Exposing the Three-Dimensional Biogeography and Metabolic States
of Pathogens in Cystic Fibrosis Sputum via Hydrogel Embedding,
Clearing, and rRNA Labeling
William H. DePas,
a
Ruth Starwalt-Lee,
a
Lindsey Van Sambeek,
a
Sripriya Ravindra Kumar,
a
Viviana Gradinaru,
a
Dianne K. Newman
a,b,c
Division of Biology and Biological Engineering, California Institute of Technology, Pasadena, California, USA
a
; Division of Geological and Planetary Sciences, California
Institute of Technology, Pasadena, California, USA
b
; Howard Hughes Medical Institute, California Institute of Technology, Pasadena, California, USA
c
W.H.D. and R.S.-L. contributed equally to this article.
ABSTRACT
Physiological resistance to antibiotics confounds the treatment of many chronic bacterial infections, motivating
researchers to identify novel therapeutic approaches. To do this effectively, an understanding of how microbes survive
in vivo
is
needed. Though much can be inferred from bulk approaches to characterizing complex environments, essential information can
be lost if spatial organization is not preserved. Here, we introduce a tissue-clearing technique, termed MiPACT, designed to re-
tain and visualize bacteria with associated proteins and nucleic acids
in situ
on various spatial scales. By coupling MiPACT with
hybridization chain reaction (HCR) to detect rRNA in sputum samples from cystic fibrosis (CF) patients, we demonstrate its
ability to survey thousands of bacteria (or bacterial aggregates) over millimeter scales and quantify aggregation of individual
species in polymicrobial communities. By analyzing aggregation patterns of four prominent CF pathogens,
Staphylococcus au-
reus
,
Pseudomonas aeruginosa
,
Streptococcus
sp., and
Achromobacter xylosoxidans
, we demonstrate a spectrum of aggregation
states: from mostly single cells (
A. xylosoxidans
), to medium-sized clusters (
S. aureus
), to a mixture of single cells and large ag-
gregates (
P. aeruginosa
and
Streptococcus
sp.). Furthermore, MiPACT-HCR revealed an intimate interaction between
Strepto-
coccus
sp. and specific host cells. Lastly, by comparing standard rRNA fluorescence
in situ
hybridization signals to those from
HCR, we found that different populations of
S. aureus
and
A. xylosoxidans
grow slowly overall yet exhibit growth rate heteroge-
neity over hundreds of microns. These results demonstrate the utility of MiPACT-HCR to directly capture the spatial organiza-
tion and metabolic activity of bacteria in complex systems, such as human sputum.
IMPORTANCE
The advent of metagenomic and metatranscriptomic analyses has improved our understanding of microbial com-
munities by empowering us to identify bacteria, calculate their abundance, and profile gene expression patterns in complex envi-
ronments. We are still technologically limited, however, in regards to the many questions that bulk measurements cannot an-
swer, specifically in assessing the spatial organization of microbe-microbe and microbe-host interactions. Here, we demonstrate
the power of an enhanced optical clearing method, MiPACT, to survey important aspects of bacterial physiology (aggregation,
host interactions, and growth rate),
in situ
, with preserved spatial information when coupled to rRNA detection by HCR. Our
application of MiPACT-HCR to cystic fibrosis patient sputum revealed species-specific aggregation patterns, yet slow growth
characterized the vast majority of bacterial cells regardless of their cell type. More broadly, MiPACT, coupled with fluorescent
labeling, promises to advance the direct study of microbial communities in diverse environments, including microbial habitats
within mammalian systems.
Received
4 May 2016
Accepted
19 August 2016
Published
27 September 2016
Citation
DePas WH, Starwalt-Lee R, Van Sambeek L, Kumar SR, Gradinaru V, Newman DK. 2016. Exposing the three-dimensional biogeography and metabolic states of
pathogens in cystic fibrosis sputum via hydrogel embedding, clearing, and rRNA labeling. mBio 7(5):e00796-16. doi:10.1128/mBio.00796-16.
Editor
Margaret J. McFall-Ngai, University of Hawaii
Copyright
© 2016 DePas et al. This is an open-access article distributed under the terms of the
Creative Commons Attribution 4.0 International license
.
Address
correspondence to Viviana Gradinaru, viviana@caltech.edu, or Dianne K. Newman, dkn@caltech.edu.
H
ost-microbe interactions are increasingly recognized as driv-
ers of health and disease in many different contexts, from the
beneficial human microbiome to deleterious bacterial infections,
such as those that chronically infect individuals living with cystic
fibrosis (CF) (1–3). In all of these cases, the relationship between
microbial and host cells is influenced by the features of the mi-
croenvironment, which change over time and can be challenging
to measure. Nevertheless, it is essential to characterize the nature
of these important associations if we seek to understand and/or
control them. Spatial organization is a defining parameter in any
environment, and it is likely that by impacting bacterium-
bacterium or bacterium-host associations, or by creating gradi-
ents of nutrients or toxins that affect bacterial growth rates, spatial
organization affects bacterial survival (4). The current toolset for
understanding microbial communities associated with animal
host environments provides limited spatial information (e.g., thin
sectioning) (5–7) or lacks it entirely (bulk measurement of abun-
dance, via metagenomics and transcriptomics) (8–10). Building
upon a tissue-embedding and clearing technique, the passive clar-
ity technique (PACT) (11–13), we developed MiPACT (
m
icrobial
RESEARCH ARTICLE
crossmark
September/October 2016 Volume 7 Issue 5 e00796-16
®
mbio.asm.org
1
mbio.asm.org
on September 27, 2016 - Published by
mbio.asm.org
Downloaded from
i
dentification after PACT) to permit the study of diverse bacterial
pathogens residing in cystic fibrosis patient sputum. While PACT
preserves spatial and molecular information and allows for effi-
cient clearing as well as protein and transcript labeling via use of
fluorescent probes, we incorporated key modifications to ensure
(i) stabilization of amorphous sputum samples, (ii) high retention
of bacteria, and (iii) efficient labeling of bacterial rRNA via hy-
bridization chain reaction (HCR) (14, 15) and fluorescence
in situ
hybridization (FISH). Though developed in the context of CF,
MiPACT-HCR can be readily applied to diverse host-microbe sys-
tems.
Patients with CF accumulate obstructive sputum plugs in their
lung airways that can harbor an array of opportunistic pathogens
(16). Sputum buildup and the resultant chronic infections lead to
severe lung damage and eventual respiratory failure (17). CF pa-
tients routinely expectorate infected sputum, which provides trac-
table samples for
in situ
analysis of pathogens (6, 18, 19). Until
recently,
Pseudomonas aeruginosa
was the most prevalent patho-
gen isolated from CF patients, and
P. aeruginosa
colonization is
well known to correlate with disease progression in CF (16, 20,
21). Therefore, the majority of studies addressing the biogeogra-
phy of CF have focused on
P. aeruginosa
. FISH analysis of thin
sections of CF lung or smears of CF sputum have revealed that
P. aeruginosa
can exist both as single cells and in large clusters and
that
P. aeruginosa
grows more slowly
in situ
than in typical labo-
ratory cultures (6, 7).
While
P. aeruginosa
plays an important role in CF pathogenic-
ity in many patients, other microbes also colonize the CF lung and
contribute to exacerbations, or increase disease severity (16). In-
deed, culture-independent studies have revealed that individuals
harbor a distinct microbial ecosystem whose species composition
can vary over time and treatment regimens (10, 22). Though re-
cent studies have attempted to gain a perspective on the distribu-
tion of particular clone types as a function of lung geography,
these studies have been herculean, requiring microdissection, cul-
tivation, and sequencing of thousands of regional isolates (5, 19,
23). Recognizing the need to study CF pathogens
in situ
to gain
information relevant to the design of accurate
in vitro
models, we
sought a method that would permit rapid scanning of large spatial
areas at various magnifications, as well as one that would permit
microbial identification and study at the single-cell level. Here, we
describe our usage of MiPACT-HCR to study three important
attributes of diverse pathogens in CF sputum: aggregation pat-
terns, bacterium-host interactions, and growth rates.
RESULTS AND DISCUSSION
We obtained seven sputum samples (numbered 1 to 4 and 5.1, 5.2,
and 5.3) with consent from five patients at the Children’s Hospital
of Los Angeles (CHLA). Samples 1, 4, 5.2, and 5.3 were collected
during an exacerbation, while samples 2, 3, and 5.1 were collected
during outpatient well visits. Disease states varied between pa-
tients, with patients 1 and 3 having FEV1% (percent forced expi-
ratory volume in 1 s, a measure of lung function) values of 48 and
44 (moderate obstruction), respectively, while the remaining pa-
tients had FEV1% values greater than 70 (mild to normal).
When fixed in paraformaldehyde (PFA; 4%) and embedded in
A
4
P
0
(4% acrylamide, 0% PFA), sputum completely dissolved
during clearing. To provide more structural stability, we replaced
acrylamide with 4% 29:1 acrylamide:bis-acrylamide (29A:1B)
4
P
0
,
providing additional cross-linking (Fig. 1a). Use of (29A:1B)
4
P
0
preserved sputum integrity and allowed for clearance in SDS
(Fig. 1b). Samples took 3 to 14 days to fully clear (Fig. 1b). Because
sputum is composed largely of host-derived DNA and mucins
(24), we labeled DNA with 4
=
,6-diamidino-2-phenylindole
(DAPI) and mucins with rhodamine-conjugated lectin (wheat
germ agglutinin [WGA]) after clearing to obtain a structural con-
text. Imaging revealed a high degree of compositional variation
between samples (Fig. 1c). For example, sputum samples from
patients 1 and 2 were composed largely of lectin-stained mucin,
with interspersed DAPI-bright host cells. Sputum 5 was composed
almost entirely of polymorphonuclear neutrophils (PMNs), con-
sistent with findings that PMNs are a major component of CF
patient sputum (25). PMN cell boundaries were outlined by a
network of extracellular DNA (Fig. 1d), potentially a result of
neutrophil extracellular traps (NETs) (26). While intersample
heterogeneity was evident, sampling different regions of a single
sputum sample revealed that intrasample composition was rela-
tively homogenous (see Fig. S1 in the supplemental material).
We next verified that the common CF pathogens
P. aeruginosa
and
Staphylococcus aureus
could be retained and visualized under
the same embedding and clearing conditions required for retain-
ing sputum integrity. There was no significant loss of DAPI-
stained logarithmic- or stationary-phase bacteria after clearing of
pure cultures embedded in (29A:1B)
4
P
0
hydrogel blocks (see
Fig. S2a in the supplemental material). FISH staining with satu-
rating probe concentrations of the universal bacterial probe
EUB338 after MiPACT (see Fig. S3a in the supplemental material)
revealed that the Gram-positive microbe
S. aureus
required treat-
ment with lysostaphin after clearing via SDS, while the Gram-
negative
P. aeruginosa
did not require lysozyme treatment (see
Fig. S3b). In sputum, autofluorescence makes bacteria, particu-
larly slowly growing cells, difficult to demarcate by FISH (see
Fig. S4 in the supplemental material). Therefore, we employed
HCR, a FISH amplifying technique which has previously been
used to fluorescently label RNA in zebrafish embryos, brain tissue,
and environmental microbes (15, 27, 56). HCR entails hybridizing
target RNA with a DNA probe that triggers amplification of fluo-
rescently labeled DNA hairpins into polymer chains via a specific
initiator region (14, 15). To directly compare FISH and HCR,
FISH with a dilabeled AlexaFluor 594 EUB338 probe and HCR
with an initiator EUB338 probe and AlexaFluor 594 hairpins were
performed separately on stationary-phase cultured cells embed-
ded in (29A:1B)
4
P
0
and cleared for 5 days. HCR increased the
average fluorescence intensity of
P. aeruginosa
cells by ~68-fold
and
S. aureus
cells by ~42-fold above levels obtained with FISH.
HCR hybridizations in sputum were optimized such that (i)
the EUB338 probe bound and nucleated hairpin polymerization,
(ii) samples did not fluoresce when incubated with both NON338,
the reverse complement of EUB338, and fluorescent hairpins, and
(iii) class/genus-specific probes did not cross-react with other rel-
evant bacteria (see Fig. S5 and S6 in the supplemental material).
The Betaproteobacteria probe BET42a, used for
Achromobac-
ter xylosoxidans
, had weak cross-reactivity with
P. aeruginosa
and
was therefore not used in
P. aeruginosa
culture-positive samples.
Some species-specific probes tested, including those specific for
A. xylosoxidans
, were excluded due to their inability to withstand
the stringent hybridization and wash conditions necessary for
HCR specificity (see Materials and Methods). Object-based colo-
calization analysis after HCR multiplexing was performed to fur-
ther validate HCR specificity. Greater than 90% of objects (dis-
DePas et al.
2
®
mbio.asm.org
September/October 2016 Volume 7 Issue 5 e00796-16
mbio.asm.org
on September 27, 2016 - Published by
mbio.asm.org
Downloaded from
crete HCR-identified cells or aggregates with a size of
4 voxels)
in sputum were concurrently identified by using two separate uni-
versal probes (see Fig. S6). Moreover,
90% of objects identified
by the class/genus-specific probes used in sputum colocalized with
EUB338 but not with other class/genus-specific probes (see
Fig. S6). HCR allowed multiscale visualization of bacteria; low
magnification (e.g.,
10) enabled broad surveying of the sample
(Fig. 1d), and increased magnification (e.g.,
25) enabled single-
cell resolution and revealed the spatial organization of bacteria
and host cells (Fig. 1d).
Once optimized for retention and identification of bacteria in
sputum, we utilized MiPACT-HCR to measure bacterial aggrega-
tion
in situ
. Bacterial aggregates are thought to contribute to the
persistence of pathogen populations in chronic infections, includ-
ing those in CF patients (1, 28–30), yet direct evidence for this is
sparse (6, 19, 31). We examined distribution patterns of
Staphy-
lococcus
sp. in sputum sample 5.1 (culture positive for
S. aureus
and
A. xylosoxidans
) by using a
Staphylococcus
-specific probe. Cul-
tured bacteria were analyzed in parallel with magnification
25
sputum surveys to calibrate our expectations for the signal size of
single bacterial cells (see Fig. S7 in the supplemental material). The
mean fluorescence volume of objects in stationary-phase cultures
of
S. aureus
was 12.1
m
3
. In sputum,
Staphylococcus
cells existed
in a range of intermediate aggregates, with only 6% of objects
being greater than 1,000
m
3
(see Fig. S7a). The
Staphylococcus
size distribution in sputum cleared for 5 or 14 days was similar,
signifying that clearing preserves a range of bacterial aggregate
sizes (see Fig. S2b in the supplemental material). Taking advan-
tage of the large-scale surveying enabled by MiPACT, we next
acquired
10 magnification
Z
-stacks of sputum sample 5.1, ana-
lyzing thousands of objects in sputum volumes of ~0.1 to
0.3 mm
3
. Like the
25 magnification surveys,
Z
-stacks at
10
magnification revealed that
Staphylococcus
was chiefly visible as
small to medium aggregates (85% of objects ranged in size from 50
FIG 1
MiPACT-HCR allows visualization of bacteria in cleared sputum samples. (a) Cartoon depicting the process of embedding and clearing sputum for
visualization of bacteria via HCR. (b) The clearing process for sputum sample 5.1. Each grid square represents 1 mm
2
. (c) Blend projections of five sputum
samples after staining with DAPI (blue) and WGA (orange) from
Z
-stacks acquired with a 10
objective. (d) HCR with a universal bacterial probe (green;
EUB338 with B1 hairpins conjugated to AlexaFluor 647) in sputum sample 5.1. The middle panel is a maximum intensity projection acquired with a 10
objective, and the right panel is a single-plane image acquired with a 25
objective. White arrows indicate PMNs.
3D Imaging of CF Sputum Bacteria after MiPACT
September/October 2016 Volume 7 Issue 5 e00796-16
®
mbio.asm.org
3
mbio.asm.org
on September 27, 2016 - Published by
mbio.asm.org
Downloaded from
FIG 2
Aggregation patterns vary between species. (a) HCR with a
Staphylococcus
-specific probe in sputum sample 5.1 (green). The first panel is a maximum
intensity projection of a
Z
-stack after HCR and staining with DAPI (blue) and WGA (orange), acquired with a 10
objective. The second panel is a maximum
intensity projection of a separate
Z
-stack acquired with a 10
objective while only collecting HCR signal (
Staphylococcus
-specific probe mix with B4 amplifier and
AlexaFluor 488-conjugated B4 hairpins) (7,910 objects analyzed). Each object identified in the second panel’s
Z
-stack was binned according to proportional
object volume (each object’s fluorescent volume relative to the total fluorescent volume for that
Z
-stack; shown in the graph on the right). The top right panel is
a maximum intensity projection of a
Z
-stack acquired with a 25
objective, highlighting a representative region from the same sputum sample. (b to d) The same
analysis was applied to sputum 5.1 using a Betaproteobacteria-specific probe with B4 amplifier and AlexaFluor 488-conjugated B4 hairpin (21,255 objects
analyzed) (b), to sputum 1 with a
Pseudomonas
-specific probe mixture with B4 amplifier and AlexaFluor 647-conjugated B4 hairpins (9,520 objects analyzed) (c),
or a
Streptococcus
-specific probe mixture with AlexaFluor 488-conjugated B4 hairpins (4,603 objects analyzed) (d).
DePas et al.
4
®
mbio.asm.org
September/October 2016 Volume 7 Issue 5 e00796-16
mbio.asm.org
on September 27, 2016 - Published by
mbio.asm.org
Downloaded from
to 1,000
m
3
) (see Fig. 2a
and S7a). In contrast, Betaproteobacte-
ria showed very little aggregation; 71% of objects fell in the small-
est bin (
50
m
3
) (Fig. 2b; see also Fig. S7b in the supplemental
material). Because
S. aureus
is the most common pathogen cul-
tured from CF patients (21), we monitored aggregation in samples
from three distinct areas of sputum 5 (5.1A, 5.1B, and 5.1C) (see
Fig. S8a in the supplemental material). Also, three temporal sam-
ples from patient 5 (5.1, 5.2 [103 days after 5.1], and 5.3 [1 day
after 5.2]) and a sample from patient 4, also culture positive for
S. aureus
, were analyzed. All samples demonstrated a similar pat-
tern of small- to medium-sized aggregates (see Fig. S8b). Next, we
took advantage of the straightforward multiplexing enabled by
HCR (15) to concurrently probe Betaproteobacteria and
Staphy-
lococcus
sp. in different regions of sputum 5.1 (see Fig. S9 in the
supplemental material). Both were present in all areas of sputum
5.1 tested, but their relative abundance differed between regions
(see Fig. S9).
The only organism for which sputum 1 was culture positive
was
P. aeruginosa
, and surveying at
10 and
25 magnifications
with a
Pseudomonas
-specific probe mixture revealed small objects
(0 to 50
m
3
) and large aggregates (
1,000
m
3
) (Fig. 2c; see also
Fig. S7c in the supplemental material), consistent with prior im-
aging of smears of CF sputum and thin sections of explanted CF
patient lungs (6, 19). While surveying sputum sample 1 with the
EUB338 probe, we unexpectedly found bacteria with a distinctive
filamentous morphology. Patient 1 had previously produced a
sputum sample that was culture positive for
Streptococcus angino-
sus
, and probing sputum 1 with a
Streptococcus
-specific probe
mixture revealed a dense bacterial population (Fig. 2d). The larg-
est proportion of
Streptococcus
signal volume (which ranged from
~10 to 300,000
m
3
at
10 magnification) came from large
(
1,000
m
3
) aggregates (Fig. 2d).
Streptococcus
is often missed
in routine clinical culturing, highlighting the gap that is often
observed between culture-dependent and culture-independent
techniques (32, 33).
An important advantage of surveying large volumes at low
magnification is the ability to quickly identify key areas that can
benefit from higher magnification. After performing
10 surveys
in sputum, we focused on large bacterial aggregates with a 25
objective (Fig. 3). Multiplexing of sputum 1 for both
Streptococcus
and
Pseudomonas
revealed that aggregates were mostly monospe-
cies, with little visible interaction (Fig. 3a).
Pseudomonas
aggre-
gates existed in a range of sizes, with large biofilms having diam-
eters up to ~50
m (Fig. 3b). PMNs could be seen surrounding,
and in some cases within, the biofilm structure (Fig. 3b). With
finer resolution, it became apparent that the large
Streptococcus
aggregates visible at
10 had morphologies indicative of associa-
tion with an interior substrate (Fig. 3c and d). To determine the
substrate, we stained samples of sputum 1 with DAPI and WGA
after HCR with
Streptococcus
-specific probes. Staining revealed
that the areas inside
Streptococcus
aggregates were in fact host cells
with single-lobed nuclei (Fig. 3d). Each host cell boundary stained
brightly with WGA, potentially indicative of polysaccharide moi-
eties on the host cell surface (Fig. 3d and e). These results exem-
plify the ability of MiPACT-HCR to identify novel bacterium-host
interactions.
While the importance of aggregative or biofilm modes of
growth in chronic infection is well appreciated (1, 3, 4, 28–30, 34),
the role of growth rate is less so. Recent studies demonstrated slow
in situ S. aureus
growth rates in CF sputum (35) and slow-growth-
specific regulation networks in
P. aeruginosa
(36), underscoring
the importance of careful growth measurements
in situ
for design-
ing
in vitro
models that faithfully recapitulate
in vivo
physiology.
Many species show a linear relationship between growth rate and
rRNA abundance (37), but a number of challenges impede the
calculation of precise growth rates in sputum from FISH data
alone: rRNA abundance can be completely decoupled from
growth rate in some species (37), at low growth rates rRNA abun-
dance ceases to linearly correlate with growth rate (7), and sputum
autofluorescence can overwhelm signals from slowly growing cells
(see Fig. S4 in the supplemental material). To address these prob-
lems, we refrained from estimating specific growth rates of indi-
vidual cells, instead opting to describe the growth rates of bacterial
populations with respect to logarithmic- and stationary-phase
standards, analyzed in parallel. We first verified that the FISH
signal of both
S. aureus
and
A. xylosoxidans
decreased in stationary
phase (see Fig. S10a in the supplemental material). We then de-
termined that FISH signal from logarithmic cells did not substan-
tially decay even after 14 days of clearing (see Fig. S10b). Lastly, we
used HCR to distinguish bacterial signals from background auto-
fluorescence and to select for the desired genus in a mixed popu-
lation. For analysis, HCR-identified objects were outlined and
EUB338 FISH fluorescence (the proxy for growth rate) within the
outlines was quantified (Fig. 4a). EUB338 was chosen as the FISH
probe due to its robustness and hybridization to a separate rRNA
locus, preventing probe competition.
The growth rate measurements described above were per-
formed on objects identified with a
Staphylococcus
-specific HCR
probe from portions of sputum 5.1 taken from distinct areas of the
sample (5.1A, -B, and -C, with subsamples 5.1A1, 5.1A2, etc.)
(Fig. 4a and b). We then calculated the percentage of objects that
crossed a threshold above which 90% of logarithmic-phase cul-
tured cells fell (corresponding to a doubling time of
1 h). Re-
gions 5.1A1, 5.1B1, and 5.1C1 demonstrated mostly low growth
rates, with 0%, 4.3%, and 0.2% of objects reaching signal thresh-
old (Fig. 4c). Interestingly, subsample 5.1C2 demonstrated an in-
crease in growth rate compared to its neighbors, with 20.8%, of
objects reaching threshold (no objects in 5.1A2 and 7.0% of ob-
jects in 5.1B2 reached threshold) (Fig. 4d). Temporal samples
from patient 5 (sputum samples 5.2 and 5.3) did not reach thresh-
old (Fig. 4e). Samples 4A and -B and -C, from a different
S. aureus
culture-positive patient, contained slowly growing bacteria as
well, with only 2.7%, 4.1%, and 2.0% of objects, respectively,
above threshold (Fig. 4f). In order to determine if aggregate size
correlated with our proxy for growth rate, we separated
S. aureus
objects (from Fig. 4c to f) into quartiles with respect to object size
and plotted against mean fluorescence intensity from FISH. Flu-
orescence intensity increased significantly with increasing object
size, signifying that larger aggregates may have experienced higher
growth rates (see Fig. S10c and d in the supplemental material).
This was possibly due to the greater susceptibility of single cells to
antibiotics (34). Interestingly, cultured, planktonic
S. aureus
cells
also showed a positive correlation between object size and fluores-
cence intensity (see Fig. S10e). This is consistent with previous
studies showing that RNA abundance and cell size increase with
higher growth rates (38). Further study would be needed to deter-
mine what, if any, correlation exists between aggregate size and
cell size
in situ
.
Lastly, as
A. xylosoxidans
is subject to the same
in vivo
condi-
tions as
S. aureus
in these samples, we assayed growth rates of
3D Imaging of CF Sputum Bacteria after MiPACT
September/October 2016 Volume 7 Issue 5 e00796-16
®
mbio.asm.org
5
mbio.asm.org
on September 27, 2016 - Published by
mbio.asm.org
Downloaded from
FIG 3
Pseudomonas
and
Streptococcus
biofilm structure. (a) A maximum intensity projection was generated after HCR was performed on sputum 1 with a
Pseudomonas
-specific probe mixture (with B1 hairpins conjugated to AlexaFluor 647) and a
Streptococcus
-specific probe mixture (with B4 hairpins conjugated
to AlexaFluor 488). (b) Maximum intensity projections showing
Pseudomonas
aggregates from sputum 1 after HCR with a
Pseudomonas
probe mixture and B4
hairpins conjugated to AlexaFluor 488 and DAPI staining. (c) Blend projection of a
Streptococcus
biofilm from sputum 1 (HCR with
Streptococcus
probe mixture
with B4 hairpins conjugated to AlexaFluor 488). (d) Blend projections showing, stepwise, a
Streptococcus
aggregate (top; green), DAPI (blue), and WGA (orange)
staining of host cells (middle), and an overlay of the two showing the arrangement of the
Streptococcus
biofilm around WGA-stained host cells (bottom). (e)
Maximum intensity projection of HCR-identified
Streptococcus
(green), DAPI (blue), and WGA (orange) staining in sputum 1.
DePas et al.
6
®
mbio.asm.org
September/October 2016 Volume 7 Issue 5 e00796-16
mbio.asm.org
on September 27, 2016 - Published by
mbio.asm.org
Downloaded from
FIG 4
Growth rate estimates of CF pathogens
in situ.
(a) Diagram showing the process of estimating growth rates
in situ
. Samples were first stained with a
species-specific B4 amplifier HCR probe, using B4 hairpins conjugated to AlexaFluor 488. Samples were then stained with the universal bacterial FISH probe
EUB338, conjugated to two Cy5 fluorophores. Masks were made based upon HCR signal, and fluorescence intensity from FISH was quantified within each mask.
(b) The basic sputum sampling technique. (c) For this and subsequent panels,
Z
-stacks of cultured cells and sputum samples were acquired with a 25
objective
in parallel. The average fluorescence intensity of the FISH channel of each object is plotted on the
x
axis as a histogram. The blue line denotes the bin above which
90% of the logarithmic objects fell (for each particular experimental set). Growth rate analysis was performed on three distinct regions of sputum sample 5.1:
5.1A1 (409 objects analyzed), 5.1B1 (697 objects analyzed), and 5.1C1 (575 objects analyzed) (c), and on 5.1A2 (418 objects analyzed) 5.1B2 (1,087 objects
analyzed), and 5.1C2 (419 objects analyzed) (d). (e) Analysis of temporal samples 5.2 (520 objects analyzed) and 5.3 (893 objects analyzed). (f) Analysis of three
distinct regions of sputum 4: 4A (1,067 objects analyzed), 4B (73 objects analyzed), and 4C (599 objects analyzed). (g) Analysis of Betaproteobacteria from
samples 5.1A2 (1,919 objects analyzed), 5.1B2 (2,351 objects analyzed), and 5.1C2 (1,523 objects analyzed).
3D Imaging of CF Sputum Bacteria after MiPACT
September/October 2016 Volume 7 Issue 5 e00796-16
®
mbio.asm.org
7
mbio.asm.org
on September 27, 2016 - Published by
mbio.asm.org
Downloaded from