advances.sciencemag.org/cgi/content/full/6/33/eaba0353/DC1
Supplementary Materials for
Metabolic multistability and hysteres
is in a model aerobe-anaerobe
microbiome community
Tahmineh Khazaei, Rory L. Williams, Said R. Bogatyrev, John C. Doyle, Christopher S. Henry, Rustem F. Ismagilov*
*Corresponding author. Email: rustem.admin@caltech.edu
Published 12 August 2020,
Sci. Adv.
6
, eaba0353 (2020)
DOI: 10.1126/sciadv.aba0353
This PDF file includes:
Figs. S1 to S9
Tables S1 to S4
Contributions of non-corresponding authors
Fig. S1. The workflow for the co
ntinuous bioreactor experiment
indicating the points at which we
inoculated with bacteria, took samples, and changed glucose con
centrations of the input feeds.
Fig. S2. Imaging of samples collected from the continuously sti
rred tank reactor experiments.
(
A
) Total
bacteria staining of each sample using DAPI (
B
) The species composition of aggregates using the GAM42a
(green) and CFB560 probe (pink) f
or Gammaproteobacteria and Bac
teroidetes, respectfully.
Fig. S3.
Bayesian parameter fitting of K and v
max
to experimental batch growth data for (
A
)
Klebsiella
pneumoniae
on glucose (
B
)
Bacteroides thetaiotaomicron
on glucose, and (
C
)
Bacteroides thetaiotaomicron
on dextran. Two or three techni
cal replicates were used for eac
h concentration of substrate examined.
Fig. S4
.
A quantitative view of regions of stability as a function of gl
ucose concentrations in the input
feed and oxygen flow rates into the reactor.
In regions of multi-stability (circles) the community can exis
t
in either a
Kp
-only state or a
Kp–Bt
(aerobe–anaerobe) state under the same conditions. In regions
of mono-
stability (triangles) the community can only exist in either a
Kp
-only or a
Kp–Bt
state.
(A)
Deviation from
yellow indicates the increase of
Kp
concentration in the
Kp–Bt
state relative to the
Kp
-only state (e.g.
concentration of
Kp
in
Kp–Bt
state divided by the concentration of
Kp
in the
Kp
-only state).
(B)
Ratio of
Kp
concentration to
Bt
concentration in the
Kp–Bt
state. Ratios were not calculated for the
Kp
-only state as
Bt
does not grow in that state (white down-pointed triangles).
Proof-of-Concept Experiment for
Community Control via SCFA
The goal of this proof-of-concept experiment was to show that w
e can jump directly from one
steady state (
Kp
-only) to an alternate steady state (
Kp-Bt
state) without changing glucose
concentration (i.e., without taking the system through a progre
ssion of glucose concentrations, from
low to high and then back to low). We predicted that we could m
ake this direct jump if we give the
system the metabolites that are exchanged in the alternate (
Kp-Bt
) steady state. We began by
flowing minimal media with a 2 mM input glucose in the CSTR. In
this condition, as demonstrated
in Fig. 4 (and Fig. S5), only
Kp
is able to grow at steady state and
Bt
gets washed out. We then
switched the media to 2 mM glucose + 7 mM acetate + 7 mM lactat
e (the metabolites identified in
the
Kp-Bt
state by transcriptomics analy
sis) and we observed that when
Bt
was inoculated, it did
not wash out but grew along with
Kp
. Finally, we switched the media back to 2 mM glucose and
observed that both
Bt
and
Kp
continued to grow at a steady state (Fig. S5). This experiment
demonstrates another method we
can use to manipulate a system t
hat exhibits MSH. We show that
we can directly provoke the system to move to a different stead
y state by adding the metabolites
exchanged by the species.
Fig S5. State switch from
Kp
-only state (state A) to the
Kp–Bt
state (state B) by the transient addition
(pulse) of short chain fatty acids (addition of 7 mM lactate +
7 mM acetate were flowed with 2 mM
glucose for 24 h).
Measurements for each species are made by digital PCR using 16
S primers specific to
each species. Error bars are S.D. of three replicates collected
from the CSTR (over 24 h after steady state
was reached, each separated by more than 1 residence time) for
each of the steady-state conditions (state A
and state B).
Fig S6. Quantification of
B. thetaiotaomicron
and
K. pneumoniae
in archived CSTR samples using
digital PCR.
Primers used targeted the 16S ribosomal RNA gene for each speci
es. Error bars are S.D. of
three replicates collected (separ
ated by >1 residence time) fro
m the CSTR for each of the eight steady-state
glucose conditions.