Fluorescence dequenching makes haem-free soluble guanylate cyclase detectable in living cells.

Fluorescence Dequenching Makes Haem-Free Soluble
Guanylate Cyclase Detectable in Living Cells
Linda S. Hoffmann1¤, Peter M. Schmidt2, Yvonne Keim1, Carsten Hoffmann3, Harald H. H. W. Schmidt4,
Johannes-Peter Stasch1,5*
1 Pharma Research Centre, Bayer HealthCare, Wuppertal, Germany, 2CSIRO Materials, Science and Engineering, Parkville, Victoria, Australia, 3 Julius-Maximilians-University
of Wuerzburg, School of Pharmacology and Toxicology, Wuerzburg, Germany, 4Department of Pharmacology, Maastricht University, Maastricht, The Netherlands,
5Martin-Luther-University, School of Pharmacy, Halle, Germany
Abstract
In cardiovascular disease, the protective NO/sGC/cGMP signalling-pathway is impaired due to a decreased pool of NO-
sensitive haem-containing sGC accompanied by a reciprocal increase in NO-insensitive haem-free sGC. However, no direct
method to detect cellular haem-free sGC other than its activation by the new therapeutic class of haem mimetics, such as
BAY 58-2667, is available. Here we show that fluorescence dequenching, based on the interaction of the optical active
prosthetic haem group and the attached biarsenical fluorophor FlAsH can be used to detect changes in cellular sGC haem
status. The partly overlap of the emission spectrum of haem and FlAsH allows energy transfer from the fluorophore to the
haem which reduces the intensity of FlAsH fluorescence. Loss of the prosthetic group, e.g. by oxidative stress or by
replacement with the haem mimetic BAY 58-2667, prevented the energy transfer resulting in increased fluorescence. Haem
loss was corroborated by an observed decrease in NO-induced sGC activity, reduced sGC protein levels, and an increased
effect of BAY 58-2667. The use of a haem-free sGC mutant and a biarsenical dye that was not quenched by haem as controls
further validated that the increase in fluorescence was due to the loss of the prosthetic haem group. The present approach
is based on the cellular expression of an engineered sGC variant limiting is applicability to recombinant expression systems.
Nevertheless, it allows to monitor sGC’s redox regulation in living cells and future enhancements might be able to extend
this approach to in vivo conditions.
Citation: Hoffmann LS, Schmidt PM, Keim Y, Hoffmann C, Schmidt HHHW, et al. (2011) Fluorescence Dequenching Makes Haem-Free Soluble Guanylate Cyclase
Detectable in Living Cells. PLoS ONE 6(8): e23596. doi:10.1371/journal.pone.0023596
Editor: Carlo Gaetano, Istituto Dermopatico dell’Immacolata, Italy
Received March 28, 2011; Accepted July 21, 2011; Published August 17, 2011
Copyright: � 2011 Hoffmann et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits
unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: The authors have no support or funding to report.
Competing Interests: Linda S. Hoffmann, Yvonne Keim and Johannes-Peter Stasch were at the time of the study employed by Bayer HealthCare and Peter M.
Schmidt is employed by CSIRO Molecular & Health Technologies. JPS holds more than 50 patent applications related to sGC stimulators, such as BAY 41-2272, and
sGC activators, such as BAY 58-2667. Details of which are available on request. The sGC stimulator riociguat (BAY 63-2521) and the sGC activator cinaciguat (BAY
58-2667) are currently in clinical development. This does not alter the authors’ adherence to all the PLoS ONE policies on sharing data and materials, as detailed
online in the guide for authors.
* E-mail: johannes-peter.stasch@bayer.com
¤ Current address: Centre for Vascular Research, School of Medical Science (Pathology) and Bosch Institute, University of Sydney, Sydney, New South Wales,
Australia
Introduction
The heterodimeric a/b haemprotein soluble guanylate cyclase
(sGC) is the physiological receptor for the endogenous gaseous
messenger nitric oxide (NO). The prosthetic haem group is non-
covalently bound to the b subunit via the haem binding motif Y-x-
S-x-R and the axial haem ligand H105 [1,2]. Binding of NO to the
reduced central iron atom of the haem moiety results in an up to
200-fold increase of the conversion rate of guanosine triphosphate
(GTP) into the second messenger cyclic guanosine monophosphate
(cGMP). cGMP is a key modulator of the cardiovascular system,
involved in processes such as smooth-muscle cell relaxation and
inhibition of platelet aggregation [3,4]. Impairment of the NO/
sGC/cGMP pathway has been linked to the development of
various cardiovascular diseases such as heart failure or arterial
hypertension [5].
The prosthetic haem group of sGC has a pivotal role in the
activation and stabilisation of the enzyme and its oxidation or
removal renders sGC insensitive to NO. Oxidative stress, i.e. the
formation of reactive oxygen species (ROS) such as O2
2, has been
associated with various cardiovascular diseases [5]. ROS are
known to interfere with the NO/sGC/cGMP signalling-pathway
via O2
2-mediated scavenging of NO and intermediate peroxyni-
trite formation, which in turn oxidizes the sGC haem to the NO-
insensitive Fe3+ state [6,7,8]. Furthermore, oxidation of the sGC
haem strongly reduces its affinity towards sGC [9,10] which leads
to subsequent loss of the haem [11,12]. There is a solid body of
evidence that sGC is redox regulated and exists in the haem-free
form in vivo [11,13,14,15]. In addition, it was shown that haem-
free sGC is prone to ubiquitin-mediated degradation [13,16].
The discovered new NO-independent drug classes, haem-
independent sGC activators and haem-dependent sGC stimulators,
have been used to distinguish between reduced haem-containing
and haem-free sGC. sGC activators such as cinaciguat (BAY 58-
2667) mimic the haem group, compete with different porphyrins,
and bind in the orphaned sGC haem pocket via the haem anchoring
residues Y135 and R139 [9,13,17,18]. This unique mode of action
allows this structural class to activate haem-free sGC opening up the
PLoS ONE | www.plosone.org 1 August 2011 | Volume 6 | Issue 8 | e23596

possibility of new mechanism-based therapies for those cardiovas-
cular diseases that are associated with oxidative stress [10,13].
Cinaciguat is currently in clinical development to treat acute
decompensated heart failure [11]. In contrast, sGC stimulators like
BAY 41-2272 or riociguat (BAY 63-2521) do not activate oxidized
or haem-free sGC but show a strong synergism with NO [12].
Despite these therapeutic breakthroughs, the physiological
existence of haem-free sGC is a matter of considerable debate
due to the lack of a direct method to detect haem-free sGC in
living cells. This has left the field with the only option to assay
cGMP levels or its functional effects upon exposure to sGC
activators when estimating haem-free sGC levels. As a first step to
overcome this scientific road block, we aimed to establish a specific
and sensitive imaging method based on fluorescence dequenching
allowing the direct tracking of the sGC haem status in living cells.
As this approach is based on the overexpression of a fluorophore-
labelled sGC in a cellular system, its general applicability for in-
vivo systems is strongly hampered. Nevertheless, the method
allowed for the first time to monitor changes in sGC haem status in
living cells under oxidative stress conditions underlining its general
usefulness. Future enhancements of the present approach could be
based on an engineered sGC-binding a fluorophore like FlAsH,
which, in combination with established knock-in techniques, might
offer the chance to track changes in the sGC haem status under
more physiological conditions.
Results
Identification of an appropriate location for the
tetracysteine motif
Under native conditions, the presence of the haem group in
sGC and its spatial vicinity to the dye are the prerequisites to
successfully apply fluorescence dequenching as described in this
paper. We assumed that the N-terminus would be in proximity of
the haem as observed in the crystal structure of the haem nitric
oxide oxygen (H-NOX) domain of Nostoc sp. [2].
To identify a position for the tetracysteine (TC, CCPGCC)
motif that will not affect the haem binding of sGC, the TC motif
was fused either to the N- or C-terminus of sGC or inserted at
different positions within the sGC b1 coding sequence. Promising
intramolecular positions were identified using two approaches.
Firstly, we used the molecular model of the sGC haem binding
domain as described by Rothkegel et al. [19] to identify positions
apparently in the proximity of the haem binding pocket. Using this
approach, the target positions aa170–175 and aa111–116 were
selected for insertion of the TC motif. Secondly, we focused on the
region of aa231–310 as previous photoaffinity labelling studies
with BAY 58-2667 indicated that this region might be in close
proximity to the haem binding pocket [14]. Within this region, we
changed positions aa243–248, aa239–244, and aa257–262 into
the TC motif. These stretches have been chosen as they already
contain a proline similar to the introduced TC motif, CCPGCC.
As the introduction of a proline can affect protein structures with
higher likelihood than other exchanges this strategy aimed to
reduce putative conformational changes to a minimum. The
described positions are illustrated in Fig. S1.
To verify whether the introduction of the TC motif at different
positions had a negative impact on sGC enzyme activity, all TC-b1
constructs were screened using the cellular cGMP reporter system
described earlier [9,20]. The dose-response curves of haem-
containing WT sGC-expressing cells are shown in Fig. 1A. BAY
58-2667 activated WT sGC under control conditions up to 10-
fold. The maximum activity was increased to 20-fold once BAY
58-2667 was combined with the haem-oxidizing sGC inhibitor
ODQ. The sGC stimulator BAY 41-2272 stimulated sGC 5-fold
and addition of NO increased stimulation to 15-fold.
From all sGC-variants tested, only the TC aa243–248 construct
(Fig. 1B) showed an sGC activation pattern very similar to WT
sGC. BAY 58-2667 activated the construct 9-fold and this was
increased to 18-fold upon addition of ODQ. BAY 41-2272
stimulated the construct 5-fold which was increased to 15-fold by
NO. Therefore, this construct was selected for further use and is
referred to as TC4 in the following. All other sGC variants showed
altered activation profiles compared to the WT enzyme indepen-
dent whether the TC motif was introduced at the N- or C-
terminus or within the sequence of the enzyme (Fig. S2).
To test whether labelling with FlAsH would impact the enzyme
activity cGMP read-out was performed as described after cells
were labelled with FlAsH (Fig. 1C). BAY 58-2667 activated
FlAsH-labeled TC4-WT-sGC up to 11-fold and this was increased
to 18-fold upon addition of ODQ. The 26-fold stimulation with
BAY 41-2272 was increased to 32-fold when combined with NO.
Both results, the synergistic effect of NO and the increase in BAY
58-2667-induced sGC-activity when combined with ODQ
indicated the presence of the haem group under native conditions
in FlAsH-labelled cGMP-reporter cells and show that neither
FlAsH nor the labeling procedure had a negative impact on the
enzyme’s function.
Optimization of concentrations of different haem
oxidizing compounds
TC4-WT sGC expressing cells were incubated with 100 nM
BAY 58-2667 and increasing concentrations of NS 2028, rotenone
or ODQ to establish the optimal concentration of the different
haem oxidants needed for an efficient oxidation of the sGC haem
moiety (Fig. S3). NS 2028 and ODQ have been described to
directly oxidize the haem group [21,22] whereas rotenone is
known to increase the intracellular ROS concentration via
inhibition of the mitochondrial complex I [23] without having a
direct effect on the sGC haem group as NS 2028 or ODQ. BAY
58-2667 activated TC4-WT sGC 6-fold which was increased upon
addition of the oxidants. Rotenone and the more effective ODQ
derivate NS 2028 were selected for use in further experiments to
study the effects of a ROS increasing compound as well as a direct
sGC inhibitor.
Effects of BAY 58-2667, NS 2028 and rotenone on
fluorescence of FlAsH-labelled TC4-WT sGC
To test the hypothesis whether loss of the sGC haem group
leads to fluorescence dequenching of FlAsH, TC4-WT sGC
expressing cGMP reporter cells were incubated for 90 min with
BAY 58-2667 which has been shown to replace the native
prosthetic group in the haem pocket [9], the sGC oxidant NS
2028, or the ROS-inducing compound rotenone. The latter two
compounds lead to oxidation of the haem moiety facilitating the
loss of the prosthetic group [8,24]. Changes in single cell
fluorescence were recorded in a time series of 120 laser scans
over 90 min. Under control conditions, 90 min of laser-induced
bleaching of FlAsH reduced the observed fluorescence to 2162%
(Fig. 2). Incubation of FlAsH-labelled cells with different
concentrations of BAY 58-2667 increased the residual FlAsH-
fluorescence to a maximum of 7264% compared to control
(Fig. 2A). Next we examined whether treatment with the haem
oxidants and subsequent loss of the sGC haem group dequenched
FlAsH fluorescence and resulted in higher fluorescence compared
to control. Treatment of cells with 1, 30 and 100 mM rotenone for
90 min led to an significant increase of residual fluorescence
Detecting Haem-Free sGC
PLoS ONE | www.plosone.org 2 August 2011 | Volume 6 | Issue 8 | e23596

intensity to 4668% (p,0.005), 3264% (p, 0.005) and 3064%
(p,0.005), respectively, when compared to control (Fig. 2B).
Treatment of cells with NS 2028 resulted in significant higher
fluorescence intensity compared to control as well (Fig. 2C). This
effect was significant for 1, 10 and 100 mM NS 2028 which
increased fluorescence intensity to 4965% (p,0.005), 3764%
(p,0.005) and 3565% (p,0.005), respectively. These data suggest
the replacement of the haem group by BAY 58-2667 or oxidation-
induced haem loss induced by rotenone or NS 2028. Represen-
tative traces of the measurement of fluorescence are shown in
Fig. 2. To measure the fluorescence intensity single cells were
marked with ‘‘regions of interest’’ (ROI) over which fluorescence
was integrated. ROIs over non-transfected cells and without cells
were used to measure background levels and subtracted from the
single cell values. The decrease in fluorescence observed after
90 min is due to bleaching effects which are the same in treated
and untreated cells. Therefore, comparison of the residual
fluorescence intensity in treated and untreated cells shows
fluorescence dequenching.
Effects of NS 2028 on FlAsH-labelled TC4-haem-free sGC
and ReAsH-labelled TC4-WT sGC
To validate that the observed increases in fluorescence were
indeed based on dequenching, two controls were used. First, the
haem-free mutant TC4-Y135A/R139A, was used to exclude that
any increases in fluorescence were due to unknown or artificial, i.e.
haem-independent mechanisms. For that purpose, TC4-Y135A/
R139A sGC expressing cells were incubated with 100 mM NS 2028
for 90 min. In contrast to TC4-WT sGC, the fluorescence was only
slightly and not significantly increased (Table 1). Second, we used
ReAsH-labelled TC4-WT sGC. As shown in Fig. S4, the emission
spectrum of ReAsH overlaps only slightly with the absorption
spectrum of haem-containing sGC. We therefore assumed that the
sGC haem would quench ReAsH to a much lesser extent than
observed for FlAsH. Again, treating the cells with 100 mM NS 2028
for 90 min did not affect residual fluorescence intensity of ReAsH
(Table 1) indicating that the observed increase for FlAsH
fluorescence has been indeed due to haem loss.
Influence of NS 2028 and rotenone pre-treatment on
activity of TC4-WT sGC
To test if treatment with haem oxidants affects the haem status
of TC4-WT sGC as implicated by increased FlAsH fluorescence,
its activity was measured 90 min after pre-treatment with 100 mM
NS 2028 or rotenone (Fig. 3). This diminished BAY 41-2272-
induced cGMP levels and abolished the effects of ODQ on BAY
58-2667-induced sGC activity as well as the synergistic effect of
NO on BAY 41-2272 induced sGC stimulation. Similar results
were obtained for WT-sGC expressing cells after pre-treatment
with NS 2028 or rotenone (Fig. S5). These data indicated that
both, the haem oxidant NS 2028, and rotenone, indeed affected
the oxidation state of the haem group.
Figure 1. Activation pattern of sGC with tetracysteine (TC) motif. Concentration response curves of WT sGC (A) and sGC with an
intramolecular tetracysteine motif without (B) or with FlAsH labelling (C) are shown. cGMP reporter cells were transiently cotransfected with WT a1
subunit and WT or a TC motif carrying b1 subunit as indicated in the respective figures. Cells were incubated with increasing concentrations of BAY
58-2667 or BAY 41-2272 alone or in combination with 10 mM ODQ or 10 nM DEA/NO (NO), respectively. sGC activity is represented as x-fold
stimulation compared to transfected control cells. Data are means 6 S.E.M. from 5–14 independent experiments, performed in duplicate. Following
basal activities were measured: (A) 10732 relative light units (RLUs), (B) 9720 RLUs, (C) 3541 RLUs.
doi:10.1371/journal.pone.0023596.g001
Detecting Haem-Free sGC
PLoS ONE | www.plosone.org 3 August 2011 | Volume 6 | Issue 8 | e23596

Figure 2. Fluorescence of FlAsH-labelled TC4-WT sGC. cGMP reporter cells were cotransfected with WT a1 sGC and TC4-WT b1 sGC and
labelled with FlAsH. The cells were then incubated with 0.1, 1 or 10 mM BAY 58-2667 (A) 1, 30 or 100 mM rotenone (B) or 1, 10 or 100 mM NS 2028 (C)
for 90 min. Fluorescence intensity of single cells was monitored in a time series of 120 laser scans and is expressed as % of start value which was
obtained before application of the test compounds. Data are means 6 S.E.M. from 9–28 single cells assayed on different days. *p,0.05; ***p,0.005:
Student’s t-test compared to control. Representative traces of the fluorescence measurement as % of start value are given in the right part of the
figure.
doi:10.1371/journal.pone.0023596.g002
Detecting Haem-Free sGC
PLoS ONE | www.plosone.org 4 August 2011 | Volume 6 | Issue 8 | e23596

Effects of NS 2028 and rotenone on TC4-WT sGC protein
levels
Haem oxidation leads also to sGC degradation [13,15,17,25].
Thus the effects of 90 min treatment with 100 mM NS 2028 or
rotenone on sGC protein levels were tested in TC4-WT sGC
expressing cells (Table 2). Although NS 2028 did not affect a1 sGC
protein levels after 90 min treatment, the haem-binding b1 subunit
of sGC was significantly reduced to 78617% of control (p,0.005).
90 min treatment with rotenone significantly decreased protein
levels of both sGC a1 and b1 (65610 and 81611%, respectively;
p,0.005). Incubating WT sGC expressing cells under the same
conditions produced similar results (Table S1). In contrast, when
TC4-Y135A/R139A expressing cells were incubated with
100 mM NS 2028 for 90 min no change in sGC protein levels
was observed (data nor shown).
Discussion
We here show that it is possible to directly monitor the sGC
haem oxidation state in intact cells. By inserting the TC motif into
the coding sequence of the b1 subunit of sGC in transiently
transfected Chinese hamster ovary cells, we were able to monitor
sGC fluorescence intensity upon addition of the haem mimetic
BAY 58-2667, the haem oxidant, NS 2028, and the ROS
generating compound, rotenone. The test compounds lead to an
increase of FlAsH fluorescence intensity when compared to
control. This was neither observed in the ReAsH-labelled negative
control nor when a haem-free sGC mutant was expressed and
labelled with FlAsH. These direct fluorescence measurements
were accompanied by indirect probing of sGC activity via the
established pharmacological tools BAY 58-2667 and BAY 41-2272
and by monitoring sGC stability as another marker for haem loss.
The NO/sGC/cGMP pathway has been shown to be impaired
in various cardiovascular diseases mainly by a reduced bioavail-
ability of NO and, in parallel, by reducing the sensitivity of sGC
towards its agonist NO [5,7,25,26,27]. Oxidation or loss of the
sGC prosthetic haem group may be the cause for this reduced
sensitivity as oxidized or haem-free sGC is unresponsive to NO
[11,13]. There is compelling evidence that haem-free sGC exists
under physiological conditions and that this pool is increased in
pathophysiological situations associated with the increased pro-
duction of ROS [13,24]. Very recently it was shown for Manduca
sexta sGC that haem oxidation leads to haem loss because the ferric
state is unstable. In contrast, the ferrous state is highly stable and
resistant to haem loss [8]. Furthermore, haem-free sGC is prone to
ubiquitin-mediated degradation and this might be at least partly
the cause for the observed decreased sGC protein levels in animal
models of cardiovascular diseases [13,16,25,28].
Although imaging methods employing FRET and non-FRET
methods have extensively been used to monitor intracellular
cGMP dynamics [29] no such method is available to investigate
the haem content of sGC. Therefore, the intracellular existence of
haem-free sGC in living cells and its regulation is still a matter of
debate.
To investigate conformational changes of proteins several
studies used the ability of the haem group to quench the
fluorescence of other chromophores if the spectra of haem and
the respective chromophore overlap. Kosarikov and coworkers
[30] used the intrinsic fluorescence of tryptophans to investigate
conformational changes of sGC. They observed that the intrinsic
fluorescence of sGC decreased upon addition of NO and suggested
that the intrinsic tryptophan fluorescence is quenched to some
extend by sGC’s haem group due to NO-induced conformational
Table 1. Effects of 100 mM NS 2028 on fluorescence intensity
of FlAsH labelled haem- free TC4-Y135A/R139A sGC (A) and
ReAsH labelled TC4-WT sGC (B) after 90 min incubation.
TC4-Y135A/R139A sGC TC4-WT sGC
FlAsH ReAsH
Control 12.6961.16 65.9967.16
100 mM NS 2028 16.5663.36 73.26611.53
cGMP reporter cells were cotransfected with WT a1 sGC and the respective b1
sGC and labelled with FlAsH or ReAsH. Fluorescence intensity of single cells was
monitored in a time series of 120 laser scans and is expressed as % of start value
which was obtained before application of the test compounds. Data are means
6 S.E.M. from 3–5 single cells assayed on different days.
doi:10.1371/journal.pone.0023596.t001
Figure 3. Effect of 90 min incubation with NS 2028 and rotenone on sGC activity. cGMP reporter cells expressing TC4-WT sGC were pre-
incubated with 100 mM NS 2028 or rotenone and then incubated with increasing concentrations of BAY 58-2667, BAY 41-2272 alone or in
combination with 10 mM ODQ or 10 nM DEA/NO (NO), respectively. Enzyme activity is expressed as x-fold stimulation compared to pre-treated but
not stimulated control. Data are means 6 S.E.M. from 3–10 independent experiments, performed in duplicate. A basal activity of 7747 relative light
units was measured. A) is identical to Fig. 1B.
doi:10.1371/journal.pone.0023596.g003
Detecting Haem-Free sGC
PLoS ONE | www.plosone.org 5 August 2011 | Volume 6 | Issue 8 | e23596