented the evidence for further reduction to the [4Fe-4S]0.
And Strop et al. (2001) obtained the first crystallographic
view of an all-ferrous [4Fe-4S]0 clus ter. It was suggested
that the solvent accessibility of the all-ferrous Av2 cluster
might play a role in its ability of formation of the oxidation
states. It is generally believed that Fe protein serves as a
one-electron donor and binds two ATPs during inter-pro-
tein electron transfer to yield an overall ATP/e- ratio of 2. If
the all-ferrous protein can function mechan istically as a
two-electron donor, its high efficiency of energy utilization
would be very intriguing.
2.2 Nitrogenase MoFe protein
Biochemical and biophysical studies have indicated that
the FeMoco and P-cluster in MoFe protein play an impor-
tant role in substrate reduction and electron/proton transfer.
More attention has been paid to their structure and function.
Since the two clusters have been neither chemically syn-
thesized nor crystallized from solutions extracted from MoFe
protein because of their novel structures and unique physi-
cal and chemical properties (Howard et al., 1996; Einsle et
al., 2002), their structure and fine composition were contro-
verted before X-ray diffraction analysis o f Av1 (Kim and
Rees, 1992).
The crystallographic structure of Av1 at 2.7Å res olu-
tion showed that the a- and b-subunits in the a2b2 tetramer
had similar polypeptide folds. The subunits consist of three
domains (designated aⅠ，aⅡ，aⅢ and bⅠ , bⅡ ，
bⅢ , res pectively). The FeMoco is completely buried
within the a-subunit, whereas the P-clus ter occurs at the
interface between a- and b-subunits. In this model of Av1,
each center consists of two bridged clusters: the FeMoco
has 4Fe:3S and 1Mo:3Fe:3S clusters bridged by three non-
protein ligands and the P-cluster contains two 4Fe:4S clus-
ters bridged by two cys teine thiol ligands. FeMoco con-
taining Mo/Fe/S/ as well as homocitrate in the proportions
1:7:9:1 is connected to the protein through the side chains
of only two residues bound to Fe and Mo sites located at
the opposite ends of the cluster. The structures of MoFe
protein from different sources are basically iden tical. The
Cp1 crystallophic structure at 3.0 Å resolution also showed
that FeMoco has 4Fe:3S and 1Mo:3Fe:3S clusters and the
P-cluster contained two [4Fe-4S] clusters (Kim J et al., 1993).
NifV- Kp1 from a nifV-mutation mutant with FeMoco
containing citrate instead of homocitrate can reduce C2H2
and H+, but its N2 reduction activity is only ~7% that of the
wild type. Ten years ago , it was not clear whether citrate
Acta Botanica Sinica 植物学报 Vol.46 No.4 2004396
took the place of homocitrate. The crystallographic struc-
ture of NifV- Kp1 showed that citrate occupied the site of
homocitrate in half of the protein, and water molecules oc-
cupied the site of homocitrate in the remainder (Mayer e t
al., 2002). Citrate differs from homocitrate by only a -CH2 ,
but the difference results in a significant functional change
of FeMoco. It indicates that the structure of FeMoco is a
highly unique one, in which the homocitrate indeed plays
an important role in reduction of N2.
In the cell, the FeMoco is synthesized as a separate, but
protein-associated entity. It is subsequently inserted into
a P-cluster-containing, FeMoco-deficient form of the MoFe
protein. The crystallographic structure of DnifB Av1 shows
that like Av1, the α and β subunits of DnifB Av1 also
consist of three domains each. Compared with Av1, except
the αⅢ, the rest remains essentially unchanged. The resi-
due rearrangements occurring in domainαⅢ create a posi-
tively charged funnel that is of sufficient size to accommo-
date insertion of the negatively charged FeMoco. The re-
sult may help to explain a special requirement of Av2 and
MgATP for DnifH Av1 to be act ivated by FeMoco. It is
possib le to suggest that the insertion funnel and accom-
modation site for FeMoco in DnifH Av1 are different from
those of some FeMoco-deficient Av1, leading to obstruc-
tion of FeMoco insertion. Only after the residue rearrange-
ments were changed with Av2 and MgATP, they are suit-
able for insertion and accommodation of FeMoco.
Unlike the earlier results, the crystallographic structures
of Cp1, Av1 and Kp1 show that the P-cluster is a [8Fe-7S]
cluster, other than [8Fe-8S] cluster (Bolin et al., 1993; Pe-
ters et al., 1997; Mayer et al., 1999). Crystallographic analy-
sis of Av1 with defined oxidation states has s hown that
only P-cluster undergoes a redox-dependent structural re-
arrangement (Peters et al., 1997; Drennan et al., 2003). This
structural rearrangement involves the exchange of two Fe
atom ligands from the shared S atom (in the reduced state)
to the polypeptide (in the oxidized state). It is shown that P-
cluster, other than FeMoco, participates in electron transfer.
It ind icates that the analysis of the redox states of the P-
cluster is also important for insights into the mechanism of
nitrogenase. Crystallographic analysis has also identified
three forms of the P-cluster, which were corresponding to
the PN, POX, Psemi-OX states, respectively (Mayer et al., 1999).
The result is consistent with the results obtained from an
extensive stud ies on spectros copes including EPR, CD,
MCD and Mös sbauer of MoFe pro tein . The studies on
spectroscopes have shown that P-cluster could exist in the
following oxidized states: the all-ferrous state (PN), the two-
electron oxidized state (POX) and the one-electron oxidized
state (Psemi -OX). Thus, the role for this cluster in coupling
electron and proton transfer in n it rogenas e could be
demonstrated.
With the improvement of growth of large single crystal
and technique of X-ray diffraction, structural information
at high resolu tion was obtained, leading to having much
accuracy st ructure of the proteins. The crystallographic
structure of Av1 at 1.16 Å resolution not only demonstrated
the [8Fe-7S] structure of P-cluster, but also showed that N
is most likely the previously unrecognized atom in FeMoco
(Einsle et al., 2002). This new advance provides either an
opportunity or a challenge for scientists to think again about
how N2 inserts and reduce in FeMoco.
2.3 Nitrogenase complex
At the protein level, the basic mechanism of nitroge-
nas e invo lves the fo llowing: (1) fo rmation of a complex
between the reduced Fe-protein with two bound ATP mol-
ecules and MoFe-protein; (2) electron transfer between the
two proteins coupled to the hydrolysis of ATP; (3) disso-
ciation o f the Fe-pro tein accompanied by reduction and
exchange of ATP for ADP; (4) repetition of this cycle until
sufficient numbers of electrons (and protons) have been
accumulated so that available substrates can be reduced
(Rees et al., 2000). It is in the transient complex between the
Fe pro tein and the MoFe pro tein that electron transfer
occurs, ultimately resulting in substrate reduction at a re-
mote site (FeMoco) in the MoFe protein. Hence, it was
speculated that nucleotide hydrolysis might serve as a regu-
lato r of conformational switching lead ing to electron
transfer.
Jang et a l. (2000) ob tained the Fe p ro tein bonding
MgADP state and determined its crystallographic structure.
The results show that switchs Ⅰ and Ⅱ underwent sig-
n ifican t s t ructu ral changes after Fe pro tein bound
nucleotide. This result provided an opportunity to differ-
entiate the individual contributions arising from nucleotide
binding, n itrogenase complex formation, and complex-de-
pendent nucleotide hydrolysis in the nitrogenase enzyme
system.
The Av1-Av2 complex stabilized with ADP·ALF4- was
crystallographicly analyzed (Schindelin et a l., 1997). The
res ult shows that Av2 underwent substantial conforma-
tional changes and Av1 d id not directly interact with the
nucleotide. It is indicated that both Av1 and Av2 played a
role in the stabilization of Av2-nucleotide intermediate in
nucleotide hydrolysis. A conformational change in Av2 re-
sults in a modified binding s urface that permits the [4Fe-
4S] cluster in Av2 to approach the Av1 P-cluster ~4 Å closer
than possible with simple van der walls contact between
ZHAO Jian-Feng et al.: Advances in Studies on Nitrogenase Crystallography 397
the individual protein components. Interactions in the com-
plex have broad implications for signal and energy trans-
duction mechanisms in the multiprotein complex. Another
complex, EDC {N-[3-(d imethy lamino)pro py l]-N’-
ethylcarbodiimide} cross-linked Av1-Av2, was als o crys-
tallographically analyzed (Schmid et al., 2002). The authors
proposed that EDC cross-linking trapped a nucleotide-in-
dependent precomplex of the nitrogenase p roteins driven
by complementary elect ros tatic in teract ions . The
precomplex subsequently rearranges in a nucleotide-de-
pendent fashion to the electron t ransfer competent state
observed in the ADP·ALF4- s tructu re. Th is gives us
dynamic insights into nitrogenase mechanism.
In the case of the Fe protein, Leu 127 and Asp 125-x2-x3-
Gly 128 in the conserved sequence motif is part of the switch
Ⅱregion. The Av2 with a deletion of residue Leu127 (DAv2)
can fo rm a tight, inactive complex with the Av1 in the ab-
sence of nucleotide (denoted DAv2-Av1 complex)(Chiu et
al., 2001). MgATP was bound to the crystal by soaking the
crystals in MgATP solution. This demonstrates that disso-
ciation of the nitrogenase complex is not required for nucle-
otide binding, whereas kinetic analysis have shown that
dissociation of the Av1-Av2 complex is the rate-determin-
ing step. Consequently, the requirement and role for com-
plex dissociation in the nitrogenase mechanism should be
reconsidered. The d ifferences in nucleotide binding be-
tween the L127 DAv2-Av1 and Av2-ADP·ALF4-Av1
complex indicate that the mechanism of nucleotide hydroly-
sis by L127 DAv2 must be different from that proposed for
the wild-type protein.
3 Prospects of Nitrogenase Crystallography
From the above discus sion, it is no t difficult to notice
that more and more attention should be paid to the follow-
ing crystallographic studies.
3.1 Nitrogenase-substrate complex
The X-ray diffraction analysis of MoFe protein and Fe
protein crystals only provides important information about
their “static” structures. In fact, the information is only a
necessary base for fully understanding the mechanism of
nitrogen fixation. The clear elucidation of the nitrogen fixa-
tion mechanism is only dependent on full knowledge about
the exact “dynamic” changes in space s tructures o f pep-
tide chains and metalloclusters in two proteins during bind-
ing and reducing substrates. The crystallographic analysis
of both Av1-Av2-ADP complex and Av1 with different re-
dox states is able to provide some “dynamic” s tructu ral
information. The crys tallographic stud ies have demon-
st rated that s ome changes in conformation of enzyme
molecule could appear after the enzyme binds substrate
（McPherson et a l., 1976）. It is reasonable to expect that
the crystallographic analysis of nitrogenase-substrate com-
plex should provide more and more information about the
“dynamic” st ructural changes of nit rogenase. The in for-
mation from the complex should be much more important
than those from individual proteins. It has also been dem-
onstrated that under some condition substrate-bound en-
zyme sometimes is easier crystallized（McPherson et a l.,
1976）. Therefore, it is undoubtedly worth to try crys talli-
zation of the nitrogenase-substrate complex when groping
about cond it ions fo r crys tallization o f n itrogenas e
components.
3.2 Mutant nitrogenase
It is certain that the changes of the metalloclusters in
MoFe protein and Fe protein could result in the changes
in space s tructures of the peptide chains around the clus-
ters when n itrogenase binds and reduces subst rate. It is
neces sary for the exact elucidation of nitrogen fixation to
obtain the information about the relation of the changes
between the clusters and peptides. The crystallographic
st ructures o f Dni fB Av1 and NifV- Kp1 show that it is
very useful for knowledge o f the s tructural changes to
comparative study on crys tallographic s tructures of ni-
trogenas e from wild-type s train and differen t mutant
strains.
3.3 Nitrogenases with the heterometal centers
From the bioinorganic point of view, it is of great impor-
tance to elucidate the influence of the heterometal center
M (Mo, V, Fe) on the structure of related protein and/or the
corresponding FeMco which may be present in different
redox states, its electric structure, especially on the selec-
tiv ity for s ubstrate reduction (Müller et al ., 1992). It is
possib le that the M in FeMco of MnFe protein and CrFe
protein could be Mn and Cr, respectively (Huang et a l.,
2001; 2002). The existence of the pro teins with different
heterometals thus provides the opportunity for better un-
derstanding of the N2 fixation mechanism by comparative
biochemical and biophysical studies.
With improving techniques of both crystal growth and
X-ray diffraction analysis, it is hopeful to perform the struc-
tural analysis of the three kinds of crystals mentioned above,
resulting in some important advances or breakthrough of
understanding mechanism of nitrogen fixation. As the crys-
tal growth is still a main “hindrance”, a realization of the
hopes is greatly dependent on the s uccessful growth of
large single crystals with good quality of the proteins men-
tioned above.
Acta Botanica Sinica 植物学报 Vol.46 No.4 2004398
References:
7th Laboratory, Institute of Botany, T he Chinese Academy of
Sciences . 1973. Purification and crystallization of MoFe pro-
tein from Aztobacter vinelandii. Acta Bot Sin , 15: 281-284.
(in Chinese with English abstract)
Bian S-M , Lü Y-B, Zhao Y , Zhao J-F, Zhou H-N , Huang J-F
. 2004. Purification and crystallization of nitrogenase MoFe
protein (DnifH Av1) from nifH deleted mutant strain of Azoto-
bacter vinelandii. Acta Bot Sin, 46:58-62.
Bolin J T, Camp obasso N, M uchmore S W, Morgan T V,
Mortenson L E. 1993. The structure and environment of the
met al clust ers in t he nitrogenase M oFe p rot ein from
Clostridium pasteurianum. Stiefel E I, Coucouvanis D, New-
ton W E. M oly bdenum Enzy mes , Cofact ors and Model
Systems. Acs Symposium Series No. 535. Washington D C:
American Chemical Society. 186-195.
Bolin J T, Ronco A E, Mortenson L E, Morgan T V, Williamson
M, Xuong N H. 1990. St ructure of the nitrogenase M oFe
protein: spatial distribution of the intrinsic metal atoms deter-
mined by X-ray anomalous scattering. Gresshoff R S, New-
ton W E. Nitrogen Fixation: Achievement s and Objectives.
London: Chapman and Hall. 117-124.
Brigle K E, Weiss M C, Newton W E, Dean D R. 1987. Products
of the iron-molybdenum cofactor-specific biosynthetic genes,
nifE and nifN, are structurally homologous to the products of
nitrogenase molybdenum-iron protein genes, nifD and nifK. J
Bacteriol, 169:1547-1553.
Burns R C, Holsten R D, Hardy R W. 1970. Isolation by crystal-
liz ation of the Mo-Fe protein of Azotobacter nitrogenase.
Biochem Biophys Res Commun, 39: 90-99.
Chiu, H J, Peters, J W, Lanzilotta W N, Ryle M J, Seefeldt L C,
Howard J B, Rees D C. 2001. MgATP-Bound and Nucle-
otide-Free Structures of a Nitrogenase Protein Complex be-
t ween Leu 127D-Fe-Protein and t he M oFe-Prot ein.
Biochemistry, 40:641-650.
Dong M-L , Huang J-F, Du H-S , Li J-H , Du S-B , Luo A-L, Jiang
Y, Zhang C-H. 1996. The application of radioactive isotope
99Mo to the reconstituted MoFe protein. J Radioanal Chem,
205:261-267.
Drennan C L, Pet ers J W. 2003. Surpris ing cofact ors in
metalloenzymes. Curr Opin Struct Biol, 13:220-226.
Drenth J, Helliwell J R, Littke W. 1987. The crystal growth of
biological materials. Walter. Fluid Science and Mat erials in
Space. Ge P-W , Wang J-T, Xu Z -Y , Li G-D , Hu W-R ,
Zhang X-M , Zhong X-R, Tang D-S , Gu B-Y, Xu S-C , Pan
H-R, Huo C-R trans. 1991. Beijing:Chinese Science and Tech-
nology Press. 358-377. (in Chinese)
Einsle O, Tezcan F A, Andrade S L A, Schmid B, Yoshida M,
Howard J B, Rees D C. 2002. The nitrogenase MoFe-protein
at 1.16 Å resolution: a central ligand in the Femo-cofactor.
Science, 297:1696-1700.
Georgiadis M M, Komiya H, Chakrabarti P, Woo D, Kornuc J J,
Rees D C. 1992. Crystallographic structure of the nitrogenase
iron protein from Azotobacter vinelandii. Science, 257:1653-
1659.
Georgiadis M M, Chakrabarti P, Rees D C. 1990. Crystal struc-
t ure of t he nitrogenase iron protein from Azotobacter
vinelandii. Roth G, Newton W E. Nitrogen Fixation: Achieve-
ments and Objectives. New York: Chapman and Hall. 117-
124.
Hales B J. 1990. Alternative nitrogenase. Adv Inorg Biochem, 8:
165-198.
Howard J B, Rees D C. 1996. Structural basis of biological nitro-
gen fixation. Chem Rev, 96:2965-2982.
Huang J-F , Wang D-Y , Dong Z-G, Wang Z-P , Li J-G, Wang Y-
P . 1999. Relationship between nifZ and the synthesis of P-
cluster in nitrogenase MoFe protein of Azotobacter vinelandii.
Acta Bot Sin , 41:1217-1220. (in Chinese with English abstract)
Huang J-F , Dong Z-G, Wang D-Y , Lü Y-B , Zhang H-F , Wang Y-
P, Han-Y , Dai X-H . 2001. Growth of the crystals of nitroge-
nase MnFe protein. Acta Bot Sin , 43: 375-379.
Huang J-F, Wang D-Y, Dong Z-G , Wang Z-P , Zhang H-F, Lü Y-
B , Xu X-M . 2001. Purification and characteristics of Mn-
containing nitrogenase component Ⅰ. Acta Bot Sin , 43:918-
922.
Huang J-F , Wang D-Y, Wang Z-P, Luo A-L , Gu S-R , Li J-G .
1998. Effect of PEG and salt on crystallization of bacteriferritin
and nitrogenase MoFe protein from Azotobacter vinelandii.
Acta Bot Sin, 40:880-882. (in Chinese with English abstract)
Huang J-F , Wang Y-P , Dong Z-G , Huang X-M, Wang D-Y , Lü
Y-B , Wang Z-P . 2000. Studies on crystalline growth of MoFe
potein from a nifZ deleted strain of Azotobacter vinelandii.
Acta Bot Sin , 42:383-387.
Huang J-F, Zhang H-F, Dong Z-G , Lü Y-B, Zhao Y, Wang Z-P .
2002. Purification and characteristics of Cr-containing nitro-
genase component. Acta Bot Sin, 44:297-300.
Huang J-F, Wang Z-P , Luo A-L,Zhong Z-P, Li J-G ). 1994. The
effect of manganese on the reconstitution of p artial metallo-
cluster-deficient nitrogenase molybdenum-iron protein. Acta
Bot Sin, 36:411-417. (in Chinese with English abstract)
Huang J-F, Wang Z-P, Luo A-L, Zhong Z-P , Li J-G. 1995. Reac-
tivation of partial metallo-cluster-deficient MoFe protein by
chromium-cont aining reconstituent solution. Acta Bot Sin ,
37:283-288. (in Chinese with English abstract)
Jang S B, Seefeldt L C, Peters J W. 2000. Insights into nucleotide
signal transduction in nitrogenase: structure of an iron protein
with MgADP bound. Biochemistry, 39:14745-14752.
ZHAO Jian-Feng et al.: Advances in Studies on Nitrogenase Crystallography 399
Jang, S B, Seefeldt L C, Peters J W. 2000. Modulating the mid-
point potential of the [4Fe-4S] cluster of the nitrogenase Fe
protein. Biochemistry, 39:641-648.
Kim J, Rees D C. 1992. Crystallographic structure and functional
implications of the nitrogenase molybdenum-iron protein from
Azotobacter vinelandii. Science, 360:553-560.
Kim J, Rees D C. 1992. Structural models for the metal clusters in
the nitrogenase molybdenum-iron protein. Science, 257:1677-
1682.
Kim J, Woo D, Rees D C. 1993. X-ray crystal structure of the
nitrogenase moly bdenum-iron p rot ein from Clos tridium
pasteur ianum at 3.0 Å resolution. Biochemistry, 32:7104-
7115.
Lü Y-B , Zhao Y, Zhao J-F , Dai X-H , Cang H-X, Wang Y-P,
Huang J-F . 2003. Growth of large single crystals of nitroge-
nase CrFe protein and MnFe protein. Acta Bot Sin , 45:289-
294.
Mayer S M, Lawson D M, Gormal C A, Roe S M, Smit h B E.
1999. New insights into s tructure-function relationships in
nitrogenase: a 1.6 angstrom resolution X-Ray crystallographic
study of Klebsiella Pneumoniae MoFe protein. J Mol Biol,
292: 871-891.
Mayer S M, Gormal C, Smit h B, Lawson D. 2002. Crystallo-
graphic analysis of the MoFe protein of nitrogenase from a
Nifv mutant of Klebsiella pneumoniae identifies citrate as a
ligand to the molybdenum of femoco. J Biol Chem, 277:35263-
35266.
McPherson A J. 1983. The growth and preliminary investigation
of protein and nuclei acid crystals for X-ray diffraction analysis.
Glick D. Methods of Biochemical Analysis. Vol. l23. Group
of Crystalline Growth, Institute of Biophysics, The Chinese
Academy of Sciences trans . Beijing: Chinese Science and
Technology Press. 249-345. (in Chinese)
Milagros S C, Tang H Y, Tucker D L, Sherman L A. 1999. Analy-
sis of the nifHDK operon and structure of the NifH protein
from the unicellular diazotrophic cyanobacterium Cyanothece
strain sp. ATTCC 51142. Biochem Biophys Acta, 1473:363-
375.
Müller A, Schneider K, Knüttel K, Hagen W R. 1992. EPR
spectroscopic characterization of an “iron-only” nitrogenase.
FEBS, 303:36-40.
Paus tian T D, Shah V K, Robert s G P. 1989. Purification and
characterization of the nifN and nif E gene products from Azo-
tobacter vinelandii mutant UW45. Proc Natl Acad Sci USA,
86: 6082-6086.
Peters J W, Stowell M H, Soltis S M, Finnegan M G, Johnson M
K, Rees D C. 1997. Redox-dependent structural changes in
the nitrogenase P-cluster. Biochemistry, 36:1181-1187.
Rangaraj P, Ryle M J, Lanzilot ta W N, Ludden P W, Shah V K.
1999. In vitro biosynthesis of iron-molybdenum cofactor and
maturation of the nif-encoded apodinitrogenase. Effect of sub-
stitution for NifH with site-specifically altered forms of NifH.
J Biol Chem, 274:19778-19784.
Rees D C, Howard J B, 1983. Crystallization of the Azotobacter
vinelandii nitrogenase iron protein. J Biol Chem, 258:12733-
12734.
Rees D C, Howard J B. 2000. Nit rogenase: standing at t he
crossroads. Curr Opin Chem Biol. 4:559-566.
Roll J, Shah V K, Dean D R, Roberts G P. 1995. Characteristics
of NIFNE in Azotobacter vinelandii st rains . J Biol Chem,
270:4432-4437.
Ryle M J, Lanzilott a W N, Seefeldt L C. 1996. Elucidating the
mechanism of nucleotide-dependent changes in the redox po-
tential of the [4Fe-4S] cluster in nitrogenase iron protein: the
role of phenylalanine 135. Biochemistry, 35:9424-9434.
Salemme F R. 1972. A free interface diffusion technique for the
crystallization of proteins for X-ray crystallography. Archives
Biochem Biophys, 151:533-539.
Schindelin H, Klsker C, Schlessman J L, Howard J B, Rees D C.
1997. Structure of ADP-AlF4-stabilized nitrogenase complex
and its implications for signal transduction. Nature, 387:370-
376.
Schlessman J L, Woo D, Joshua-Tor L, Howard J B, Rees D C.
1998. Conformational variability in structures of the nitroge-
nase iron proteins from Azotobacter vinelandii and Clostridium
pasteurianum. J Mol Biol, 280:669-685.
Schmid B, Ribbe M W, Einsle O, Yoshida M, Thomas L M, Dean
D R, Rees D C, Burgess B K. 2002. Structure of a cofactor-
deficient nitrogenase MoFe protein. Science, 296:352-356.
Schmid B, Einsle O, Chiu H J, Willing A, Yoshida M, Howard J B,
Rees D C. 2002. Biochemical and structural characterization
of the crosslinked complex of nitrogenase: comparison to the
ADP-AlF4 s tabilized st ruct ure. Biochemistry, 41:15557-
15565.
Schmid B, Ribbe M W, Einsle O, Yoshida M, Thomas L M, Dean
D R, Rees D C, Burgess B K. 2002. Structure of a cofactor-
deficient nitrogenase MoFe protein. Science, 296:352-356.
Shah V K, Brill W J. 1973. Nit rogenase. IV. Simple method of
purification to homogeneity of nitrogenase components from
Azotobacter vinelandii. Biochim Biophys Acta, 305:445-454.
Shah V K, Brill W J . 1977. Isolation of an iron-molybdenum
cofactor from nitrogenase. Proc Natl Acad Sci USA, 74:3249-
3253.
Sorlie M, Christiansen J, Lemon B J, Peters J W, Dean D R, Hales
B J. 2001. Mechanistic features and structure of the nitroge-
nase Alpha-Gln195 MoFe protein. Biochemistry, 40:1540-
1549.
Strop P, Takahara P M, Chiu H-J, Angove H C, Burgess B K,
Acta Botanica Sinica 植物学报 Vol.46 No.4 2004400
Rees D C. 2001. Crystal structure of the all-ferrous [4Fe-4S]0
form of the nitrogenase iron p rotein from Azotobacter
vinelandii. Biochemistry, 40:651-656.
Tal S, Chun T W, Gavini N, Burgess B. 1991. The DnifB (or
DnifE) FeMoco-deficient MoFe protein is different from the
DnifH protein. J Biol Chem, 266:10654-10657.
Watt G D, Reddy K R N. 1994. Formation of an all ferrous Fe4S4
cluster in the iron protein-component of Azotobacter vinelandii
nitrogenase. J Inorg Biochem, 53:281-294.
Weininger M S, Mortenson L E. 1982. Crystallographic proper-
ties of the MoFe proteins of nitrogenase from Clostridium
pasteurianum and Azotobacter vinelandii. Proc Natl Acad Sci,
79:378-380.
Zhang H-F, Wang Y-P,Lü Y-B ,Zhao Y, Dai X-H, Dong Z -G,
Huang J-F . 2002. Crystalline growth of nitrogenase CrFe
protein. Acta Bot Sin , 44:400-404.
Zhao J-F, Zhao Y, Wang Z -P , Lü Y-B , Qian Z-X, Huang J -F.
2003. Purification and activation in vitro of MoFe Prot ein
from nifE deleted mutant strain of Azotobacter vinelandii. Acta
Bot Sin, 45: 815-819.
Zhao Y, Zhao J-F, Lü Y-B, Wang Z-P , Wang Y-P, Huang J -F.
2003. Crystallization of nitrogenase MoFe protein from nifE
deleted mutant strain of Azotobacter vinelandii. Acta Bot Sin ,
45:427-431.
Zhao Y, Zhao J-F, Lü Y-B, Huang J-F, Wang Y-P,Cang H-X, Bi
R-C. 2003. Crystal growth of nitrogenase CrFe prot ein and
MnFe protein in space. Acta Bot Sin, 45:1043-1048.
Zhao Y, Lü Y-B, Zhao J -F, Zhou J-X, Qian Z-X, Wang Y-P,
Huang J-F. 2003. Crystallizat ion of nitrogenase MoFe pro-
tein (NifB-Av1) from a nifB mutated strain UW45 of Azoto-
bacter vinelandii. Acta Bot Sin, 45:820-824.
Zhong Z-P, Hu C-Z , Wang J-W , Huang J-F, Luo A-L, Li J-G.
1996. Purificat ion and properties of nitrogenase MoFe pro-
tein from a nifZ deleted strain of Azotobacter vinelandii. Acta
Bot Sin, 38:605-611. (in Chinese with English abstract)
(Managing editor: HE Ping)