The cardiac proteome in patients with congenital ventricular septal defect: A comparative study between right atria and right ventricles

Abstract

Right ventricle (RV) remodelling occurs in neonatal patients born with ventricular septal defect (VSD). The presence of a defect between the two ventricles allows for shunting of blood from the left to right side. The resulting RV hypertrophy leads to molecular remodelling which has thus far been largely investigated using right atrial (RA) tissue. In this study we used proteomic and phosphoproteomic analysis in order to determine any difference between the proteomes for RA and RV. Samples were therefore taken from the RA and RV of five infants (0.34 ± 0.05 years, mean ± SEM) with VSD who were undergoing cardiac surgery to repair the defect. Significant differences in protein expression between RV and RA were seen. 150 protein accession numbers were identified which were significantly lower in the atria, whereas none were significantly higher in the atria compared to the ventricle. 19 phosphorylation sites (representing 19 phosphoproteins) were also lower in RA. This work has identified differences in the proteome between RA and RV which reflect differences in contractile activity and metabolism. As such, caution should be used when drawing conclusions based on analysis of the RA and extrapolating to the hypertrophied RV.

Significance
RV hypertrophy occurs in neonatal patients born with VSD. Very little is known about how the atria responds to RV hypertrophy, especially at the protein level. Access to tissue from age-matched groups of patients is very rare, and we are in the unique position of being able to get tissue from both the atria and ventricle during reparative surgery of these infants. Our findings will be beneficial to future research into heart chamber malformations in congenital heart defects.

Full text

Contents lists available at ScienceDirect
Journal of Proteomics
journal homepage: www.elsevier.com/locate/jprot
The cardiac proteome in patients with congenital ventricular septal defect: A
comparative study between right atria and right ventricles
Bond A.R.
a
, Iacobazzi D.
a
, Abdul-Ghani S.
a
, Ghorbel M.T.
a
, Heesom K.J.
b
, George S.J.
a
,
Caputo M.
a,c
, Suleiman M.-S.
a
, Tulloh R.M.
a,c,⁎
a
Bristol Heart Institute, Research Floor Level 7, Bristol Royal Infirmary, Marlborough Street, Bristol BS2 8HW, United Kingdom
b
Proteomics Facility, University of Bristol, Bristol BS8 1TD, United Kingdom
c
Department of Congenital Heart Disease, King David Building, Upper Maudlin Street, Bristol BS2 8BJ, United Kingdom
ARTICLE INFO
Keywords:
Congenital heart disease
Ventricular septal defect
Right ventricle
Right atria
Proteomics
ABSTRACT
Right ventricle (RV) remodelling occurs in neonatal patients born with ventricular septal defect (VSD). The
presence of a defect between the two ventricles allows for shunting of blood from the left to right side. The
resulting RV hypertrophy leads to molecular remodelling which has thus far been largely investigated using right
atrial (RA) tissue. In this study we used proteomic and phosphoproteomic analysis in order to determine any
difference between the proteomes for RA and RV. Samples were therefore taken from the RA and RV of five
infants (0.34 ± 0.05 years, mean ± SEM) with VSD who were undergoing cardiac surgery to repair the defect.
Significant differences in protein expression between RV and RA were seen. 150 protein accession numbers were
identified which were significantly lower in the atria, whereas none were significantly higher in the atria
compared to the ventricle. 19 phosphorylation sites (representing 19 phosphoproteins) were also lower in RA.
This work has identified differences in the proteome between RA and RV which reflect differences in contractile
activity and metabolism. As such, caution should be used when drawing conclusions based on analysis of the RA
and extrapolating to the hypertrophied RV.
Significance: RV hypertrophy occurs in neonatal patients born with VSD. Very little is known about how the atria
responds to RV hypertrophy, especially at the protein level. Access to tissue from age-matched groups of patients
is very rare, and we are in the unique position of being able to get tissue from both the atria and ventricle during
reparative surgery of these infants. Our findings will be beneficial to future research into heart chamber mal-
formations in congenital heart defects.
1. Introduction
Ventricular septal defect (VSD), a hole in the septum between the
left and right ventricles of the heart, is the most common acyanotic
congenital cardiac malformation affecting 30% of babies born with
congenital heart disease each year (0.36% of all live births [1,2]). Not
all infants with VSD require treatment, many closing by themselves as
the child grows. However, larger VSDs cause the flow of oxygenated
blood from the left ventricle (LV) back into the right ventricle (RV),
causing an increase in right ventricular pressure, and increased blood
flow to the lungs via the pulmonary artery; pulmonary hypertension
and right ventricular hypertrophy can ensue.
The role of the atria is to fill the ventricles with blood, for sub-
sequent ejection from the heart either to the lungs (RV) or around the
systemic circulation (LV). The ventricles therefore contract much more
forcefully than the atria. In the right heart, these chamber differences
are reflected by the differences in pressures; RA mean
pressure < 5 mm Hg and RV systolic pressure often 25 mm Hg, RV
diastolic pressure < 5 mm Hg [3]. In patients with a large VSD the RV
must generate a much larger systolic pressure (~80 mm Hg; similar to
arterial systolic pressure) in response to the increased pressure exerted
by the LV, resulting in cardiac remodelling. The cellular ultrastructure
of atrial and ventricular cardiomyocytes has many similarities. How-
ever, they have very different calcium pattern in response to depolar-
isation; atrial myocytes have a shorter duration of action potential than
ventricular (150 ms vs. 250 ms) [4] and they contract at a faster rate
with rapid repolarization, with both chambers showing peak contrac-
tion within tens of milliseconds [5]. This is thought to partly be due to
differences in transverse (T-tubule) organization in atria [6]; previously
thought to be lacking T-tubules [7]. Calcium signals tend to be found at
the cell periphery, and as such the atria rely on hormones for the in-
ward movement of Ca
2+
to the contractile machinery in the adult heart
https://doi.org/10.1016/j.jprot.2018.03.022
Received 1 November 2017; Received in revised form 8 March 2018; Accepted 19 March 2018
⁎
Corresponding author at: Professor Robert Tulloh, Bristol Heart Institute, Research Floor Level 7, Bristol Royal Infirmary, Marlborough Street, Bristol BS2 8HW, United Kingdom.
E-mail address: robert.tulloh@bristol.ac.uk (R.M. Tulloh).
Journal of Proteomics 191 (2019) 107–113
Available online 20 March 2018
1874-3919/ © 2018 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/BY-NC-ND/4.0/).
T

[8]. Recent studies (in a dog model) have now shown that there is a
sparse network of T-tubules in the atria but there are differences be-
tween the left and right heart [9]. It should be noted that no comparison
of cell ultrastructure has been carried out in infants.
Comparisons have been made of individual proteins between the
right atria and ventricles from different pathologies (normoxaemic e.g.
VSD/ASD (atrial septal defect), and hypoxaemic e.g. tetralogy of fallot
(TOF) [10,11]), but these were not performed using current proteomics
methods, and as such had to target certain protein groups (e.g. con-
traction and extracellular matrix). Previous proteomic studies have
looked at the proteins implicated in atrial fibrillation (AF) comparing
the right and left atria of adult hearts [12], the normal human fetal atria
and ventricles [13] and more recently comparing the RV from different
pathologies (VSD vs. TOF) [14]. Genomics studies have compared the
gene profile from different chambers [15–18] but little has been re-
ported on the differences between the whole proteome of the RA and
RV from the same patient, and none from infants with VSD. This is an
issue which needs addressing as researchers have in the past drawn
conclusions on ventricular remodelling at the molecular level based on
atrial data, due to ease of tissue access [19]. Atrium biopsy genomics
are useful for a wide array of functions, but are not useful for studying
contractility; ventricular biopsies should be used instead [16].
Finally, there is the problem of availability of control/normal car-
diac tissue. The only way to obtain fresh “control”cardiac tissue is from
age-matched children with a normal heart having just been declared
dead in a hospital setting. These situations are extremely rare and have
stringent ethical considerations. Tissues from heart transplant or from
recent death cannot be considered as control since there will be sig-
nificant protein turnover due to storage and handling. Therefore, we
wished to determine differences between the RA and RV at the protein
level in a group of age-matched neonates with VSD, which would im-
pact on future studies.
2. Patients and methods
Infants with RV hypertension who underwent cardiopulmonary
bypass were recruited to the RVENCH (Right Ventricle Function in
Children) study between 01.07.15 and 31.03.17. The subset of patients
described herein (n = 5, Table 1) had right ventricular pressures at
systemic levels, and hence high pulmonary artery pressure as a result of
a large ventricular septal defect (VSD). One patient had Trisomy 21 but
was still included in the analysis as the ventricular function of these
patients is not known to be any different. Fully informed parental
consent was obtained prior to inclusion in study. Ethical approval was
granted by the National Research Ethics Service number 14/NW/1256,
IRAS 143683.
Right atrial and right ventricular biopsy samples were taken from
the patients at the time of surgery. To maintain protein integrity, all
tissue samples were obtained as soon as possible after instituting car-
diopulmonary bypass. Biopsies were immersed in Allprotect tissue re-
agent (Qiagen, UK) overnight, before being stored at −80 °C until
protein extraction.
2.1. Proteomics methods and analysis
Proteomics methods and analysis have been described elsewhere
[14]. Briefly, for proteomic analysis (performed by the proteomics fa-
cility, BioMedical Sciences Building, University of Bristol), samples
were labelled with Tandem Mass Tag (TMT) 10Plex reagents and ana-
lysed by LCMS using an Orbitrap Fusion Tribrid mass spectrometer
running an SPS-MS3 acquisition. For phospho-proteomic analysis,
phosphorylated TMT-labelled peptides were enriched prior to LCMS
analysis using titanium oxide-based enrichment. Raw data files were
processed and quantified using Proteome Discover software v1.4
(Thermo Scientific) and searched against the Uniprot Human database
(downloaded 18/04/16: 134169 sequences) using the SEQUEST algo-
rithm.
For comparison between chambers, proteins not found in both the
RA and RV of all patients were excluded from the main analysis.
However, proteins only detected in either the RA or RV were also in-
vestigated. Fold changes (a fold decrease/increase > 1.3) between
chambers were determined and a log-2 transformation applied.
Significant changes between the right atria and ventricle were de-
termined by a paired t-test, and a –log 10 transformation of the p-value
applied. Significant fold-changes (p < 0.05) were deemed of biological
importance.
Accession numbers were converted to protein symbol (gene name)
either using the Uniprot database mapping tool (converting ‘UniprotKB
AC/ID’to ‘Gene name’) or by extracting the relevant name from the
protein description assigned by Proteome Discoverer, and these shall be
used synonymously with protein name throughout. Gene names were
not found by these methods for only ~200 accession numbers.
Significantly altered proteins (including cDNA with high similarity se-
quences to proteins) between pathologies were inputted into Ingenuity
Pathway Analysis software (IPA, v39480507, Qiagen) to determine
significantly enriched canonical pathways, and diseases and functions
(calculated by Fisher's exact test right-tailed). Significantly changed
proteins and phosphoproteins were also analysed in the ‘Gene Ontology
enRIchment anaLysis and visuaLizAtion’tool (GOrilla, Database update
v.Feb 4 2017, [20]) to determine enriched gene ontology (GO) terms,
versus a background list comprised of all proteins detected during the
proteomics and phosphoproteomics analysis. A p-value threshold of
0.001 was set, and enrichment false discovery rate q-value
threshold < 0.05 (q-value: correction of p-value for testing of multiple
gene ontology terms).
3. Results
3.1. Comparison between right atrium and right ventricle in VSD
3605 protein accession numbers (representing 3336 different pro-
teins), were detected in all right atria and right ventricle samples from
VSD patients. 150 protein accession numbers were significantly dif-
ferent between heart chambers, all of which were lower in the right
atria (Fig. 1A). 414 phosphorylated proteins were identified, each
containing at least one translation modification at serine, threonine, or
tyrosine (resulting in a total of 700 phosphorylation site matches (as
assigned by Proteome Discoverer Software v1.4)) (Fig. 1B). 19 phos-
phorylation sites, representing 19 unique proteins were significantly
lower in the right atria, compared to the RV. Mean (and standard error)
values for significantly changed proteins and phosphoproteins are
shown in Supplementary Tables A & B, and heat-maps for these are in
Supplementary Tables C & D.
Ingenuity pathway analysis of the significantly changed proteins
determined that 33 canonical pathways were enriched (p < 0.05)
(Supplementary Tables E) and the top three were those for tRNA
charging, tricarboxylic acid cycle II and dopamine degradation (p-va-
lues 3.65E−07, 2.41E−05, and 7.16E−05 respectively). Of the ten
most significantly enriched diseases and functions (Table 2)five
Table 1
Patient demographics and pre-operative data displayed as mean
(SEM).
Mean (SEM)
Age (years) 0.34 (0.05)
Weight (kg) 4.83 (0.22)
Oxygen saturation (%) 97.0 (1.14)
Male:Female 4:1
Systolic blood pressure (mm Hg) 86.4 (4.50)
Fractional shortening (%) 34.6 (0.75)
A.R. Bond et al. Journal of Proteomics 191 (2019) 107–113
108

Fig. 1. Volcano plot of entire set of proteins (A) and phosphoproteins (B) quantified in the right atria (RA) versus right ventricle (RV) in patients with ventricular
septal defect (VSD) (a negative log2(fold change) indicates higher expression in right ventricular samples). Each point represents the difference in expression (log2
fold difference) between the groups, and the associated significance of this change (independent paired samples t-test). Proteins significantly altered ( ± 1.3-fold,
p < 0.05) are found within the grey shaded boxes.
Table 2
Diseases or functions with enriched protein expression for significantly higher total protein levels in the right ventricle compared to the right atria, of VSD patients.
Diseases or functions annotation p-Value Molecules # Molecules
Mitochondrial disorder 5.43E−12 BCS1L, CARS2, COA6, DLAT, ECHS1, ETFA, FARS2, IBA57, MIPEP, NDUFS1, PDHB, PDHX, SCO2, TACO1,
TSFM
15
Lactic acidosis 5.29E−08 COA6, DLAT, PDHB, PDHX, SCO2, TSFM 6
Oxidative phosphorylation deficiency 2.18E−07 CARS2, COA6, FARS2, MIPEP, SCO2, TSFM 6
Cell spreading of keratinocytes 4.14E−07 ITGA5, ITGAV, ITGB1 3
Enzymopathy 8.71E−07 ALDH3A2, BCS1L, CLPB, CPT1A, DLAT, ETFA, HEXA, MCCC2, NDUFS1, PDHB, PDHX, SCO2, TACO1 13
Pyruvate decarboxylase deficiency 4.09E−06 DLAT, PDHB, PDHX 3
Autosomal recessive disease 4.31E−06 ACO2, CARS2, CLPB, COA6, COCH, CSRP3, DHTKD1, ECHS1, EPS8, ETFA, FARS2, GCSH, HARS2, HEXA,
HMGCL, IBA57, ITGB1, KLHL41, MCCC2, MIPEP, MMAB, PDHX, PHYH, PSMB8, SCN5A, SPG7, SUCLA2,
TNIK, TSFM, VLDLR
30
Primary dilated cardiomyopthy 5.23E−06 CSRP3, FLNC, LDB3, SCN5A, SCO2, TSFM 6
Hypertrophic cardiomyopathy 1.82E−05 COA6, CPT1A, CSRP3, FLNC, LDB3, SCN5A, SCO2 7
Hereditary myopathy 1.89E−05 BCS1L, CSRP3, DHTKD1, ECHS1, FLNC, IBA57, ITGB1, KLHL41, LDB3, MTM1, NDUFS1, SCN5A, SCO2,
SPG7, SUCLA2, TAX1BP3
16
A.R. Bond et al. Journal of Proteomics 191 (2019) 107–113
109

metabolic diseases were identified, of which mitochondrial disorder
was the most enriched (the others being lactic acidosis, oxidative
phosphorylation deficiency, enzymopathy, and pyruvate decarboxylase
deficiency). Cardiomyopathy was identified twice, along with cell
spreading of keratinocytes, autosomal recessive disorder and hereditary
myopathy.
Analysis of significantly changed phosphorylated proteins between
chambers showed that there was significant enrichment of those in-
volved in the regulation of cell communication by electrical coupling
(GO:0010649, 3 proteins; HRC-Ser311, CAV1-Tyr42, and CASQ2-
Ser339) and the negative regulation of ion transport (GO:0043271, 4
proteins; HRC-Ser311, CAV1-Tyr42, and CASQ2-Ser339, PACSIN3-
Ser319).
3.2. Heart contraction
Proteins annotated in the Gene Ontology terms ‘Cardiac Muscle
Contraction’(GO:0060048), ‘Heart Contraction’(GO:0060047) and
‘Regulation of heart contraction’(GO:0008016), and including calcium
signalling proteins (Supplementary Table F) were compared to our
dataset. Of the 274 proteins, 107 accession numbers (representing 99
unique proteins) were detected in our samples, but only 4 (CXADR,
CSRP3, GSK3A, and SCN5A) were significantly different between
chambers (all lower in RA).
26 phosphorylated proteins were detected (some with multiple
phosphorylation sites) but only 3 were significantly higher in RV
samples (HRC-Ser311, CAV1-Tyr42, and CASQ2-Ser339).
3.3. Structural and extracellular matrix proteins
Structural proteins in human skeletal muscle [21] were cross-re-
ferenced against protein lists found in our samples.
In the RA and RV of patients with VSD, 106 accession numbers
(representing 94 proteins) were detected, and these were also found in
skeletal muscle. Of these, only five (5.3%) were significantly different
between heart chambers (β-endolase (ENO3; involved in calcium sig-
nalling) and microtubule-associated protein (MAP4; promotes micro-
tubule assembly), and FLNC, CSRP3 and LDB3), all being lower in the
right atria. Phospho-peptide enrichment suggested significantly lower
expression of four distinct phosphorylated proteins in the right atria:
striated muscle preferentially expressed protein kinase (SPEG) –Serine
(Ser)-2448, obscurin (OBSCN) –Ser-5563, EIF5B protein (fragment;
EIF5B) –Ser-164 and sarcoplasmic reticulum histidine-rich calcium-
binding protein (HRC) –Ser-311, which are involved in myocyte cy-
toskeletal assembly, sarcoplasmic M-band, protein biosynthesis and
calcium signalling respectively.
3.4. Mitochondrial proteins
GOrilla enrichment analysis found that significantly altered proteins
were found to be particularly enriched in the mitochondria
(GO:0005739-Mitochondrion, 47 proteins; GO:0005759-Mitochondrial
matrix, 22 proteins) so the list of proteins and phospho-proteins derived
from the analysis was searched using the search terms ‘mitochondria’
and ‘mitochondrial’.The list was also searched using gene lists for
‘Mitochondrial respiratory chain complexes’and ‘Mitochondrial re-
spiratory chain complex assembly factor’, obtained from the HGNC
(HUGO Gene Nomenclature Committee) database [22].
401 mitochondrial proteins were detectable in both the right atria
and right ventricle of patients with VSD, of which 30 were significantly
different between chambers (Table 3); all being higher in the RV. The
majority of these (17 proteins, 57%) have catalytic activity (hydrolase,
isomerase, ligase, lyase, oxidoreductase, and transferase activity) [23]
(Table 3). However, only seven phosphorylated mitochondrial proteins
were detectable (NDUFB4-Ser26, TOMM70A-Ser91, TOMM20-Ser138,
MFF-Ser17, CLUH-Ser702, BCKDHA-Ser308 and -Ser318, BCKDK-
Ser31, PUS1-Thr133), none of which changed between chambers.
4. Discussion
To our knowledge, this is the first comparison of the proteome of the
right atria and right ventricle of infants with ventricular septal defect. It
is surprising to discover that there were not more changes in the pro-
teins involved in heart contraction, as the primary function of the
ventricles is to pump blood out of the heart, whereas the main me-
chanical purpose the atria serves is to be a minimally contractile re-
servoir. The atria of our cohort of patients are mostly normal, however
the proteins associated with contraction of the ventricle are probably
changed in the presence of the VSD due to marked hypertrophy brought
on by exposure to left ventricular pressure in systole. Previous studies
have shown that ventricles have a higher expression of genes associated
with contractility [24], albeit studies tend to be carried out in adults.
Very few studies are carried out in infants, so much of what we know
must be extrapolated from adult or animal data, and a lack of healthy
control tissue could make interpretation of our results difficult.
Genomics studies have been carried out on heart tissue however
these cannot necessarily be directly compared to proteomics. Out of a
set of 13 genes described by Barth et al. [24] as known to be expressed
differently between atria and ventricular samples, only three were also
found in all of our samples (natriuretic peptide A, NPPA; myosin light
chain 3, MYL3; cardiac phospholamban, PLN). Of these, none changed
significantly, although there was a non-significant > 1.3 fold-change in
MYL3 and PLN. They also showed that genes related to metabolism and
mitochondria dominated the expression in ventricular myocardium, as
we have shown here. A genomics study comparing the RA and LV from
adults undergoing aortic valve replacement showed that genes with
higher expression in the LV were mainly associated with contractility
[16]. Proteome differences between chambers in fetal hearts obtained
from elective terminations of healthy pregnancies could perhaps be
seen as control healthy hearts [13]. They identified a number of mar-
kers higher in atrial (MYL7, NPPA, GJA5, PAM, WBP11, and GNAO1)
and ventricular (MYL2, MYL3, MYL5, MYH7, GJA1, RPL3L) tissue. All
of these, with the exception of GJA1 and MYL5, were detected in our
samples, however none were significantly altered; MYL2, MYH7 and
RPL3L (cDNA) were increased non-significantly in our ventricular
samples. It is unclear whether the similar expression between chambers
is caused by VSD, or due to the small sample size. In the study by Lu
et al. [13]they do not appear to have separated the chambers into the
left versus right heart, so any differences may be confounded.
Our analysis showed that a large proportion of the proteins which
were higher in the right ventricle, were associated with the mitochon-
dria, whose primary function is ATP production, via the oxidative
phosphorylation (OXPHOS) pathway. For normal contraction and basal
metabolism, the human (adult) heart consumes 30 kg of ATP per day
[25], and to maintain this demand, ventricular myocardium contains a
large volume of mitochondria (~35% of cardiomyocyte volume in rats)
[26]. The number and size of ventricular mitochondria increase with
age; a change not seen in atrial cardiomyocytes. Adult atrial myocytes
are smaller than ventricular ones, being five times shorter (20 μm and
100 μm respectively) and have a smaller diameter [27], and lower
mitochondrial content leading to reduced activity of oxidative enzymes,
including succinic dehydrogenase (respiratory Complex II) which par-
ticipates in the citric acid cycle and electron transport chain [28,29].
The abundance of glycogen in atrial myocytes correlates well with the
higher activity of phosphorylase, transglycosidase and glycogen syn-
thetase [29]. Incorporation of
3
H-leucine, a measure of the level of
protein synthesis at most stages of cardiomyogenesis, is of the same
order in both atria and ventricular, except during the perinatal period
where ventricular labelling is 10–25% higher [29]. The mitochondrial
and nuclear genome are able to produce the ~90 proteins necessary for
OXPHOS [30]. Aminoacyl-tRNA synthetases (ARSs) are pivotal sub-
strates in the translation of mRNA to the correct amino acid sequences
A.R. Bond et al. Journal of Proteomics 191 (2019) 107–113
110

found in these proteins, by catalysing the direct aminoacetylation of
tRNAs with the correct anticodon sequence [31], and were found to be
significantly enriched in the RV compared to RA of our patients. Mi-
tochondrial phenylalanine-tRNA synthetase (FARS2), cysteine-tRNA
synthase (CARS2), and glutaminyl-tRNA synthase (glutamine-hydro-
lyzing)-like 1 (QRSL1), two subunits of the cytoplasmic form, FARSA
(alpha catalytic subunit) which forms a tetramer with FARSB (beta
regulatory subunit), and cDNA with sequence similarity to HARS2
(histidyl-tRNA-synthetase 2) and WARS2 (tryptophanyl tRNA synthe-
tase 2) were all increased in RV. Four further proteins were also up-
regulated in the RV which are also known to be involved in mi-
tochondrial translational elongation in protein synthesis (elongation
factor Ts, the small mitochondrial ribosomal 28S subunits 6 and 36, and
the large 39S subunit 34). The lower levels of these proteins seen in the
RA correlates well with the reduced energy demands in this chamber.
This is also reflected by the lower protein level of the cytoplasmic en-
zymes aconitase-1 and -2, dehydrogenase E1 and transketolase, and
succinate-CoA ligase ADP-forming (ACO1, ACO2, DHTKD1 and
SUCLA2 respectively) which are all involved in the TCA cycle which is
at the core of cell metabolism. Of the 30 mitochondrial proteins higher
in the RV, 22 have known catalytic activity [32] but the lack of sig-
nificantly altered phosphorylated mitochondrial proteins suggests there
is no change in the activation of these enzymes.
Atrial cardiomyocytes produce a potent natriuretic peptide (NPPA)
which increases in response to atrial stretch (increased blood volume)
but also increases in the RV in response to hypertrophy [33]. NPPA was
detected in our patients, however surprisingly there is no difference
between chambers. In the absence of control tissue it is hard to say
whether the atrial levels are lower than normal, or the ventricular ex-
pression is higher. One patient appears to have much higher levels of
NPPA in the right ventricle and lower levels in the atria (RV:RA ratio of
18.9), compared to the other patients (RV:RA ratio of 0.85 ± 0.78;
mean ± SEM). When excluded, this leads to levels in the RV being
1.41-fold lower (albeit non-significantly) than those in the atria.
Table 3
Significantly increased mitochondrial proteins in the right ventricle (vs. right atria).
Accession Gene Description Enzyme Activity
Q9P1A0 NDUFS1 NDUFS1 protein OS = Homo sapiens GN = NDUFS1 PE = 2 SV = 1 - [Q9P1A0_HUMAN] Oxidoreductase
Q9BQ48 MRPL34 39S ribosomal protein L34, mitochondrial OS = Homo sapiens GN = MRPL34 PE = 1 SV = 1 -
[RM34_HUMAN]
P43897 TSFM Elongation factor Ts, mitochondrial OS = Homo sapiens GN = TSFM PE = 1 SV = 2 - [EFTS_HUMAN]
Q96HY7 DHTKD1 Probable 2-oxoglutarate dehydrogenase E1 component DHKTD1, mitochondrial OS = Homo sapiens
GN = DHTKD1 PE = 1 SV = 2 - [DHTK1_HUMAN]
Oxidoreductase
Q6UWS5 PET117 Protein PET117 homolog, mitochondrial OS = Homo sapiens GN=PET117 PE = 3 SV = 1 - [PT117_HUMAN]
B2RBJ8 QRSL1 Glutamyl-tRNA(Gln) amidotransferase subunit A, mitochondrial OS = Homo sapiens GN = QRSL1 PE = 2
SV = 1 - [B2RBJ8_HUMAN]
Hydrolase/Ligase
P30837 ALDH1B1 Aldehyde dehydrogenase X, mitochondrial OS = Homo sapiens GN = ALDH1B1 PE = 1 SV = 3 -
[AL1B1_HUMAN]
Oxidoreductase
Q9BT30 ALKBH7 Alpha-ketoglutarate-dependent dioxygenase alkB homolog 7, mitochondrial OS = Homo sapiens
GN = ALKBH7 PE = 1 SV = 1 - [ALKB7_HUMAN]
Q6QN92 GCSH Mitochondrial glycine cleavage system H-protein (Fragment) OS = Homo sapiens PE = 2 SV = 1 -
[Q6QN92_HUMAN]
P83111 LACTB Serine beta-lactamase-like protein LACTB, mitochondrial OS = Homo sapiens GN = LACTB PE = 1 SV = 2 -
[LACTB_HUMAN]
Hydrolase
Q9HA77 CARS2 Probable cysteine—tRNA ligase, mitochondrial OS = Homo sapiens GN = CARS2 PE = 1 SV = 1 -
[SYCM_HUMAN]
Ligase
P11177 PDHB Pyruvate dehydrogenase E1 component subunit beta, mitochondrial OS = Homo sapiens GN = PDHB PE = 1
SV = 3 - [ODPB_HUMAN]
Lyase/Oxidoreductase
P82932 MRPS6 28S ribosomal protein S6, mitochondrial OS = Homo sapiens GN = MRPS6 PE = 1 SV = 3 - [RT06_HUMAN]
P27144 AK4 Adenylate kinase 4, mitochondrial OS = Homo sapiens GN = AK4 PE = 1 SV = 1 - [KAD4_HUMAN]
Q96HS1 PGAM5 Serine/threonine-protein phosphatase PGAM5, mitochondrial OS = Homo sapiens GN = PGAM5 PE = 1
SV = 2 - [PGAM5_HUMAN]
Q9NUJ1 ABHD10 Mycophenolic acid acyl-glucuronide esterase, mitochondrial OS = Homo sapiens GN = ABHD10 PE = 1
SV = 1 - [ABHDA_HUMAN]
Q5JTJ3 COA6 Cytochrome c oxidase assembly factor 6 homolog OS = Homo sapiens GN = COA6 PE = 1 SV = 1 -
[COA6_HUMAN]
P13804 ETFA Electron transfer flavoprotein subunit alpha, mitochondrial OS = Homo sapiens GN = ETFA PE = 1 SV = 1 -
[ETFA_HUMAN]
Oxidoreductase
P82909 MRPS36 28S ribosomal protein S36, mitochondrial OS = Homo sapiens GN = MRPS36 PE = 1 SV = 2 -
[RT36_HUMAN]
O43819 SCO2 Protein SCO2 homolog, mitochondrial OS = Homo sapiens GN = SCO2 PE = 1 SV = 3 - [SCO2_HUMAN] Oxidoreductase
Q9HCC0 MCCC2 Methylcrotonoyl-CoA carboxylase beta chain, mitochondrial OS = Homo sapiens GN = MCCC2 PE = 1
SV = 1 - [MCCB_HUMAN]
Ligase
Q7Z434 MAVS Mitochondrial antiviral-signalling protein OS = Homo sapiens GN = MAVS PE = 1 SV = 2 - [MAVS_HUMAN]
P30405 PPIF Peptidyl-prolyl cis-trans isomerase F, mitochondrial OS = Homo sapiens GN = PPIF PE = 1 SV = 1 -
[PPIF_HUMAN]
Isomerase
Q5T440 IBA57 Putative transferase CAF17, mitochondrial OS = Homo sapiens GN = IBA57 PE = 1 SV = 1 -
[CAF17_HUMAN]
P30084 ECHS1 Enoyl-CoA hydratase, mitochondrial OS = Homo sapiens GN = ECHS1 PE = 1 SV = 4 - [ECHM_HUMAN] Isomerase/Ligase/Lyase/
Oxidoreductase
Q9Y305 ACOT9 Acyl-coenzyme A thioesterase 9, mitochondrial OS = Homo sapiens GN = ACOT9 PE = 1 SV = 2 -
[ACOT9_HUMAN]
Hydrolase
Q99797 MIPEP Mitochondrial intermediate peptidase OS = Homo sapiens GN = MIPEP PE = 1 SV = 2 - [MIPEP_HUMAN] Hydrolase
Q7Z4Y4 AK3 GTP:AMP phosphotransferase AK3, mitochondrial OS = Homo sapiens GN = AK3 PE = 2 SV = 1 -
[Q7Z4Y4_HUMAN]
A0A0K0K1H7 HEL-S-284 Aconitate hydratase, mitochondrial OS = Homo sapiens GN = HEL-S-284 PE = 2 SV = 1 -
[A0A0K0K1H7_HUMAN]
O95363 FARS2 Phenylalanine—tRNA ligase, mitochondrial OS = Homo sapiens GN = FARS2 PE = 1 SV = 1 -
[SYFM_HUMAN]
Ligase
A.R. Bond et al. Journal of Proteomics 191 (2019) 107–113
111

The main analysis performed was on proteins found in all patient
samples. However, the data was also interrogated to find whether any
proteins were only detected in either the RV or RA of the majority of the
patients (detected in either RV or RA of ≥3 patients). Alpha-protein
kinase 2 (ALPK2; a kinase that recognises phosphorylation sites in
which surrounding peptides have an alpha-helical conformation) was
found in the RV, but not the RA, of four patients, and the proteins
Cytochrome C Oxidase Subunit 1 (COX1; involved in energy metabo-
lism), Collagen-XXI alpha chain (COL21A1; involved in collagen bio-
synthesis), Protein FAM81A (FAM81A) and Drebrin-like protein (DBNL;
has a role in actin cytoskeleton reorganization) were only found in the
RV of three of the five patients (albeit not the same three each time).
Gelsolin (GSN) was found only in the RA of three patients. Gelsolin has
previously been shown to have increased expression in failing hearts
[34].
Despite our patients being age-matched, paired atrial and ven-
tricular samples being taken, and experience of the surgeon, an un-
known variable is the exact location within the chamber wall from
which the biopsies were taken, with slight discrepancies perhaps
leading to unexpected differences. The right atria is made up of three
separate sections (a venous component, an appendage and a vestibule)
each derived differently embryonically [35] and as such may have very
different proteomic characteristics. Increasing the patient sample size in
the future would hopefully enable us to statistically minimise the effect
of this variation.
5. Conclusions
We can conclude that there are differences in the proteome of the
RA and RV of infants with VSD. As discussed, there were large differ-
ences in mitochondrial proteins, which are probably linked to the en-
ergy demands of the ventricle. We did not see much evidence for
changes in proteins associated with hypertrophy in the right ventricle,
or for differences in the contractile mechanisms between the two
chambers studied. The findings described here will be beneficial to the
future study of multiple congenital heart defects, where there are sus-
pected changes in the atrial and ventricular function, such as atrial
septal defect, or TOF. It is however clear that the right atria is not a
good model of protein expression in the right ventricle, despite its re-
lative ease of access. It is hoped that the findings described here will be
beneficial to future research into heart chamber malformations in
congenital heart defects.
Funding
Funding was obtained from Sparks, the Childrens charity *
13BTL01. This study was supported by the NIHR Biomedical Research
Centre at the University Hospitals Bristol NHS Foundation Trust and the
University of Bristol.
Appendix A. Supplementary data
Supplementary data to this article can be found online at https://
doi.org/10.1016/j.jprot.2018.03.022.
References
[1] J.I.E. Hoffmann, S. Kaplan, The incidence of congenital heart disease, J. Am. Coll.
Cardiol. 39 (12) (2002) 1890–1900.
[2] M.S. Minette, D.J. Sahn, Ventricular septal defects, Circulation 114 (20) (2006)
2190–2197.
[3] R.A. Bloomfield, H.D. Lauson, A. Cournand, E.S. Breed, D.W. Richards, Recording of
right heart pressures in normal subjects and in patients with chronic pulmonary
disease and various types of cardio-circulatory disease, J. Clin. Invest. 25 (4) (1946)
639–664.
[4] V.A. Barnett, Cardiac myocytes, in: P.A. Iaizzo (Ed.), Handbook of Cardiac
Anatomy, Physiology and Devices, Humana Press Inc., Totowa, NJ, 2005, pp.
113–122.
[5] I. Luss, P. Boknik, L.R. Jones, U. Kirchhefer, J. Knapp, B. Linck, H. Luss, A. Meissner,
F.U. Muller, W. Schmitz, U. Vahlensieck, J. Neumann, Expression of cardiac calcium
regulatory proteins in atrium v ventricle in different species, J. Mol. Cell. Cardiol.
31 (6) (1999) 1299–1314.
[6] A.W. Trafford, J.D. Clarke, M.A. Richards, D.A. Eisner, K.M. Dibb, Calcium sig-
nalling microdomains and the t-tubular system in atrial mycoytes: potential roles in
cardiac disease and arrhythmias, Cardiovasc. Res. 98 (2) (2013) 192–203.
[7] M. Frisk, J.T. Koivumaki, P.A. Norseng, M.M. Maleckar, O.M. Sejersted, W.E. Louch,
Variable t-tubule organization and Ca2+ homeostasis across the atria, Am. J.
Physiol. Heart Circ. Physiol. 307 (4) (2014) H609–20.
[8] M.D. Bootman, D.R. Higazi, S. Coombes, H.L. Roderick, Calcium signalling during
excitation-contraction coupling in mammalian atrial myocytes, J. Cell Sci. 119 (Pt
19) (2006) 3915–3925.
[9] R. Arora, G.L. Aistrup, S. Supple, C. Frank, J. Singh, S. Tai, A. Zhao, L. Chicos,
W. Marszalec, A. Guo, L.S. Song, J.A. Wasserstrom, Regional distribution of T-tu-
bule density in left and right atria in dogs, Heart Rhythm. 14 (2) (2017) 273–281.
[10] R. Heying, M. Qing, K. Schumacher, M. Sokalska-Duhme, J.F. Vazquez-Jimenez,
M.C. Seghaye, Myocardial cardiotrophin-1 is differentially induced in congenital
cardiac defects depending on hypoxemia, Futur. Cardiol. 10 (1) (2014) 53–62.
[11] V. Pelouch, M. Milerova, B. Ostadal, B. Hucin, M. Samanek, Differences between
atrial and ventricular protein profiling in children with congenital heart disease,
Mol. Cell. Biochem. 147 (1995) 43–49.
[12] H. Liu, G. Chen, H. Zheng, H. Qin, M. Liang, K. Feng, Z. Wu, Differences in atrial
fibrillation associated proteins between the left and right atrial appendages from
patients with rheumatic mitral valve disease: a comparative proteomic analysis,
Mol. Med. Rep. 14 (5) (2016) 4232–4242.
[13] Z.Q. Lu, A. Sinha, P. Sharma, T. Kislinger, A.O. Gramolini, Proteomic analysis of
human fetal atria and ventricle, J. Proteome Res. 13 (12) (2014) 5869–5878.
[14] A.R. Bond, D. Iacobazzi, S. Abdul-Ghani, M. Ghorbel, K. Heesom, M. Wilson,
C. Gillett, S.J. George, M. Caputo, S. Suleiman, R.M.R. Tulloh, Changes in con-
tractile protein expression are linked to ventricular stiffness in infants with pul-
monary hypertension or right ventricular hypertrophy due to congenital heart
disease, Open Heart 5 (1) (2018) e000716.
[15] M. Tsubakihara, N.K. Williams, A. Keogh, C.G. dos Remedios, Comparison of gene
expression between left atria and left ventricles from non-diseased humans,
Proteomics 4 (1) (2004) 261–270.
[16] J. Asp, J. Synnergren, M. Jonsson, G. Dellgren, A. Jeppsson, Comparison of human
cardiac gene expression profiles in paired samples of right atrium and left ventricle
collected in vivo, Physiol. Genomics 44 (1) (2012) 89–98.
[17] P. Ellinghaus, R.J. Scheubel, D. Dobrev, U. Ravens, J. Holtz, J. Huetter, U. Nielsch,
H. Morawietz, Comparing the global mRNA expression profile of human atrial and
ventricular myocardium with high-density oligonucleotide arrays, J. Thorac.
Cardiovasc. Surg. 129 (6) (2005) 1383–1390.
[18] S. Kaab, A.S. Barth, D. Margerie, M. Dugas, M. Gebauer, L. Zwermann, S. Merk,
A. Pfeufer, K. Steinmeyer, M. Bleich, E. Kreuzer, G. Steinbeck, M. Nabauer, Global
gene expression in human myocardium-oligonucleotide microarray analysis of re-
gional diversity and transcriptional regulation in heart failure, J. Mol. Med. (Berl)
82 (5) (2004) 308–316.
[19] H. Zhang, D.X. Gong, Y.J. Zhang, S.J. Li, S. Hu, Effect of mitochondrial aldehyde
dehydrogenase-2 genotype on cardioprotection in patients with congenital heart
disease, Eur. Heart J. 33 (13) (2012) 1606–1614.
[20] E. Eden, R. Navon, I. Steinfeld, D. Lipson, Z. Yakhini, GOrilla: a tool for discovery
and visualization of enriched GO terms in ranked gene lists, BMC Bioinformatics 10
(2009) 48.
[21] K. Hojlund, B.P. Bowen, H. Hwang, C.R. Flynn, L. Madireddy, T. Geetha,
P. Langlais, C. Meyer, L.J. Mandarino, Z. Yi, In vivo phosphoproteome of human
skeletal muscle revealed by phosphopeptide enrichment and HPLC-ESI-MS/MS, J.
Proteome Res. 8 (11) (2009) 4954–4965.
[22] H.D. [Internet]. HGNC Database, HUGO Gene Nomenclature Committee (HGNC),
EMBL Outstation - Hinxton, European Bioinformatics Institute, Wellcome Trust
Genome Campus, Hinxton, Cambridgeshire, CB10 1SD, UK.
[23] H. Mi, X. Huang, A. Muruganujan, H. Tang, C. Mills, D. Kang, P.D. Thomas,
PANTHER version 11: expanded annotation data from gene ontology and reactome
pathways, and data analysis tool enhancements, Nucleic Acids Res. 45 (D1) (2017)
D183–D189.
[24] A.S. Barth, S. Merk, E. Arnoldi, L. Zwermann, P. Kloos, M. Gebauer, K. Steinmeyer,
M. Bleich, S. Kaab, A. Pfeufer, P. Uberfuhr, M. Dugas, G. Steinbeck, M. Nabauer,
Functional profiling of human atrial and ventricular gene expression, Pflugers Arch.
450 (4) (2005) 201–208.
[25] R. Ferrari, S. Censi, F. Mastrorilli, A. Boraso, Prognostic benefits of heart rate re-
duction in cardiovascular disease, Eur. Heart J. Suppl. 5 (G) (2003) G10–G14.
[26] E. Page, L.P. McCallister, Quantitative electron microscopic description of heart
muscle cells. Application to normal, hypertrophied and thyroxin-stimulated hearts,
Am. J. Cardiol. 31 (2) (1973) 172–181.
[27] M.J. Legato, Ultrastructure of the atrial, ventricular, and Purkinje cell, with special
reference to the genesis of arrhythmias, Circulation 47 (1) (1973) 178–189.
[28] H. Plattner, F. Tiefenbrunner, W. Pfaller, Cytomorphometric and biochemical dif-
ferences between the muscle cells in atria and ventricles of the guinea pig heart,
Anat. Rec. 167 (1) (1970) 11–16.
[29] P.P. Rumyantsev, B.M. Carlson, Growth and Hyperplasia of Cardiac Muscle Cells,
Harwood Academic, London, 1991.
[30] P. Smits, J. Smeitink, L. van den Heuvel, Mitochondrial translation and beyond:
processes implicated in combined oxidative phosphorylation deficiencies, J Biomed
Biotechnol 2010 (2010) 737385.
[31] M. Ibba, D. Soll, Aminoacyl-tRNA synthesis, Annu. Rev. Biochem. 69 (2000)
617–650.
A.R. Bond et al. Journal of Proteomics 191 (2019) 107–113
112

[32] D. Szklarczyk, A. Franceschini, S. Wyder, K. Forslund, D. Heller, J. Huerta-Cepas,
M. Simonovic, A. Roth, A. Santos, K.P. Tsafou, M. Kuhn, P. Bork, L.J. Jensen, C. von
Mering, STRING v10: protein-protein interaction networks, integrated over the tree
of life, Nucleic Acids Res. 43 (database issue) (2015) D447–52.
[33] H. Ruskoaho, J. Leppaluoto, Immunoreactive atrial natriuretic peptide in ventricles,
atria, hypothalamus, and plasma of genetically hypertensive rats, Circ. Res. 62 (2)
(1988) 384–394.
[34] J. Yang, C.S. Moravec, M.A. Sussman, N.R. DiPaola, D. Fu, L. Hawthorn,
C.A. Mitchell, J.B. Young, G.S. Francis, P.M. McCarthy, M. Bond, Decreased SLIM1
expression and increased gelsolin expression in failing human hearts measured by
high-density oligonucleotide arrays, Circulation 102 (25) (2000) 3046–3052.
[35] R.H. Anderson, A.C. Cook, The structure and components of the atrial chambers,
Europace 9 (Suppl. 6) (2007) vi3–9.
A.R. Bond et al. Journal of Proteomics 191 (2019) 107–113
113