Biology Reference
In-Depth Information
Additionally, we computed the growth yields of
P. putida
on sole sources of three
other important elements—Nitrogen (N), Phosphorous (P), and Sulfur (S)—and com-
pared these with published experimental data from continuous cultivations [37], as
shown in Table 4. Since biomass composition can play a role in the effi ciency of
in
silico
usage of basic elements, this analysis can aid in assessing how well the biomass
equation, which is equivalent to the
E. coli
biomass reaction, reproduces the true bio-
mass composition of
P. putida
. The yield on nitrogen differs only by 10% between
in
silico
and
in vivo
experiments, which suggests that the associated metabolic network
for nitrogen metabolism is well characterized in the iJP815 reconstruction. The yields
on phosphorous and sulfur, however, differ by more than a factor of two between the
in vivo
and
in silico
analyses, suggesting that there may be signifi cant differences
between the biomass requirements and the metabolic networks of
P. putida
and
E.
coli
for these components. The differences in yields, however, may be also caused by
the change of the
in vivo
biomass composition, which decreases the fraction of com-
pounds containing the limited element, when compared to the biomass composition
while the bacterium is grown under carbon-limitation. Such changes were observed
experimentally in
P. putida
for nitrogen and phosphate limitations [47]. Thus, the bio-
mass composition of
P. putida
needs to be determined precisely in the future. Howev-
er, for the purpose of this work and since the global effect of the biomass composition
on the outcome of the simulations is negligible (as shown above), we considered the
use of the original biomass equation to be justifi ed.
Table 4.
Comparison of the
in silico
predicted growth yields (in g
DW
g
Element
−
1
) with experimental
continuous culture data.
Limiting Element
Yield - Experimental
Yield- Model
C
0.47
0.61
N
5.74
6.67
P
84.95
34.92
S
268.75
130.18
Analysis of Blocked Reactions: The Quest for Completeness
As described above, iJP815 contains 289 unconditionally (i.e., not dependent on ex-
ternal sources) blocked reactions (i.e., reactions unable to function because not all
connections are made), corresponding to 33% of the metabolic network. In previously
published genome-scale metabolic reconstructions, the fraction of blocked reactions
varies between 10 and 70%t [48]. Blocked reactions occur in reconstructions mostly
due to knowledge gaps in the metabolic pathways. Accordingly, the blocked-reactions
set can be divided into two major groups; (1) reactions with no connection to the set of
non-blocked reactions, and (2) reactions that are either directly or indirectly connected
to the operating core of the
P. putida
model. The first group of reactions includes
members of incomplete pathways that, with increasing knowledge and further model
refinement, will gradually become connected to the core. This subset comprises 108
reactions (35% of blocked-reaction set). The second group of reactions comprises also
members of incomplete pathways, but many of them belong to pathways that are
