Validation of E-C protocol using dynamic growth of the
coculture
The E-C protocol is not applicable to the static coculture with known
concentrations, as it is based on the growth stoichiometry of individual
microorganisms. Therefore, in this subsection, we use coculture batch
growth experiments to demonstrate and validate the E-C protocol. With
the validity of the cell counting method established, the individual
biomass concentration obtained from the cell counting method were used
to validate the E-C protocol. Figure 3 (a) and (b) plot the total OD of
the coculture over 3 days for the salt water pair and fresh water pair,
respectively; and Figure 3 (c) and (d) plot the gas phase composition
for each coculture pair for one inoculum ratio (1:10 for the salt water
pair and 1:2 for the fresh water pair), respectively. The gas
compositions for the other inoculum ratios are provided in Supplementary
Material Figure S2. For the fresh water methanotroph-microalgae pair,
higher inoculum concentration of the microalgae resulted in better
growth of the coculture. This is because the microalgae grows much
slower than the methanotroph, so the methanotroph growth is limited by
O2 availability. Therefore, more microalgae in the
inoculum enabled better growth of the methanotroph. For the salt water
pair, higher inoculation concentration of the cyanobacteria did not have
much impact on coculture growth. This is because the cyanobacteria grew
much faster than the methanotroph, and the methanotroph growth is
limited by mass transfer of CH4 from gas phase.
Figure 4 compares the individual biomass concentration measured through
the cell counting approach and the E-C protocol for both coculture
pairs, where each point represents one of the duplicates, and the error
bar represents the standard deviation from three cell counting
measurements for the same sample. As can be seen from these figures, the
results obtained from the two approaches correlated very well,
particularly at low biomass concentrations. The R2 for
the linear relationship between the results from the E-C protocol and
cell counting approach ranges 0.90 – 0.98, which validates the results
obtained from the E-C protocol.
However, Figure 4 also shows that the agreement between the cell
counting approach and the E-C protocol deteriorates at higher
concentrations after coculture growth. To determine which approach
performs better, we calculated the total OD for each sample using the
measured individual biomass concentrations, and plotted them against the
measured total OD. The results are shown in Figure 5 (a) and (b) for the
salt water pair and the fresh water pair respectively. Both figures
showed that the total OD calculated from the E-C protocol were almost
exactly the same as the measured total OD. On the other hand, the total
OD calculated from the cell counting approach showed larger deviation
from the measured total OD, particularly at higher concentrations. The
bar chart of the mean squared error (MSE) of predictions in the total OD
based on six experimental runs (three inoculum concentrations with
duplicates) are plotted in Figure 5. The error bar represents one
standard deviation of MSE’s. Student’s t -test shows that the
MSE’s of the cell counting is statistically significantly larger than
that of the E-C protocol, with a p-value of 0.0158 for the salt water
pair and 0.0030 for the fresh water pair.
Besides obtaining individual biomass concentration for each
microorganism in the coculture accurately and quickly, the E-C protocol
also provides estimates of individual substrate consumption rates and
product excretion rates. Figure 6 (a) and (b) plot the individual
consumption and production rates of O2 and
CO2 respectively by M. alcaliphilum 20ZR
and S. sp. PCC7002 over a three-day period for the inoculum ratio
of 1:10, and Figure 6 (c) and (d) plot those values for M.capsulatus – C. sorokiniana , for the inoculum ratio of
1:2.
Figure 6 shows that although for many cases very small amounts of
O2 were detected in the gas phase (e.g., day 2 and 3 for
the salt water pair and all 3 days for the fresh water pair),
significant amount of O2 was produced by the
photoautotroph, which was completely consumed in situ by the
methanotroph. Similarly, Figure 6 shows that the actual amount of
CO2 consumed by the photoautotroph was much larger than
what was directly measured in the experiment, because the
CO2 produced by the methanotroph would be preferably
consumed by the photoautotroph, as it was produced in situ and
did not involve the mass transfer resistance from gas to liquid.