Molecular and biochemical characterization of the bicarbonate-sensing soluble adenylyl cyclase from a bony fish, the rainbow trout Oncorhynchus mykiss

Abstract

Soluble adenylyl cyclase (sAC) is a
HC

O
3

-

-stimulated enzyme that produces the ubiquitous signalling molecule cAMP, and deemed an evolutionarily conserved acid-base sensor. However, its presence is not yet confirmed in bony fishes, the most abundant and diverse of vertebrates. Here, we identified sAC genes in various cartilaginous, ray-finned and lobe-finned fish species. Next, we focused on rainbow trout sAC (rtsAC) and identified 20 potential alternative spliced mRNAs coding for protein isoforms ranging in size from 28 to 186 kDa. Biochemical and kinetic analyses on purified recombinant rtsAC protein determined stimulation by
HC

O
3

-

at physiologically relevant levels for fish internal fluids (EC50 ∼ 7 mM). rtsAC activity was sensitive to KH7, LRE1, and DIDS (established inhibitors of sAC from other organisms), and insensitive to forskolin and 2,5-dideoxyadenosine (modulators of transmembrane adenylyl cyclases). Western blot and immunocytochemistry revealed high rtsAC expression in gill ion-transporting cells, hepatocytes, red blood cells, myocytes and cardiomyocytes. Analyses in the cell line RTgill-W1 suggested that some of the longer rtsAC isoforms may be preferentially localized in the nucleus, the Golgi apparatus and podosomes. These results indicate that sAC is poised to mediate multiple acid-base homeostatic responses in bony fishes, and provide cues about potential novel functions in mammals.

Full text

royalsocietypublishing.org/journal/rsfs
Research
Cite this article: Salmerón C et al. 2021
Molecular and biochemical characterization of
the bicarbonate-sensing soluble adenylyl
cyclase from a bony fish, the rainbow trout
Oncorhynchus mykiss.Interface Focus 11:
20200026.
https://doi.org/10.1098/rsfs.2020.0026
Accepted: 9 December 2020
One contribution of 13 to a theme issue
‘Carbon dioxide detection in biological
systems’.
Subject Areas:
biochemistry, environmental science,
biocomplexity
Keywords:
cAMP, CO
2
, Golgi, microdomain,
pH sensing, sAC
Author for correspondence:
Martin Tresguerres
e-mail: mtresguerres@ucsd.edu
Electronic supplementary material is available
online at https://doi.org/10.6084/m9.figshare.
c.5260354.
Molecular and biochemical
characterization of the bicarbonate-
sensing soluble adenylyl cyclase from
a bony fish, the rainbow trout
Oncorhynchus mykiss
Cristina Salmerón1,2, Till S. Harter1, Garfield T. Kwan1, Jinae N. Roa1, Salvatore
D. Blair3,4, Jodie L. Rummer5, Holly A. Shiels6, Greg G. Goss3, Rod W. Wilson7
and Martin Tresguerres1
1
Scripps Institution of Oceanography, University of California San Diego, La Jolla, CA, USA
2
Department of Pharmacology, University of California San Diego, San Diego, CA, USA
3
Department of Biological Sciences, University of Alberta, Edmonton, Alberta, Canada
4
Department of Biology, Winthrop University, Rock Hill, SC, USA
5
Australian Research Council Centre of Excellence for Coral Reef Studies, James Cook University, Townsville,
Queensland, Australia
6
Division of Cardiovascular Sciences, Faculty of Biology, Medicine and Health, University of Manchester, UK
7
Department of Biosciences, University of Exeter, Exeter, UK
MT, 0000-0002-7090-9266
Soluble adenylyl cyclase (sAC) is a HCO 
3-stimulated enzyme that produces
the ubiquitous signalling molecule cAMP, and deemed an evolutionarily
conserved acid–base sensor. However, its presence is not yet confirmed in
bony fishes, the most abundant and diverse of vertebrates. Here, we ident-
ified sAC genes in various cartilaginous, ray-finned and lobe-finned fish
species. Next, we focused on rainbow trout sAC (rtsAC) and identified 20
potential alternative spliced mRNAs coding for protein isoforms ranging
in size from 28 to 186 kDa. Biochemical and kinetic analyses on purified
recombinant rtsAC protein determined stimulation by HCO 
3at physiologi-
cally relevant levels for fish internal fluids (EC
50
∼7 mM). rtsAC activity was
sensitive to KH7, LRE1, and DIDS (established inhibitors of sAC from other
organisms), and insensitive to forskolin and 2,5-dideoxyadenosine (modulators
of transmembrane adenylyl cyclases). Western blot and immunocytochemistry
revealed high rtsAC expression in gill ion-transporting cells, hepatocytes,
red blood cells, myocytes and cardiomyocytes. Analyses in the cell line
RTgill-W1 suggested that some of the longer rtsAC isoforms may be
preferentially localized in the nucleus, the Golgi apparatus and podosomes.
These results indicate that sAC is poised to mediate multiple acid–base
homeostatic responses in bony fishes, and provide cues about potential
novel functions in mammals.
1. Background
The enzyme soluble adenylyl cyclase (sAC, adcy10) is directly stimulated by
bicarbonate ions (HCO 
3) to produce the ubiquitous messenger molecule
cyclic adenosine monophosphate (cAMP) [1]. In the presence of carbonic anhy-
drase (CA), carbon dioxide (CO
2
), proton (H
+
) and HCO 
3are in a near
instantaneous equilibrium, and this enables sAC to act as a general acid–base
sensor for multiple physiological processes (reviewed in [2,3]).
The original biochemical studies in the 1970s reported a ‘soluble’source of
cAMP in rat testes that was distinct from the traditional transmembrane ACs
(tmACs) based on the potent stimulation induced by Mn
2+
and its insensitivity
© 2021 The Author(s) Published by the Royal Society. All rights reserved.

to hormones and fluoride [4,5]. However, it took about
20 years until the Mn
2+
-stimulated AC enzyme was finally
identified as sAC [6]. It was simultaneously discovered that
the two catalytic domains of sAC, C1 and C2, were more
similar to ACs from cyanobacteria than to the tmACs that
first appeared in metazoans, indicating that sAC is an evolu-
tionarily conserved enzyme [6]. Shortly after, sAC activity
was found to be directly stimulated by HCO 
3[1], which
led to a flurry of studies on its potential role as an acid–
base sensor in diverse organisms. Initially, the lack of sAC
genes in the genomes of Drosophila and Caenorhabditis elegans
was interpreted as evidence that sAC had been lost in most
animal phyla [7]; however, the subsequent identification
and characterization of sAC in sea urchin [8], shark [9] and
coral [10] confirmed that sAC indeed is an evolutionarily
conserved acid–base sensing enzyme.
The stimulating effect of HCO 
3on the activity of sAC is
species-specific. In mammalian sAC, the HCO 
3half maximal
stimulation (EC
50
) is approximately 20 mM [1,11,12], whereas
that of shark sAC is only approximately 5 mM [9]; and these
values closely match average HCO 
3concentrations in the
internal fluids of these organisms. As a result, cAMP pro-
duction by sAC is maximally modulated by small HCO 
3
fluctuations around species-specific homeostatic set points
[13,14]. In addition, in vitro characterization of sAC cAMP-
producing activity revealed a potent stimulation by millimolar
[Mn
2+
] (reviewed in [15]). However, the physiologically rel-
evant catalytic metals that sustain HCO 
3stimulation in vivo
are Mg
2+
and Ca
2+
for mammalian sAC [12], and Mg
2+
and
Mn
2+
for shark sAC [9]. Advances from pharmacological
studies have provided essential tools for studying sAC activity
in vivo. Both mammalian and shark sACs are inhibited by
derivatives of catechol estrogen (dCE) and by the small mol-
ecules KH7 and LRE1 [9,16–19], and are insensitive to the
widely used tmAC stimulator, forskolin, and to the tmAC
inhibitor, 2,2-dideoxyadenosine (DDA) [16,20]. Other work
has described the presence of numerous alternative promoters
in the sAC gene as well as extensive alternative splicing, which
have been characterized to different extents in some mammals
[21–25], corals [10] and sharks [26].
As a result of the most expansive adaptive radiations among
vertebrates, ray-finned fishes (Actinopterygii) make up half of
all vertebrates on Earth and are of great ecological and economi-
cal importance. As is typical for water-breathing animals, fishes
may routinely experience severe acid–base disturbancessuch as
exercise-induced acidosis, postprandial alkaline tide and
environmental hypercapnia that can induce large fluctuations
in plasma [HCO 
3] ranging nearly 0 to over 50 mM (reviewed
in [27,28]). Although sAC is an ideal candidate to sense
acid–base disturbances and trigger compensatory responses
in ray-finned fishes, the apparent absence of sAC genes in the
zebrafish, puffer fish and medaka genomes initially suggested
that sAC had been lost in ray-finned fishes, and that they may
use fundamentally distinct mechanisms for acid–base sensing
[9]. However, the apparent absence of sAC genes in teleost
fishes could be explained by the high structural complexity of
the sAC gene that may confound its identification using bioin-
formaticstools (reviewedin [15]). Indeed, the presence of sAC in
ray-finned fishes is supported by sAC-like immunoreactivity in
the intestine of the toadfish (Opsanus beta) [29], in Atlantic
salmon (Salmo salar) sperm [30], and in Pacific chub mackerel
(Scomber japonicus) inner ear epithelium [31]. Furthermore, the
reported sensitivity of NaCl and water absorption in the
intestine of marine fish to pharmacological inhibitors of sAC
[29,32] provided functional evidence for its presence. More
recently, partial sAC coding sequences were identified by
BLAST searches in gar, rainbow trout and salmon [14]. How-
ever, a detailed molecular and biochemical characterization of
sAC genes and proteins from ray-finned fishes is still absent,
which is currently hindering functional studies about the
roles of sAC.
Therefore, in the current study, we explored and ident-
ified nucleotide sequences coding for sAC in genomic and
transcriptomics databases from multiple ray-finned fish
species. Of those, we focused on sAC from the rainbow
trout (Oncorhynchus mykiss), a teleost fish that is widely
used as a model species for physiological, ecological and
applied studies in aquaculture. Subsequently, we: (i) cloned
multiple rainbow trout sAC (rtsAC) variants, (ii) character-
ized the enzymatic activity of recombinant purified rtsAC
protein, and (iii) determined the presence and subcellular
localization of rtsAC in different tissues. The novel molecular,
biochemical and immunocytochemical findings of this study
provide a solid foundation for future functional studies and
will enable investigating the role of sAC as an acid–base
sensor in ray-finned fishes.
2. Methods
2.1. Experimental animals
Oncorhynchus mykiss were maintained in freshwater aquaria at the
Bamfield Marine Science Center (BC, Canada), the Aquatic
Resources Centre at University of Exeter (UK), or the University
of Manchester (UK). Procedures were approved by animal care pro-
tocols from the University of Alberta (AUP00001126 and
AUP00000072), University of Exeter (PPL P88687E07), and
University of Manchester (HO lic 40/3584). After euthanasia,
samples were stored at −80°C, in RNAlater, or in fixative for immu-
nostaining(see electronic supplementary material formore details).
2.2. Bioinformatics and cloning of rtsAC splice variants
Phylogenetic analysis was performed using Phylogeny.fr with
100 bootstraps (http://www.phylogeny.fr/) [33] (electronic
supplementary material, figure S1). Protein domains, tetratrico-
peptide repeat (TPR) motifs and intracellular targeting
sequences were, respectively, identified on NCBI’s Conserved
Domain Database, TPRpred (https://toolkit.tuebingen.mpg.de/
#/tools/tprpred), and the ngLOC database (http://genome.unmc.
edu/ngLOC/index.html). Cloning was performed on purified
mRNA from trout testis and RTgill-W1 cells using standard tech-
niques. Primers included Oligo(dT)
20
and 2 µM gene-specific
primers (electronic supplementary material, table S1). RT-PCR was
performed using Phusion High-Fidelity PCR Master Mix (New Eng-
land Biolabs, Ipswick, MA, USA). PCR products were purified and
cloned into pCR 2.1 TOPO vectors (Thermo Fisher Scientific). Plas-
mids were heat-shock transformed into One Shot TOP10
Chemically Competent E. coli cells (Thermo Fisher Scientific), puri-
fied and sequenced using M13 (-20) FW, M13 RV or rtsAC-specific
primers. The exon organization,splicing mechanisms, and otherfea-
tures of the cloned rtsAC isoforms are shown in electronic
supplementary material, figure S3.
2.3. Production of recombinant rtsAC
The cDNA coding for a 53 kDa rtsAC protein (GenBank
no. MF034901, named as rtsAC
T
; see Results) was cloned into
pDEST17 thus adding a 6X His-tag at the N-terminus
royalsocietypublishing.org/journal/rsfs Interface Focus 11: 20200026
2

(His-rtsAC
T
). Plasmids were amplified in One Shot TOP10
Chemically Competent E. coli, miniprep purified, sequenced
and transformed into One Shot BL21-AI E. coli strain. His-
rtsAC
T
protein production was induced with 0.2% L-arabinose
for 3 h at approximately 30°C (electronic supplementary
material, figure S4). Cells were harvested by centrifugation and
the pellet was stored at −80°C. After thawing, pellets were resus-
pended in lysis buffer (10 mM imidazole; 1 mg ml
−1
lysozyme
(Amresco), 1 : 800 benzonase nuclease (Novagen), 1 mM PMSF,
1mM Na
3
VO
4
, 1 mM NaF and protease inhibitor cocktail
(SIGMA); in PBS, pH 7.4). After sonication (five cycles of 10 s
each on ice), lysates were incubated at room temperature on a
horizontal shaker, centrifuged (21 000g, 30 min, 4°C). The super-
natant was passed through a HisPur™Ni-NTA resin (Thermo
Fisher Scientific). His-rtsAC
T
was eluted with 5 ml PBS with
250 mM imidazole; pH 7.4.
2.4. Characterization of rtsAC activity
Purified recombinant rtsAC
T
protein (approx. 0.2 µg µl
−1
)was
incubated in triplicate with assay buffer (100 mM Tris pH 7.4,
1 mM DTT, 150 mM NaCl, 250 µM IBMX, 2.5 mM ATP, and
either 5 mM MnCl
2
or 15 mM MgCl
2
+ 1 mM CaCl
2
, as indicated
below). The drugs KH7, LRE1, forskolin or 2,2-DDA were dis-
solved in DMSO (final concentration of 0.02% v/v) and their
IC
50
values were calculated in assay buffer containing 5 mM
MnCl
2
, and controls were run with DMSO alone. The EC
50
for
HCO 
3was measured in assay buffer containing 15 mM MgCl
2
and 1 mM CaCl
2
. Finally, the effects of KH7, LRE1 and
4,40-diisothiocyano-2,2’-stilbenedisulfonic acid (DIDS) were
measured in the presence of 20 mM NaHCO
3
in assay buffer con-
taining 15 mM MgCl
2
and 1 mM CaCl
2
, to determine their effects
on rtsAC baseline activity and on the HCO 
3-stimulated portion.
These conditions were selected based on extensive optimization
efforts that revealed that Mg
2+
and Ca
2+
were essential for sus-
taining HCO 
3-stimulated rtsAC activity, but did not induce a
clear dose–response on sAC activity by themselves under our
assay conditions.
Reactions were run for 1 h at room temperature and were
stopped by addition of an equal volume of 0.2 N HCl and
cAMP was quantified using the Direct cyclic AMP ELISA kit
(Arbor Assays, Ann Harbor, MI, USA). Each treatment was per-
formed using His-rtsAC
T
from two to five different protein
purifications.
Data were analysed and plotted using GraphPad Prism 7
(La Jolla, California, USA). The EC
50
for HCO 
3was determined
using a variable slope four parameters curve fit of log ([HCO 
3])
versus [cAMP]. The IC
50
values for KH7 and LRE1 were deter-
mined using a variable slope four- or three-parameter curve fit
of log [KH7] or [LRE1] versus [cAMP]. Sensitivity of control
and 20 mM HCO 
3-stimulated rtsAC to KH7, LRE1 and DIDS
was analysed by two-way ANOVA followed by Tukey’s multiple
comparisons test. Statistical significance was set at p< 0.05. Data
are presented as mean ± standard error of the mean (s.e.m.).
2.5. Cell culture
An immortalized cell line derived from rainbow trout gill, RTgill-
W1 cell line (ATCC CRL-2523) [34] was cultured in modified Lei-
bovitz’s L-15 medium containing 2.05 mM L-glutamine (Hyclone,
Logan, UT, USA), 10% v/v fetal bovine serum (Thermo Fisher
Scientific) and 1% v/v penicillin/streptomycin solution
(Thermo Fisher Scientific), and maintained at 18°C.
2.6. Western blotting and immunofluorescence
The following stocks of commercially available antibodies were
used: mouse anti-Na
+
/K
+
-ATPase (NKA) [35], mouse anti-actin
and mouse anti-α-tubulin (12G10) (all from Developmental Studies
Hybridoma Bank, The University of Iowa, IA, USA), mouse
anti-sarcomeric α-actinin (EA-53) (abcam, Cambridge, MA, USA),
mouse anti-Golgi matrix protein of 130 kDa (GM-130) (BD
Biosciences) (2.5 µg ml
−1
); anti-rabbit and anti-mouse Alexa 555
and Alexa 488 (Invitrogen). In addition, we generated custom-
made rabbit polyclonal antibodies against the peptide LSSKKGY-
GADELTRC in the C1 catalytic domain (anti-rtsAC-C1, 3 µg ml
−1
stock) and against the peptide SVEREEGYPLLGREC at the
beginning of the P-loop motif (anti-rtsAC-FL, 1.2 µg ml
−1
stock)
(GeneScript USA, Inc). Validation of anti-rtsAC-C1 antibodies by
Western blotting using His-rtsAC
T
as the positive control and
peptide-preabsorption controls for the various tissues is shown in
electronic supplementary material, figure S4.
For Western blots, tissues were processed as previously
described [20,26,36], and RTgill-W1 cells were lysed in RIPA
buffer containing protease inhibitors and centrifuged (30 min,
23 000g, 4°C) [36]. Anti-rtsAC-C1 antibodies were applied at
1 : 5000 dilution. For immunohistochemistry, tissue samples were
processed as described in [26]; staining of cardiomyocytes, white
muscle, red blood cells (RBCs) and RTgill-W1 cells required
optimization (detailed in electronic supplementary material).
3. Results and discussion
3.1. sAC genes are present throughout fish lineages
BLAST searches on nucleotide database confirmed the pres-
ence of sAC genes in species from all major animal phyla,
including cartilaginous (Chondrichthyes), ray-finned (Acti-
nopterygii) and lobe-finned (Sarcopterygii) fishes (electronic
supplementary material, figure S1). Within the ray-finned
fishes, sAC-like genes were identified in species from the sub-
classes Chondrostei (i.e. sturgeon) and Neopterigii (i.e.
Holostei (gar) and Teleostei (teleosts)). Within the teleosts,
complete or partial sAC genes were found in many species
from the superorders Protacanthopterygii, Ostariophysi, Clu-
peomorpha and Stenopterygii (electronic supplementary
material, table S2). However, we failed to obtain definitive
evidence for the presence of sAC genes in Cypriniformes (a
large order within the Ostariophysi that includes zebrafish),
or in any Neoteleostei. These results are puzzling when con-
sidering that sAC activity seems essential for proper sperm
function across a broad range of phyla, including sea
urchin [37], ascidian [35], salmon [30] and human [17],
which may be expected to exert a strong positive selective
pressure for the retention of functional sAC genes. Interest-
ingly, biochemical and functional evidence suggests sAC
may be present in three Neoteleostei species: toadfish [29],
sea bream [32] and mackerel [31], which belong to three
different orders within the superorder Acanthopterygii
(Batrachoidiformes, Percifomes and Scombriformes, respect-
ively). However, a definite confirmation will await detailed
studies in more fish species. Evidence for alternative splicing
of sAC mRNA was found in salmonids, northern pike, ayu,
channel and shark catfishes, piranha, herring and allis shad
(electronic supplementary material, table S2). However, auto-
mated computational analyses align mRNA fragments with
annotated gene extrons and generate the longest possible
mRNA sequence, and do not necessarily reflect real mRNA
splice variants. Thus, alternative splicing in trout sAC
(rtsAC) was further investigated using traditional RT-PCR.
3.2. rtsAC splice variants
Due to their importance for sperm function, testis typically
express a high level and heterogeneity of sAC mRNAs
royalsocietypublishing.org/journal/rsfs Interface Focus 11: 20200026
3

[6,38] and thus we chose this organ as the mRNA source for
our cloning efforts. We were able to clone the complete open
reading frame (ORF) of six rtsAC mRNA splice variants
(figure 1) (from the eukaryotic Kozak consensus translation
initiation sequence to the stop codon). The rtsAC splice var-
iants are the result of different splicing mechanisms
including exon skipping, alternative 30splice site selection
(30SSS), alternative 50splice site selection (50SSS), intron
retention and combinations thereof. We named the six
rtsAC splice variants with complete ORF as full-length
rtsAC (rtsAC
FL
), truncated rtsAC (rtsAC
T
), C1-only rtsACs
a, b and c (rtsAC
C1a−c
) and C2-only rtsAC (rtsAC
C2
).
rtsAC
FL
is the longest splice variant, with an ORF of
4953 bp coding a 1650 amino acid protein with a predicted
molecular weight (MW) of 186 kDa (figure 1a; electronic sup-
plementary material, figure S2). The N-terminus contains the
two catalytic domains C1 and C2 that are essential for cAMP
production and HCO 
3stimulation, followed by a STAND
nucleotide-binding P-loop motif that also contains three
Golgi apparatus targeting sequences. In the most C-terminal
region, rtsAC
FL
contains five TPR motifs predicted to mediate
protein–protein interactions and the assembly of multiprotein
complexes [6,10,37]. For example, sAC in sea urchin sperm
was reported to co-immunoprecipitate with at least 10 other
proteins including a Na
+
/H
+
exchanger, a cyclic nucleotide-
gated channel, dynein tubulins, creatine kinase, a guanylyl
cyclase and a GMP-specific phosphodiesterase [39]. In the
C-terminus region, rtsAC
FL
also shows similarity to a novel
haem-binding domain identified in mammalian sAC
FL
,
which could confer activation by NO and CO gases [40].
rtsAC
T
has a 1431 bp ORF that codes for a 476 amino acid
and approximately 53 kDa protein, which contains C1 and
C2 but none of the C-terminus motifs present in rtsAC
FL
(figure 1b). This rtsAC isoform resembles the mammalian
‘truncated sAC’(sAC
T
) [6]; however, while mammalian
sAC
T
results from the skipping of exon 11 [24], rtsAC
T
orig-
inates from the retention of the intron after exon 11 that
introduces a premature stop codon.
rtsAC
C1a−c
code for proteins containing the entire C1 but
with partial or absent C2. They result from a combination of
splicing mechanisms that introduce stop codons before the P-
loop and TPR motifs and the Golgi targeting sequences.
rtsAC
C1a
codes for a 28 kDa protein that contains exons 1–
6, but contains an intron retention before exon 7 that changes
the reading frame and introduces a stop codon early in exon
7. This sequence lacks part of the C1–C2 linker and the entire
C2 (figure 1c). rtsAC
C1b
codes for a 35 kDa protein and orig-
inates by alternative 50SSS at exon 6, skipping of exons 7, 8
and 9 (which code for the C1–C2 linker and part of C2), 30
SSS of exon 10, and intron retention after exon 11 introducing
a stop codon (figure 1d). The third C1-only rtsAC variant is
rtsAC
C1c
, which codes for a 47 kDa protein that is the result
of intron retention after exon 10 that introduces a stop codon
(figure 1e). Compared to rtsAC
T
, it lacks exon 11, which is
the most C-terminal exon of C2 and contains the residues
that bind ATP.
rtsAC
C2
has a 1143 bp ORF and codes for a 43 kDa
protein. It skips exons 2–4, which contain residues in C1
essential for binding of catalytic cations, as well as ATP, and
HCO 
3. It contains exons 5–11, which code for the entire C2
but retains the intron after exon 11, which introduces an
early stop codon. In summary, rtsAC
C2
contains C2 but
lacks most of C1 (figure 1f).
3.3. Potential significance of C1- and C2-only rtsACs
All metazoan Class III ACs (i.e. sAC and tmACs) character-
ized to date contain two types of catalytic domains, C1 and
(a)
(b)
(f)
(d)
(c)
(e)
Figure 1. Schematic organization of rtsAC isoforms with complete Open Reading Frames cloned. Sequences (a–f) with GenBank accession nos. MF034909, MF034901,
MF034907, MF034906, MF034908 and MF034905, respectively. C1 and C2: catalytic domains; green boxes: exons; red boxes: contain stop codons; blue boxes: skipped
exons; inverted (Y): epitopes recognized by anti-sAC-C1 and anti-sAC-FL antibodies; P-loop: P-loop domain; TPR: tetratricopeptide repeat (TPR)-like domain (1 to 5); the
red arrows in (a) indicate Golgi apparatus targeting sequences. IR: intron retention. 30and 50SSS: alternative 30or 50splice site selection. The black arrows indicate the
primer sets used to clone each sequence (F, forward; R, reverse) and are listed in table S1 (electronic supplementary material).
royalsocietypublishing.org/journal/rsfs Interface Focus 11: 20200026
4

C2, which are sequentially arranged within a single protein
chain. C1 and C2 form an internal pseudoheterodimer with
one catalytic pocket that coordinates the binding of ATP
and catalytic metals enable the cyclization of ATP into
cAMP [41]. By contrast, bacterial and protozoan Class III
ACs typically have a single type of catalytic domain, and
cAMP-forming catalytic activity depends on the formation
of a homodimer between catalytic domains from two distinct
protein chains [41]. While C2-only sAC isoforms are
expressed in mammals [21,22], C2 homodimers do not pos-
sess the required complementary set of key amino acid
residues [41] and thus do not possess cAMP-producing
activity [21,23]. However, experiments on airway epithelial
cells from C2 knock-out mice suggested that a C2-only sAC
isoform could heterodimerize with yet-unidentified C1-only
sAC isoforms and sustain cAMP production [21]. Our
findings in rainbow trout provide the first evidence that
C1-only sAC isoforms indeed exist, and we are currently
investigating whether they can sustain functional AC activity
as homodimers, or combine with C2-only sAC isoforms to
form functional heterodimers.
3.4. Multiple other rtsAC splice variants
We identified 14 additional potential rtsAC splice variants
comprising different combinations of exons and introns that
result from a variety of splicing mechanisms (electronic sup-
plementary material, figure S3). Unfortunately, we were
unable to clone their cDNAs as single amplicons due to limit-
ations in primer design and primer binding to multiple rtsAC
cDNA sequences (see [15] for a detailed explanation). Impor-
tantly, all of the RT-PCR products were sequenced, follow
established splicing mechanisms, are missing complete
exons while maintaining an ORF, and/or include intronic
regions. In combination, these factors strongly suggest that
the cDNAs with atrial ORFs are bona fide rtsAC splice variants
and not the result from PCR artefacts.
Sequence analyses indicate that these putative rtsAC var-
iants are different from rtsAC
FL
, rtsAC
T
, rtsAC
C1a−c
and
rtsAC
C2
. Three of these splice variants were deduced from
overlapping RT-PCRs, and putatively code for additional
C1- or C2-only rtsACs (electronic supplementary material,
figure S3G-I). Some of the other potential rtsAC splice variants
would completely or partially lack a variable numberof exons
and could code for rtsAC isoforms of diverse sizes (electronic
supplementary material, figure S3 J-T); however, large por-
tions of their N- or C-terminus remain unknown which
precluded us from definitely assessing their sizes. Interest-
ingly, many of these putative rtsAC isoforms would lack
some or all of the P-loop or TPR domains, suggesting different
regulatory properties and protein–protein interactions
compared to rtsAC
FL
.
Overall, our results indicate the existence of extensive
alternative splicing, potentially resulting in as many as 20
rtsAC isoforms. These results are consistent with multiple
previous reports from a variety of other tissues and organ-
isms and may be related to the differential subcellular
localization of sAC and their association with other proteins
[10,21–23,26,42–44].
3.5. Characterization of rtsAC activity
Enzymatic cAMP-producing activity was characterized using
His-rtsAC
T
. Truncated sAC proteins are ideal for this type of
analyses because they generally are the most active sAC iso-
forms [1,6,11,17] and this facilitates more accurate cAMP
measurements. In addition, they allow the characterization
of catalytic activity by C1 and C2 without interference from
putative regulatory domains that might be present in the
C-terminus region. And although different sAC isoforms
might have different V
max
, the affinity for ATP, HCO 
3and
catalytic metals is solely determined by amino acid residues
in C1 and C2 [11].
Preliminary assays revealed that His-rtsAC activity was
maximal at 2.5 mM ATP. This relatively low affinity for ATP
is one of the characteristics that differentiates sAC from
tmACs [5,6]; however, a detailed ATP dose–response for
rtsAC was not performed. Production of cAMP by His-
rtsAC
T
was increased approximately 20-fold by 5 mM
Mn
2+
, which is another hallmark of sAC activity. Mn
2+
-
stimulated rtsAC activity was sensitive to the sAC-specific
small molecule inhibitors KH7 and LRE1 (IC
50
∼5 and
18 µM, respectively), but insensitive to DDA, a specific
inhibitor of tmACs (figure 2). In the presence of more physio-
logically relevant 15 mM Mg
2+
and 1 mM Ca
2+
as catalytic
metals, His-rtsAC
T
activity was stimulated by HCO 
3with
an EC
50
of approximately 7 mM (figure 3a). This value is
within the normal range of [HCO 
3] in rainbow trout
plasma [45] and closely matches the approximately 5 mM
reported for sAC from the dogfish shark [9], a cartilaginous
fish. These kinetic characteristics make rtsAC a suitable
HCO 
3sensor under physiologically relevant acid–base con-
ditions typically found in the internal fluids of fish, where
[HCO 
3] ranges from nearly 0 to approximately 20 mM,
(a)(b)(c)
0
0
IC50 ˜ 5 mM
(5 mM Mn2+)
IC50 ˜ 18 mM
(5 mM Mn2+)
(KH7) mM
0.1 1 10 100 1000 0
(LRE1) mM
0.1 1 10 100 1000
0.2
0.4
0.6
relative sAC activity
0.8
1.0
1.2
0
0.2
0.4
0.6
relative sAC activity
0.8
1.0
1.2
(5 mM Mn2+)
0
(DDA) mM
0.1 1 10 100 1000
0
0.2
0.4
0.6
relative sAC activity
0.8
1.0
1.2
Figure 2. Effect of various inhibitors on recombinant His-rtsAC
T
protein. cAMP production in the presence of 5 mM Mn
2+
, the sAC inhibitors (a) KH7 or (b) LRE1, or
(c) the transmembrane adenylyl cyclase inhibitor 2,2-DDA. The data are shown as mean ± s.e.m. and are representative of two to four different protein purifications,
each run in two to four replicates. Values are expressed as fold-stimulation relative to baseline. IC
50
: half maximal inhibitory concentration.
royalsocietypublishing.org/journal/rsfs Interface Focus 11: 20200026
5

and pH from approximately 7.4 to approximately 8.3 [14]. In
addition to Mg
2+
, HCO 
3-stimulated rtsAC activity required
1mM Ca
2+
as the second catalytic metal. However, 1 mM
Ca
2+
alone did not stimulate basal rtsAC activity, nor did
we find a consistent dose–response to [Ca
2+
] under the exper-
imental conditions used in our assays. Thus further in vitro
and in vivo experiments are required to confirm whether
rtsAC may also act as a Ca
2+
and ATP sensor as reported
for mammalian sAC [46–48]. Similar to mammalian sAC
[6], His-rtsAC
T
was unaffected by the widely used tmAC
stimulator, forskolin (figure 3b). These results validate the
use of KH7, LRE1, DDA and forskolin as pharmacological
agents to study cAMP signalling pathways in fish (see [15]
for additional considerations).
In addition, we examined the effects of 25 and 50 µM
KH7 and LRE1 on the activity of His-rtsAC
T
that was stimu-
lated by 20 mM HCO 
3(which is close to the highest
physiologically relevant [HCO 
3] in fish internal fluids [14]),
in the presence of Mg
2+
and Ca
2+
as catalytic metals. While
KH7 significantly inhibited both basal and HCO 
3-stimulated
His-rtsAC
T
activity ( figure 3c), LRE1 only inhibited the
HCO 
3-stimulated component and had no effect on basal
His-rtsAC activity (figure 3d). These results are consistent
with more detailed pharmaco-kinetic analyses on mammalian
sAC, which suggested KH7 is a mixed competitive–
uncompetitive inhibitor [16,17,19], and indicated that LRE1 is
a competitive inhibitor with HCO 
3[18].
Finally, we tested the effect of DIDS on basal and
HCO 
3-stimulated His-rtsAC
T
activity. This disulfonic stil-
bene compound is an inhibitor of anion exchanger proteins
and is commonly used to study HCO 
3transport processes
in cells and epithelia, including many aquatic animals (e.g
[49–56]). However, DIDS was recently found to also
competitively inhibit mammalian sAC with
[HCO 
3]-dependent IC
50
between approximately 40 and
130 µM [57]. Here, we demonstrate that DIDS also inhibits
His-rtsAC
T
activity with an apparent IC
50
of approximately
50 µM (figure 3e). Based on DIDS potential off-target effects
on sAC-dependent signal transduction pathways, we rec-
ommend revisiting previous studies that reported effects of
DIDS on HCO 
3-dependent processes and discourage its
use in future studies.
3.6. rtsAC expression and localization in various rainbow
trout tissues
The presence of multiple rtsAC splice variants together
with a generally low mRNA abundance in tissues other
than testis greatly complicated the identification of sAC
splice variants using RT-PCR or other transcriptomic tech-
niques [15]. Therefore, we probed selected rainbow trout
tissues using Western blotting with anti-rtsAC-C1 anti-
bodies, and visualized the cellular localization of rtsAC
by immunohistochemistry and immunocytochemistry.
Western blotting revealed protein bands of varied sizes,
which demonstrated tissue-specific expression patterns.
Some of the bands were consistent with the rtsAC splice
variants with complete ORF shown in figure 1 (i.e. 186,
53, 47, 43, 35 and 28 kDa). Other bands hinted at the
presence of rtsAC isoforms of approximately 70, 90 and
100–110 kDa that could correspond to some of the splice
variants shown in electronic supplementary material,
figure S3, and additionally matched the sizes of known
sAC isoforms in mammals [21,23], coral [10] and sharks
[9], respectively. Because the immunoreactive bands were
absent in the peptide preabsortion controls (electronic
(a)
(c)(d)(e)
(b)
EC50 ˜ 7 mM
(15 mM Mg2+; 1 mM Ca2+)(15 mM Mg2+; 1 mM Ca2+)
(HCO3–) mM
0
(forskolin) mM
0.1 1 10 100 1000
relative sAC activity
relative sAC activity
relative sAC activity
0 10203040
1.0
1.1
1.2
1.3
1.4
1.5
1.6
1.7
1.8
0.2
0
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
0.2
0
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
2.0
2.2
a
b
control
25 KH7
50 KH7
B20
B20+25 KH7
B20+50 KH7
b
bb
c
relative sAC activity
0.2
0
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
2.0
2.2
aa
a
control
25 LRE1
50 LRE1
B20
B20+25 LRE1
B20+50 LRE1
a
a
b
relative sAC activity
0.2
0
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
2.0
2.2
a
a,b
b
control
50 DIDS
500 DIDS
B20
B20+50 DIDS
B20+500 DIDS
a
b
c
Figure 3. Biochemical characterization of recombinant His-rtsAC
T
protein. cAMP production in the presence of 15 mM Mg
2+
and 1 mM Ca
2+
:(a) stimulation by
HCO 
3;(b) effect of the transmembrane adenylyl cyclase agonist, forskolin. Effect of the sAC inhibitors (in μM) (c) KH7 or (d) LRE1, and (e) DIDS on His-rtsAC
T
activity stimulated with or without 20 mM NaHCO
3
(B20). Values are expressed as fold-stimulation relative to baseline. The data are shown as mean ± s.e.m. and are
representative of four to five different protein purifications, each run in two to four replicates. Different letters (a, b or c) indicate p< 0.05. EC
50
: half maximal
effective concentration.
royalsocietypublishing.org/journal/rsfs Interface Focus 11: 20200026
6

supplementary material, figure S4) and the epitope recog-
nized by the anti-rtsAC antibodies is not preset in any
other protein found in trout databases, we are confident
these bands represent authentic rtsAC isoforms; however,
this should be confirmed in future research. Similarly
establishing the functional significance of the observed
tissue-differential expression of multiple rtsAC isoforms
across tissues awaits future experiments to determine the
roles of C1- and C2-only rtsAC isoforms, and of the differ-
ent sAC domains in rtsACs or of homologous domains in
sACs from other organisms.
The gills are the main acid–base regulatory organs in
many aquatic organisms, and their function is in many
ways analogous to that of the mammalian kidney [58]. Pre-
vious studies have established that sAC is a physiological
acid–base sensor in ion-transporting cells from both elasmo-
branch fish gills [9,20] and the mammalian kidney [59–61].
Here, we show that rainbow trout gills may express four
rtsAC isoforms of approximately 45, approximately 50,
approximately 90 and approximately 110 kDa (figure 4a).
Furthermore, rtsAC is preferentially present in Na
+
/K
+
-
ATPase-rich gill ionocytes as well as in another epithelial
cell subtype with reduced or absent Na
+
/K
+
-ATPase
expression (figure 4b–e). Presumably, these are the acid- and
base-secreting cells [62], suggesting rtsAC plays important
roles in both processes as well as in NaCl uptake for osmor-
egulatory purposes (unlike the gills from marine sharks, the
gills of freshwater fishes play a dual role in acid–base and
osmotic regulation).
The expression of rtsAC was highest in heart and white
muscle tissue. Both tissues predominantly expressed an
approximately 50 kDa protein, most likely rtsAC
T
(figure 5a).
However, there were some differences in the expression levels
of some other rtsAC isoforms, most notably an approximate
90 kDa isoform in heart and an approximate 70 kDa isoform
in white muscle. Immunocytochemistry on isolated cardio-
myocytes revealed that rtsAC is present in distinct
subcellular regions including the sarcomere M- and Z-lines,
and potentially the sarcoplasmic reticulum (figure 5c–f). We
have previously reported the presence of sAC in the heart
of shark [26] and Pacific hagfish [63], and in the white
muscle of shark [26]. In the hagfish heart, sAC activity is
involved in sustaining tachycardia during recovery from
anoxia [63]. In the mammalian heart, sAC was originally
found to participate in the mitochondrial apoptosis pathway
[64], but more recent studies have revealed that sAC
T
plays
additional roles in modulating cardiac contractility [65] and
cardiac hypertrophy [66]. Our results suggest that sAC may
play similar functions in the teleost heart; however, the lack
of t-tubules in non-mammalian vertebrates [67] is likely
associated with species-specific sAC-dependent regulation
that will be determined through future research. In white
muscle (figure 5b), rtsAC exhibits a similar localization pat-
tern compared to that in cardiomyocytes. The lack of
previous studies on sAC in mammalian white muscle pre-
vents us from proposing potential roles. However, white
muscle is by far the most abundant fibre type in fish and
relies heavily on glycolytic ATP production to power rapid
muscle contraction. Whether sAC activity plays a functional
role in fish locomotion and whether this putative modulation
is in response to HCO 
3,Ca
2+
and/or ATP remains to be
studied.
In the rainbow trout liver, the most abundant rtsAC iso-
forms were approximately 50 kDa (consistent with rtsAC
T
)
and approximately 70 kDa (electronic supplementary
material, figure S5A). The localization of rtsAC was fairly
homogeneous throughout the hepatocyte cytoplasm, with
some hints of nuclear localization in a minority of cells
(electronic supplementary material, figure S5B,C). Based
on reports from mammalian liver cells, potential functions
~90 kDa
~50 kDa
~45 kDa
gillkDa
250
150
100
75
50
37
25
(a)(c)
(d)
(b)
(e)
~110 kDa
Figure 4. The presence and localization of rtsAC protein in trout gill cells. (a) Western blot of gill tissue probed with anti-rtsAC-C1 antibodies detected four bands
ranging from approximately 45 to approximately 110 kDa. Gill filament and lamellae co-immunostained with (b) anti-rtsAC-C1 (green), and (c)Na
+
-K
+
-ATPase (NKA,
red) antibodies. (d) sAC was highly expressed in NKA-rich ionocytes, which are cells specialized for acid/base and ionic regulation (arrows). Co-localization of sAC and
NKA antibodies appears yellow. (e) Immunofluorescence image merged with differential interference contrast (DIC). Nuclei were labelled with Hoechst 33342 (blue).
royalsocietypublishing.org/journal/rsfs Interface Focus 11: 20200026
7

in fish liver may include the regulations of the transcription
factor CREB [68] and of Cl
−
and HCO 
3driven cholangio-
cyte fluid secretion [69]. Western blotting additionally
revealed rtsAC presence in the trout intestine and inner
ear (electronic supplementary material, figure S4).
We also found robust expression of rtsAC within RBCs,
which seemed to exclusively express an approximately
110 kDa rtsAC isoform ( figure 6a). Immunocytochemical
analyses revealed rtsAC is present throughout the RBC cyto-
plasm, but also in association with the plasma membrane and
both around and within the nucleus (figure 6b). To our
knowledge, the only previous report of sAC in RBCs was in
shark [2], where a protein of similar size is recognized by
specific anti-shark sAC antibodies. However, evidence for
any functional roles of sAC in RBCs is lacking. Given the
cyclical fluctuations in acid–base conditions experienced by
250
M-line?
Z-line
SR?
colocalization
rtsAC + a-actinin
nucleus
10 mm
a-actinin
rtsAC
(C1)
*
(a)(c) (d)
(e) ( f)
100
75
50
37
150
25
~70 kDa
~50 kDa
~45 kDa
~90 kDa
white muscleheart
kDa
20 mm
SR?
rtsAC
(C1)
actin
nucleus
(b)
Figure 5. The presence and localization of rtsAC protein in trout striated muscle. (a) Western blot of heart and white muscle tissue probed with anti-rtsAC-C1
antibodies detected different bands ranging from approximately 45 to approximately 90 kDa. (b) Isolated white muscle fibres co-immunostained with anti-rtsAC-C1
(red) and anti-actin (green) antibodies. (c–f) Isolated cardiomyocytes were co-immunostained with anti-rtsAC-C1 (red) and anti-α-actinin (green) antibodies, the
latter labels sarcomeric Z-lines. The merged images (e,f) show sAC presence in Z-lines (yellow arrows, f), in a region consistent with M-lines (red arrows, f), and in a
region consistent with the sarcoplasmic reticulum (SR, asterisk, c). Nuclei were labelled with Hoechst 33342 (blue).
(a)(b)
~110 kDa
RBCs
kDa
160
110
80
60
50
40
30
Figure 6. The presence and localization of rtsAC protein in trout RBCs. (a) Western blot of rainbow trout RBCs probed with anti-rtsAC-C1 antibodies detected a band
at approximately 110 kDa. (b) RBCs immunostained with anti-rtsAC-C1 antibodies (red) showed the presence of sAC in the cytoplasm (+), plasma membrane (arrow
head) and nucleus (arrow). Nuclei were labelled with Hoechst 33342 (blue).
royalsocietypublishing.org/journal/rsfs Interface Focus 11: 20200026
8

RBCs with every pass through the circulatory system and the
role of intracellular pH in modulating haemoglobin O
2
and
CO
2
binding (reviewed in [70]), sAC may play important
roles in RBC physiology and whole-animal gas exchange.
3.7. Subcellular localization of rtsAC isoforms in
RTgill-W1 cells
Finally, we studied the subcellular localization of rtsAC iso-
forms in more detail in RTgill-W1 cells, an immortalized
cell line derived from rainbow trout gills [34]. These cells
are widely used for in vitro studies in fish toxicology, virology
and biochemistry [71–73] and their large size and flat mor-
phology facilitate microscopy studies. RT-PCR detected
mRNA for rtsAC
FL
and rtsAC
T
(figure 7a) as well as several
additional amplicons indicating multiple other rtsAC splice
variants. Western blots with anti-rtsAC-C1 antibodies
(figure 7b) revealed protein bands consistent with rtsAC
FL
(approx. 180 kDa) and rtsAC
T
(approx. 50 kDa) as well as
approximately 100 and approximately 70 kDa bands that
resemble those seen in gills and white muscle, respectively.
Additionally, there was a protein band of very high MW
(greater than 250 kDa), which might have resulted from
SDS-resistant complexes between rtsAC with other proteins,
or from protein aggregation artefacts.
Immunocytochemistry on RTgill-W1 cells using anti-
rtsAC-C1 antibodies detected rtsAC presence in various
intracellular compartments including the cytoplasm, podo-
some-like structures, the nucleus and the Golgi apparatus
(figure 7d,e). Based on a lack of co-localization with mito-
tracker (electronic supplementary material, figure S6), rtsAC
isoforms with the C1 motif do not appear to be present in
RTgill-W1 mitochondria; however, this may be re-evaluated
after optimization of alternative fixation protocols. In an
attempt to study differential subcellular localization of
rtsAC isoforms, we generated other custom-made polyclonal
antibodies, these ones against a peptide located in the begin-
ning of the P-loop (anti-rtsAC-FL; figure 7c). Unfortunately,
these antibodies did not work for Western blotting; however,
they intensely stained the nucleus, the Golgi apparatus and
(b)(a)(c)
~180 kDa
~100–120 kDa
~70 kDa
~50 kDa
RTgill-W1
rtsACFL
rtsACFL
rtsACT
rtsACT
rtef1a
C1 C2
C1 C2
P-loop TPRs TPRs
sAC-C1
Y
Y
sAC-C1
sAC-FL
kDa
250
NT RT PC
150
100
75
50
37
25
(d1) (d2) (d3) (d4)
(e1) (e2) (e3) (e4)
(f1) (f2) (f3) (f4)
Y
Figure 7. The presence and localization of rtsAC protein in RTgill-W1 cells. (a) RT-PCR detected the two rtsAC splice variants, full-length (rtsAC
FL
) and truncated
(rtsAC
T
). Elongation factor 1 alpha (ef1α) served as internal control. NT, no template control; RT, no reverse transcriptase control; PC, PCR control. (b) Western blot of
RTgill-W1 cells using anti-rtsAC-C1 antibodies detected bands ranging from approximately 50 to greater than 250 kDa. (c) Schematic of the rtsAC
FL
and rtsAC
T
proteins; C1 and C2: catalytic domains. P-loop: P-loop domain. TPRs: tetratricopeptide repeat (TPR)-like domains. Inverted (Y): epitopes recognized by anti-
sAC-C1 and anti-sAC-FL antibodies. Arrows: Golgi apparatus targeting sequences. Immunostaining of RTgill-W1 cells using anti-rtsAC-C1 antibodies (red) (panels
d1, d3, d4 and e1, e3, e4) with alpha-tubulin (Tubulin, green) (panels d2–d4) or with Golgi apparatus marker (Golgi matrix protein of 130 kDa or GM130,
green; panels e2–e4); or anti-rtsAC-FL (red; panels f1, f3, f4) with GM-130 ( panels f2–f4). Nuclei were labelled with Hoechst 33342 (blue). Co-localization is
seen as yellow (Merged). Arrowheads indicate Golgi apparatus, wide arrows podosome-like structures, thin arrows nuclei and (+) cytoplasm. Scale bars = 20 µm.
royalsocietypublishing.org/journal/rsfs Interface Focus 11: 20200026
9

podosomes of RTgill-W1 cells (figure 7f). These results
suggest that longer rtsAC isoforms are preferentially
expressed in these compartments; however, the epitope that
is recognized by the anti-rtsAC-FL antibodies is present in
at least nine of the cloned rtsAC isoforms (electronic sup-
plementary material, figure S3), precluding further
speculation. Nonetheless, it is worth mentioning that not all
those rtsAC variants possess the three Golgi apparatus target-
ing sequences, which further narrows down the options for
rtsAC isoforms that might be present in that compartment.
Research on mammals has reported sAC localization in
the nucleus of hepatocytes, COS-7, Huh7 and HeLa cells
[68,74], and in the peri-nuclear region (described as ‘discrete
dotlike Golgi staining’) of some benign melanomas [75]
and some virally infected keratinocytes [76]. However, to
our knowledge this is the first conclusive report about the
association of sAC with the Golgi apparatus, based on co-
localization with the specific marker GM-130 (figure 7e,f).
Similarly, while sAC-dependent cAMP signalling was
deemed important for leucocyte transendothelial migration
[77] and for neutrophil TNF activation [48], the experiments
described here are the first to visualize the presence of sAC
in podosome-like structures.
4. Conclusion
This study definitely establishes that sAC is present and func-
tional in a subset of ray-finned fishes, and we hope that the
molecular and biochemical characterization of rtsAC pre-
sented herein will help advance our knowledge about the
roles of sAC in rainbow trout and other ray-finned fishes.
The ability of sAC to respond to changes in intracellular
HCO 
3concentration and its downstream cAMP signalling
pathway may allow sAC to function as a general sensor of
intracellular acid–base status across a variety of different
tissues in rainbow trout. In addition, we report novel obser-
vations on the presence and localization of sAC in gills
from freshwater fish, in myocytes, and in RBCs. Whether
sAC plays a role in modulating transepithelial NaCl uptake
for hyperosmoregulation, muscle contraction during a
fatigue-induced acidosis or pH-sensitive oxygen transport
within RBCs remains to be investigated. Ultimately, the
roles of sAC will depend on the sAC variants that are
expressed within each cell type and on their functional associ-
ation with upstream proteins that may affect sAC stimulation
(i.e. CAs, HCO 
3transporters and CO
2
channels), other com-
ponents of the cAMP pathway (i.e. PKA, EPAC, cyclic
nucleotide-gated channels), and downstream effector pro-
teins. In addition, some of these putative novel functions,
as well as those played by sAC in podosomes and the
Golgi apparatus, may also apply to mammals and may
guide future biomedical research.
Ethics. Procedures involving fish were approved by animal care proto-
cols from the University of Alberta (AUP00001126 and
AUP00000072), University of Exeter (PPL P88687E07) and University
of Manchester (HO lic 40/3584).
Data accessibility. All data are available in the paper and electronic
supplementary material.
Authors’contributions. M.T. conceived the project. C.S. performed all the
molecular and cell culture work, and some of the other experiments.
M.T., T.S.H., J.N.R. and G.T.K. performed some of the Western blots
and immunolocalization experiments. S.D.B., G.G.G., H.S. and R.W.
provided the trout tissue samples. J.L.R., H.A.S., G.G.G. and
R.W.W. provided advice on potential roles of sAC in different trout
cell types. M.T. and C.S. analysed the results and wrote the paper.
All authors edited the manuscript and approved the final version.
Competing interests. The authors declare no competing interests.
Funding. This work was supported by grants from the National Science
Foundation (USA) to M.T. (NSF IOS 1354181 and 1754994) and from
the Biotechnology and Biological Sciences Research Council (UK) to
R.W.W. (BBSRC BB/N013344/1), a UCSD Chancellor’s Research
Excellence Scholarship (CRES) to M.T. and C.S., a Company of
Biologists Travel grant to C.S. (JEBTF-161115).
Acknowledgements. We are grateful to Dr Niels Bols and Dr Nathan Vo
from the University of Waterloo (Canada) for providing the RTgill-
W1 cell line. Angus Thies provided useful advice for the phylogenetic
analysis. We also thank Dr Sébastien Santini (CNRS/Aix-Marseilles
Université) for maintaining the Phylogeny.fr website.
References
1. Chen Y, Cann MJ, Litvin TN, Iourgenko V, Sinclair
ML, Levin LR, Buck J. 2000 Soluble adenylyl cyclase
as an evolutionarily conserved bicarbonate sensor.
Science 289, 625–628. (doi:10.1126/science.289.
5479.625)
2. Tresguerres M, Barott KL, Barron ME, Roa JN. 2014
Established and potential physiological roles of
bicarbonate-sensing soluble adenylyl cyclase (sAC)
in aquatic animals. J. Exp. Biol. 217, 663–672.
(doi:10.1242/jeb.086157)
3. Tresguerres M, Levin LR, Buck J. 2011 Intracellular
cAMP signaling by soluble adenylyl cyclase.
Kidne Int. 79, 1277–1288. (doi:10.1038/ki.
2011.95)
4. Braun T, Dods RF. 1975 Development of a Mn
2+
-
sensitive, ‘soluble’adenylate cyclase in rat testis.
Proc. Natl Acad. Sci. USA 72, 1097–1101. (doi:10.
1073/pnas.72.3.1097)
5. Braun T. 1991 Purification of soluble form of
adenylyl cyclase from testes. Methods Enzymol. 195,
130–136. (doi:10.1016/0076-6879(91)95160-L)
6. Buck J, Sinclair ML, Schapal L, Cann MJ, Levin LR.
1999 Cytosolic adenylyl cyclase defines a unique
signaling molecule in mammals. Proc. Natl Acad.
Sci. USA 96,79–84. (doi:10.1073/pnas.96.1.79)
7. Roelofs J, Van Haastert PJM. 2002 Deducing the
origin of soluble adenylyl cyclase, a gene lost in
multiple lineages. Mol. Biol. Evol. 19, 2239–2246.
(doi:10.1093/oxfordjournals.molbev.a004047)
8. Nomura M, Beltrán C, Darszon A, Vacquier VD. 2005
A soluble adenylyl cyclase from sea urchin
spermatozoa. Gene 353, 231–238. (doi:10.1016/j.
gene.2005.04.034)
9. Tresguerres M, Parks SK, Salazar E, Levin LR, Goss
GG, Buck J. 2010 Bicarbonate-sensing soluble
adenylyl cyclase is an essential sensor for acid/base
homeostasis. Proc. Natl Acad. Sci. USA 107,
442–447. (doi:10.1073/pnas.0911790107)
10. Barott KL, Barron ME, Tresguerres M. 2017 Identification
of a molecular pH sensor in coral. Proc. R. Soc. B 284,
20171769. (doi:10.1098/rspb.2017.1769)
11. Chaloupka JA, Bullock SA, Iourgenko V, Levin LR,
Buck J. 2006 Autoinhibitory regulation of soluble
adenylyl cyclase. Mol. Reprod. Dev. 73, 361–368.
(doi:10.1002/mrd.20409)
12. Litvin TN, Kamenetsky M, Zarifyan A, Buck J, Levin
LR. 2003 Kinetic properties of ‘soluble’adenylyl
cyclase. Synergism between calcium and
bicarbonate. J. Biol. Chem. 278, 15 922–15 926.
(doi:10.1074/jbc.M212475200)
13. Buck J, Levin LR. 2014 The role of soluble adenylyl
cyclase in health and disease. Biochim. Biophys.
Acta. 1842, 2533–2534. (doi:10.1016/j.bbadis.2014.
09.009)
royalsocietypublishing.org/journal/rsfs Interface Focus 11: 20200026
10

14. Tresguerres M. 2014 sAC from aquatic organisms as
a model to study the evolution of acid/base
sensing. Biochim. Biophys. Acta 1842, 2629–2635.
(doi:10.1016/j.bbadis.2014.06.021)
15. Tresguerres M, Salmerón C. 2018 Molecular,
enzymatic, and cellular characterization of soluble
adenylyl cyclase from aquatic animals. Methods
Enzymol. 605, 525–549. (doi:10.1016/bs.mie.2018.
02.022)
16. Bitterman JL, Ramos-Espiritu L, Diaz A, Levin LR,
Buck J. 2013 Pharmacological distinction between
soluble and transmembrane adenylyl cyclases.
J. Pharmacol. Exp. Ther. 347, 589–598. (doi:10.
1124/jpet.113.208496)
17. Hess KC et al. 2005 The ‘soluble’adenylyl cyclase in
sperm mediates multiple signaling events required
for fertilization. Dev. Cell. 9, 249–259. (doi:10.1016/
j.devcel.2005.06.007)
18. Ramos-Espiritu L et al. 2016 Discovery of LRE1 as a
specific and allosteric inhibitor of soluble adenylyl
cyclase. Nat. Chem. Biol. 12, 838–844. (doi:10.
1038/nchembio.2151)
19. Steegborn C. 2014 Structure, mechanism, and
regulation of soluble adenylyl cyclases: similarities
and differences to transmembrane adenylyl cyclases.
Biochim. Biophys. Acta 1842, 2535–2547. (doi:10.
1016/j.bbadis.2014.08.012)
20. Roa JN, Tresguerres M. 2016 Soluble adenylyl
cyclase is an acid-base sensor in epithelial base-
secreting cells. Am. J. Physiol. Cell Physiol. 311,
C340–C349. (doi:10.1152/ajpcell.00089.2016)
21. Chen X, Baumlin N, Buck J, Levin LR, Fregien N,
Salathe M. 2014 A soluble adenylyl cyclase form
targets to axonemes and rescues beat regulation in
soluble adenylyl cyclase knockout mice.
Am. J. Respir. Cell Mol. Biol. 51, 750–760. (doi:10.
1165/rcmb.2013-0542OC)
22. Farrell J, Ramos L, Tresguerres M, Kamenetsky M,
Levin LR, Buck J. 2008 Somatic ‘soluble’adenylyl
cyclase isoforms are unaffected in Sacy
tm1Lex$
/
Sacy
tm1Lex$
‘knockout’mice. PLoS ONE 3, e3251.
(doi:10.1371/journal.pone.0003251)
23. Geng W, Wang Z, Zhang J, Reed BY, Pak CYC, Moe
OW. 2005 Cloning and characterization of the human
soluble adenylyl cyclase. AJP Cell Physiol. 288,
C1305–C1316. (doi:10.1152/ajpcell.00584.2004)
24. Jaiswal BS, Conti M. 2001 Identification and
functional analysis of splice variants of the
germ cell soluble adenylyl cyclase. J. Biol. Chem.
276, 31 698–31 708. (doi:10.1074/jbc.M011698200)
25. Schmid A et al. 2007 Soluble adenylyl cyclase is
localized to cilia and contributes to ciliary beat
frequency regulation via production of cAMP. J. Gen.
Physiol. 130,99–109. (doi:10.1085/jgp.200709784)
26. Roa JN, Tresguerres M. 2017 Bicarbonate-sensing
soluble adenylyl cyclase is present in the cell
cytoplasm and nucleus of multiple shark tissues.
Physiol. Rep. 24, e13090. (doi:10.14814/phy2.
13090)
27. Tresguerres M, Milsom WK, Perry SF. 2019 CO
2
and
acid-base sensing. In Fish physiology (eds M Grosell,
PL Munday, AP Farrell, CJ Brauner), pp. 33–68.
San Diego, CA: Academic Press.
28. Heisler N. 1988 Acid-base regulation. In Physiology
of elasmobranch fishes (ed. TJ Shuttleworth), pp.
215–252. Berlin, Germany: Springer.
29. Tresguerres M, Levin LR, Buck J, Grosell M. 2010
Modulation of NaCl absorption by [HCO
3
−
] in the
marine teleost intestine is mediated by soluble
adenylyl cyclase. AJP Regul. Integr. Comp. Physiol.
299, R62–R71. (doi:10.1152/ajpregu.00761.2009)
30. von Schalburg KR, Gowen BE, Leong JS, Rondeau
EB, Davidson WS, Koop BF. 2018 Subcellular
localization and characterization of estrogenic
pathway regulators and mediators in Atlantic
salmon spermatozoal cells. Histochem.
Cell Biol. 149,75–96. (doi:10.1007/s00418-017-
1611-3)
31. Kwan GF, Smith TR, Tresguerres M. 2020
Immunological characterization of two types of
ionocytes in the inner ear epithelium of Pacific
Chub Mackerel (Scomber japonicus). J. Comp.
Physiol. B 190, 419–431. (doi:10.1007/s00360-
020-01276-3)
32. Carvalho ESM, Gregório SF, Power DM, Canário AVM,
Fuentes J. 2012 Water absorption and bicarbonate
secretion in the intestine of the sea bream are
regulated by transmembrane and soluble adenylyl
cyclase stimulation. J. Comp. Physiol. B 182,
1069–1080. (doi:10.1007/s00360-012-0685-4)
33. Dereeper A et al. 2008 Phylogeny.fr: robust
phylogenetic analysis for the non-specialist. Nucleic
Acids Res. 36(Web Server issue), W465–W469.
(doi:10.1093/nar/gkn180)
34. Bols NC, Barlian A, Chirino-Trejo M, Caldwell SJ,
Goegan P, Lee LEJ. 1994 Development of a cell line
from primary cultures of rainbow trout,
Oncorhynchus mykiss (Walbaum), gills. J. Fish Dis.
17, 601–611. (doi:10.1111/j.1365-2761.1994.
tb00258.x)
35. Shiba K, Inaba K. 2014 Distinct roles of soluble and
transmembrane adenylyl cyclases in the regulation
of flagellar motility in Ciona sperm. Int. J. Msol. Sci.
15, 13 192–13 208. (doi:10.3390/ijms150813192)
36. Roa JN, Munévar CL, Tresguerres M. 2014 Feeding
induces translocation of vacuolar proton ATPase and
pendrin to the membrane of leopard shark (Triakis
semifasciata) mitochondrion-rich gill cells. Comp.
Biochem. Physiol. A Mol. Integr. Physiol. 174,
29–37. (doi:10.1016/j.cbpa.2014.04.003)
37. Vacquier VD, Loza-Huerta A, García-Rincón J,
Darszon A, Beltrán C. 2014 Soluble adenylyl cyclase
of sea urchin spermatozoa. Biochim. Biophys. Acta
1842, 2621–2628. (doi:10.1016/j.bbadis.2014.
07.011)
38. Sinclair ML, Wang XY, Mattia M, Conti M, Buck J,
Wolgemuth DJ, Levin LR. 2000 Specific expression
of soluble adenylyl cyclase in male germ cells. Mol.
Reprod. Dev. 56,6–11. (doi:10.1002/(SICI)1098-
2795(200005)56:1<6::AID-MRD2>3.0.CO;2-M)
39. Nomura M, Vacquier VD. 2006 Proteins associated
with soluble adenylyl cyclase in sea urchin sperm
flagella. Cell Motil. Cytoskeleton 63, 582–590.
(doi:10.1002/cm.20147)
40. Middelhaufe S, Leipelt M, Levin LR, Buck J,
Steegborn C. 2012 Identification of a haem domain
in human soluble adenylate cyclase. Biosci. Rep. 32,
491–499. (doi:10.1042/BSR20120051)
41. Linder JU, Schultz JE. 2003 The class III adenylyl
cyclases: multi-purpose signalling modules. Cell.
Signal. 15, 1081–1089. (doi:10.1016/S0898-
6568(03)00130-X)
42. Chen J, Martinez J, Milner TA, Buck J, Levin LR.
2013 Neuronal expression of soluble adenylyl
cyclase in the mammalian brain. Brain Res. 1518,
1–8. (doi:10.1016/j.brainres.2013.04.027)
43. Sun XC, Cui M, Bonanno JA. 2004 [HCO
3
−
]-regulated
expression and activity of soluble adenylyl cyclase in
corneal endothelial and Calu-3 cells. BMC Physiol. 4,
8. (doi:10.1186/1472-6793-4-8)
44. Geng W, Hill K, Zerwekh JE, Kohler T, Müller R, Moe
OW. 2009 Inhibition of osteoclast formation and
function by bicarbonate: role of soluble adenylyl
cyclase. J. Cell. Physiol. 220, 332–340. (doi:10.1002/
jcp.21767)
45. Cameron JN, Randall DJ. 1972 The effect of increased
ambient CO
2
on arterial CO
2
tension, CO
2
content and
pH in rainbow trout. J. Exp. Biol. 57,673–680.
46. Ramos LS, Zippin JH, Kamenetsky M, Buck J, Levin
LR. 2008 Glucose and GLP-1 stimulate cAMP
production via distinct adenylyl cyclases in INS-1E
insulinoma cells. J. Gen. Physiol. 132, 329–338.
(doi:10.1085/jgp.200810044)
47. Zippin JH et al. 2013 CO
2
/HCO
3
−
- and calcium-
regulated soluble adenylyl cyclase as a physiological
ATP sensor. J. Biol. Chem. 288, 33 283–33 291.
(doi:10.1074/jbc.M113.510073)
48. Han H et al. 2005 Calcium-sensing soluble adenylyl
cyclase mediates TNF signal transduction in human
neutrophils. J. Exp. Med. 202, 353–361. (doi:10.
1084/jem.20050778)
49. Parks SK, Tresguerres M, Goss GG. 2007 Interactions
between Na
+
channels and Na
+
-HCO
3
−
cotransporters in the freshwater fish gill MR cell: a
model for transepithelial Na
+
uptake.
Am. J. Physiol. Cell Physiol. 292, C935–C944.
(doi:10.1152/ajpcell.00604.2005)
50. Hailey DW, Roberts B, Owens KN, Stewart AK, Linbo
T, Pujol R, Alper SL, Rubel EW, Raible DW. 2012
Loss of Slc4a1b chloride/bicarbonate exchanger
function protects mechanosensory hair cells from
aminoglycoside damage in the zebrafish mutant
persephone. PLoS Genet. 8, e1002971. (doi:10.1371/
journal.pgen.1002971)
51. Nikinmaa M, Huestis WH. 1984 Adrenergic swelling
of nucleated erythrocytes: cellular mechanisms in a
bird, domestic goose, and two teleosts, striped bass
and rainbow trout. J. Exp. Biol. 113, 215–224.
52. Tresguerres M, Parks SK, Sabatini SE, Goss GG,
Luquet CM. 2008 Regulation of ion transport by pH
and [HCO
3
−
] in isolated gills of the crab Neohelice
(Chasmagnathus) granulata.AJP Regul. Integr.
Comp. Physiol. 294, R1033–R1043. (doi:10.1152/
ajpregu.00516.2007)
53. Wilson JM, Moreira-Silva J, Delgado ILS, Ebanks SC,
Vijayan MM, Coimbra J, Grosell M. 2013
Mechanisms of transepithelial ammonia excretion
and luminal alkalinization in the gut of an intestinal
air-breathing fish, Misgurnus anguilliacaudatus.
royalsocietypublishing.org/journal/rsfs Interface Focus 11: 20200026
11

J. Exp. Biol. 216, 623–632. (doi:10.1242/jeb.
074401)
54. Jagadeshwaran U, Onken H, Hardy M, Moffett SB,
Moffett DF. 2010 Cellular mechanisms of acid
secretion in the posterior midgut of the larval
mosquito (Aedes aegypti). J. Exp. Biol. 213,
295–300. (doi:10.1242/jeb.037549)
55. Furimsky M, Moon TW, Perry SF. 2000 Evidence for
the role of a Na
+
/HCO
3
−
cotransporter in trout
hepatocyte pHi regulation. J. Exp. Biol. 203,
2201–2208.
56. Tambutté E, Allemand D, Mueller E, Jaubert J. 1996
A compartmental approach to the mechanism of
calcification in hermatypic corals. J. Exp. Biol. 199,
1029–1041.
57. Kleinboelting S, Diaz A, Moniot S, van den Heuvel J,
Weyand M, Levin LR, Buck J, Steegborn C. 2014
Crystal structures of human soluble adenylyl cyclase
reveal mechanisms of catalysis and of its activation
through bicarbonate. Proc. Natl Acad. Sci. USA 111,
3727–3732. (doi:10.1073/pnas.1322778111)
58. Larsen EH, Deaton LE, Onken H, O’Donnell M,
Grosell M, Dantzler WH, Weihrauch D. 2014
Osmoregulation and excretion. Compr. Physiol. 4,
405–573. (doi:10.1002/cphy.c130004)
59. Gong F, Alzamora R, Smolak C, Li H, Naveed S,
Neumann D, Hallows KR, Pastor-Soler NM. 2010
Vacuolar H
+
-ATPase apical accumulation in kidney
intercalated cells is regulated by PKA and AMP-
activated protein kinase. Am.J.Physiol.RenalPhysiol.
298, F1162–F1169. (doi:10.1152/ajprenal.00645.2009)
60. Hallows KR et al. 2009 Regulation of epithelial Na
+
transport by soluble adenylyl cyclase in kidney
collecting duct cells. J. Biol. Chem. 284, 5774–5783.
(doi:10.1074/jbc.M805501200)
61. Paunescu TG, Da Silva N, Russo LM, McKee M, Lu
HAJ, Breton S, Brown D. 2008 Association of soluble
adenylyl cyclase with the V-ATPase in renal
epithelial cells. Am. J. Physiol. Renal Physiol. 294,
F130–F138. (doi:10.1152/ajprenal.00406.2007)
62. Galvez F, Reid SD, Hawkings G, Goss GG. 2002
Isolation and characterization of mitochondria-rich
cell types from the gill of freshwater rainbow trout.
AJP Regul. Integr. Comp. Physiol. 282, R658–R668.
(doi:10.1152/ajpregu.00342.2001)
63. Wilson CM, Roa JN, Cox GK, Tresguerres M, Farrell
AP. 2016 Introducing a novel mechanism to control
heart rate in the ancestral pacific hagfish. J. Exp.
Biol. 219, jeb138198. (doi:10.1242/jeb.138198)
64. Ladilov Y, Appukuttan A. 2014 Role of soluble
adenylyl cyclase in cell death and growth. Biochim.
Biophys. Acta 1842, 2646–2655. (doi:10.1016/j.
bbadis.2014.06.034)
65. Espejo MS, Orlowski A, Ibañez AM, Di Mattía RA,
Velásquez FC, Rossetti NS, Ciancio MC, De Giusti VC,
Aiello EA. 2020 The functional association between
the sodium/bicarbonate cotransporter (NBC) and
the soluble adenylyl cyclase (sAC) modulates cardiac
contractility. Pflügers Arch. Eur. J. Physiol. 472,
103–115. (doi:10.1007/s00424-019-02331-x)
66. Schirmer I et al. 2018 Soluble adenylyl cyclase: a
novel player in cardiac hypertrophy induced by
isoprenaline or pressure overload. PLoS ONE 13,
e0192322. (doi:10.1371/journal.pone.0192322)
67. Shiels HA, Galli GLJ. 2014 The sarcoplasmic
reticulum and the evolution of the vertebrate heart.
Physiol. Bethesda Md. 29, 456–469. (doi:10.1152/
physiol.00015.2014)
68. Zippin JH, Farrell J, Huron D, Kamenetsky M, Hess
KC, Fischman DA, Levin LR, Buck J. 2004
Bicarbonate-responsive ‘soluble’adenylyl cyclase
defines a nuclear cAMP microdomain. J. Cell Biol.
164, 527–534. (doi:10.1083/jcb.200311119)
69. Strazzabosco M, Fiorotto R, Melero S, Glaser S, Francis
H, Spirli C, Alpini G. 2009 Differentially expressed
adenylyl cyclase isoforms mediate secretory functions
in cholangiocyte subpopulation. Hepatol. Baltim. Md.
50,244–252. (doi:10.1002/hep.22926)
70. Tresguerres M, Clifford AM, Harter TS, Roa JN, Theis
AB, Yee D, Brauner C.2020 Evolutionary links
between intra- and extracellular acid–base
regulation in fish and other aquatic animals. J. Exp.
Zool. 333, 449–465. (doi:10.1002/jez.2367)
71. Lee LEJ, Dayeh VR, Schirmer K, Bols NC. 2009
Applications and potential uses of fish gill cell lines:
examples with RTgill-W1. Vitro Cell Dev. Biol. Anim.
45, 127–134. (doi:10.1007/s11626-008-9173-2)
72. Fischer M et al. 2019 Repeatability and
reproducibility of the RTgill-W1 cell line assay for
predicting fish acute toxicity. Toxicol. Sci. 169,
353–364. (doi:10.1093/toxsci/kfz057)
73. Schnell S, Stott LC, Hogstrand C, Wood CM, Kelly
SP,PärtP,OwenSF,BuryNR.2016Proceduresfor
the reconstruction, primary culture and
experimental use of rainbow trout gill epithelia.
Nat. Protoc. 11, 490–498. (doi:10.1038/nprot.
2016.029)
74. Zippin JH, Chen Y, Nahirney P, Kamenetsky M,
Wuttke MS, Fischman DA, Levin LR, Buck J. 2003
Compartmentalization of bicarbonate-sensitive
adenylyl cyclase in distinct signaling microdomains.
FASEB J. 17,82–84. (doi:10.1096/fj.02-0598fje)
75. Magro CM, Yang S-E, Zippin JH, Zembowicz A. 2012
Expression of soluble adenylyl cyclase in lentigo
maligna: use of immunohistochemistry with anti–
soluble adenylyl cyclase antibody (R21) in diagnosis
of lentigo maligna and assessment of margins.
Arch. Pathol. Lab. Med. 136, 1558–1564. (doi:10.
5858/arpa.2011-0617-OA)
76. Zippin JH, Chadwick PA, Levin LR, Buck J, Magro CM.
2010 Soluble adenylyl cyclase defines a nuclear cAMP
microdomain in keratinocyte hyperproliferative skin
diseases. J. Invest. Dermatol. 130,1279–1287.
(doi:10.1038/jid.2009.440)
77. Watson RL, Buck J, Levin LR, Winger RC, Wang J,
Arase H, Muller WA. 2015 Endothelial CD99 signals
through soluble adenylyl cyclase and PKA to
regulate leukocyte transendothelial migration.
J. Exp. Med. 212, 1021–1041. (doi:10.1084/jem.
20150354)
royalsocietypublishing.org/journal/rsfs Interface Focus 11: 20200026
12