Biology Reference
In-Depth Information
communities, giving rise to various beneficial interactions.
These phenomena have been modeled for multistrain
[145]
and multispecies
[135,146,147]
networks, with applica-
tions in such diverse areas as biofuel engineering and
diagnostics
[79,135]
.
The general process in this area is to create metabolic
models for each organism of interest and then integrate the
models via reactions facilitating the transport of metabo-
lites from one organism to the other
[135,145
e
147]
.For
example, a host
e
pathogen model was recently recon-
structed representing the interactions between the human
alveolar macrophage and M. tuberculosis
[135]
, thereby
allowing the simulation of the metabolic interactions
between the cells after M. tuberculosis invaded the
macrophage. Evaluation of feasible steady-state solutions
for the integrated model showed increases in flux through
the glyoxalate shunt, decreased glycolysis, and increased
gluconeogenesis. In addition, condition-specific models
were created for various stages of infection and provided
evidence that currently proposed drugs targeting the poly-
prenyl metabolic pathway
[148,149]
might only prove
effective during latent stages of infection. This is because
the pathway is not utilized during later infection stages.
In another study, intraspecies cooperation between
auxotrophic strains of E. coli was investigated by finding
synergistic growth pairs between conditionally lethal
mutants
[145]
. Co-culture of the strains was found to result
in synergistic growth when both strains received a benefit
from the other via secreted products, and the secreted
products provided a high benefit to the recipient strain
while also not posing a large burden to the donor strain.
Computationally, this corresponded to high shadow prices
for received metabolites and low shadow prices for donated
metabolites (i.e., the metabolites were highly beneficial to
the recipient and did not significantly influence biomass
production in the donor). Although the quantitative growth
rate predictions were generally higher than experimental
values, it was interesting that the prediction of cooperation
could be computed purely from the stoichiometric matrices
of the two strains, making the approach a valuable quali-
tative tool.
The development of multi-cell models is a relatively
new area of research in constraint-based modeling.
However, it is expected that it will continue to develop and
play an important role, since community structure influ-
ences metabolism in the environment
[150]
, human
microbiomes
[151]
, normal physiology
[152]
, and patho-
physiology
[135]
.
eukaryotes. Thus, for a more holistic view, goals exist to
develop an integrated 'OME' model, consisting of the
transcriptional regulatory network (O for operon), metab-
olism (M for metabolism), and transcription and trans-
lational processes behind enzymes (E for expression). It is
expected that such a model would have greater predictive
power than any of its components or combinations thereof.
Reconstruction of metabolic networks is a well-established
process and an E-matrix reconstruction has recently been
developed for E. coli
[153]
. However, the reconstruction of
transcriptional regulatory processes remains a relatively
unrefined area of research.
E-Matrix Reconstructions
The procedure for E-matrix reconstruction is fairly similar
to the protocol described earlier in this chapter, but the
focus is on genes and reactions that are necessary for
transcription and translation. The resulting model
(
Figure 12.6
C) contains information describing the many
processes
[153]
, including transcription, mRNA degrada-
tion, translation, protein maturation and folding, protein
complex formation, ribosome assembly, RNA processing,
rRNA/tRNA modification, and tRNA charging.
These reconstructions represent another perspective on
cellular processes beyond metabolism. Moreover, these can
be computationally modeled in a fashion similar to meta-
bolic reconstructions to investigate the effects of drugs,
mutations, etc.
Efforts are being made to integrate the E-matrix with
the metabolic network for an organism (ME-matrix)
[154]
,
which is accomplished fairly readily on a purely mathe-
matical level by combining metabolites shared between the
matrices into a single entry (row) and requiring that
enzymes be synthesized by the transcription and translation
machinery for a metabolic reaction to occur.
Specifically, the reaction fluxes of the E-matrix and the
metabolic network are coupled together, such that if
a metabolic reaction is needed, the flux through translation
reactions for its associated enzyme must have a non-zero
flux
[153
e
155]
. Thus reaction fluxes for transcription of
the associated mRNA must also be non-zero, and reaction
fluxes for ribosomal synthesis must also carry flux. Thus,
these integrated models are able to more accurately predict
phenotypes as they explicitly account for most cellular
costs, including the costs of using the molecular machinery.
There are numerous potential applications of models
that couple transcription, translation, and metabolism.
These include (1) improvements in metabolic engineering,
as the models explicitly account for recombinant protein
synthesis cost in terms of metabolic burden, as well as (2)
a network-level understanding of the effects of antibiotics,
many of which
[156
e
159]
act by blocking the functionality
of E-matrix components such as bacterial rRNA.
FUTURE DIRECTIONS
Metabolic networks are central to almost all cellular
processes. However, they represent only about one-third of
genes in prokaryotes and an even smaller fraction in higher
