I heard from my colleagues that my Ph.D. thesis was a topic of discussion
on this news group some weeks ago (while I was away on holiday and had no
internet access), and going through the archives now I came across a
request to publish the abstract here. Well, here it comes.
I took up a lecturing position at the University of Stellenbosch, South
Africa (in Jannie Hofmeyr's department), at the beginning of July. Because
of the usual starting-up chaos I have only recently started to follow
bionet.metabolic-reg again. For those interested, I can now be reached at:
Dept. of Biochemistry
University of Stellenbosch
Private Bag X1
7602 Matieland
SOUTH AFRICA
Fax: +27-21-8083022
Email: jr at land.sun.ac.za
I obtained my Ph.D. from the University of Amsterdam, where I studied the
control and regulation of the bacterial phosphotransferase system under the
supervision of Hans Westerhoff and Pieter Postma. The work described in my
thesis is relevant to recent discussions on this newsgroup, especially to
the application of MCA to non-ideal metabolic systems. In Chapter 6, I show
that MCA can be used to gain mechanistic information about the nature of
the channelling process in the bacterial PTS (see summary for details). By
measuring the sum of the flux-response coefficients of the PTS enzymes
(these have also been called enzyme control coefficients, but the
distinction between control- and response coefficients is another topic
addressed in the thesis --- see below), one can determine whether
long-living complexes exist between the different PTS enzymes and the bound
phosphoryl group or not. At first sight it appears that the summation
theorem for the flux-control coefficients (defined in terms of enzyme
concentration) is violated. but it is precisely this violation that gives
us additional information, as the coefficients and concentrations of the
enzyme-enzyme complexes are related mathematically. Furthermore, when the
control coefficients are defined with reference to local rates of reactions
of the biochemical system (and not with reference to enzyme concentration),
the summation theorems still hold.
In conclusion, I feel that MCA, if applied carefully to a metabolic system,
can yield a wealth of information about its behaviour and control and may
even give us some mechanistic detail which is otherwise not measurable
directly.
Sorry for this rather long initial post! I invite anyone interested to
respond in this thread or contact me personally by email.
Best regards
Johann Rohwer
===========================================================
Ph.D. Thesis of Johann M. Rohwer:
"Interaction of functional units in metabolism.
Control and regulation of the bacterial
phosphoenolpyruvate-dependent phosphotransferase system"
Summary
This thesis describes the interactions between different functional units
in the carbohydrate and free-energy metabolism of the enteric bacteria
Escherichia coli and Salmonella typhimurium, using the framework of
metabolic control analysis. The phosphoenolpyruvate:carbohydrate
phosphotransferase system (PTS) of these bacteria, which catalyses the
uptake and concomitant phosphorylation of several carbohydrates, is studied
as experimental model system. The PTS is a group-transfer pathway: a
phosphoryl group derived from phosphoenolpyruvate (PEP) is transferred
along a series of enzymes to the carbohydrate molecule in a process called
group translocation.
Metabolic control analysis describes how the steady-state behaviour of a
metabolic system depends on its (bio)chemical constitution. In contrast to
the enzymes of `ideal' and normal metabolic pathways, group-transfer
enzymes cannot be assigned to one specific reaction because they
participate in two reactions, i.e. the one accepting and the one donating
the transferred group. To facilitate the distinction between control by
group-transfer enzyme concentrations on the one hand, and control by the
chemical processes they are involved in on the other hand, a new
nomenclature system for metabolic control analysis is proposed in Chapter
1. We distinguish between control coefficients, which describe the
dependence of a steady-state variable on the activity of an independent
step, and response coefficients, which describe the sensitivity of a
steady-state variable to a system parameter. The `non-ideal' nature of
group-transfer pathways (discussed in Chapter 1), interferes with the
correspondence between the response coefficient towards a group-transfer
enzyme concentration and a control coefficient. Moreover, a group-transfer
enzyme can be defined to catalyse both reactions in which it participates.
Chapter 1 also introduces the experimental systems under investigation: the
PTS and the H+-ATPase, which plays a central role in free-energy
transduction.
In the analysis of intracellular metabolic networks, it is often useful to
discern subsystems (modules) of which the metabolites are only produced or
consumed by reactions within that subsystem or by a limited number of
fluxes crossing the borders of the subsystem. In Chapter 2, we derive
general conditions, valid for metabolic systems of any complexity, for a
subsystem to behave as a unit with respect to its effect on the remainder
of the system. A subsystem is a `monofunctional unit' if the relative
response of any two metabolic variables outside that subsystem (as
quantified by their co-response coefficient) to a perturbation of the
subsystem does not depend on which reaction inside the subsystem is
perturbed. For a subsystem to be a monofunctional unit, three conditions
suffice: (i) the reactions outside the subsystem are not affected directly
by metabolites belonging to the subsystem; (ii) there are no conservation
relations linking the subsystem to the rest; and (iii) the subsystem is
linked to the remainder of the system only via one degree of freedom in
fluxes. Different relative responses in two metabolic variables to two
parameter perturbations indicate that these perturbations do not originate
from the same monofunctional unit. This can serve to identify unsuspected
regulatory interactions in the metabolic network. An advantage of
identifying monofunctional units is that they can be treated as single
well-coupled enzymes. This reduces the complexity of the analysis and of
the system.
During PTS-mediated uptake of glucose or methyl alpha-D-glucopyranoside
(MeGlc, an analogue of glucose), a phosphoryl group derived from PEP is
transferred sequentially along four PTS proteins, Enzyme I (EI), HPr,
IIAGlc and IICBGlc, to the glucose or MeGlc molecule. Because this process
requires an input of free energy (delivered by the conversion of PEP to
pyruvate), we investigated the influence of the cellular free-energy state
on MeGlc uptake by the E. coli PTS (Chapter 3). The intracellular
[ATP]/[ADP] ratio was modulated by changing the expression of the atp
operon, which codes for the H+-ATPase, or by adding an uncoupler of
oxidative phosphorylation or an inhibitor of respiration. Initial MeGlc
uptake rates decreased when the intracellular [ATP]/[ADP] ratios were
lowered by any of these methods. These results suggest that the cellular
free-energy state, as reflected in the intracellular [ATP]/[ADP] ratio,
plays a significant role in regulating the activity of the PTS.
Because the phosphoryl donor for PTS-mediated uptake is PEP, and not ATP,
we investigated whether the changes in PTS uptake activity, described
above, were mediated by the intracellular PEP concentration, or the
[PEP]/[pyruvate] ratio (Chapter 4). After the development of accurate and
sensitive assays, the intracellular PEP and pyruvate levels were measured
in response to changes in the cellular free-energy state, brought about by
the above methods. Neither the intracellular PEP level nor the
[PEP]/[pyruvate] ratio correlated with the [ATP]/[ADP] ratio or with the
initial MeGlc uptake rate. Effects of changes in the free-energy state were
analysed further by assaying PTS phosphorylation activity in vitro with
different [PEP]/[pyruvate] and [ATP]/[ADP] ratios. The variations in PTS
activity were insufficiently large, under a wide variety of conditions, to
account for the variation in PTS uptake rates observed in vivo. This
suggests that an additional regulatory mechanism must be responsible for
the effects of modulations in the cellular free-energy state on
PTS-mediated uptake activity.
Apart from catalysing the uptake of carbohydrates, the PTS plays an
important role in signal transduction by mediating catabolite repression.
The presence or absence of PTS carbohydrates is sensed by IIAGlc and
reflected in its phosphorylation state. Upon transport of PTS
carbohydrates, IIAGlc is dephosphorylated. This form of the enzyme inhibits
the uptake systems for lactose, maltose and melibiose, as well as the
enzyme glycerol kinase, by the formation of stoichiometric complexes with
these proteins, preventing the entry of these carbon sources into the cell
and the induction of their corresponding operons. This is the so-called
`inducer exclusion' effect. When PTS carbohydrates are absent, on the other
hand, IIAGlc is phosphorylated; this form of the enzyme activates adenylate
cyclase and thus increases intracellular cyclic AMP levels. Cyclic AMP is
required for the expression of numerous genes, amongst others the catabolic
operons for many carbon sources. Because unphosphorylated IIAGlc forms
stoichiometric complexes with its target proteins during inducer exclusion,
we investigated whether the PTS could be retro-inhibited as a result of
IIAGlc sequestration into such complexes, which are inactive in the
phosphotransfer reaction of the PTS (Chapter 5). MeGlc uptake into S.
typhimurium could be inhibited by pre-incubation of the cells with
glycerol, and required the pre-induction of both the glucose and glycerol
catabolising systems. Larger extents of inhibition by glycerol correlated
with higher intracellular levels of glycerol kinase. Controls with lactate,
which is not subject to inducer exclusion, seem to exclude metabolic
effects as the origin of the observed inhibition. In vitro, PTS activity
was also inhibited by glycerol kinase. This inhibition depended on the
addition of macromolecular crowding agents known to promote macromolecule
associations. We conclude that uptake systems, which have hitherto only
been regarded as targets of IIAGlc, also have the capacity to regulate the
PTS flux by sequestration of the available IIAGlc in a process which we
propose to call `reverse inducer exclusion'.
Because group-transfer enzymes participate in two reactions, as described
above, they cannot be regarded individually as monofunctional units.
Metabolic control analysis predicts that interactions between them can lead
to special control properties, such as that the summed flux-response with
respect to the enzyme concentrations and boundary metabolites exceeds one;
the exact value of this sum depends on the degree of complex formation
between the enzymes: it decreases from two, in the absence of complexes,
towards one as complex formation increases. We determined the effect of
different total enzyme concentrations on the flux through the PTS in vitro
(Chapter 6). The dependence of the flux on total PTS enzyme concentration
was more than linear but less than quadratic, as quantified in terms of a
combined flux-response coefficient of the four glucose PTS enzymes, which
decreased with increasing protein concentration. Addition of macromolecular
crowding agents led to a sharper decrease of this combined response with
increasing protein concentration. The data could be explained with a simple
kinetic model. We conclude that the different PTS enzymes may well form
complexes in the cell that live long on the time-scale of their turnover,
and that intracellular macromolecular crowding may strongly determine the
control of and by important transport and regulatory systems such as the
PTS.
In an attempt to answer as yet unresolved questions on the metabolic
behaviour of the glucose PTS, and in particular to reconcile differences in
the dependence of the flux in vivo and in vitro on the concentrations of
the constituent enzymes, we constructed a detailed kinetic model, based on
literature data (Chapter 7). This model could describe the experimental
dependence of the PTS-mediated uptake rate on the enzyme concentrations in
vivo, as well as the dependence of the flux in vitro on the total PTS
enzyme concentration, solely on the basis of the higher protein
concentrations in vivo. The model suggests a new interpretation of
experimental results hitherto not understood: in vivo, the major fraction
of the PTS enzymes may exist as complexes with other PTS enzymes or with
boundary metabolites, whereas in most in vitro experiments, the fraction of
complexed enzymes is much smaller.
Finally, in Chapter 8 we relate the results presented in this thesis to the
analysis of other systems. The delineation of monofunctional units in
metabolic systems may prove useful in characterising genes of unknown
function by localising the effects of mutations. The non-ideal metabolic
behaviour of the PTS has implications for the analysis of similar
group-transfer systems, i.e. the electron transport chain of mitochondria,
two-component regulatory systems in bacteria, and signal transduction
networks in mammals.
--
J.M. Rohwer
Dept. of Biochemistry, Univ. of Stellenbosch
E-mail replies: remove SPAMBLOCK from address
--
