Bacterial Energy Metabolism

Chapter 11
Bacterial Energy Metabolism

Bruce Ward
School of Biological Sciences, The University of Edinburgh, Edinburgh, UK
INTRODUCTION TO ENERGY energy generation is coupled to ATP synthesis and utili-

METABOLISM zation. This will be followed by a discussion of ATPases
and energy generation by fermentation. Bacterial aerobic
Energy metabolism is integrated with other metabolic pro- and anerobic respiration will be discussed with an
cesses such as chemotaxis, nutrient uptake, secretion of emphasis on the terminal cytochrome oxidases in aerobic
polymers, efflux of waste metabolites and toxic com- respiration denitrification in anaerobic respiration and
pounds. The central component in most bacteria is a pro- ATPases. The main systems by which regulatory proteins
ton (H1) translocating ATPase. In fermentative bacteria control aerobic and anaerobic respiration are covered
this converts ATP derived from substrate-level phosphor- through examples of one and two component regulators,
ylation into a proton motive force (pmf), composed of followed by consideration of the switch from aerobic to
two components: an electrochemical proton gradient ΔpH anaerobic metabolism in E. coli and Paracoccus denitri-
and a membrane potential Δμ. In bacteria that derive ficans. At the end of the chapter some examples/case
energy by respiration electrons are passed down a respira- studies of energy metabolism of bacteria are provided
tory chain consisting of cytochromes and flavoproteins; at and related to the medical interest in these bacteria.
specific points in this process there is a charge separation
and protons are translocated across the cytoplasmic mem-
brane, creating a pmf. The H1-ATPase is reversible so it
is possible to use the pmf to generate ATP.
CHEMIOSMOSIS
Bacteria show an astounding diversity in their ability Bacteria can gain energy by a number of processes: aero-
to use metabolites for energy generation and to switch bic respiration, anaerobic respiration, fermentation and
from one mode of energy generation to another. This is photosynthesis. In all these processes energy-yielding
most obviously shown in facultative aerobes such as metabolism is coupled to the formation of ATP. Prior to
Escherichia coli that can use aerobic respiration with the proposal of the chemiosmotic hypothesis by Peter
oxygen as the final electron acceptor but can also use Mitchell in the 1960s the coupling of energy-yielding
alternative electron acceptors such as nitrate, fumarate, reactions such as oxidation of organic substrates to
trimethylamine oxide (TMAO) and dimethylsulphoxide energy-utilizing reactions such as ATP synthesis or solute
(DMSO). E. coli also can fine tune the respiratory chain uptake was poorly understood and researchers sought in
by using alternative cytochrome oxidases according to the vain for an ‘energy-transducing intermediate’. The ele-
oxygen concentration surrounding the cell. gance of the chemiosmotic hypothesis is that it provides a
Other bacteria are more limited. Anaerobes such as simple and unifying explanation for the links between
Bifidobacterium longum, a colonizer of the human gut, multiple cellular processes (e.g. active transport, flagella
gain energy by fermentation. rotation, electron transport, etc.) and fitted in with the
known features of respiration and oxidative phosphory-
lation (e.g. the breakdown of the control of respiration by
SCOPE OF CHAPTER ATP by uncouplers). Peter Mitchell proposed in 1961
In this chapter we will consider first the theory of chemi- that proton translocation across the membrane was the
osmosis which forms a testable basis to explain how link between the different processes and made four
Molecular Medical Microbiology. DOI: http://dx.doi.org/10.1016/B978-0-12-397169-2.00011-1

© 2015 Elsevier Ltd. All rights reserved. 201
202 PART | 2 Bacterial Cell Function
hypothetical postulates [1] which are summarized below membrane and the Δψ term is due to the difference in
in modern terminology. electrical potential.
Biologists tend to use the term Δp rather than Δμ~ H1
1. The ATP synthase is a reversible proton-motive
but the two are simply related by the equation:
ATPase with characteristic H1/ATP stoichiometry.
2. Respiratory and photosynthetic electron transport Δp 5 Δ~μH1 =F 5 Δψ 2 2:3RTðpHin 2 pHoutÞ=F
chains translocate protons with characteristic H1/2e2
5 Δψ 2 2:3RT ΔpH=F
stoichiometry.
3. Membranes should possess solute transport and ion (11.1)
exchange systems utilizing proton-linked carriers.
4. Energy-transducing membranes should have low per- By convention the two sides of the membrane are
meability to protons (or other translocating ion). termed the P-side and the N-side (positive and negative)
such that the direction of proton translocation is from the
In order to ensure efficient coupling between generation N-side to the P-side. For example, in bacteria aerobic res-
and utilization of the pmf it is important to ensure that pro- piration protons are pumped from the cytoplasmic side of
tons are rapidly utilized and not dissipated by loss into the the membrane to the periplasmic side creating a pmf of
external medium. Bacteria utilize several means of doing about 170 mV with a Δψ value of c. 140 mV. A ΔpH of
this: increased efficiency of proton binding in the H1- 0.5 units gives a contribution of c. 30 mV as it has a
ATPase, close localization of ion producers and ion utili- numerical value of approximately 60 mV per pH unit.
zers, minimizing loss of protons to the hydrophilic phase Much effort has been devoted to measuring the Δψ and
and maximizing retention of protons at the external face of ΔpH components, to showing the different ways in which
the cytoplasmic membrane, and the use of membrane inva- net proton translocation can be obtained for specific redox
ginations. Such consideration has led to the concepts of systems, to measuring the stoichiometry of proton translo-
organization of bioenergetic complexes into microcompart- cation, to identifying the channels for proton movement,
ments (reviewed in [2]). and to identifying the detailed mechanisms by which redox
The term chemiosmotic was chosen by Mitchell catalysis and proton translocation occur for specific sys-
because the reactions were vectorial (directional) and tems, e.g. cytochrome oxidase. Two mechanisms by which
involved both a chemical component (transport of chemi- proton translocation could theoretically happen are the
cal groups) and osmotic (translocation of a solute across a redox loop and the proton pump [36].
membrane). The process of proton translocation generates In the redox loop system (Fig. 11.1) oxidation of the
a proton gradient across the membrane termed the proton substrate, D at the N-side is coupled to reduction of the
electrochemical gradient Δ~μH1 composed of two compo- acceptor A in such a way that the protons are released
nents, ΔpH and Δψ, the membrane potential. The ΔpH finally from the intermediate carrier (e.g. UQH2) at
term is due to the difference in concentration (more accu- the P-side and the final reduction step occurs by transfer of
rately activity) of protons between the two sides of the electrons and utilization of protons at the N-side
(a) 2H+ (b) FIGURE 11.1 Mechanisms for the generation of a proton motive force:
2H+ (a) redox loop, (b) proton pump, (c) combination of redox loop and pro-
P
ton pump as in cytochrome caa3 oxidase and (d) ATP hydrolysis by an
2e QH2 H1-ATPase. In each case a transmembrane electrochemical proton gradi-
QH2
N Q Q ent is created between the P-side (positive) and the N-side (negative) of
the membrane. Note other arrangements of components in the membrane
SH2 S A+2H+ A SH2 S A AH2 are possible. Adapted from [3,4,7].
(c) (d)
4 cyt cred 4 cyt cox
4H+ 3H+
4e P
O2
2H2O
4H+
N
4H+ 4H+
ATP + H2O ADP
Chapter | 11 Bacterial Energy Metabolism 203
(2H1 1 2e2 1 A - AH2). There should be strict stoichi-

ometry of 2H1 translocated per 2e2 transferred. There are a
number of variants of this redox loop system [35]. In the
proton pump model protons are pumped from the N- to the
P-side in a series of steps but overall in a unidirectional
manner. The proton carriers in this case are mainly
protonatable residues of amino acids in the complexes,
which can be experimentally tested by site-directed muta-
genesis studies. There are two important requirements for
this process: first the pKa of key residues must change so
the proton affinity (binding strength) will change to allow
uptake and release of protons. The pKa of amino acids
respond to conformational changes but also to changes in
local pH. Secondly the entry and exit ‘channels’ for protons
must be gated so that when the pathway to the N-side is
open, that to the P-side is closed and vice versa [5]. An
example of a redox loop system is transfer from cytochrome
c to the P860 reaction centre in bacterial photosynthesis; FIGURE 11.2 Representation of the bacterial F1F0 H1-ATPase. The F0
the change of pKa of cofactors in response to the redox state component ab2c1015 is membrane-bound and the F1 component α3β3γδε
(oxidized or reduced) and to the pH (which will be different is cytoplasmic. F0 has an oligomeric c-ring and is linked to the F1 α3β3
on the N- and P-sides of the membrane) is important to catalytic units in two ways: centrally by the γ subunit, which together
allow proton binding and release [6]. There can be cases with the ε subunit forms a foot connecting to the c-ring, and peripherally
by the elongated stalk δab2, which interact with α3β3 through the N-
where there is a combination of redox loop and proton terminal domains of the δ and b subunits. In the rotary motor the rotor
pumping; examples are the fumarate reductase system in comprises the c-ring and γε subunit and the stator is the δab2 stalk that
Wolinella succinogenes [5] and the haemcopper bacterial prevents rotation of the α3β3 module [8].
cytochrome caa3 oxidases. Studies on the latter have identi-
fied two proton input pathways [6], termed the D channel
and the K channel after key residues in each (an aspartate D a mutant of the ε subunit (S65D). The strategy behind this
and a lysine K); this is discussed further in the section on approach was that the E. coli ATPase can adopt both
cytochrome oxidases. extended and compact conformations. Introduction of the
mutation destabilized the compact conformation mutation
and dramatically increased crystallization success and
ATPase reproducibility [10].
As discussed above the existence of a proton translocating Structural and biochemical studies have clearly estab-
ATPase is a key tenet of Mitchells’ chemiosmotic hypothe- lished that the ATP synthase functions as a rotary molecular
sis. For all organisms these ATPases are central to energy motor with c10 acting as a rotor and the peripheral stalk
metabolism because they convert the pmf generated by δab2 as the main component of the stiff stator [8,11,12]. The
energy-yielding reactions to ATP, which can then be used α and β subunits assemble as an α3β3 (subassembly), com-
to drive a multitude of cellular processes such as chemo- mon to all ATP synthases, with the α and β subunits alter-
taxis. The H1-ATPase in both mitochondria and bacteria nating and clustering around the γ subunit, which also
has two main components: a hydrophilic F1 component and extends into the membrane and together with the bacterial
a hydrophobic F0 component linked by a slender stalk ε subunit (or functionally equivalent mitochondrial δ sub-
(Fig. 11.2). The F1 component, which can be solubilized unit) bind to the F0 portion. During catalysis the asymmetric
from cells using low ionic strength buffers, consists of five γ subunit rotates in steps of 120 (with two substeps) and the
subunits α, β, γ, δ, ε in the ratio 3:3:1:1:1 and can be seen three β subunits change successively between three different
as spheres of 85100 Å in electron micrographs. The F0 conformational states: βDP the ADP inhibited state, βTP
component contains integral membrane proteins a, b and c and βE. This latter state is known as the ‘open’ state as there
in the ratio 1:2:1014 and forms the base and stalk of the is no ADP or ATP bound and results from the γ subunit
complex. The crystal structure of the bovine ATPase was causing part of the nucleotide-binding domain and the
solved by Walker and colleagues [9] and since then other attached C-terminal domain to hinge outwards. As the rotor
studies of bovine heart and yeast mitochondrial ATPase turns ADP and phosphate bind at the βE-site as it closes,
have yielded structures, whereas few bacterial ATPases ATP formation occurs in the βDP-conformation and ATP is
have been successfully crystallized. Recently a structure of released as the βTP-site opens and converts back into the
the E. coli ATPase was obtained at 3.15 Å resolution using βE-conformation. Each 360 rotation of the central stalk
leads to the formation of an ATP molecule from each of the b was sufficient to give an F0 complex capable of H1
three β subunits. During ATP hydrolysis rotation is counter- translocation. Topological studies were used to show that
clockwise, whereas it is clockwise during ATP formation. subunit a forms five TMHs [15]. It is also thought that
The above description omits the nature and mechanism TMH4 of subunit a packs in parallel to TMH2 of subunit
of proton translocation. Rotation of the γ subunit is c. Specific residues thought to form part of the aqueous
thought to occur in conjunction with rotation of the c sub- channel were determined by substituting residues with
unit in the hydrophobic F0 portion (due to binding between Cys and then testing with N-ethylmaleimide and Ag1.
the c subunits and the γε subunits such that γεc10 acts as a Reaction of Cys with these reagents was predicted to
rotor with α3β3δab2 as the stator [11]. Rotation has been occur in aqueous environments. These studies led to the
demonstrated by immobilizing the F1ATPase to a solid conclusion that TMH2, TMH4 and TMH5, which cross-
surface and then using a reporter molecule (e.g. fluorescent linking studies show are in contact with each other, con-
actin filament or gold bead) fused to one of the rotating tribute to the two aqueous channels and a model was pro-
subunits (e.g. γ or c subunit). The F0 portion comprises a posed in which swivelling of TMH2 of subunit c, brought
complex of ab2c10 (E. coli) where the c10 oligomer is Asp61 of subunit c and Arg210 of subunit a together with
arranged in a ring and each c subunit has two transmem- proton transfer from Asp61 to Arg210 and then exit of the
brane helices (TMHs) (Fig. 11.2). The α-helices of the c proton to the cytoplasm via an aqueous channel involving
subunits are ordered such that the N-terminal portions TMH4. Further conformational changes would move
form an inner ring in the core of the cylinder, and the C- Arg210 of subunit a away from Asp61 of subunit c and
terminal α-helices provide the external surface of the ring. allow movement of protons from the periplasm through
An aspartyl residue (D61) in subunit c (glutamate in some the aqueous channel involving TMH2 at the periplasmic
species) plays a key role in binding protons and is postu- face to the centre of the membrane where residues in
lated to release protons to an aspartyl residue (D210) in mid-helix of TMH4 and TMH5 might contribute to the
the a subunit. This residue is in the C-terminal α-helix on proton pathway. Other later models propose that the exit
the outside of the ring circumference of the cylinder. pathway is along the a and c subunit interface.
Models for proton translocation through the lipid bilayer
for H1-ATPase share common features with those for the
Na1-ATPase and the flagellar motor [13]. It is assumed
Na1-ATPase
that the aspartyl (glutamate) residue in the c subunit is Although proton translocation to create a pmf is prevalent
always protonated when facing the hydrophobic lipid core in most bacteria the possibility of sodium ion transloca-
of the membrane (as a charged residue would be energeti- tion to create a sodium-motive force (smf) exists.
cally unfavourable), that the exchange of protons with Theoretically the use of sodium ion pumping in akali-
either bulk phase is allowed only for those copies of c in philes would seem a sensible mechanism for these bacte-
contact with subunit a and that the entry and exit proton ria: at high pH the pmf would be low, and the smf could
channels in subunit a are not colinear [13]. The model be used in Na1-coupled systems for solute uptake and
described in [13] as applied to bacteria proposes that pro- motility. Other advantages would be in keeping the inter-
tons external to the cytoplasmic membrane enter a channel nal environment closer to neutral pH and the osmotic
and bind to the aspartyl residue in the c subunit thus giving strength low by reducing proton translocation to the
a neutral residue which can rotate in contact with the outside of the cell and sodium accumulation inside.
hydrophobic lipid bilayer until the residue reaches a sec- However in practice most alkaliphiles such as the extreme
ond site in contact with the a subunit, where conditions alkaliphile B. pseudofirmus OF4 use a H1-ATPase.
permit release of the proton and exit through a second site Comparison of sequences between the ATPase subunits
on the cytoplasmic side of the membrane. Each complete of Gram-positive alkaliphiles and neutrophiles, e.g.
turn of the c-ring yields three ATP and the models predict Bacillus spp., shows conserved differences in residues in
that the number of H1 that move across to the c-side of the a and c subunits. In neutrophiles cross-linking and
the membrane during a full turn is equal to the number of mutational studies indicate Gly218 and Glu219 in TMH4
c subunits in the ring. Therefore the H1/ATP ratio varies of subunit c lie close to His245 in TMH5 and play a role
among different bacterial strains [13] according to the in proton translocation. The corresponding residues in the
number of c subunits in the ATPase [8,1115]. For E. coli alkaliphile B. pseudofirmus firmus are Lys180, Glu181
with 10 c subunits this ratio would be 3.3. and Gly212. There is evidence that the conserved Lys180
The entry and exit of protons through the membrane and Gly212 play an essential role in alkaliphile
involves aqueous channels and subunit a is thought to B. pseudofirmus conduction of protons to the periplasmic
form part of these channels. Reconstitution experiments surface at high pH (reviewed by [14]).
with preformed ac subcomplexes from purified a and c However, there are a limited number of bacteria that
subunits with the 34-aa transmembrane domain of subunit utilize a physiologically active Na1-ATPase under
optimal growth conditions. Examples are Propiogenum Na1-ATPase but could function as a H1-ATPase under
modestum, Ilyobacter tartaricus, Acetobacter woodii and conditions favourable to proton coupling, such as low pH
Methanobrevibacter ruminantium. The common feature and low levels of sodium ions.
appears to be that these are all anaerobic bacteria that Although H1-ATPases predominate in terms of occur-
obtain energy by enzyme reactions directly coupled to rence in microorganisms there is evidence that the ances-
Na1 translocation to the periplasm and this is then tral ATPase was a Na1-ATPase. Ion coupling also
coupled to Na1 re-entry via the Na1-ATPase. For requires membranes impermeable to the translocating ion
P. modestum and I. tartaricus the primary Na1 pumps are (see section on chemiosmosis). Sodium ions are intrinsi-
decarboxylases (S-methylmalonyl-CoA decarboxylase and cally much less permeable than protons and organisms
oxaloacetate decarboxylase) involved in fermentation of have evolved different means of reducing the proton leak-
succinate and tartrate, respectively. The Na1-ATPases of iness of membranes such as restricting hydrocarbon chain
these bacteria have the same basic structure of an F1 mobility, modification of side chains and addition of
α3β3γ and an Fo portion with ab2c11. The structure of the minor constituents that enable tighter packing.
c11-ring in I. tartaricus and the Na1-binding sites have
been determined (see [11]). Each c subunit consists of
two TMHs separated by a hairpin loop in the middle such
FERMENTATION
that the N-terminal helices, with a conserved GxGxGxGx In the absence of oxygen or an alternative electron acceptor
motif that allows tight packing between subunits, forms like nitrate, bacteria must metabolize carbon substrates by
an inner ring surrounding a cavity and the C-terminal fermentation and ATP is generated by substrate phosphory-
helices contribute predominantly to the outer ring with a lation (phosphorylation directly linked to a chemical reac-
more complex shape. In the centre of the membrane the tion but not involving ion translocation), e.g. the conversion
outer helix approaches the inner ring most closely, fitting of succinyl-CoA to succinate. In many fermentations, e.g.
into the groove between two inner helices with a kink the homolactic fermentation of glucose to lactate, both
near the position of the key binding site Glu65 in the cen- reductant and oxidant are produced in the course of the fer-
tre of the membrane. The Na1 is bound between adjacent mentative metabolism of a single substrate that serves as
helices (Gln32, Glu65, of one subunit and Val63 and their common precursor. Thus the first part of the pathway
Ser66 of the next). Additional hydrogen bonds stabilize is oxidative and the reducing equivalents accumulate as
this conformation and ensure that Na1 rather than H1 NADH or FADH2. In the second part of the pathway, the
binds. For H1-ATPases H1 is bound at the equivalent reducing equivalents are reoxidized. In this example pyru-
Glu or Asp site (Asp61 in E. coli) but the basic hairpin vate acts as the electron acceptor and is reduced to lactate;
architecture of the c subunits is the same. the reaction utilizes NADH 1 H1, which regenerates
From the above considerations it is clear that there are NAD1. There are two interconnected consequences of this:
variations in how generation of ion gradients is coupled first, partly because the carbon source is only partially oxi-
to ATP synthesis. In most neutrophilic bacteria respiration dized, the energy yields are lower than in aerobic respira-
generates a pmf through H1 translocation that is con- tion, and secondly, the reduced products such as succinate,
verted to ATP through an H1-ATPase. In a few anaerobic butyrate, acetate, ethanol and H2 can be used for further
bacteria an smf produced by Na1 translocation through metabolism by other bacteria. This occurs in many natural
membrane-bound decarboxylases is converted to ATP environments such as the ruminant gut and in wetlands. For
through a Na1-ATPase. Despite the low pmf generated example the methanogenic archaea produce methane from
by alkaliphiles, these bacteria use an H1-ATPase to gen- either acetate or H2 1 CO2.
erate ATP but use Na1-coupled processes for solute Although fermentations are based on simple chemical
uptake and motility. In Vibrio cholerae where aerobic res- reactions, the subject area is extremely wide because of
piration using the Na1-translocating NADH_ubiquinone the wide variety of substrates that can be fermented, the
reductase complex (Na1NQR) is possible, the enormous number of different fermenting bacteria, varia-
H1-ATPase is also used. Conditions may alter specificity; tions in fermentation pathways due to genotype, environ-
for example the Na1-ATPase of P. modestum, can use mental conditions, etc., adaptations by some bacteria to
H1 for coupling to ATP synthesis when Na1 concentra- remove reducing equivalents by alternative pathways and
tions are low. Some methanogenic bacteria like the fact that in nature most fermentations occur in com-
Methanobrevibacter ruminantium, that convert H2 and plex microbial communities, e.g. the human gut micro-
CO2 to methane, appear to use both pmf and smf genera- biome. The natural occurrence of fermenting microbial
tion for growth as growth was inhibited in the presence of consortia allows the waste product of one bacterium to be
either protonophores or sodium ionophores that collapse excreted (e.g. lactate efflux) and to be utilized by another
the ion gradients generated. Biochemical characterization bacterium, thus preventing product accumulation and inhi-
of the archaeal A1A0 ATPase showed it was essentially a bition of growth.
Interest in microbial fermentations has received fresh and then to xylulose-5-phosphate. A lyase reaction
impetus from two rather different standpoints. Firstly, then splits D-xylulose-5-phosphate into acetyl
interest in examining the links between human health and phosphate 1 glyceraldehyde-3-phosphate, which can be fer-
nutrition has been advanced greatly by genomics and the mented to ethanol and lactate, respectively, in a heterolac-
characterization of the human gut microbiome in people tate fermentation (see later section).
from different dietary and genetic backgrounds. Secondly, The EMP pathway is present in very many bacteria
industrial fermentations involving genetic engineering to including E. coli, homofermentative Lactobacillus spp.,
provide products such as butanol, biogas, etc. are being Bacteroides fragilis, Bacillus spp. and Staphylococcus spp.
developed to meet the predicted future world require- There are relatively few bacteria that utilize the
ments. In this chapter discussion focuses mainly on non- ED pathway as the sole route for glycolysis (e.g.
industrial fermentation of sugars and polysaccharides but Pseudomonas aeruginosa and Zymomonas mobilis) but
there are some points in industrial fermentations that are some bacteria use it as a secondary pathway, e.g. metabo-
very important in this context but also have relevance to lism of gluconate in E. coli is a modified ED pathway. The
fermentations in natural environments. PK pathway is found in heterolactate fermenters such as
Leuconostoc spp. The major role of the PP pathway is in
biosynthesis, e.g. fatty acid biosynthesis; the PP pathway is
Fermentations of Sugars and Polysaccharides found in many bacteria but is thought to make a minor
Fermentations of sugars and polysaccharides can be con- contribution to glycolysis (e.g. 26% in B. subtilis).
ceptually divided into three steps: breakdown of poly- The presence of key enzymes in each of the four
meric polysaccharides into disaccharides and pathways (EMP fructose bisphosphate aldolase, ED
monosaccharides, intermediary metabolism of these sim- KDPG aldolase, PP phosphogluconate dehydrogenase
pler sugars, and terminal steps in the fermentation. The and PK phosphoketolase) has led to the suggestion that
first step produces glucose or other monosaccharides such the presence of a particular pathway in a specific bacterium
as xylose, fructose which are then metabolized by the can be deduced from the presence of the encoding gene in
classical pathways of glycolysis: the Embden-Meyerhoff the genome through a BLAST search. Whilst this is a nec-
pathway (EMP), the Entner-Doudoroff (ED) pathway, the essary condition it is safer to check out the entire pathway
pentose phosphate cycle (PP) and the phosphoketolase using a tool such as KEGG (Kyoto Encyclopedia of Genes
(PK) pathway. These are well described in older textbooks and Genomes) pathways software (www.genome.ad.jp/
(e.g. [16,17]). kegg/) with the caveats pertaining to metabolic reconstruc-
Briefly, the EMP pathway metabolizes glucose to pyru- tion (e.g. alternative genes fulfilling the same function,
vate with formation of two ATP and generation modification of gene function, etc.).
of two NADH molecules. The key enzyme is fructose-1,6-
bisphosphate aldolase which converts (splits) fructose-
1,6-biphosphate to glyceraldehyde-3-phosphate and
dihydroxyacetonephosphate. In the other three pathways
Fermentation Pathways
glucose-6-phosphate is converted to 6-phosphogluconate As emphasized above the terminal steps in fermentation
rather than fructose-6-phosphate as in the EMP pathway. In regenerate reducing equivalents and in some pathways
the ED pathway the key enzyme is 2-keto-3-deoxy-6-phos- additional ATP may be generated by SLP in the terminal
phogluconate aldolase (KDPG aldolase) which converts steps in addition to that during metabolism to pyruvate.
2-keto-3-deoxy-6-phosphogluconate to pyruvate and The ATP generated is either coupled to oxidative reac-
glyceraldehyde-3-phosphate. In the PP pathway the tions or to a lyase reaction [16]. An example of the for-
6-phosphogluconate is converted to the 5-carbon sugar mer is the conversion of 2-oxoglutarate to succinate.
ribulose-5-phosphate and CO2 in a reaction that yields Oxidation of 2-oxoglutarate to succinyl-CoA catalysed by
NADPH, required for biosynthetic reactions. Ribulose-5- 2-oxoglutarate dehydrogenase is followed by conversion
phosphate is then cycled in a series of transaldolase and of succinyl-CoA to succinate catalysed by succinate thio-
transketolase reactions that are both energy- and redox- kinase with ATP generated in the final step. An example
neutral; the benefit to the cell is to yield a variety of sugar of a lyase reaction is the conversion of pyruvate to acetate
phosphates useful for intermediary metabolism and using pryuvate:formate lyase of Pfl (lyases are enzymes
biosynthesis of nucleic acids. Amongst these intermediates that cleave bonds, CaC, CaO, CaN etc. by reactions
glyceraldehyde-3-phosphate and fructose-6-phosphate can other than oxidation or hydrolysis). Pfl produces acetyl-
be tapped off for fermentation. In the PK pathway CoA and formate (methanoate) from pyruvate and the
(hexose monophosphate shunt) the 6-phosphogluconate is acetyl-CoA is converted first to acetyl phosphate by phos-
converted by phosphogluconate dehydrogenase in an photransacetylase and then to acetate via acetate kinase
oxidative decarboxylation reaction to ribulose-5-phosphate yielding ATP.
Some of the major end pathways of fermentation from glucose 1 2ATP 1 2NADH 1 2H1 3
pyruvate are shown in Figure 11.3 and examples of 2 glycerol 1 2ADP 1 2Pi 1 2NAD1
fermentations are given in the subsequent sections.
The same ethanolic fermentation when undertaken by
the bacterium Zymomonas mobilis proceeds via a different
Ethanol Fermentations route (the ED pathway) which gives rise to only l ATP/
This is one of the simplest and best-known reactions in glucose. Genetic engineering of ethanol fermentation in
which one molecule of glucose is fermented to two mole- Z. mobilis for potential industrial application has been
cules of ethanol with release of two molecules of carbon carried out to increase substrate range and to introduce
dioxide. The brewing yeast Saccharomyces cerevisiae alternative pathways for NAD1 regeneration.
utilizes the EMP pathway to pyruvate and yields 2ATP/
glucose. Pyruvate is converted to acetaldehyde by pyru-
vate carboxylase and to ethanol using alcohol dehydro- Lactate Fermentations
genase. This step regenerates NAD1. In wild-type yeast Fermentation to lactate is carried out by many bacteria, in
no lactate fermentation is possible because the lactate this fermentation NAD1 is regenerated by a NAD1-
dehydrogenases are not NAD1-linked enzymes. Even linked lactate dehydrogenase in the conversion of pyru-
under aerobic conditions S. cerevisiae shows partial fer- vate to lactate (Fig. 11.3).
mentative metabolism unless glucose is limiting; this is
known as the Crabtree effect. S. cerevisiae also regener- CH3 COUCOOH 1 NADH1 1 H1
ates NAD1 by other means including fermentation to 3CH3 CHOHUCOOH 1 NAD1
glycerol, though this reaction requires ATP.
Fermentations may be homofermentative or
heterofermentative:
Homolactic fermentation: glucose - 2 lactate
Heterolactic fermentation: glucose - lactate 1 ethanol 1 CO2
Lactose-utilizing bacteria in the dairy industry such as
Lactobacillus bulgaricus and Streptococcus thermophilus
use a homolactic fermentation with lactate as the principal
product that proceeds via the EMP pathway. Other bacte-
ria such as Lactobacillus casei and Leuconostoc spp. are
heterofermentative and utilize the PK pathway to produce
lactate and ethanol. Lactobacilli are also found in the
human gut but comprise a relatively minor component of
the jejunum/ileum and colon at numbers estimated at
106108 CFU/g; they are also found in saliva at
c. 105 CFU/g.
An alternative heterolactic fermentation is used in
Bifidobacteria (Fig. 11.4) such as Bifidobacterium longum
characterized by the enzyme fructose-6-phosphate phospho-
ketolase (F6PPK). This variation is known as the bifid or
fructose-6-phosphate shunt and yields 2.5 ATP/glucose fer-
mented [18]. The overall reaction is: 2 glucose - 2 lactate
1 3 acetate.
Glucose-6-phosphate is not converted to 6-
phosphogluconate as in the ED and PK pathways but
to fructose-6-phosphate which is cleaved by the TFPK
(F6PPK/XFP) enzyme into erythrose-4-phosphate and
FIGURE 11.3 Fermentation pathways showing end products from pyru- acetyl phosphate. Erythrose-4-phosphate and another mol-
vate with regeneration of NAD1. The full lines show the pathways in the
ecule of fructose-6-phosphate can then undergo a similar
mixed-acid fermentation of E. coli and the dotted lines reactions occur-
ring in other bacteria. For clarity acetone and butanol formation have set of transaldolase/transketolase reactions as in the PP
been omitted. Reactions that fix CO2 into organic acids (e.g. conversion pathway leading to xylulose-5-phosphate which is then
of PEP to oxaloacetate) are important to maintain levels of intermediary split to glyceraldehyde-3-phosphate (to lactate) and acetyl
metabolites. phosphate (to acetate) catalysed by xylulose-5-phosphate
FIGURE 11.4 The bifid shunt of Bifidobacterium longum. The dual sub-
strate xylulose-5-P/fructose-6-P phosphoketolase (TFPK) splits fructose-6-
P to erythrose-4-P and acetyl-P and xylulose-5-P to glyceraldehyde-3-P
and acetyl-P. Transketolase and transaldolase reactions convert fructose-6-
P to xylulose-5-P. This provides an alternative route for fructose-6-P for
bacteria lacking the ability to produce fructose-1,6-di-P as in the EMP
pathway.
FIGURE 11.5 Fermentation pathway to butyrate from glucose and amino
acids for human gut bacteria based on genomic data and knowledge of the
pathway in Clostridium kluvyeri. (a) Crotonyl/butyryl-CoA cycle with buty-
rate as end product and (b) coupling of the butyryl-CoA dehydrogenase/
transketolase (Fig. 11.4). Diagrams of such pathways can electron transferring flavoprotein (BCD/ETF) complex to production of a
now be produced from genomic data for any sequenced proton motive force. The BCD/ETF complex utilizes NADH for the con-
bacterium using systems such as KEGG. comitant conversion of crotonyl-CoA to butyryl-CoA and the production of
reduced ferredoxin (FdR). Reduced ferredoxin can be reoxidized to FdO via
two pathways: production of hydrogen by a soluble ferredoxin-dependent
hydrogenase or H1 translocation via the membrane-bound NADH:ferre-
Mixed Acid Fermentations doxin oxidoreductase complex (Rnf complex). The distribution between
Mixed acid fermentations occur when bacteria utilize two these pathways depends on several factors including NAD1 availability and
utilization of H2 by other organisms. Adapted from [19,20].
or more different pathways in the terminal steps of fer-
mentation. Enteric Gammaproteobacteria like E. coli
and Enterobacter aerogenes are facultative anaerobes and
so use aerobic respiration in the presence of oxygen and two- to three-fold more lactic, succinic and acetic acid than
anaerobic respiration if there are suitable alternative does Enterobacter aerogenes. At low pH E. coli converts
electron acceptors such as nitrate, fumarate and TMAO formate to CO2 1 H2 using the enzyme formate hydrogen
that they are able to utilize. In the absence of any such lyase (Fhl) system that not all Enterobacteriaceae possess.
acceptor then these bacteria will ferment glucose to a mix- This system consists of hydrogenase 3 and formate
ture of acetate, formate, lactate and succinate acids plus dehydrogenase-H.
ethanol as shown in Figure 11.3. The relative amounts of Clostridial species carry out mixed acid fermentations
each product may vary with growth conditions and with in which butyrate is one of the products (Figs 11.3,
bacterial species involved. For example E. coli produces 11.5a). For example Clostridium butyricum ferments
glucose producing butyrate, acetate, carbon dioxide and carbon should match the fermentation product carbon and
hydrogen with the following stoichiometry: each product is measured experimentally to give a carbon
balance. Equally, because there is no net oxidation/reduc-
glucose - 0:4 acetate 1 0:8 butyrate 1 0:4 acetate tion in fermentation, the amount of oxidized product
1 2:0CO2 1 2:4H2 should be equal to the amount of reduced carbon. In bio-
logical terms the fermentation equation above can be
The ratio of products varies with bacterial strain, interpreted as almost equal amounts of the glucose being
growth phase, pH and partial pressure of hydrogen. metabolized by the two pathways to lactate and ethanol.
The EMP pathway is used for conversion of glucose The small amounts diverted to acetate and ethanol proba-
to pyruvate. The branch to acetate utilizes phosphotrans- bly enable a small amount of extra ATP to be generated
acetylase and acetate kinase while the branch to butyrate by SLP and some regeneration of NAD1. The latter has
involves phosphotransbutyrylase and butyrate kinase. The been shown to be important in yeast.
equations for the two branches are: In medical microbiology these terms are not com-
monly used but the concepts are useful in dealing with
glucose - 2 acetate 1 4H2 1 2CO2 1 4ATP
the fact that microbial consortia will produce varying fer-
glucose - butyrate 1 2H2 1 2CO2 1 3ATP mentation products dependent on microbiota, nutritional
feedstock and environmental conditions, e.g. butyrate
Thus energy yield is higher for acetate production and
metabolism in the human gut.
during exponential growth this is the major pathway.
At the end of exponential growth, the organisms
slow down acetate production and take up excreted ace- Butyrate Metabolism in the Human Gut
tate, converting it into butyrate. Other ‘solventogenic’
In natural environments there are multiple species of bacte-
Clostridia such as C. acetobutylicum have a biphasic
ria coexisting in the same environment. Metagenomic stud-
metabolism during growth. In the first ‘acidogenic’ phase
ies of the human microbiome [23] have shown over 1000
the mixed acid fermentation described above occurs but
bacterial species are capable of colonizing the human gut
below pH 5 there is a switch to a solventogenic phase
and that each individual harbours .160 species. The com-
in which acetone, butanol and ethanol are produced.
bined genetic pool provides an interconnecting set of meta-
The pathways to acetone and butanol both involve forma-
bolic pathways. In terms of energy metabolism bacteria
tion of acetoacetyl-CoA from acetyl-CoA; metabolism of
ferment complex carbohydrates and plant polysaccharides,
acetoacetyl-CoA to acetolactate yields acetone whilst the
including cellulose and xylans to short-chain fatty acids:
pathway to butyrate proceeds via butyryl-CoA. This is
acetate, proprionate and butyrate in particular. The major
discussed further in the section on fermentation in the
microbially produced short-chain fatty acids detected in
human gut as the majority of the microbial enzymes are
ileal effluents were acetate, propionate and butyrate in the
homologous.
molar proportions of 20:1:4.
Butyrate, butanol and acetone are important industri-
Butyrate production in the colon is important as it is
ally and are covered well in recent reviews [21,22].
readily transferred across the epithelial cell wall and used
From the discussion above on multiple pathways in
by colonocytes for energy. Butyrate producers in the gut
lactate and butyrate formation with varying ratios of pro-
are strict anaerobes and use the pathways shown in
ducts it is clear that the stoichiometries of the overall con-
Fig. 11.5a.
version are non-integral, though of course the individual
There is recent evidence that in addition to SLP,
biochemical pathways have integer numbers:
energy may be derived by generation of a pmf coupled to
e:g: 1 glucose - 1 lactate 1 1 ethanol 1 1CO2 conversion of crotonyl-CoA to butyryl-CoA during the
butyrate fermentation in the human gut and this system
1 glucose - 2 acetate 1 2CO2 is shared between phylogenetically distinct groups [19].
This mechanism was first discovered in Clostridium kluy-
In the fermentation of Lactobacillus pentoaceticus the
veri [20]. C. kluyveri fermentation on ethanol and acetate
stoichiometry as measured by product formation is [17]:
leads to formation of butyrate, caproate and H2 as
1 glucose - 0:96 lactate 1 0:86 ethanol 1 0:89CO2 products with energy derived by SLP. Conversion of etha-
1 0:07 acetate 1 0:07 glycerol nol or acetate to acetyl-CoA is followed by a cyclic con-
version to acetyl-CoA, hydroxybutyryl-CoA, crotonyl-
So in practice fermentation scientists use the concept CoA, butyryl-CoA and back to acetyl-CoA (Fig. 11.5a).
of carbon balances and redox balances. In fermentation, The conversion of crotonyl-CoA to butyryl-CoA via
because energy yield is low, the majority of carbon is the cytoplasmic butyryl-CoA dehydrogenase (BCD com-
converted to product rather than cell mass; so the input plex) is energetically very favourable and essentially
unidirectional and can be used to create additional energy Xylan Utilization

as a pmf. The BCD reaction is coupled to an electron-
transferring flavoprotein (ETF) which both regenerates Hemicelluloses are plant cell wall polysaccharides with
NAD1 from NADH and reduces ferredoxin (Fd), an β1-4-linked saccharides and include xylans, xyloglu-
ironsulphur protein used in electron transfers at low cans, mannans and glucomannans. In xylans the β1-4-
potential to produce reduced ferredoxin: FdH2. This is linked xylose backbone may have substitutions, e.g.
coupled to a proton-translocating NADH-ferredoxin oxi- arabinofuranose at O2, acetyl groups at O3, ferulate
doreductase complex (RnfA-E) as shown in Fig. 11.5b. esters at O5. The variations in substituent group and
The genes for the BCD and ETF complexes and the position of substitution give a wide variety of xylans
pathway from acetyl-CoA to butyryl-CoA are common to [27]. Moreover there may be cross-linking with other
all human colonic butyrate producers (which are Gram- hemicelluloses. Hence microbial degradation of xylans
positive Firmicutes, but phylogenetically diverse) though requires a repertoire of enzymes to cleave both substi-
the gene arrangements differ [19]. Thus it seems this tuents and to degrade the xylose backbone; these
additional mechanism of energy generation is present in include endoxylanases, β-xylosidases, α-arabinofurano-
bacteria that are taxonomically quite different. sidases and ferulic acid esterases [28]. It has been esti-
mated that c. 5070% of xylans in the human gut are
degraded by the microbiota with Bacteroides spp.
UTILIZATION OF COMPLEX being frequently isolated xylan utilizers. Genome anal-
CARBOHYDRATES ysis of B. thetaiotaomicron showed that there were
multiple homologues (c. 90) of the susCD outer-
The first step in the fermentation of polysaccharides is membrane transport system involved in starch metabo-
their breakdown into disaccharides and monosaccharides, lism often present in large gene clusters called PULs
which are then further metabolized by the intermediary for polysaccharide utilization locus with the likelihood
glycolytic pathways (EMP, PP, ED and PK). The main some were associated with xylan degradation [26,29].
polysaccharides utilized by bacteria are plant: starch, pec- For the rumen bacterium Prevotella bryantii transcrip-
tin, hemicelluloses, celluloses and xylans. Thus bacteria- tomic studies have identified six sus-like genes in a
degrading plant materials, whether in the environment, in cluster that were up-regulated by growth on wheat ara-
the animal rumen or in human gut, require an array of binoglycan; this cluster has been named xus for xylan
enzymes and the ability to degrade these substrates in a utilization system [30]. Further work in this area is
complex insoluble matrix. The number of genes encoding likely to identify specific xus clusters in human gut
glycoside hydrolases (GH) reflects this: the celluloytic bacteria.
bacterium Fibrobacter succinogenes encodes 104 GHs
whilst the xylan- and starch-utilizing Bacteroides thetaio-
taomicron encodes 246 [24]. Cellulose Degradation
Aerobic cellulose-degrading bacteria characteristically use
a free cellulase mechanism in which multiple secreted
Starch Utilization enzymes act synergistically. Anaerobic cellulose degra-
Starch, a mixture of straight-chain amylose molecules and ders, which account for 510% of all cellulose decompo-
amylopectin molecules, is broken down relatively easily sition, are known to have two mechanisms for cellulose
by many bacteria. Bacteroides spp. encode an 8 gene decomposition. The first is the use of a cellulosome com-
termed starch utilization system (Sus) that includes cell plex of cellulolytic enzymes, first characterized in clos-
envelope carbohydrate-binding domains (CBMs), most tridial species but also shown to occur in the human gut
studied in B. thetaiotaomicron. SusE and SusF are outer- bacterium Ruminococcus flavefaciens in a more complex
membrane lipoproteins that bind starch using tandem organization [24]. The cellulosome complex is attached to
CBMs [25,26]. SusG encodes an amylase but also has the outer cell envelope of the bacterium and contains
CBM domains; with an α1-4 glycosidase activity that acts proteins with cohesion domains (two types) that bind
on amylose, amylopectin and pullulan. Current models strongly to the dockerin domains of the cellulolytic
suggest that SusE, SusF, SusG and SusD act in concert enzymes. In addition the scaffoldin often has a separate
and oligosaccharides produced by action of SusG enter cellulose-binding domain (CBM) for binding to the celul-
the periplasm by the outer-membrane porin, SusC. SusA losic substrate. The second is the system used by
is a periplasmic neopullulanase and is assumed to be Fibrobacter succinogenes (formerly Bacteroides succino-
responsible for hydrolysing oligosaccharides in the peri- genes) in which cellulose is attached to the outer mem-
plasm. Mutant studies have shown that susC, susD and brane through adhesins, e.g. fibro-slime proteins and
susG are essential for growth on starch. possibly type IV pilins.
AEROBIC RESPIRATION (a) CH=CH2

CH3
CH3
–
CH2–(CH2–CH=C–CH2)3–H
–
Components of Respiratory Chains H3C CH=CH2
HO–CH CH3
N N
Most eukaryotes, including humans and yeasts, only use Fe H3C CH=CH2
N N N N
aerobic respiration and have a ‘classical’ respiratory chain H3C CH3 Fe
N N
consisting of four protein complexes (IIV). The entry H3 C CH3
CH2 CH2
points to this chain are NADH and FADH2 generated by CH2 CH2
CH2 CH2
the metabolism of sugars, carboxylic acids and amino C=O C=O
CH2 CH2
acids through the tricarboxylic acid cycle. The respiratory OH OH
C=O C=O
chain from NADH to oxygen utilizes a number of redox OH OH
Haem b
components that carry electrons either just as electrons Haem o
(e2) or as bound hydrogen (H1 1 e2). Cytochromes and
CH3
ironsulphur proteins carry electrons whilst flavoproteins
–
CH2–(CH2–CH=C–CH2)3–H
–
have bound FAD or FMN, which is reduced to FADH2 or HO–CH CH3
FMNH2. A key mobile hydrogen carrier under aerobic O
HOOC-CH2 CH3 COOH H3C CH=CH2
conditions is ubiquinone (UQ), a lipid molecule with iso- CH2
N N
O Fe
prenoid side chains, which is reduced to ubiquinol N N CH3
N N
(UQH2 2H is equivalent to 2H 1 2e2); some anaerobic
Fe
N N HC CH3
=
H3C CH3 O
respiratory chains utilize menaquinone (MQ), methyl- CH2 CH2
CH2 CH2
menaquinone (MMQ) or demethylmenaquinone (DMK). CH2 CH2
CH2 CH2
C=O C=O
The ironsulphur proteins contain ironsulphur clusters C=O C=O OH OH
OH OH
that are linked to the apoprotein via Cys residues;
[2Fea2S] and [4Fea4S] are the most common but Cys Haem d1 Haem a
[3Fea4S] also occurs. Cytochromes are electron transfer CH3 Cys

S
proteins containing one or more haem groups based on a CH3 (b) L1
S N N
tetrapyrrole structure; the haem group determines both the H3C
N N Fe Low spin
type of cytochrome and the main component of the CH3
Fe N N
absorption spectrum. Due to the importance of cyto- N N L2
H3C CH3
chromes as intermediates in respiratory pathways and
their role as components of cytochrome oxidases some CH2 CH2 L1
CH2 CH2 N N
background on their structures is provided. C=O C=O High spin
Fe
The central atom of the haem ring in cytochromes is OH OH
N N
Fe and electron transfer results because of a single elec- H2O
tron transfer between the Fe(II) and Fe(III) states. Haem c
Cytochromes are grouped into four major groups (a, b, c

FIGURE 11.6 Cytochrome classification based principally on the
and d) based on the nature of the haem group and the structure of the haem group: (a) structures for haem b, haem o, haem d1
presence or absence of a covalent bond between the pro- and haem a with differences in the peripheral chemical groups circled
tein and the haem group (Fig. 11.6); experimentally the and haem c with covalent attachment to the protein via Cys residues;
absorption spectra of cytochromes are useful in determin- (b) coordination to Fe in high-spin and low-spin haem groups.
ing the group as the alpha peaks occur in distinct regions
(a 580590 nm, b 556558 nm, c 549553 nm and ligands (5th and 6th ligand). In low spin the 5th and 6th
d 600620 nm). Cytochromes a have the haem a as pros- ligands are ‘strong axial’ ligands such as His or Met and
thetic group characterized by a long isoprenoid side chain not easily displaced. For high-spin cytochromes the sixth
on one of the peripheral vinyl groups. Cytochromes b ligand is ‘weak’, e.g. water, and can be displaced
contain protohaem and cytochromes d have more satu- (Fig. 11.6). This enables ligands such as O2, to bind
rated variants. Cytochromes c contain a covalent link important for oxidases (also NO or CO). A high-spin
between the protein and the haem through one or two cytochrome b should be referred to as cytochrome b0 but
thioether linkages. the practice of utilizing the term cytochrome o still per-
Nomenclature of cytochromes, particularly cyto- sists and is widely used. Cytochromes may be involved in
chrome oxidases, can be confusing as it is influenced by electron transport chains as intermediary electron carriers,
custom and the history of discovery. The Fe atom in as components of oxidases in the final complex of an
haems is coordinated by four nitrogen atoms from each of anaerobic respiratory chain or as part of an enzyme, e.g.
the pyrrole rings leaving two available spaces for further lactate dehydrogenase or fumarate reductase complex.
E. COLI
H+ translocating Yes Yes No
Oxygen affinity Low High High
Cytochrome bo3 Cytochrome bd-l Cytochrome bd-Il

oxidase oxidase oxidase
(Complex I)
NADH NADH dehydrogenase
NDH-I
UQ8/UQH2
Succinate Succinate dehydrogenase pool
SDH
(Complex II)
bc1 complex (Complex III)
Cytochrome c
pool
Cytochrome ba3 Cytochrome aa3 Cytochrome cbb3

oxidase oxidase oxidase
H+ translocating Yes Yes Yes

Oxygen affinity Low Low High
P. DENITRIFICANS
FIGURE 11.7 Aerobic respiratory chains of (a) Escherichia coli and (b) Paracoccus denitrificans in upper and lower sections respectively, showing
oxygen affinities of the terminal oxidases. Other dehydrogenases connect to the electron transport chain through the quinol/quinone pool or via
NADH for NADH-dependent dehydrogenases like malate dehydrogenase. Both NDH-1 and succinate dehydrogenase are proton translocating; E. coli
has a second non-proton translocating NADH dehydrogenase, NDH-2 (not shown). Low affinity for oxygen for P. denitrificans cytochrome ba3 oxi-
dase is based on it being a class A haemcopper oxidase.
These respiratory components are assembled into com- the mitochondrial electron transport chain though the
plexes. For example the aerobic mitochondrial respiratory number of components is fewer (Fig. 11.7a). However,
chain comprises Complex I (NADH dehydrogenase), bacteria can have alternative aerobic respiratory chains as
Complex II (succinate dehydrogenase) and glycerol-3- shown for E. coli in Fig. 11.7b. Two important points
phosphate dehydrogenases and ETF-ubiquinone oxidore- need to be made as this stage. First, evidence from native
ductase all feeding into the ubiquinone/ubiquinol pool polyacrylamide gel electrophoresis of solubilized mem-
and thence to Complex III (cytochrome bc1 complex). branes suggests that mitochondrial, E. coli and other bac-
Soluble cytochrome c provides the link between Complex terial respiratory chains are arranged in supercomplexes
III and Complex IV (cytochrome caa3 oxidase) where [31]. Secondly, there are points in the respiratory chain
oxygen is reduced to water. The aerobic electron transport where net translocation of protons across the membranes
chain of the bacterium Paracoccus denitrificans to cyto- occurs to create a pmf (see section on chemiosmosis). In
chrome caa3 oxidase has the same basic organization as the mitochondrial electron transport chain this occurs at
Complex I, Complex III and Complex IV and much centre for O2 reduction with haem a, haem a3 and CuB,
recent work has been concentrated on the channels by arranged so electrons are transferred singly from cyto-
which protons enter and exit during this process. chrome c through subunit II which contains a CuA centre
Utilization of this pmf by the proton-translocating (Figs 11.8, 11.9a) with the overall reaction being the four
ATPase creates the ATP for cellular metabolism. electron reduction of O2 to H2O:
This brief description of aerobic respiration only out-
lines the nature of respiratory carriers, their organization 4 cytc21 1 4H1 1 O2 - H2 O 1 4 cytc31
in the membrane and the associated proton translocation. There is a thermodynamic barrier to O2 reduction which
For more details the reader is referred to some excellent is overcome by using both CuB and the Fe atom of haem a3.
texts and reviews (overall [4,32]; ironsulphur proteins Two electrons convert Cu(II) to Cu (I) and Fe (III) to Fe (II)
[33,34]; quinones [35]; proton pumping [5,6]). The focus allowing O2 to bridge between the Cu (I) and Fe (II) atoms.
here is on cytochrome oxidases because of their impor- The third electron is used to break the OaO bond and form
tance in the growth and survival of pathogenic bacteria in a transient oxygenated intermediate, termed PM, followed
and on host tissue. by a fourth electron to release the remaining O bound to
haem a3 as an Fe(IV)aO22 complex [3]. The overall pro-
Cytochrome Oxidases cess requires protons both for the series of reactions and for
proton pumping. These are supplied through two distinct
The availability of oxygen to bacteria can vary widely so proton-conducting channels, termed the D and K channels,
many bacteria have evolved to have cytochrome oxidases that permit uptake of protons from the N-side of the mem-
that allow growth both in aerobic and microaerobic brane and their delivery to the redox centres (Fig. 11.8);
environments. Oxidases with a low affinity for oxygen comparative analysis allows these to be traced by the pat-
(Km c. 200 nM) but high respiratory capacity are tern of conserved polar amino acid residues within the pro-
suitable for aerobic growth whereas high-affinity oxidases tein. The D pathway transfers most of the protons, both for
have low Km values (210 nM). Recent evidence sug- oxygen reduction and pumping, and the K pathway is used
gests that the high-affinity oxidases that are proton trans- only for protons that are taken up during reduction of the
locating generate a pmf with half the efficiency of the active site. Further details of these channels and
low-affinity oxidases (0.51.0 H1/e2). This has implica- proton pumping models can be found in the following
tions for the mechanisms of proton pumping. The cyto- references [3,5,6].
chrome oxidases fall into two major classes: those that
are in the haemcopper oxidase (HCO) superfamily and
those that are quinol oxidases.
Cytochrome cbb3 Oxidases
The subgrouping of HCOs into A, B and C families Cytochrome cbb3 oxidases were first identified in the
based on structural and phylogenetic evidence is mirrored nitrogen-fixing bacterium Bradyrhizobium japonicum, but
by their functional roles and oxygen affinities. The A have since been found in other environmental bacteria
family, e.g. cytochrome caa3 oxidase, has low oxygen which can grow in microaerobic environments such as
affinity and is used normally in aerobic conditions, Paracoccus denitrificans and the phototroph Rhodobacter
whereas cytochrome cbb3 from the C family and cyto- sphaeroides. They have also been found in the pathogens
chrome ba3 of the extremophile Thermus thermophilus Campylobacter jejuni, Helicobacter pylori, Neisseria
from the B family are preferred under microaerobic con- meningitidis and in Pseudomonas spp. Where measured
ditions [36]. For the quinol oxidases cytochrome bo3 is a these oxidases have been shown to have a high affinity
low-affinity oxidase that is proton translocating whereas for oxygen.
cytochrome bd is a high-affinity non-proton-translocating The cbb3 complex belongs to the C family of copper-
oxidase. In the next section the properties of the different containing oxidases and shares structural similarities with
cytochrome oxidases are described. caa3 oxidases, which are in the A family (Fig. 11.9).
It consists of four subunits encoded usually in an operon
cco(fix)NOQP. Subunit I (CcoN) is the catalytic subunit
Cytochrome caa3 Oxidases
and contains both high-spin (b3) and low-spin haem (b).
Mitochondrial cytochrome c oxidase, also known as cyto- The low-spin haem donates electrons to the binuclear
chrome caa3 oxidase, is a member of a large superfamily active site which contains high-spin haem and a copper
called the haemcopper oxidase (HCO) superfamily. centre, CuB (Fig. 11.9c). Subunit II (CcoO) is a
Early experimental work concentrated on the bovine heart membrane-anchored protein but has a soluble c-type cyto-
complex, which has 13 subunits. However the ‘business chrome domain; unlike the corresponding caa3 oxidase
end’ of cytochrome caa3 oxidase is composed of two sub- subunit it lacks a CuA centre. Subunit III (CcoP) contains
units: subunit I and II. Subunit I contains the catalytic two c-type haems in two separate soluble domains, whilst
FIGURE 11.8 Comparison of crystal structures of cytochrome caa3 and bo3 from P. denitrificans and E. coli respectively. Left panel is the bo3 qui-
nol oxidase normal (top) and parallel (bottom) to the membrane. Right panel corresponding views for cytochrome caa3. Electron donating substrates
are shown as a dotted circle. For both oxidases subunits I, II, III and IV are shown in yellow/green, green, blue and pink, respectively. Haems b and
o3 are shown in red and blue respectively. The blue spheres in subunits I and II are the CuB and CuA centres respectively. Reprinted by permission
from MacMillan Publishers Ltd: Nature [33], copyright (2000).
subunit IV (CcoQ) is postulated to stabilize the interac- assembly of subunits IIV and the incorporation of sub-
tions of subunit III with the core complex of subunits I units into the membrane (reviewed in [38] for Rodobacter
and II and thus prevent proteolytic degradation; it is not capsulatus). For several bacteria (R. capsulatus, R. sphaer-
present in all cytochrome cbb3 oxidases. The crystal struc- oides, P. denitrificans, B. japonicum) it has been shown
ture of cbb3 is known for the P. stutzeri complex [37] and that the ccoGHIS operon is located downstream of the
this has enabled the electron transfer pathway to be pre- ccoNOQP operon and is involved in correct cbb3 assembly.
dicted. Electrons from a soluble c-type cytochrome are Blue native polyacrylamide gel electrophoresis (BN-
transferred to the first (outer) haem c centres in subunit PAGE) showed that the active complex in R. capsulatus is
III and then to the second inner centre which forms an a 230-kDa complex which contains subunits IIV but that
interface with the core complex. So electrons can be CcoH was associated with intermediate complexes and that
transferred on via the haem c in subunit II to haem b in haem c’s were incorporated into CcoO and CcoP prior to
subunit I and thence to the binuclear centre (b3CuB) in subunit assembly. Mutant studies are being used to eluci-
subunit I. In contrast to cytochrome caa3 and bo3 there is date the genes involved in cbb3 assembly but at the time of
only one proton channel for catalysis and proton translo- writing the process is only partially understood. Mutants in
cation (the K channel) and there is no pattern of con- ccoS are defective in haem b and haem b3 incorporation
served residues that could form a D channel (Fig. 11.9). into CcoN whilst mutants in ccoI, which encodes a P1B-
Other genes are involved in cbb3 assembly for the type ATPase (normally involved in Cu1 efflux), show very
uptake and incorporation of copper into subunit I, the low levels of cbb3.
FIGURE 11.9 Comparison of redox and catalytic

centres in cytochrome oxidases showing subunits I and
II. Cytochrome caa3, cbb3 and the quinol oxidase bo3
all belong to the haemcopper oxidase family with
CuB forming part of a haemcopper catalytic centre.
Cytochrome bd-I has no copper centre but a haem b
in subunit I and haem d1 and a second haem b proba-
bly at the interface between subunits I and II. Dotted
lines represent the region for D and K proton channels.
The D pathway starts at a D residue and ends near the
CuBaFe bimetallic centre. The K pathway starts at a
K residue and ends below the high-spin haem at the
catalytic centre. aOa, haem; closed circle, CuA; open
circles, CuB.
with the use of reporter gene fusions such as alkaline phos-

Cytochrome bo3 Oxidase
phatase, β-galactosidase and β-lactamase. The CyoA,
Cytochrome bo3 is a membrane-bound ubiquinol oxidase CyoB, CyoC, CyoD and CyoE subunits were predicted to
consisting of four subunits, which are encoded by the have 2, 15, 5, 3 and 7 TMHs, respectively [39]. CyoA has a
cyoABCDE operon. It has been most studied in E. coli but large periplasmic loop and a small cytoplasmic loop
is widely found in the Proteobacteria with over 60% of whereas the topology of CyoB indicated a hydrophobic
the reported sequences found in the Alpha-, Beta- and integral membrane protein. These findings have been sub-
Gammaproteobacteria. Selected examples of bacteria con- stantiated by the crystal structure obtained for the
taining this oxidase are Citrobacter freundii, Klebsiella CyoABCD complex; crystallization was achieved in a
pneumoniae, P. aeruginosa and Shigella sonnei. The four buffer containing 1% (w/v) of the non-ionic detergent octyl
structural subunits assemble in a 1:1:1:1 stoichiometry to glucoside [40] (Fig. 11.8). Assembly of the cytochrome oxi-
form a functional complex (Figs 11.8, 11.9). Subunit I is dase subunits into a functional complex requires an ordered
the main catalytic component and contains the redox cen- pathway. The order of assembly appears to be subunits III
tres: haem b, haem o3 and CuB (for historical reasons, the and IV first, followed by subunit I and finally subunit II.
oxygen-binding haem in the o3aCuB binuclear centre is
designated with a subscript ‘3’). Oxygen binds at the
binuclear centre and is reduced to water. The electron
Cytochrome bd Oxidase
transport pathway can be envisaged as: Cytochrome bd oxidase, like cytochrome bo oxidase, is a
quinol oxidase but has a higher affinity for oxygen and is
QLH2 - QH - haem b - haem o3 -CuB - O2
expressed preferentially under microaerobic conditions
Ubiquinol (UQH2) binds to cytochrome bo3 at two [41]. It is found in many bacteria, including E. coli,
sites, a low-affinity site QL and a high-affinity site QH, Shigella flexneri, many Staphylococcus spp. and Bacillus
where it is tightly bound and which transfers electrons to spp., and in pathogens such as Brucella suis,
haem b. Based on evidence from mutagenesis studies in Mycobacterium tuberculosis, and even in anaerobes like
which ubiquinone binding was lost on changing key resi- Bacteroides fragilis. Cytochrome bd oxidase has several
dues the QH site lies in subunit I close to haem b. During distinct properties: it is not part of the haemcopper super-
the reduction of UQH2 protons are translocated across the family of oxidases, it contains haem d rather than CuB
cytoplasmic membrane creating a pmf (see section on che- at the oxygen-binding site, and it shows cyanide resistance.
miosmosis and proton pumping). The crystal structure indi- The enzyme comprises two subunits CydA (subunit I) and
cates two proton-pumping channels, D and K, similar to CydB (subunit II) in a heterodimeric complex. There are
those found in cytochrome caa3 oxidase [33] (Fig. 11.9). three haem groups: b558, b595 and d. The overall reaction is
Due to difficulties in getting crystals of membrane- 2QH2 1 O2 - 2Q 1 2H2O with transferring two electrons
bound proteins, early work on the arrangement of mem- from each molecule of QH2 the electron transport pathway
brane subunits in membranes was done by a combination of QH2 - b558 - b595 1 d - O2. In the absence of a crystal
predicting TMHs from hydropathy plots in combination structure models for the structure are based on topology
studies and knowledge of haem-binding ligands. The mem- metabolic regulation; in the presence of oxygen the genes
brane topology of E. coli cytochrome bd oxidase was encoding denitrifying enzymes are repressed. In addition
determined by β-lactamase fusions. On the basis of this nitrous oxide reductase is sensitive to the presence of
evidence subunit I has nine TMHs with all three haems molecular oxygen, which causes inactivation of the enzyme
bound on the periplasmic side of the membrane with the by altering the ligand structure at the catalytic copper centre
low-spin haem b558 binding to two ligands (His186 and Cuz. At low oxygen tensions both aerobic respiration and
Met393) and the high-spin haem b595 to one His ligand denitrification may coexist and evidence for ‘aerobic deni-
(His19). The exact locations of haems b595 and d are not trification’ has been reported in a number of studies. The
known but they may bind in the interface between subunits two processes share the first part of the electron transport
I and II (Fig. 11.8d); spectroscopic studies show haem d chain from NADH to ubiquinol but measurement of respira-
binds O2 with a high affinity. There is also a periplasmic tion rates indicates the major flow is to oxygen.
Q-loop between TMH6 and TMH7, so-called because it is
thought that QH2 binds in this region as shown by
The Reductases Involved in Denitrification
antibody-binding studies and trypsin digestion experiments
which inhibited ubiquinol oxidase activity. The length of Details of the electron transport chains, structure of the ter-
the Q-loop can either be short or long. minal reductases and bioenergetics involved in denitrifica-
The oxidation of quinol at the periplasmic side of the tion have been studied principally in three bacteria:
membrane releases protons to the outside (P-side) and for P. denitrificans, P. aeruginosa and P. stutzeri. With the
proton translocation this must be matched by uptake of availability of bacterial genome sequences the presence of
protons from the cytoplasmic (N-side) and implies a denitrifying pathways in a much wider range of bacteria
proton-conducting channel from the N-side to the catalytic can be deduced by sequence comparison but this requires
site. Flash photolysis experiments at low temperature have physiological, biochemical and genetic studies to sub-
been used to show that reduction of oxygen at the catalytic stantiate, a process now underway for several bacteria, e.g.
site yields a membrane potential. In these experiments CO for the photosynthetic bacterium Roseobacter denitrificans.
is first bound to haem d and then released by flash photoly-
sis under conditions where oxygen can then bind. Nitrate Reductase Nar and Nitrate/Nitrite
Transporter NarK
ANAEROBIC RESPIRATION The nitrate reductase complex of P. denitrificans and
In the absence of oxygen and in the presence of an alterna- P. aeruginosa consists of three subunits αβγ (NarGHI)
tive electron acceptor the process of anaerobic respiration with the site of nitrate reduction on the periplasmic side.
allows transfer of reducing equivalents to the final electron The complex works in tandem with a NarK transporter
acceptor and the regeneration of NAD1 from NADH. In which imports nitrate and exports nitrite to the periplas-
anaerobic environments, which are strongly reducing, alter- mic side. The crystal structure of the nitrate/nitrite trans-
native electron acceptors are scarce; bacteria have adapted porter NarK has been solved [42]. NarK is composed of
to use a number of such acceptors including nitrate, two domains each of which has six TMHs and the
TMAO, DMSO, fumarate and sulphate. An emphasis is put substrate-binding domain is formed between the two
here on denitrification. domains such that four helices from each domain contrib-
ute to it. Two of these helices, TMH5 and TMH11, con-
tain glycine-rich signature sequences only found in the
Denitrification family of nitrate transporters. Two conserved arginine
In denitrifying bacteria, which reduce nitrate to nitrous residues (in TMH2 and TMH8) contribute to substrate
oxide and molecular nitrogen, the intermediates also act binding and are essential for transporter function.
as anaerobic electron acceptors (valency in brackets): Nitrate reduction by NarGHI involves the transfer of
electrons from ubiquinol through NarI and NarH to the
NO2 2
3 ð15Þ - NO2 ð13Þ - NOð12Þ - N2 Oð11Þ - N2 ð0Þ catalytic subunit NarG, which contains the molybdenum
Nitrate nitrite nitric oxide nitrous oxide dinitrogen cofactor bisMGD: UQH2 - 2H1 1 2e2. The two protons
are released at the periplasmic face and the two electrons
Thus denitrification from nitrate to nitrogen requires are consumed at the cytoplasmic face (Fig. 11.10) so that
four electron transport pathways to nitrate reductase, to there is a net translocation of positive charge to the
nitrite reductase, to nitric oxide reductase and to nitrous exterior, i.e. an electrogenic process that creates a pmf.
oxide reductase. Most denitrifiers also utilize aerobic respi- The composition of the NarGHI complex and the associ-
ration to obtain energy but the processes of aerobic respira- ated cofactors is as follows; NarI is a small dihaem pro-
tion and denitrification are normally distinct because of tein which accepts electrons from UQH2 and donates
them to the NarH subunit which is a larger ironsulphur either pseudoazurin (a blue copper protein) or cytochrome
protein containing three [4Fea4S] and one [3Fea4S] c550 can act as intermediates in the electron transfer from
centres. Finally electrons are passed from NarH to the the cytochrome bc1 complex to the cytochrome cd1 nitrite
molybdenum cofactor in NarG. To date the only available reductase. A double mutant deficient in both pseudoazurin
crystal structure for a membrane-bound bacterial nitrate and cytochrome c550 grew poorly in anaerobic conditions
reductase is that from E. coli, a non-denitrifying bacte- with nitrate as electron acceptor and nitrite accumulated
rium (see [43]). For E. coli NarGHI the crystal structure in the medium.
allows the probable electron transport pathway through NirK is the copper-containing Nir; each subunit of the
the redox centres of NarI, NarG and NarH to be deduced, trimer has one type I Cu centre and one type II copper
the total distance across the eight redox centres is 75 Å centre. Type I Cu centres confer a blue-green colour
and the typical distances between individual centres is to proteins and typically but not invariably have three
c. 13 Å, which is compatible with the known distances for ligands: one Cys and two His residues; they are only
electron transfer in electron transport proteins. The simi- involved in single electron transfer reactions. Type II cop-
larities in subunit and cofactor compositions between per centres are colourless with three to four ligands,
E. coli NarGHI and the NarGHI complexes of denitrifying where one site is available to an exogenous ligand such as
bacteria make it reasonable to conclude that the pathways oxygen or nitrite [44]. In NirK the type I Cu site has four
of electron transport are similar [43]. ligands and acts as the intermediate electron transfer cen-
tre between a donor protein (e.g. azurin) and the type II
site, which is located between two monomers, where
Nitrite Reductases NO2 2 is reduced to NO. The type II site in NirK has three
Nitrite reductases in denitrifying bacteria are of two types, His ligands with the fourth occupied by a water molecule.
either cytochrome cd1 or copper nitrite reductases, often The two sites are linked by a Cys-His bridge and are close
designated cd1NIR and CuNIR (naming analogous to NOR enough for rapid electron transfer. If switched to aerobic
for nitric oxide reductase). For example, cytochrome cd1 is conditions O2 (an analogue of NO) can bind and be
found in P. denitrificans, P. aeruginosa and P. stutzeri reduced to H2O2 causing inactivation of NirK through fur-
whereas a copper nitrite reductase is found in many ther production of reactive oxygen species (ROS).
Neisseria spp. including N. gonorrhoeae and N. meningiti- The above discussion centres on nitrite reduction in
dis, Rhodobacter sphaeroides and Moraxella catarrhalis. denitrification. However, dissimilatory nitrite reductases in
1
Cytochrome cd1 is encoded by the nirS gene and func- which NO2 2 is reduced directly to NH4 , a 6-electron,
tions as a dimer with each subunit containing one haem c reduction, is potentially very useful for bacteria as the elec-
and one haem d1. The properties of haem d1, a haem with tron sink allows regeneration of NAD1 from NADH. Two
modified A and B tetrapyrrole rings, enable the binding classes of enzymes catalyse this reaction: a cytoplasmic sir-
of both NO2 ohaem nitrite reductase (NirB or NirB-D) and a periplasmic
2 and NO to the reduced haem d1 [4]. In
P. denitrificans mutants studies have established that Nrf complex (nitrite reduction by formate). Two types of
Nrf complex have been characterized: NrfA-G as in E. coli
and NrfHAIJ as in W. succinogenes, an epsilon proteobac-
terium. A nrfHA gene order is also found in the related
organism C. jejuni (see [45]) and there is evidence that
NrfA may play a role in relief of nitrosative stress as the
enzyme can reduce the nitric oxide radical NO to NH41.
G
The NrfA subunit is a pentahaem cytochrome a. In the δ-

and ε-Proteobacteria the electron donor for NrfA is NrfH, a
tetrahaem cytochrome whereas in the γ-Proteobacteria the
donor is NrfB, a pentahaem cytochrome. For further infor-
mation on Nrf complexes see [45,46].
Nitric Oxide Reductases

FIGURE 11.10 Nitrate reductase complex NarGHI of E. coli.
Ubiquinol (UQH2) reduces haem b in NarI and electrons are passed suc- Nitric oxide reductases (NORs) catalyse the two-electron
cessively down the electron transport chain through ironsulphur centres reduction of nitric oxide to nitrous oxide:
in NarH and NarG to the molybdenum cofactor centre, MGD, in NarG.
Reduction of nitrate at the cytoplasmic (N-side) and proton release at the 2NO 1 2H1 1 2e2 - N2 O 1 H2 O
periplasmic (P-side) results in generation of a proton motive force.
aOa, haem b;grey box, [3Fea4S] centre; grey hatched box, [4Fea4S] As mentioned in the section on cytochrome oxidases,
centre; filled circle, molybdenum cofactor (MGD). NORs are related both structurally and by evolution and
belong to the same family of copper oxidases. There are This reaction is thermodynamically favourable but the
three variants of NOR: cNOR, qNOR and qCuANOR. A molecule is chemically very inert and so requires a large
comparison of these structures is shown in Figure 11.11. activation energy to drive the reaction. Biologically the
cNOR is found in many different species of denitrifying reaction is catalysed by a homodimeric copper-containing
bacteria including Pseudomonas spp., P. denitrificans [47] enzyme which contains two types of Cu centre CuA and
and has two subunits NorB (c. 475 aa) and NorC (c. 145 CuZ, the CuA centre is involved in electron transfer and
aa). The NorB subunit has two haem b centres: one with the unusual CuZ centre in catalysis [48]. The majority of
two His ligands (haem b low spin) and one with one His nitrous oxide reductases are of this type and hence have
ligand (haem b3 high spin). The NO reduction centre has been termed NosZ nitrous oxide reductases; an exception
FeB rather than CuB found in the cytochrome caa3 oxidases. is the NOS from W. succinogenes. The crystal structures
The basic features of NorB in cNOR are found in qNOR of the enzyme from several bacterial species have been
and qCuANOR. Though not as widespread in occurrence as determined and show the same basic structure (e.g.
cNOR, qNOR often referred to as large NorB, NorB2 or [49,50]). This clearly shows that the C-terminal domain
NorZ (c. 760 aa) is found in a number of bacteria of med- with the CuA centre is folded in the same manner as blue
ical interest including Haemophilus spp., Corynebacterium copper proteins (cupredoxin fold) and that the CuZ centre
diphtheriae and N. gonorrhoeae. Large NorB has a c. 280 is in the N-terminal domain which is folded as a seven-
N-terminal periplasmic domain which has quinol oxidase bladed beta propeller (Fig. 11.12). The two copper atoms
activity. qCuANOR has so far only been characterized in in the CuA centre are linked by two cysteine bridges and
the Gram-positive bacterium Bacillus azotoformans and coordinated to four other amino acids (2His/1Trp/1Met).
contains two subunits, NorB and a membrane-bound sub- The CuZ centre from anaerobically purified P. stutzeri has
unit with two copper atoms (CuA). A membrane-bound a [4Cu:2S] ratio rather than the [4Cu:S] ratio found in the
cytochrome c551 has been shown to act as an electron donor aerobically purified enzyme from P. denitrificans [42,43].
in addition to MQH2. Early work on purification of the enzyme gave different
forms which were dependent on the conditions used and
mainly reflected differences in oxidation state, ligand
binding and loss of either copper or sulphur atoms from
Nitrous Oxide Reductases
the CuZ centre. The active [4Cu:2S] form from P. stutzeri
Nitrous oxide reductases (NOSs) catalyse the two- is purple with seven coordinated His residues from six/
electron reduction of nitrous oxide to dinitrogen. seven propeller blades, whereas aerobically prepared pre-
parations are pink; aerobically purified P. denitrificans is
N2 O 1 2H1 1 2e2 - N2 1 H2 O blue. The crystallographic and spectroscopic data (from
(a) c CuA
SII c MQH2
a N b
b FeB
o
CuB FeB
r
a3 b3 b3
SI C NorB NorB
caa3 cNoR qNoR
(b)
FIGURE 11.11 Comparison of the nitric oxide reductases cNOR and qNOR with cytochrome caa3 oxidase. (a) Schematic diagram and (b) depiction
of tertiary structures. Ribbon structures reproduced with kind permission of M. Pereira [31]. aOa, haem; closed circle, CuA.
(c) (d)
FIGURE 11.12 Structure of the nitrous oxide reductase from Pseudomonas stutzeri. (a) Ribbon structure. The monomers of the homodimer are
arranged head to tail. (b) CuA-CuZ distances. (c) The catalytic CuZ site, a [4Cu2S] structure liganded to neighbouring His residues is in the N-
terminal domain. (d) The CuA site in the C-terminal domain. See text for further details. Reprinted by permission from MacMillan Publishers Ltd:
Nature [43], copyright (2000).
Raman, MCD and EPR) show that the pink form is due to and so has leader sequences for export from the cyto-
loss of sulphur in the CuZ centre to give [4Cu:S], an inac- plasm to the periplasm by the Tat (twin arginine translo-
tive form previously described as CuZ . The blue form is cation) export machinery. The Tat system exports folded
also described as [4Cu:S] but is not identical to the pink proteins and the leader peptide which is cleaved after
form. export has three regions: an N-terminal basic region con-
A dimer is essential for enzyme activity and this is taining the twin arginine conserved motif SRRXFLK, a
because the distance between the CuA and the CuZ centre central hydrophobic region and a C-terminal polar c
is too far for electron transfer. However, in the dimer the region with a recognition sequence for the cleaving signal
CuA centre from one monomer is just 10 Å from the CuZ peptidase [51]. Only three proteins, TatA, TatB and TatC,
centre of the second monomer (this has been termed are required; briefly the TatBC complex binds the protein
domain swapping). The dimer form is thought to be stabi- via the signal peptide and then multiple TatA monomers
lized by calcium ions at the interface between monomers are recruited and polymerize to build the transport com-
[49]. A model for N2O binding is that the molecule orien- plex. Although detergent-isolated TatA complexes have a
tates itself by coordination with a methionine residue at ring structure with a central aqueous cavity, it is not
the CuZ centre and that the product, N2, leaves by a known exactly whether this forms a membrane pore chan-
hydrophobic channel to the exterior surface [50]. The nel. The indications are that the structure is dynamic/
nitrous oxide reductase is located in the periplasmic space transient rather than permanent.
As well as the requirement for proton translocation anchored to the outside of the cytoplasmic membrane by
there is a need for assembly of the Cu centres and their an N-terminal hydrophobic domain. As quinones (usually
insertion into the apoprotein. MK) donate electrons to TMAO reductase this enables a
number of primary electron donors (formate, glycerol-3-
phosphate hydrogen, lactate and hydrogen) to be utilized.
Fumarate and Trimethylamine Oxide A minor additional TMAO reductase is TorZYZ but its
physiological contribution to TMAO reduction is minimal
Reductases and the enzyme is not induced by TMAO.
Fumarate and TMAO are alternative electron acceptors to The substrate specificity of E. coli TMAO reductase is
oxygen but are not as energetically favourable as nitrate. broad with a number of N-oxides and S-oxides utilized; this
Fumarate is reduced to succinate, the reverse of the TCA includes DMSO. It should be noted that there are alternative
cycle oxidation of succinate to fumarate that occurs aero- enzymes for DMSO reduction: DMSO reductase (e.g.
bically. Under anaerobic conditions in the presence of DmsABC in E. coli and DorCA in Rhodobacter sphaer-
fumarate in E. coli a specific membrane-bound quinol: oides), which can also use TMAO as substrate. Whilst there
fumarate reductase complex is induced, encoded by four are differences between these groups they are broadly simi-
genes frdA-D, that is distinct from the succinate dehydro- lar in being periplasmic molybdoenzymes [55].
genase complex, encoded by sdhA-D. Although the A number of observations support the idea that
SdhAB subunits are structurally similar to the FrdAB sub- TMAO reduction does not follow the pattern of hierarchi-
units the regulation of the fumarate and succinate dehy- cal utilization of alternative electron acceptors whereby
drogenases are quite different. sdhA-D is induced oxygen represses anaerobic electron acceptors that yield
aerobically, repressed anaerobically and subject to catabo- less energy and nitrate represses fumarate reductase.
lite repression by glucose. By contrast, frdA-D is TMAO reductase appears regulated by TMAO and not
repressed aerobically, in the presence of nitrate but nitrate or oxygen; TMAO respiration can be observed aer-
induced anaerobically in the presence of fumarate (this obically and is not controlled by the anaerobic regulator
regulation through Fnr and ArcA is discussed later). Fnr (see section on regulation). Thus there is possibly an
The catalytic centre is FrdA, the largest subunit, a flavo- alternative physiological role for TMAO respiration.
protein with bound FAD that accepts electrons from a
smaller ironsulphur protein FrdB with three distinct
ironsulphur clusters: [2Fea2S], [4Fea4S] and
REGULATION OF AEROBIC AND
[3Fea4S] [52]. The crystal structure of fumarate reduc-
tase [52] confirms that the FrdAB subunits are soluble in ANAEROBIC METABOLISM
the cytoplasm and anchored to the membrane by the two The Requirement for Regulation of Aerobic
small subunits FrdC and FrdD, each of which has three
TMHs. The structure also confirmed two sites for mena-
and Anaerobic Energy Metabolism
quinone binding, on opposite sides of the membrane Bacteria vary in their diversity of electron transport
(MKD and MKP). A clear pathway for electron transfer chains. Some, like P. denitrificans, are metabolically
from quinol to FAD is also evident from the structure diverse with three alternative cytochrome oxidases (caa3,
(MKD-MKP-[3Fea4S]-[4Fea4S]-[2Fea2S]FAD). The cbb3 and ba3) and the ability to respire with a number
factors governing whether ubiquinone or menaquinone of alternative electron acceptors. E. coli has two well-
are used in quinol:succinate/fumarate reductases are com- characterized aerobic oxidases: the cytochrome bo3 and
plex but menaquinone appears to be the physiological bd (bd-I) complexes (Fig. 11.7). For both bacteria the
substrate for fumarate reductase despite the fact that iso- pathways of aerobic electron transport are dependent on
lated preparations use ubiquinone efficiently [52]. By the availability of oxygen and are regulated both kineti-
contrast to E. coli the fumarate reductase system in cally and through transcriptional regulation. For P. deni-
C. jejuni has dual functionality, acting as a succinate trificans at 20% oxygen cytochrome caa3, which has a
dehydrogenase under microaerobic conditions and a lower affinity for oxygen, is expressed whilst the high
fumarate reductase under anaerobic conditions [53]. oxygen affinity cytochrome cbb3 is expressed predomi-
TMAO reductase catalyses the reduction of TMAO nantly at a low oxygen concentration; measurement of
((CH3)3NO) to trimethylamine ((CH3)3N) and like for- promoter activities under aerobic, microaerobic and
mate dehydrogenase is a molybdoenzyme containing the anaerobic conditions showed that cytochrome caa3 activ-
molybdenum cofactor [54,55]. In E. coli it is encoded by ity was highest under aerobic conditions and decreased
the torCAD operon and the complex is located in the peri- substantially in microaerobic conditions and again in
plasm. TorA is the catalytic centre for TMAO reduction anaerobic conditions. Cytochrome cbb3 was highest under
and receives electrons from TorC, a c-type cytochrome microaerobic conditions whilst the lower activity of
the cytochrome ba3 promoter compared to that for cyto- well as genetic regulation. For example, in P. denitrifi-
chrome caa3 and the similar activities under aerobic and cans it has been shown that the redox state of the electron
microaerobic conditions led to the conclusion that in transport chain regulates electron flow through kinetic
wild-type cells under aerobic conditions the predominant effects [56], shown by a comparison of electron flow in
route is through cytochrome caa3. These results show that wild-type and mutants defective in the cytochrome bc1
aerobic electron transport is subject to close regulation. complex (fbcC) and in the quinol oxidase (qoxB) and in
Similarly for E. coli, cytochrome bo3 predominates at the presence of respiratory inhibitors. Respiratory rates
high oxygen tension but cytochrome bd-I provides the were measured by measuring oxygen consumption and the
major electron transport pathway under microaerophilic redox states of quinones by extracting quinones from the
conditions. A third oxidase, cytochrome bd-II (AppBC/ membranes and measuring the ratio of quinol (QH2) and
CyxAB/CbdAB) is induced under conditions of carbon quinone (Q) to obtain the fraction of reduced quinone. The
and phosphate starvation and can provide energy and results showed that in wild-type electrons flow was pre-
support aerobic growth in mutants lacking the other respi- dominantly through the bc1 complex to cytochome oxi-
ratory complexes that generate a pmf under aerobic condi- dases caa3 and cbb3 and that only at high ratios of QH2/Q
tions [39,54]. was there significant flow through the quinol oxidase.
Under anaerobic growth conditions P. denitrificans Regulation is also needed for biosynthesis of cofactors
cytochrome caa3 oxidase is very low, the activity of cyto- and their insertion into apoproteins. Two examples are
chrome cbb3 lower than under microaerobic conditions cytochromes and ironsulphur proteins. For cytochromes,
and the cell induces the four enzyme complexes required regulation of biosynthesis of the haem groups, incorpo-
for denitrification from nitrate to nitrogen. For E. coli ration of the haem into the apoprotein and export of
under anaerobic conditions in the absence of an exoge- periplasmic-located cytochromes is needed. Similarly for
nous electron acceptor energy production is by the mixed ironsulphur proteins, which are oxygen labile, the bio-
acid fermentation according to the equation: synthesis of ironsulphur clusters and their transfer and
incorporation into ironsulphur proteins needs regulation
glucose-2 pyruvate-acetate 1 ethanol 1 formate that responds to iron availability and oxidative stress.
1 lactate 1 2CO2 Ironsulphur clusters are present in many redox pro-
teins including ferredoxins, hydrogenases, dehydrogenases
This requires a change in metabolism: the TCA cycle and sensor proteins [57]. For example, in E. coli the
which is fully functional under aerobic conditions becomes fumarate and nitrate reductase complexes, the hydroge-
split into two partial routes: from citrate to oxoglutarate nases, formate dehydrogenase and the stress regulators
and from oxaloacetate to succinate (often called the reduc- IscR, NsrR and OxyR all contain ironsulphur clusters.
tive branch). Two key dehydrogenases are repressed under Ironsulphur proteins are easily damaged and are sensi-
anaerobic conditions: oxoglutarate dehydrogenase, which tive to oxidative stress and nitrosative stress; in oxidative
catalyses conversion of 2-oxoglutarate to succinate, and stress hydrogen peroxide produced through respiratory
pyruvate dehydrogenase, which converts pyruvate to processes generates highly reactive oxygen species (ROS)
such as hydroxyl radicals (OH ) and superoxide (O2
G
acetyl-CoA and CO2 (Fig. 11.13). Under anaerobic condi- 2 ) is

tions synthesis of pyruvate formate lyase (Pfl) is induced also produced in aerobic respiration; while in nitrosative
and Pfl converts pyruvate to acetyl-CoA and formate. In stress, nitric oxide (NO), produced either by denitrifica-
the absence of nitrate formate hydrogenlyase (Fhl a sys- tion or by host phagocytes, produces reactive nitrogen
G
tem involving formate dehydrogenase L and hydrogenase) species (RNS) such as nitric oxide radical NO . Iron
converts formate to CO2 and H2, which prevents excessive metabolism in cells is highly regulated for connected rea-
accumulation of acid. In the presence of nitrate fhl is sons; in many habitats iron is scarce and present in sidero-
repressed and the electron transport chain from formate to phores where it is tightly bound because uncomplexed
nitrate functions. This dissimilatory nitrate reduction gener- iron is highly reactive. Iron regulation involves the Fur
ates energy and provides an electron transport sink; regulator [58] and a regulatory sRNA (RyhB), a system
enzymes of the formate-nitrate pathway are induced. found in many other bacteria [59]. Iron influences the vir-
Nitrate represses electron transport pathways to other alter- ulence of many bacterial pathogens. Thus, there are many
native electron acceptors like fumarate and TMAO. interconnections between energy metabolism, iron
These two examples from E. coli and P. denitrificans homeostasis, oxidative/nitrosative stress and biogenesis of
indicate that as environmental conditions change from ironsulphur proteins. As an example the role of the reg-
aerobic to microaerobic to anaerobic there is a coordi- ulator IscR (ironsulphur cluster regulator) will be
nated set of regulatory changes involving both induc- discussed.
tion and repression of enzymes associated with aerobic/ The initial steps in the biogenesis of ironsulphur
anaerobic energy metabolism. This may involve kinetic as clusters involves cysteine desulphurases that convert
3(3 S\UXYDWH
)+/
S\UXYDWH 3)/ &2+
IRUPDWH
3'+ DFHW\O&R$ 1$'++
DFHW\O&R$ &R 1$' &R$
DFHWDOGHK\GH
FLWUDWH 1$'++
R[DORDFHWDWH FLWUDWH
R[DORDFHWDWH $'+
1$'
PDODWH LVRFLWUDWH HWKDQRO
PDODWH LVRFLWUDWH
IXPDUDWH IXPDUDWH
6'+ R[RJOXWDUDWH )5 R[RJOXWDUDWH
VXFFLQDWH 2'+ VXFFLQDWH
VXFFLQ\O&R$
D E
FIGURE 11.13 Changes in pyruvate metabolism and the TCA cycle in the switch from aerobic (a) to anaerobic (b) metabolism. FR, fumarate reduc-
tase; PDH, pyruvate dehydrogenase complex; PFL, pyruvate formate lyase; FHL, formate hydrogen lyase; SDH, succinate dehydrogenase. AH, alcohol
dehydrogenase; ODH, 2-oxoglutarate dehydrogenase.
cysteine to sulphur, an iron donor and a source of reduc- system is shut off as the induced sRNA RyhB binds to
ing power to reduce S0 to S22; the ironsulphur cluster is the isc operon causing degradation of the iscSUA
assembled on a scaffold system and transferred via an mRNA; but in such a way that IscR is still produced.
intermediate carrier protein to the target apoprotein This is important as IscR regulates at least 40 other
(for details see [33,34]). Two assembly systems have genes including some involved in anaerobic respiration.
been characterized in E. coli: the Isc system (a five- The suf operon is activated by the stress regulator OxyR
component complex encoded by the iscRSUAHscABfdx and is also repressed through the Fur regulator at normal
operon) and the Suf system (encoded by the sufABCDSE levels of iron.
operon). Regulation is complex and involves Fur, RyhB
(a small RNA) and OxyR in addition to IscR. The Isc sys- Basic Principles of Regulatory
tem has been thought to be the housekeeping system that
operates under normal iron availability and the Suf sys-
Systems in Bacteria
tems reponds to stress situation such as iron starvation The regulatory machinery includes sigma factors (includ-
and oxidative/nitrosative stress. ing extracytoplasmic function (ECF) sigma factors and
IscR is a one component regulator with three Cys resi- antisigma factors), transcriptional regulators and sRNAs.
dues that are thought to be involved in the binding of the Some bacteria like B. subtilis use sigma factors as a major
[2Fea2S]11 in the sensor domain. IscR is able to respond control strategy, e.g. in the regulation of sporulation.
to the availability of iron and ironsulphur clusters Whilst E. coli has alternative sigma factors this organism
through its low affinity for ironsulphur clusters which is noted for the number of transcriptional regulators that
affects the balance between the apoform of IscR and the alter transcriptional activity and in particular for two-
functional holoform with a [2Fea2S]11 cluster. Under component systems in which a pair of proteins, a sensor
conditions of iron limitation IscR exists mainly in its and a transducer, effect a change in transcription pattern
apoform which is an inducer of the suf operon. Under (see below). Other bacteria such as N. gonorrhoeae have
normal conditions the Isc system is active and responds relatively few two-component systems.
to ironsulphur availability within the cell; autoregula- Two-component systems (TCS) in bacteria consist of a
tion is achieved by IscR repression at the iscR promoter sensor protein that senses the level of a metabolite
and RyhB is not induced. Under iron limitation the Isc directly or indirectly and as result is autophosphorylated.
FIGURE 11.14 Two-component systems (TCS).

(a) Signalling between sensor and regulator in a
simple system and in a one-component regulator.
(bd) Domain structures of the sensor proteins ArcB,
NarL and Fnr (one-component system). HK, histidine
kinase domain; RR, response regulator receiver
domain; HPT, histidine-containing phosphotransfer
domain.
The phosphorylated form of the sensor protein acts on a

second protein a transcriptional activator protein
NarX-NarL Two-Component System
(Fig. 11.14). The transcriptional activator protein has at NarX is the sensor protein (c. 65 kDa; 598 aa) and NarL
least two domains: a protein-binding domain to interact the transcriptional regulator (c. 120 kDa). NarX binds
with the sensor protein and a DNA-binding domain for nitrate through the N-terminal periplasmic domain (c.
interacting with the promoter. Interaction of the phos- 115 aa) for which a crystal structure has been obtained
phorylated sensor protein with the transcriptional regula- [60]. Nitrate is bound to the homodimer of NarX and
tor results in phosphorylation of the regulator. The binds to two aspartate and two glycine residues (R54,
phosphorylated form of the regulator binds usually to a R540 , G51, G510 ) and to a water molecule. The signal
promoter region of the regulated gene and for positive from the sensor domain is transmitted through a linker
regulation (induction) this enhances the interaction of region to the kinase domain where the conserved H399
the RNA polymerase/sigma factor complex resulting in residue is phosphorylated (Fig. 11.14). The NarL
enhanced transcription. The sensor kinases (SK) are histi- response regulator comprises two domains: an N-terminal
dine kinases whereas the transcriptional regulators, in this receiver domain (often termed REC) and a DNA-binding
context termed response regulators (RR), are aspartate domain with a helix-turn-helix (HTH) domain. The
kinases. As the histidine and aspartate residues are the receiver domain of NarL has a conserved Asp residue
sites for the phosphorylation these are necessarily con- (A59). The phosphorylated form of NarL binds to pro-
served residues. moter regions of target genes usually in the 240 to 250
Two examples of two-component systems involved in region upstream of the transcription start site. The bind-
energy metabolism (NarX-NarL and ArcB-ArcA) will be ing sequences at individual promoters vary but from
considered before looking at one example of a one- analysis of many promoter regions it is possible to
component system (Fnr). The principles discussed will deduce the consensus-binding site. For NarL, the consen-
then be used to indicate the breadth of bacterial regulators sus has been determined as TACYYMT (Y is C or T; M
and how they fit into regulatory networks. is A or T) and the data can be conveniently shown as a
ArcB proteins showed that ArcB is a membrane-bound

histidine kinase sensor protein and ArcA a cytoplasmic
response regulator. The domain structure of ArcB is more
complex than that of a basic regulator (Fig. 11.14). Early
studies indicated that there was an extra domain with two
aspartate residues, D533 and D576, akin to a receiver
(REC) module in a response regulator. This has been con-
firmed by later analysis which shows ArcB has a sensor
domain and three further domains: a transmitter (kinase)
domain, a receiver domain and a phosphotransfer domain.
The N-terminal portion of ArcB has a very short seven-
amino-acid periplasmic loop with two adjacent TMHs (1
and 2). The sensor domain is a PAS domain (named after
the Per/Arnt/Sim proteins in which it is found); in general
this domain often binds cofactors such as FAD and small
molecules like citrate and structurally consists of a central
five-stranded antiparallel β-sheet flanked by α-helices on
either side. In ArcB the PAS domain contains two con-
served Cys residues, C180 and C241. It is now understood
FIGURE 11.15 Position weight matrix for (a) NarL, (b) ArcA and that these respond to the redox state of quinones in the
(c) Fnr. Compiled from current data on regulon DB (E. coli)
respiratory chain though the exact details have been under
considerable debate [63]. A reporter gene strategy, in
position weight matrix in which the percentage occur- strains defective for ubiquinone biosynthesis (ubiCA), has
rence of each base is depicted graphically (Fig. 11.15). been used to indicate whether ArcB was functional; the
The binding sequence relates to the manner in which results showed that ubiquinone was required for ArcB
residues of the NarL protein bind to the DNA. NarL is silencing on transition from anaerobic to aerobic condi-
thought to bind as a dimer; in full-length NarL, the rec- tions. Furthermore, using mutants defective in menaqui-
ognition helix is blocked but receiver phosphorylation of none biosynthesis (menFHDB) it was shown that
the receiver domain causes conformational changes allow menaquinone was needed for reactivation of ArcB on shift
DNA binding. Originally NarL was considered to be a from aerobic to anaerobic conditions [63]. To reconcile
local regulator of nitrate reductase but is now known to the differing views on ArcB kinetic regulation the authors
be a global regulator. In E. coli it is an activator of 51 proposed a model in which both ubiquinone and menaqui-
operons and a repressor for 41 operons [61]. This agrees none compete for interaction with ArcB Cys residues.
with the hierarchy of regulators discussed above. In Under aerobic conditions ubiquinone (redox potential
E. coli, but not other bacteria, there is a second very sim- 1100 mV) is dominant and oxidizes the cysteines with
ilar system NarQ-NarP that diverged from NarX-NarL disulphide bond formation. Under anaerobic conditions the
following gene duplication. The binding specificities to lower redox potential menaquinone (2 41 mV) can break
promoters for the two are slightly different. the disulphide bonds and restore ArcB functionality. An
alternative explanation is that fermentation products cause
mediated inhibition of ArcB phosphatase activity.
Regulation at the Global Level by a Once the sensor domain of ArcB is phosphorylated
at H292 (in a conserved sequence ISHELRTPL also found
Two-Component System: ArcB-ArcA in the other sensor molecules) the phosphate moiety needs
One of the major regulators of energy metabolism in to be transferred to the cognate regulator ArcA. This occurs
E. coli is the ArcB-ArcA system. arcA and arcB mutants via a phosphorelay process with sequential transfer to
were originally isolated by a screen that detected high D576 of the ArcB receiver domain to His717 of the phos-
levels of succinate dehydrogenase using a sdh:lacZ fusion photransfer domain and then to D54 of the receiver domain
strain and measuring β-galactosidase activity as red of ArcA (Fig. 11.14). ArcA has a simple two-domain struc-
papilla in colonies; fermentation of lactose to lactic acid ture of a REC domain and a regulator domain that folds as
in these strains turned the indicator dye from colourless to winged helix-turn-helix (wHTH). The crystal structure of
red [62]. These mutants were termed Arc (Aerobic the regulator domain [64] shows that ArcA forms dimers
Respiratory Control) and arcA and arcB were shown not in solution which fits in with the known binding pattern
be colocated on the genome. Analysis of the ArcA and of other proteins in the OmpR/PhoB subfamily of wHTH
regulators that bind in tandem to DNA direct repeat interaction is not simply just binding of the regulator
half-sites and the presumed consensus sequence of to the DNA at the binding site in the promoter region.
(t)GTTAAttAAaTGTTA [65]. This consensus sequence The mechanisms by which transcriptional regulators
agrees well with that we compiled by bioinformatics enhance transcription have been grouped according to the
(MEME) from RegulonDB (Fig. 11.15) but it is prudent contacts made with the domains of the RNA polymerase
to check consensus sequences in public databases. and the number of binding sites in the promoter region.
The RNA polymerase α subunit has two distinct domains:
α-NTD (N-terminal domain) and α-CTD (C-terminal
Fnr a One-Component System domain) joined by a flexible domain. The two copies of
One-component regulators are proteins in which the sensor the α-CTD can interact with both the promoter DNA
domain and the regulator domain are within a single pro- and the transcriptional activator in a number of defined
tein [66]. Examples are OxyR and the Crp-Fnr family of arrangements. In addition most promoters involved in
regulators including Crp and Fnr [67]. Fnr-like regulators regulation of energy metabolism are complex both in
are present in many Gram-negative bacteria (Table 11.1) binding more than one regulator and in having multiple
but very few Gram-positive bacteria. Whilst the Fnr in E. binding sites, which may have different binding affinities.
coli is noted for its role in regulation of energy metabolism This means that the position and binding sequences of
and the related phenomenon of nitrosative stress, it may specific activators may vary both between different genes
regulate the physiology of bacteria in other ways: two and at different sites giving a range of actual binding
examples are its role in regulation of toxin production in B. sequences that differ from the consensus sequence.
pertussis and in bioluminescence in Vibrio fischeri. Fnr
(named for fumarate and nitrate reductase) mutants were Diversity of Bacterial Energy Metabolism
isolated in several laboratories as pleiotropic mutants
defective in most pathways for anaerobic respiration.
Regulators
Characterization of the protein showed that Fnr has an N- The Fnr and ArcB-ArcA systems have been discussed in
terminal sensor domain with four cysteine residues in the detail as illustrations of regulators of energy metabolism.
sequence 16-Cys-X3-Cys-X2-Cys-X5-Cys-29 that together However, the specific regulators used by different species
form a functional [4Fea4S] cluster and a DNA-binding of bacteria vary. Fnr is part of a superfamily of Crp-Fnr
domain located at the C-terminal end (Fig. 11.14) [68,69]. transcriptional regulators that includes Crp (catabolite
At first sight it may seem odd that Fnr with its major repression), FnrN (nitrogen fixation/photosynthesis), FixK
role as a regulator of anaerobic respiration is synthesized (nitrogen fixation), NnrR (response to NO in denitrifica-
constitutively and has a high copy number (estimated at tion and nitrosative stress) and Dnr (dissimilatory nitrate
2400 per cell). On shift from anaerobic to aerobic conditions respiration regulator) [67]. Fnr-like proteins are found in
the [4Fea4S]21 cluster is destabilized by oxygen and con- many Proteobacteria but not in Helicobacter pylori or
verted to a [2Fea2S]21 cluster causing destabilization of C. jejuni.
the Fnr dimer and thus inactivation of its activity. Fnr is also Other regulators include the TCS regulators of the
sensitive to NO. Thus the absence of oxygen and presence RegBA family (PrrBA/ActSR/RegSR/RoxSR) first identified
of reducing conditions seems sufficient to allow the reverse in photosynthetic bacteria; the sensor RegB is thought to
process and rapid activation of Fnr to an active form to reg- respond to the redox state of the ubiquinone pool [71].
ulate anaerobic energy metabolism. Transcriptomic studies Knowledge of energy metabolism in Gram-positive
have shown that in E. coli Fnr controls at least 103 and pos- bacteria is increasing rapidly: key regulators include:
sibly 115 operons [61]. The consensus binding site or Fnr SrrAB, a TCS that regulates anaerobic enzymes, Rex,
box has the sequence TTGAT-N4-ATCAA (Fig. 11.15) and a redox sensor protein that responds to the cellular
in transcriptional activation of fnr regulated genes the Fnr NADH/NAD1 levels, and CcpA that regulates genes
box is often located at 241.5 with respect to the transcrip- involved in sugar uptake, fermentation and amino acid
tion start site. Fnr binds as a dimer and may bind at tandem metabolism.
sites [70]. Early studies showed Fnr induced genes for
anaerobic metabolism and repressed genes involved in aero-
Regulatory Networks
bic metabolism [62].
The exact mechanism by which response regulators Early work made much use of reporter gene fusions to
interact with the transcriptional machinery and the differ- study gene regulation in particular pathways. Now using
ences between activation and repression by the same reg- a combination of genomic analysis, bioinformatics and
ulator at different sites are beyond the scope of this modelling it is possible to make regulatory network mod-
chapter. However a few points should be noted. First the els for the whole cell. The early view of global and local
TABLE 11.1 Selected Examples of Fnr Homologues and Paralogues
Bacterial Protein Residues Similarity Similarity to Sensor Moiety Regulation

Species to E. coli P. denitrificans.
Fnr (%) Nnr (%)
Escherichia coli Fnr 250 100 Ironsulphur cluster Large regulon (.150
genes): represses genes
for aerobic metabolism
and induces genes for
anaerobic metabolism
Bordetella Btr 242 59.2 (Ironsulphur cluster)

pertussis
Paracoccus FnrP 249 37.5 Ironsulphur cluster Nitrate reductase, cbb3

denitrificans oxidase, nitrous oxide
reductase plus other
genes
Nnr 224 37.0 100 Haem group Nitrite and nitric oxide
reductases
NarR 234 30.1 Unknown Nitrate reductase
Neisseria Fnr 244 58.3 Ironsulphur cluster Small regulon (,20

gonorrhoeae genes) inc. nitrite
reductase AniA
Pseudomonas Anr 244 71.6 Ironsulphur cluster Activation of genes for

aeruginosa anaerobic metabolism
( . 50 genes)
Dnr 227 36.8 49.4 Haem group Limited regulon

(c. 10 genes) for
denitrification
Rhodobacter FnrL 248 44.2 Ironsulphur cluster Regulon of c. 27 genes

sphaeroides involved in anaerobic
metabolism
Nnr 233 40.2 41.2 (Haem group)
Brackets denote evidence is from sequence analysis only. Quantitation of genes in a regulon depends on several factors including whether proteomics or
transcriptomics was used for the determination, on stringency used and whether counting direct or indirect regulation.
regulators was that global regulators controlled several and IHF. Two of these proteins, Fis (factor for inversion
operons in different metabolic pathways (pleoiotropic) stimulation) and IHF (integration host factor), are nucleoid
whilst local regulators controlled only a specific set of proteins that might be expected to bind non-specifically.
genes. This has been modified to reflect the hierarchical However immunoprecipitation studies (CHiP) reveal that
nature of regulation, and the connectivity of regulatory 60% of the binding sites for these nucleoid proteins and a
networks with the objective of producing criteria that third one H-NS (histone-like nucleoid-structuring protein)
define global regulators [72,73]. For example for E. coli it are in intergenic regulatory regions; this together with their
has been calculated that there are c. 300 transcriptional known ability to bend DNA sharply makes it understand-
regulators in the 4405 ORFs identified; such analysis indi- able why these proteins in association with either global or
cated seven global regulators: Crp, IHF, Fnr, Fis, ArcA, local regulators affect regulation of specific genes or oper-
Lrp and H-NS. It has been estimated these global regula- ons. In the mechanism of activation known as antirepres-
tors (excluding H-NS and including NarL) modulate sion the second activator blocks the repressive effect of
expression of 51% of the genes in the genome [72]. In fur- the repressor on the first activator [74]. An example of this
ther analysis supporting this hierarchical model of cell is regulation at the nir promoter (the nir operon encodes
regulation it was proposed that there were ten top-tier reg- genes for NADH-nitrite reductase); here the first activator
ulators: Crp, Fnr, NarL, ArcA, Lrp, FruR, PurR, ArgR, Fis is Fnr and the second activator is NarL and the repressors
are Fis and IHF. Fis binds at 2142, IHF at 288 cycle [45]. There are also pathways for anaerobic respira-
and 2115 and NarL at 269.5. Binding of NarL is suffi- tion with the following electron acceptors: fumarate,
cient to displace IHF from the 288 site and to enable nitrate, nitrite, TMAO and DMSO. The cbb3 oxidase is
dependent transcription [74]. essential for viability in laboratory media and this has
been confirmed in infection studies of the chicken
caecum, where mutants singly defective in one of the ter-
ENERGY METABOLISM OF SELECTED minal reductases were inoculated into week-old chicks
BACTERIAL PATHOGENS and the caecal contents analysed two weeks later [75].
Mutants deficient in either nitrate reductase (napA) or
Our knowledge of the variety and composition of electron
nitrite reductase (nrfA) showed lower colonization
transport chains in bacteria is now well advanced and our
(approximately ten-fold).
understanding of gene regulation is well understood in
A similar study with mutants deficient in five respira-
several bacteria and developing rapidly in many more.
tory proteins was carried out to determine the effect of
Genomic sequences can be used to deduce the likely elec-
these mutations on motility (essential for colonization),
tron pathways in a specific bacterium and this can be fol-
survival under different conditions and infection of intes-
lowed up by microarray, proteomic and other analyses to
tinal cell lines (INT407 and PIC). The results showed
show the quantitative induction of electron transport pro-
a complex dependence on temperature and oxygen but
teins under different growth conditions in wild-type and
indicated an impairment of infection in mutants defective
in specific mutants. This to a considerable extent replaces
in hydrogenase (hydB) or formate dehydrogenase (fdhA).
previous methods in which induction and repression of
This is consistent with the role of hydrogen and formate
individual genes and operons were studied using reporter
as the main energy sources for C. jejuni.
genes, antibody assays, etc.
A quantitative proteomic study compared the levels
This position enables us to study the microbial physi-
of electron transport proteins in cells grown in vitro and
ology of bacteria in infection and in the environment and
after infection within COS-1 mammalian cells [76].
to take a wider approach to determining the importance
General stress response proteins and oxidative stress pro-
and contribution of particular pathways to the survival
teins were similar under the two growth conditions but
and growth of bacteria in specific hosts. This approach is
analysis of metabolic pathways indicated a significant
proving promising in examining the contribution of
metabolic downshift within host cells 20 h post-infection
energy metabolism to the infection, colonization growth
to what was described as ‘a dormant state’. The expres-
and survival of pathogens in hosts. This section gives a
sion levels of the main oxidase involved in aerobic
few examples illustrating this approach for the pathogens
respiration decreased markedly as did levels of nitrate
C. jejuni, Helicobacter pylori, N. gonorrhoeae, P. aerugi-
reductase and TMAO reductase. By contrast the levels of
nosa and Staphylococcus aureus describing both physio-
fumarate reductase (FrdA) remained constant and quite
logical studies and quantitative studies.
high. Further experiments with mutants showed that frdA
and aspA (aspartase) mutants were deficient in intracellu-
lar survival. This confirmed earlier work showing AspA
Campylobacter jejuni contributed fumarate for anaerobic respiration and con-
C. jejuni is the major food-poisoning agent and cause of tributed to survival of bacteria in the chicken intestine.
bacterial gastroenteritis in most countries (see Together these results show that the fumarate reductase
Chapter 67). Energy metabolism is a contributing factor complex FrdABC is essential for host colonization; a sim-
in the establishment of these bacteria in chickens ilar observation has been made for H. pylori (see next sec-
(a natural host and a major reservoir) and in colonization tion). Interestingly, another anaerobically induced
of human epithelial cells during infection. C. jejuni is a fumarate reductase, methylmenaquinol fumarate reduc-
microaerophile and grows preferentially in conditions of tase, MfrA, does not seem to be required for intracellular
510% (v/v) O2 and 510% (v/v) CO2. For aerobic res- survival. MfrA is encoded on the mfrABE operon (incor-
piration C. jejuni has a cbb3-type cytochrome oxidase rectly annotated as sdhAB in the C. jejuniNCTC 11168
(suitable for growth at low oxygen tension with Km of genome sequence) and is a periplasmic enzyme that can
40 nM) and a cyanide-resistant insensitive, low-affinity utilize fumarate, crotonate and mesaconate as substrates
oxidase now designated CioAB (distinct to the bd-type [77]. It has been suggested that in addition to contributing
quinol oxidase E. coli cydAB as originally annotated in to fumarate reduction under anaerobic conditions MrfA
the genome with Km of 0.8 μM). Aerobic growth involves may also act to metabolize crotonate and mesaconate, fer-
utilization of amino acids such as aspartate, glutamate, mentation products of other gut bacteria; these substrates
proline and serine which are catabolized by the TCA are impermeable to the cytoplasmic membrane.
For colonizing bacteria, seeking environmental micro- E. coli with the conversion of pyruvate to acetyl-CoA and of
niches where conditions are optimum for energy metabo- 2-oxoglutarate to succinate being catalysed by pyruvate
lism can be important and bacteria have been shown to oxidoreductase (POR) and 2-oxoglutarate oxidoreductase
move towards these in a process known as ‘energy taxis’. (OOR), respectively, instead of pyruvate dehydrogenase and
This process utilizes two proteins, a sensor protein and a 2-oxoglutarate dehydrogenase [82]. Genome analysis indi-
transducer protein, that in tandem signal the motility cates the TCA cycle is incomplete.
machinery to navigate the bacterial cell (reviewed by [78]). For successful colonization high motility is important
There is some overlap between chemotaxis and energy in enabling penetration of cells into the mucus layer sur-
taxis but the latter is characterized by sensor proteins rounding gastric cells and non-motile cells are attenuated
that detect either redox or pH gradients or changes in in colonization. Chemotaxis may also be beneficial in per-
the intracellular energy status of the cell, e.g. reduced sistent infection. Two of the four Tlp proteins in H. pylori
electron transport. C. jejuni utilizes a two-component have been studied in detail and enable cells to move to
sensor protein system CetA/CetB consisting of a PAS- niches that are more conducive to growth: TlpD responds
domain sensory protein (CetB) and an associated trans- to electron transport/ATP levels whilst TlpB senses pH
ducer protein (CetA) with HAMP and HCD domains [78,83]. In the presence of the electron transport inhibitors
[79]. These domains are found in a wide variety of bac- myxathiazol and HQNO cells showed lower intracellular
terial proteins (PAS domain named after three proteins ATP levels and exhibited smooth swimming in contrast to
Per, ARNT and Sim; HAMP domain named for its smooth swimming with frequent stops to change direction
occurrence in histidine kinases, adenylyl cyclases, MCPs in the wild-type. In addition wild-type cells did not
and phosphatises). The sensory protein (CetB) contains a migrate significantly from a reservoir with no electron
FAD cofactor which detects the energy/redox status of transport inhibitor to another containing inhibitor whereas
the cell. TlpD mutant cells migrated to the second reservoir. TlpD
A study on chemotaxis and attraction to amino acids mutants showed also smooth running in normal media.
and organic acids (which have a role in energy metabo- The conclusion drawn from these observations is that the
lism either as carbon sources or electron donors/acceptors) TlpD contributes toward the ability of the bacteria to
compared taxis and infective ability of wild-type and avoid energy-poor environments and that wild-type bacte-
strains carrying mutations in 5/10 genes for putative ria show a repellent response when any adverse energy
MCP-like proteins (also termed Tlp proteins transducer gradient is detected. The membrane-bound TlpB sensor is
like proteins) [80]. Invasion of human epithelial cells required for movement of motile bacteria away from acid
and chicken embryo cells was reduced in 4/5 mutants but a role for TlpB in colonization has not been convinc-
though chemotaxis was not affected. The authors con- ingly demonstrated.
cluded that energy taxis is important in helping C. jejuni
cells migrate to the optimum niches for energy generation Neisseria gonorrhoeae and Neisseria
and colonization.
meningitidis
Both N. gonorrhoeae and Neisseria meningitidis colonize
Helicobacter pylori human epithelial mucosa. N. gonorrhoeae is a pathogen
H. pylori is a long-term colonizer of the human stomach responsible for the sexually transmitted infection gonor-
with c. 30% of healthy carriers. This bacterium has specific rhoea whilst N. meningitides is a commensal organism of
mechanisms for pH adaptation in the acidic conditions of the nasopharynx that can cause meningitis and sepsis if it
the stomach including expression of a very potent urease. crosses the epithelial barrier and enters the bloodstream
H. pylori is a major pathogen of humans with strong associ- (see Chapters 82 and 98).
ation with ulcerative conditions and with the development These pathogenic Neisseria have a partial denitrifica-
of both gastric cancer and mucosa-associated lymphoid tion pathway from nitrite to nitrous oxide with a nitrite
tissue lymphoma in humans (see Chapter 68). This has reductase and a nitric oxide reductase but no nitrate
led it to be declared a ‘definite carcinogen’ by the World reductase or nitrous oxide reductase, which contrasts with
Health Organization’s International Agency for Research the situation for most commensal Neisseria, which pos-
on Cancer. sess the complete pathway for denitrification [84]. The
H. pylori, like C. jejuni, is a member of the epsilon nitrite reductase of N. gonorrhoeae, termed AniA, is a
group of bacteria and is a microaerophile which grows copper nitrite reductase lipoprotein showing similarity to
in O2 concentrations of 25% (v/v) and requires CO2 the NirK nitrite reductases in the catalytic domain but
for growth (510% v/v). Aerobic respiration using only differs in having an N-terminal lipid domain and a
oxidase (cytochrome cbb3 oxidase) is the only proven means C-terminal glycosylation domain that varies between
to obtain energy [81] but the TCA cycle differs from that in Neisseria spp. [84]. It is induced during anaerobic growth
on nitrite and the fact that antibodies to AniA have been was observed in keeping with an anaerobic lifestyle.
detected in the sera of patients indicates that the bacte- Another adaption of N. gonorrhoeae energy metabolism
rium utilizes anaerobic respiration on nitrite during gono- allows multiple routes of electron transfer to AniA which
coccal infection. The nitric oxide reductase is a qNOR is located in the outer membrane: the cytochrome oxidase
(NorB). CcoP subunit (a c-type cytochrome) has an extra haem
Under aerobic conditions N. gonorrhoeae and domain and electrons for nitrite reduction can be trans-
N. meningitidis utilize a cytochrome cbb3 oxidase, an oxi- ferred via this domain or via cytochrome c5 [85]. By con-
dase suited for growth under microaerobic conditions. trast a model for the evolution of N. meningitidis has been
This involves a branched electron transport pathway that proposed in which it is evolving towards an aerobic life-
allows high oxygen utilization rates enabling toxic ROS style to take advantage of oxygen-rich habitats by loss of
levels to be kept low. From NADH electrons are trans- nitrite reductase and by utilizing a capsular form more
ferred to a cytochrome bc1 complex and then by two par- suited to planktonic than biofilm growth [86]. For this
allel routes through cytochrome c4 and c5 to the terminal change adaptation to cope with oxidative stresses whilst
oxidase [85]. However under anaerobic conditions both maintaining the acquired defences against NO is required.
bacteria can grow in the presence of nitrite. Indeed deni- Thus energy metabolism can be one component of
trification appears to be important for the pathogenicity of pathogenicity. This goes alongside other adaptations such
both species and to alleviating nitrosative stress caused by as molecular mimicry in surface glycoproteins to enable
RNS species formed from macrophage-produced NO. successful colonization.
In N. meningitidis the qNOR is thought to contribute to
intracellular survival through removal of NO and thereby
interferes with NO-dependent signalling processes such
Pseudomonas aeruginosa
as host cytokine responses and inhibition of apoptosis Pseudomonads are opportunistic pathogens (see
[84,86]. Similarly N. gonorrhoeae is able to reduce NO Chapter 41) that inhabit a wide variety of environments
levels in host cells from pro-inflammatory levels to anti- and share with Paracoccus denitrificans a similar versatil-
inflammatory levels. In addition a cytochome c0 may also ity in bioenergetics; both have alternate aerobic oxidases
play a role in NO detoxification, probably as a holding and the ability to denitrify. For aerobic respiration there
defence prior to qNOR induction [47]. are two quinol oxidases cytochrome bo3 (low affinity for
For these reasons denitrification and its regulation in oxygen), the high-affinity cytochrome CIO (for cyanide)
N. gonorrhoeae and N. meningitidis have been investi- and three haemcopper oxidases: cytochrome caa3 (low
gated by several laboratories [8486]. As might be affinity), cbb3-1 and cbb3-2 (both presumed high affinity).
expected the regulators are mainly those discussed previ- The two ccoNOQP operons c for the two cbb3 oxidases
ously for other Gram-negative bacteria: Fnr, NarQ-NarP are located adjacently in P. aeruginosa PAO1, presum-
and NsrR (an NO-sensing regulator) but with some impor- ably as a result of gene duplication. Measurement of
tant minor differences. The aniA promoter has binding growth of P. aeruginosa strains under different oxygen
sites for NarP, NsrR and Fnr. The nitrite reductase is only concentrations in the planktonic state showed that the two
induced once oxygen inactivation of Fnr is relieved. The cbb3 forms were differentially expressed. Surprisingly,
NarQ-NarP system in N. gonorrhoeae (equivalent to cbb3-1 was expressed at high aeration and the levels chan-
NarX-NarL in E. coli) acts slightly differently from ged little when the oxygen concentration was decreased;
E. coli where the sensor protein NarQ responds to nitrite in addition, expression of cbb3-1 repressed that of the
and activates NarP. In N. gonorrhoeae NarP is produced CIO oxidase. By contrast ccb3-2 activity was higher under
constitutively and can bind to and activate the aniA pro- microaerobic conditions and was induced by the Crp-Fnr
moter but NarP activation is not absolutely required for family regulator Anr [87]. This conclusion was confirmed
aniA induction. Importantly, aniA can be activated solely by later mutant studies [88] using both wild-type and
by derepression of NsrR in the presence of NO. Nitric mucoid strains of P. aeruginosa. Other transcriptomic
oxide reductase activity is activated similarly by derepres- studies showed the cytochrome caa3 oxidase was only
sion of norB caused by inactivation of NsrR by NO. induced under starvation conditions/stationary phase and
These findings suggest a dual role for the partial denitrifi- that genes for the low-affinity oxidases were up-regulated
cation pathway in Neisseria: one in anaerobic growth and by the RoxRS TCS. Together these results indicate that
the other in detoxification of NO. The regulatory system oxidases bo3 and cbb3-1 operate at high oxygen tension in
described allows the bacterium to respond differentially exponential growth and that all three high-affinity oxi-
to the presence of nitrite and nitric oxide. This is related dases contribute at low oxygen tensions.
to the evolution of Neisseria as a mucosal pathogen; in P. aeruginosa is also a denitrifier, reducing nitrate to
biofilms where population densities are high and oxygen nitrogen and utilizing a cytochrome cd1 nitrite reductase.
scarce up-regulation of aniA and norB in N. gonorrhoeae Regulation of anaerobic metabolism is complex with
different Fnr-Crp-type regulators fulfilling different roles recently been on their virulence factors and mechanisms
[89]. P. aeruginosa contains two Fnr proteins: Anr, simi- of infection. It is now realized that bacterial physiology is
lar to E. coli Fnr with an ironsulphur centre that reacts related to growth and survival in the host.
with molecular oxygen, and Dnr which contains a The Gram-positive bacterium S. aureus is an impor-
haem group in the sensor domain that reacts with NO. tant nosocomial pathogen (see Chapter 37) with concern
However, the DNA-binding site domains of both Anr and over antibiotic-resistant strains in hospitals. S. aureus is a
Dnr are indistinguishable. Anr is a global transcriptional facultative anaerobe that grows aerobically on glucose
regulator that regulates aerobic respiratory enzymes (e.g. through glycolysis and oxidation of acetyl-CoA through
positive regulator of genes for cbb3-2) and up-regulates the TCA cycle. Under anaerobic conditions glucose is fer-
genes for anaerobic metabolism, e.g. those involved in mented by a mixed acid fermentation theoretically yield-
fermentation of arginine whereas Dnr is a local regulator ing acetate, lactate, formate and ethanol and also a
of denitrification that regulates transcription of the nitrite butanediol fermentation theoretically yielding acetoin and
reductase and nitric oxide reductase operons [67,71,89]. butanediol (Fig. 11.1). Modern molecular techniques
Anr regulates anaerobic respiration in a hierarchical man- allow the overall pathways to be deduced. Genomic anal-
ner by regulating transcription of Dnr and hence indirectly ysis and metabolic reconstruction show the possibilities of
regulates the genes for denitrification [89]. the above metabolism, which was then measured by tran-
These results are of interest to the study of the physiol- scriptomics (microarray), proteomics and end product for-
ogy of P. aeruginosa as a pathogen in cystic fibrosis. mation [90]. As in E. coli the expected changes in
P. aeruginosa can exist both as a planktonic state but also enzyme levels associated with aerobic and anaerobic
converts to a sessile state in biofilms in a process for which metabolism occur, e.g. for pyruvate metabolism Pdh was
30 ,50 -cyclic diguanylic acid (c-di-GMP) is a central regula- repressed and PFL induced under anaerobic conditions.
tor and that leads to changes in expression levels of many The main end products were acetate and lactate with little
genes. In the CF lung P. aeruginosa grows to high densities butanediol or ethanol detected. S. aureus can also grow
and the local oxygen concentration is very low (c. 7 μM); anaerobically on nitrate or nitrate using the nitrate reduc-
thus high-affinity oxidases may be important in the slow tase complex NarGHI encoded by narGHJI and a NADH-
growth of P. aeruginosa in the CF lung. However, a note nitrite reductase NirBD encoded by nirBD (nasDE).
of caution should be exercised. Firstly, bacterial strains are A number of regulators for the aerobic/anaerobic
under a high selection pressure and multiple mutations switch in S. aureus have been identified [91]. These
have been shown to occur in vivo. Secondly, although include YhcSR , SrrAB (Srr for staphylococcal respiratory
mutants defective in all three high-affinity oxidases failed response, also named as SrhSR), and NreBC, a TCS sys-
to produce normal biofilms, this is due to a general defect tem for nitrate and nitrite reduction in S. aureus in
in energy metabolism under these conditions. Other work response to oxygen. SrrAB is a TCS homologous to the
has shown that Anr is important for biofilm formation in ResDE system of B. subtilis, where SrrA is the response
anaerobic conditions but phospholipase C may contribute regulator and SrrB, the membrane-bound sensor protein
to Anr activation in oxic conditions through the products that responds to oxygen availability. SrrA is activated
choline and glycine betaine. The role of nitrate respiration under anaerobic conditions and positively regulates genes
and denitrification for growth in the cystic fibrosis lung is involved in fermentation.
uncertain. Evidence for a role is that the respiratory nitrate There is a complex relationship between bacterial
reductase is required for growth in CF sputum and that physiology, biofilm formation, quorum sensing and viru-
nitrate levels in sputum can support denitrification (see lence [92]. Cell density is the key factor by which Agr
[86]) but this may not be sufficient for the high cell densi- (accessory gene regulator) quorum regulator works; the
ties observed. The observations that nitric oxide reductase quorum-sensing response is triggered when the concentra-
activity is needed both for formation of anaerobic biofilms tion of secreted autopeptides reaches a certain threshold
and for survival in activated macrophages may be due to value. Quorum sensing is important in biofilm formation,
its role in NO detoxification. Nevertheless, the physiology in secretion of phenol-soluble modulins, in biofilm matu-
of P. aeruginosa in the cystic fibrosis lung confirms the ration and detachment of biofilms for further dissemina-
connections that exist between bacterial energy metabo- tion [92]. In biofilms local oxygen concentrations will
lism, biofilm formation and virulence. vary greatly and will influence bacterial metabolism; for
example, biofilm formation is enhanced under anaerobic
conditions through induction of polysaccharide intracellu-
Staphylococcus aureus lar adhesion. There are likely to be interactions between
In comparison to the Gram-negative bacteria discussed the regulators of energy metabolism and the regulators
above, the physiology of Gram-positive bacterial patho- both for virulence and biofilm formation. One example is
gens is less well studied and the emphasis has until SrrAB, identified as a regulator of anaerobic energy
metabolism that also negatively regulates some agr- [15] Wada T, Long JC, Zhang D, Vik SB. A novel labeling approach
dependent virulence factors under anaerobic conditions; supports the five-transmembrane model of subunit a of the
overexpression of srrAB also decreased biofilm formation. Escherichia coli ATP synthase. J Biol Chem 1999;274:173537.
There are many other examples of connections between [16] Hamilton WA. Energy transduction in anaerobic bacteria.
In: Anthony CA, editor. Bacterial Energy Transduction. London:
energy metabolism and bacterial pathogenicity. One is the
Academic Press; 1988. p. 83149.
use of different oxidases by Brucella suis in the phagocytic
[17] Moat AG, Foster JW, Spector MP. Microbial Physiology. 4th ed.
vacuole of macrophages [93]. Another is the metabolism of New York: John Wiley & Sons, Inc.; 2002.
ironsulphur proteins in Mycobacterium tuberculosis, [18] Fushinobu S. Unique sugar metabolic pathways of bifidobacteria.
which has many ironsulphur proteins, 40% of which are Biosci Biotechnol 2010;74:237484.
involved in energy metabolism. In moving from the dor- [19] Louis P, Flint HJ. Diversity, metabolism and microbial ecology of
mant state inside hypoxic granulomas to the active state in butyrate-producing bacteria from the human large intestine. FEMS
the oxygen-rich upper lobe of the human lung, the host Microbiol Lett 2009;294:18.
may target these proteins for inactivation. In turn M. tuber- [20] Seedorf H, Fricke WF, Veith B, Brueggemann H, Liesegang H,
culosis has stress sensors, such as WhiB3, that may have a Strittimatter A, et al. The genome of Clostridium kluyveri, a strict
role in a nitrosative stress response [94]. anaerobe with unique metabolic features. Proc Natl Acad Sci USA
2008;105:212833.
[21] Dwidar M, Park JY, Mitchell RJ, Sang B-I. The future of butyric
acid in industry. Science 2012;2012:471417.
REFERENCES [22] Luetke-Eversloh T, Bahl H. Metabolic engineering of Clostridium
[1] Mitchell P. Vectorial chemiosmotic processes. Annu Rev acetobutylicum: recent advances to improve butanol production.
Biochem 1977;46:9961005. Curr Opin Biotechnol 2011;22:63447.
[2] Busch KB, Deckers-Hebestreit G, Hanke GT, Mulkidjanian AY. [23] Qin JJ, Li RQ, Raes J, Arumugam M, Burgdorf KS, Manichanh C,
Dynamics of bioenergetic microcompartments. J Biol Chem 2013; et al. A human gut microbial gene catalogue established by meta-
394:63188. genomic sequencing. Nature 2010;464:5965.
[3] Brzezinski P, Gennis RB. Cytochrome c oxidase: exciting progress [24] Flint HJ, Bayer EA, Rincon MT, Lamed R, White BA.
and remaining mysteries. J Bioenerg Biomembr 2008;40:52131. Polysaccharide utilization by gut bacteria: potential for new insights
[4] Nicholls DG, Ferguson SJ. Bioenergetics. 4th ed. San Diego, CA: from genomic analysis. Nat Rev Microbiol 2008;6:12131.
Academic Press; 2013. [25] Cameron EA, Maynard MA, Smith CJ, Smith TJ, Koropatkin NM,
[5] Al-Attar S, de Vries S. Energy transduction by respiratory metallo- Martens EC. Multidomain carbohydrate-binding proteins involved
enzymes: from molecular mechanism to cell physiology. Coord in Bacteroides thetaiotaomicron starch metabolism. J Biol Chem
Chem Rev 2013;257:6480. 2012;287:3461425.
[6] Gunner MR, Amin M, Zhu XY, Lu J. Molecular mechanisms for [26] Flint HJ, Scott KP, Duncan SH, Louis P, Forano E. Microbial deg-
generating transmembrane proton gradients. Biochim Biophys radation of complex carbohydrates in the gut. Gut Microbes
Acta, Bioenerg 2010;1827:892913. 2012;3:289306.
[7] Morris JG. Vital energy for bacterial growth and reproduction. [27] Scheller HV, Ulvskov P. Hemicelluloses. Ann Rev Plant Biol
In: Susmann M, editor. Molecular Medical Microbiology. 1st ed. 2010;61:26389.
San Diego, CA: Academic Press; 2002. p. 23356. [28] Dodd D, Mackie RI, Cann IK. Xylan degradation, a metabolic
[8] Walker JE. The ATP synthase: the understood, the uncertain and property shared by rumen and human colonic Bacteroidetes. Mol
the unknown. Biochem Soc Trans 2013;41:116. Microbiol 2011;79:292304.
[9] Abrahams JP, Leslie AGW, Lutter R, Walker JE. Structure at 2.8 Å [29] Martens EC, Koropatkin NM, Smith TJ, Gordon JI. Complex gly-
resolution of F1-ATPase from bovine heart mitochondria. Nature can catabolism by the human gut microbiota: the Bacteroidetes
1994;370:6218. Sus-like paradigm. J Biol Chem 2009;284:246737.
[10] Roy A, Hutcheon ML, Duncan TM, Cingolani G. Improved crys- [30] Dodd D, Moon Y-H, Swaminathan K, Mackie RI, Cann IK.
tallization of Escherichia coli ATP synthase catalytic complex Transcriptomic analyses of xylan degradation by Prevotella
(F-1) by introducing a phosphomimetic mutation in subunit epsi- bryantii and insights into energy acquisition by xylanolytic
lon. Acta Crystallogr, Sect F: Struct Biol Cryst Commun 2012;68: Bacteroidetes. J Biol Chem 2010;285:3026173.
122933. [31] Sousa FL, Alves RJ, Pereira-Leal JB, Teixeira M, Pereira MM. A
[11] Nakanishi-Matsui M, Futai M. Stochastic rotational catalysis of bioinformatics classifier and database for heme-copper oxygen
proton pumping F-ATPase. Philos Trans R Soc, B 2008;363: reductases. PLoS One 2011;6:e19117.
213542. [32] Zubay GL. Biochemistry. 4th ed. WC Brown Publishers; 1998.
[12] von Ballmoos C, Wiedenmann A, Dimroth P. Essentials for ATP [33] Lill R. Function and biogenesis of iron-sulphur proteins. Nature
synthesis by F1F0 ATP synthases. Ann Rev Biochem 2009;78: 2009;460:8318.
64972. [34] Roche B, Aussel L, Ezraty B, Mandin P, Py B, Barras F. Iron/sulfur
[13] Junge W. ATP synthase and other motor proteins. Proc Natl Acad proteins biogenesis in prokaryotes: formation, regulation and diver-
Sci USA 1999;96:47357. sity. Biochim Biophys Acta, Bioenerg 2013;1827:92337.
[14] Hicks DB, Liu J, Fujisawa M, Krulwich TA. F1F0-ATP synthases [35] Nowicka B, Kruk J. Occurrence, biosynthesis and function of iso-
of alkaliphilic bacteria: lessons from their adaptations. Biochimica prenoid quinones. Biochim Biophys Acta, Bioenerg 2010;1797:
Biophys Acta 2010;1797:136277. 1587605.
[36] Morris RL, Schmidt TM. Shallow breathing: bacterial life at low O2. [54] Gennis RB, Stewart V. Respiration. In: Neidhardt FC, Curtiss III
Nat Rev Microbiol 2013;11:20512. R, Ingraham JL, Lin ECC, Low KB, Magasanik B, et al., editors.
[37] Buschmann S, Warkentin E, Xie H, Langer JD, Ermler U, Michel. H. Escherichia coli and Salmonella: Cellular and Molecular Biology.
The structure of cbb3 cytochrome oxidase provides insights into pro- 2nd ed. Washington, DC: ASM Press; 1996. p. 21761.
ton pumping. Science 2010;329:32730. [55] McCrindle SL, Kappler U, McEwan AG. Microbial dimethyl-
[38] Ekici S, Pawlik G, Lohmeyer E, Koch H-G, Daldal F. Biogenesis sulfoxide and trimethylamine-N-oxide respiration. In: Poole RK,
of cbb3-type cytochrome c oxidase in Rhodobacter capsulatus. editor. Advances in Microbial Physiology, vol. 50. Elsevier-
Biochim Biophys Acta, Bioenerg 2012;1817:898910. Academic Press; 2005. p. 14798.
[39] Chepuri V, Lemieux L, Hill J, Alben JO, Gennis RB. Recent stud- [56] Otten MF, Reijnders WNM, Bedaux JJM, Westerhoff HV, Krab K,
ies of the cytochrome o terminal oxidase complex of Escherichia Van Spanning RJM. The reduction state of the Q-pool regulates the
coli. Biochim Biophys Acta, Bioenerg 1990;1018:1247. electron flux trough the branched respiratory network of
[40] Abramson J, Riistama S, Larsson G, Jasaitis A, Svensson-Ek M, Paracoccus denitrificans. Eur J Biochem 1999;261:76774.
Laakkonen L, et al. The structure of the ubiquinol oxidase from [57] Fleischhacker AS, Kiley PJ. Iron-containing transcription factors
Escherichia coli and its ubiquinone binding site. Nat Struct Biol and their roles as sensors. Curr Opin Chem Biol 2011;15:33541.
2000;7:9107. [58] Spiro S, D’Autreaux. Non-heme iron sensors of reactive oxygen
[41] Borisov VB, Murali R, Verkhovskaya ML, Bloch DA, Han HZ, and nitrogen species. Antioxid Redox Signal 2012;17:126476.
Gennis RB, et al. Aerobic respiratory chain of Escherichia coli is [59] Oglesby-Sherrouse AG, Murphy ER. Iron-responsive bacterial
not allowed to work in fully uncoupled mode. Proc Natl Acad Sci small RNAs: variations on a theme. Metallomics 2013;5:27686.
USA 2011;108:173204. [60] Cheung J, Hendrickson WA. Structural analysis of ligand stimula-
[42] Zheng HJ, Wisedchaisri G, Gonen T. Crystal structure of a nitrate/ tion of the histidine kinase NarX. Structure 2009;17:190201.
nitrite exchanger. Nature 2013;497:64751. [61] Constantinidou C, Hobman JL, Griffiths L, Patel MD, Penn CW,
[43] Richardson DJ, Van Spanning RJM, Ferguson SJ. The prokary- Cole JA, et al. A reassessment of the FNR regulon and transcrip-
otic nitrate reductases. In: Bothe H, Ferguson SJ, Newton WE, tomic analysis of the effects of nitrate, nitrite, NarXL, and NarQP
editors. Biology of the Nitrogen Cycle. Amsterdam: Elsevier as Escherichia coli K12 adapts from aerobic to anaerobic growth.
BV; 2007. p. 2135. J Biol Chem 2006;281:480215.
[44] Macpherson S, Murphy MEP. Type-2 copper-containing enzymes. [62] Lynch AS, Lin ECC. Responses to molecular oxygen.
Cell Mol Life Sci 2007;64:288799. In: Neidhardt FC, Curtiss III R, Ingraham JL, Lin ECC, Low KB,
[45] Kelly DJ. Complexity and versatility in the physiology and metab- Magasanik B, et al., editors. Escherichia coli and Salmonella:
olism of Campylobacter. In: Nachamkin I, Szymanski CM, Blaser Cellular and Molecular Biology. 2nd ed. Washington, DC: ASM
MJ, editors. Campylobacter. 3rd ed. Washington, DC: ASM Press; Press; 1996. p. 152638.
2008. p. 4161. [63] Alvarez AF, Rodriguez C, Georgellis D. Ubiquinone and mena-
[46] Mohan SB, Cole J. The dissimilatory reduction of nitrate to ammo- quinone electron carriers represent the yin and yang in the redox
nia in anerobic bacteria. In: Bothe H, Ferguson SJ, Newton WE, regulation of the ArcB sensor kinase. J Bacteriol 2013;195:
editors. Biology of the Nitrogen Cycle. Amsterdam: Elsevier BV; 305461.
2007. pp. 93106. [64] Toro-Roman A, Mack TR, Stock AM. Structural analysis and
[47] Zumft WG. Nitric oxide reductases of prokaryotes with emphasis solution studies of the activated regulatory domain of the response
on the respiratory, heme-copper oxidase type. J Inorg Biochem regulator ArcA: a symmetric dimer mediated by the α4-β5-α5
2005;99:194215. face. J Mol Biol 2005;349:1126.
[48] Zumft WG, Kroneck PMH. Respiratory transformation of nitrous [65] McGuire AM, De Wulf P, Church GM, Lin ECC. A weight matrix
oxide (N2O) to dinitrogen by Bacteria and Archaea. In: Poole RK, for binding recognition by the redox-response regulator ArcA-P of
editor. Advances in Microbial Physiology, vol. 52. Escherichia coli. Molec Microbiol 1999;32:21921.
Elsevier-Academic Press; 2007. p. 107227. [66] Ulrich LE, Koonin EV, Zhulin IB. One-component systems domi-
[49] Haltia T, Brown K, Tegoni M, Cambillau C, Saraste M, Mattila K, nate signal transduction in prokaryotes. Trends Microbiol
et al. Crystal structure of nitrous oxide reductase from Paracoccus 2005;13:526.
denitrificans at 1.6 angstrom resolution. Biochem J [67] Körner H, Sofia HJ, Zumft WG. Phylogeny of the bacterial super-
2003;369:7788. family of Crp-Fnr transcription regulators: exploiting the meta-
[50] Pomowski A, Zumft WG, Kroneck PMH, Einsle O. N2O binding bolic spectrum by controlling alternative gene programs. FEMS
at a (4Cu:2S) copper-sulphur cluster in nitrous oxide reductase. Microbiol Rev 2003;27:55992.
Nature 2011;477:2348. [68] Spiro S. The FNR family of transcriptional regulators. Antonie
[51] Palmer T, Berks C. The twin-arginine translocation (Tat) protein Van Leeuwenhoek 1994;66:2336.
export pathway. Nat Rev Microbiol 2012;10:48396. [69] Kucera I, Mazoch J. Control of gene expression by FNR-like
[52] Maklashina E, Cecchini G, Dikanov SA. Defining a direction: proteins in facultatively anaerobic bacteria. Folia Microbiol
electron transfer and catalysis in Escherichia coli complex II 2002;47:95103.
enzymes. Biochim Biophys Acta, Bioenerg 2013;1827:66878. [70] Barnard AML, Green J, Busby SJW. Transcription regulation by
[53] Weingarten RA, Taveirne ME, Olson JW. The dual-functioning tandem-bound FNR at Escherichia coli promoters. J Bacteriol
fumarate reductase is the sole succinate:quinone reductase in 2003;185:59936004.
Campylobacter jejuni and is required for full host colonization. [71] Bueno E, Mesa S, Bedmar EJ, Richardson DJ, Delgado MJ.
J Bacteriol 2009;191:5293300. Bacterial adaptation of respiration from oxic to microoxic and
anoxic conditions: redox control. Antioxid Redox Signal [83] Schweinitzer T, Mizote T, Ishikawa N, Dudnik A, Inatsu S,
2012;16:81952. Schreiber S, et al. Functional characterization and mutagenesis of
[72] Martinez-Antonio A, Collado-Vides J. Identifying global regula- the proposed behavioural sensor TlpD of Helicobacter pylori.
tors in transcriptional regulatory networks in bacteria. Curr Opin J Bacteriol 2008;190:324455.
Microbiol 2003;6:4829. [84] Barth KR, Isabella VM, Clark VL. Biochemical and genomic
[73] Seshasayee AS, Fraser GM, Babu MM, Luscombe NM. Principles analysis of the denitrification pathway within the genus Neisseria.
of transcriptional regulation and evolution of the metabolic system Microbiology 2009;155:4093103.
in E. coli. Genome Res 2009;19:7991. [85] Cole JA. Response of Neisseria gonorrhoeae to oxygen limita-
[74] Browning DF, Grainger DC, Busby SJW. Organisation of com- tion and excess. In: Kidd SP, editor. Stress Response in
plex Escherichia coli promoters. In: Lee SY, editor. Systems Pathogenic Bacteria, Advances in Molecular and Cellular
Biology and Biotechnology of Escherichia coli. Dordrecht: Microbiology, vol. 19. Wallingford, UK: CAB International;
Springer; 2009. p. 25771. 2011. p. 22142.
[75] Weingarten R, Grimes J, Olson J. Role of Campylobacter jejuni [86] Moir JWB. Bacterial nitrogen cycling in the human body. In: Moir
respiratory oxidases and reductases in host colonization. Appl JWB, editor. Nitrogen Cycling in Bacteria: Molecular Analysis.
Environ Microbiol 2008;74:136775. Norfolk, UK: Caister Academic Press; 2011. p. 23348.
[76] Liu X, Gao B, Novik V, Galan JE. Quantitative proteomics of [87] Comolli JC, Donohue TJ. Differences in two Pseudomonas
intracellular Campylobacter jejuni reveals metabolic reprogram- aeruginosa cbb3 cytochrome oxidases. Mol Microbiol 2002;51:
ming. PLoS Pathog 2012;8:e1002562. 1193203.
[77] Guccione E, Hitchcock A, Hall S, Mulholland F, Shearer N, [88] Alvarez-Ortega C, Harwood CS. Responses of Pseudomonas aeru-
van Vliet AHM, et al. Reduction of fumarate, mesaconate and ginosa to low oxygen indicate that growth in the cystic fibrosis
crotonate by Mfr, a novel oxygen-regulated periplasmic reduc- lung is by aerobic respiration. Mol Microbiol 2007;65:15365.
tase in Campylobacter jejuni. Environ Microbiol 2010;12: [89] Schreiber K, Krieger R, Benkert B, Eschbach M, Arai H, Schobert M,
57691. et al. The anaerobic regulatory network required for Pseudomonas
[78] Schweinitzer T, Josenhans C. Bacterial energy taxis: a global aeruginosa nitrate respiration. J Bacteriol 2007;189:43104.
strategy? Arch Microbiol 2010;192:50720. [90] Fuchs S, Pane-Farre J, Kohler C, Hecker M, Engelmann S.
[79] Elliott KT, DiRita VJ. Characterization of CetA and CetB, a Anaerobic gene expression in Staphylococcus aureus. J Bacteriol
bipartite energy taxis system in Campylobacter jejuni. Mol 2007;189:427589.
Microbiol 2008;69:1091103. [91] Hall JW, Ji Y. Sensing and adapting to anaerobic conditions by
[80] Vegge CS, Brondsted L, Li YP, Bang DD, Ingmer H. Energy taxis Staphylococcus aureus. Adv Appl Microbiol 2013;84:125.
drives Campylobacter jejuni toward the most favorable conditions [92] Otto M. Staphylococcal infections: mechanisms of biofilm matu-
for growth. Appl Environ Microbiol 2009;75:530814. ration and detachment as critical determinants of pathogenicity.
[81] Meyers JD, Kelly DJ. Respiratory electron transport in Helicobacter Annu Rev Med 2013;64:17588.
and Campylobacter. In: Zannoni D, editor. Respiration in Archaea [93] Köhler S, Foulongne V, Ouahrani-Bettache S, Bourg G, Teyssier J,
and Bacteria, vol.2: Diversity of Prokaryotic Respiratory Systems. Ramuz M, et al. The analysis of the intramacrophagic virulome of
Advances in Photosynthesis and Respiration, vol. 16. Dordrecht: Brucella suis deciphers the environment encountered by the patho-
Springer; 2004. p. 6380. gen inside the macrophage host cell. Proc Natl Acad Sci USA
[82] Kelly DJ, Hughes NJ. The citric acid cycle and fatty acid biosyn- 2002;99:157116.
thesis. In: Mobley HLT, Mendz GL, Hazell SL, editors. [94] Saini V, Farhana A, Glasgow JN, Steyn AJC. Iron sulfur cluster
Helicobacter pylori: Physiology and Genetics. Washington, DC: proteins and microbial regulation: implications for understanding
ASM Press; 2001. p. 13546. tuberculosis. Curr Opin Chem Biol 2012;16:4553.

Bacterial Energy Metabolism

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Bacterial Energy Metabolism

Uploaded by

Copyright:

Available Formats

Chapter 11

Bacterial Energy Metabolism

INTRODUCTION TO ENERGY energy generation is coupled to ATP synthesis and utili-

Molecular Medical Microbiology. DOI: http://dx.doi.org/10.1016/B978-0-12-397169-2.00011-1

(2H1 1 2e2 1 A - AH2). There should be strict stoichi-

unidirectional and can be used to create additional energy Xylan Utilization

AEROBIC RESPIRATION (a) CH=CH2

[3Fea4S] also occurs. Cytochromes are electron transfer CH3 Cys

Cytochromes are grouped into four major groups (a, b, c

Cytochrome bo3 Cytochrome bd-l Cytochrome bd-Il

bc1 complex (Complex III)

Cytochrome ba3 Cytochrome aa3 Cytochrome cbb3

H+ translocating Yes Yes Yes

FIGURE 11.9 Comparison of redox and catalytic

with the use of reporter gene fusions such as alkaline phos-

The NrfA subunit is a pentahaem cytochrome a. In the δ-

Nitric Oxide Reductases

caa3 cNoR qNoR

acetyl-CoA and CO2 (Fig. 11.13). Under anaerobic condi- 2 ) is

FIGURE 11.14 Two-component systems (TCS).

The phosphorylated form of the sensor protein acts on a

ArcB proteins showed that ArcB is a membrane-bound

TABLE 11.1 Selected Examples of Fnr Homologues and Paralogues

Bacterial Protein Residues Similarity Similarity to Sensor Moiety Regulation

Bordetella Btr 242 59.2 (Ironsulphur cluster)

Paracoccus FnrP 249 37.5 Ironsulphur cluster Nitrate reductase, cbb3

NarR 234 30.1 Unknown Nitrate reductase

Neisseria Fnr 244 58.3 Ironsulphur cluster Small regulon (,20

Pseudomonas Anr 244 71.6 Ironsulphur cluster Activation of genes for

Dnr 227 36.8 49.4 Haem group Limited regulon

Rhodobacter FnrL 248 44.2 Ironsulphur cluster Regulon of c. 27 genes

Nnr 233 40.2 41.2 (Haem group)

You might also like