Proteomics analysis of the p.G849D variant in neurexin 2 alpha may reveal insight into Parkinson’s disease pathobiology

Abstract

Parkinson’s disease (PD), the fastest-growing neurological disorder globally, has a complex etiology. A previous study by our group identified the p.G849D variant in neurexin 2 ( NRXN2 ), encoding the synaptic protein, NRXN2α, as a possible causal variant of PD. Therefore, we aimed to perform functional studies using proteomics in an attempt to understand the biological pathways affected by the variant. We hypothesized that this may reveal insight into the pathobiology of PD. Wild-type and mutant NRXN2α plasmids were transfected into SH-SY5Y cells. Thereafter, total protein was extracted and prepared for mass spectrometry using a Thermo Scientific Fusion mass spectrometer equipped with a Nanospray Flex ionization source. The data were then interrogated against the UniProt H. sapiens database and afterward, pathway and enrichment analyses were performed using in silico tools. Overexpression of the wild-type protein led to the enrichment of proteins involved in neurodegenerative diseases, while overexpression of the mutant protein led to the decline of proteins involved in ribosomal functioning. Thus, we concluded that the wild-type NRXN2α may be involved in pathways related to the development of neurodegenerative disorders, and that biological processes related to the ribosome, transcription, and tRNA, specifically at the synapse, could be an important mechanism in PD. Future studies targeting translation at the synapse in PD could therefore provide further information on the pathobiology of the disease.

Full text

Frontiers in Aging Neuroscience 01 frontiersin.org
Proteomics analysis of the
p.G849D variant in neurexin 2
alpha may reveal insight into
Parkinson’s disease pathobiology
Katelyn Cuttler
1, Suereta Fortuin
2, Amica Corda
Müller-Nedebock
1,3, Maré Vlok
4, Ruben Cloete
5 and
Soraya Bardien
1,3*
1 Division of Molecular Biology and Human Genetics, Faculty of Medicine and Health Sciences,
Stellenbosch University, Cape Town, South Africa, 2 Faculty of Medicine and Health Sciences,
African Microbiome Institute, Stellenbosch University, Cape Town, South Africa, 3 South African
Medical Research Council/Stellenbosch University Genomics of Brain Disorders Research Unit,
Cape Town, South Africa, 4 Mass Spectrometry Unit, Central Analytical Facilities, Stellenbosch
University, Cape Town, South Africa, 5 South African Medical Research Council Bioinformatics Unit,
SouthAfrican National Bioinformatics Institute, University of the Western Cape, Cape Town, South
Africa
Parkinson’s disease (PD), the fastest-growing neurological disorder globally,
has a complex etiology. A previous study by our group identified the p.G849D
variant in neurexin 2 (NRXN2), encoding the synaptic protein, NRXN2α, as
a possible causal variant of PD. Therefore, weaimed to perform functional
studies using proteomics in an attempt to understand the biological pathways
aected by the variant. We hypothesized that this may reveal insight into
the pathobiology of PD. Wild-type and mutant NRXN2α plasmids were
transfected into SH-SY5Y cells. Thereafter, total protein was extracted and
prepared for mass spectrometry using a Thermo Scientific Fusion mass
spectrometer equipped with a Nanospray Flex ionization source. The data
were then interrogated against the UniProt H. sapiens database and afterward,
pathway and enrichment analyses were performed using in silico tools.
Overexpression of the wild-type protein led to the enrichment of proteins
involved in neurodegenerative diseases, while overexpression of the mutant
protein led to the decline of proteins involved in ribosomal functioning. Thus,
weconcluded that the wild-type NRXN2α may beinvolved in pathways related
to the development of neurodegenerative disorders, and that biological
processes related to the ribosome, transcription, and tRNA, specifically at the
synapse, could be an important mechanism in PD. Future studies targeting
translation at the synapse in PD could therefore provide further information
on the pathobiology of the disease.
KEYWORDS
neurexin 2α (NRXN2), Parkinson’s disease, proteomics, mass spectrometry, p.G849D,
synaptic translation, mitochondrial dysfunction, ribosomal functioning
TYPE Brief Research Report
PUBLISHED 30 November 2022
DOI 10.3389/fnagi.2022.1002777
OPEN ACCESS
EDITED BY
Sasanka Chakrabarti,
Maharishi Markandeshwar University,
India
REVIEWED BY
Bharat Singh,
Maharishi Markandeshwar University,
India
Phalguni Anand Alladi,
National Institute of Mental Health and
Neurosciences, India
*CORRESPONDENCE
Soraya Bardien
sbardien@sun.ac.za
SPECIALTY SECTION
This article was submitted to
Parkinson’s Disease and Aging-related
Movement Disorders,
a section of the journal
Frontiers in Aging Neuroscience
RECEIVED 25 July 2022
ACCEPTED 08 November 2022
PUBLISHED 30 November 2022
CITATION
Cuttler K, Fortuin S, Müller-Nedebock AC,
Vlok M, Cloete R and Bardien S (2022)
Proteomics analysis of the p.G849D variant
in neurexin 2 alpha may reveal insight into
Parkinson’s disease pathobiology.
Front. Aging Neurosci. 14:1002777.
doi: 10.3389/fnagi.2022.1002777
COPYRIGHT
© 2022 Cuttler, Fortuin, Müller-Nedebock,
Vlok, Cloete and Bardien. This is an open-
access article distributed under the terms
of the Creative Commons Attribution
License (CC BY). The use, distribution or
reproduction in other forums is permitted,
provided the original author(s) and the
copyright owner(s) are credited and that
the original publication in this journal is
cited, in accordance with accepted
academic practice. No use, distribution or
reproduction is permitted which does not
comply with these terms.

Cuttler et al. 10.3389/fnagi.2022.1002777
Frontiers in Aging Neuroscience 02 frontiersin.org
Introduction
Parkinson’s disease (PD) is an incurable neurodegenerative
disorder which primarily aects movement, resulting in
bradykinesia, rigidity, postural instability, resting tremor, and a
range of neuropsychiatric symptoms. Notably, it has been reported
to bethe fastest-growing neurological disorder globally (Feigin
etal., 2017), aecting over 6 million people (Dorsey etal., 2018).
Over the past two and a half decades, several genetic causes of PD
have been identied, implicating various biological processes
including mitochondrial dysfunction, toxic protein accumulation,
dysfunctional vesicle recycling, and synaptic dysfunction in PD
development (Panicker etal., 2021).
Recently, wereported the nding of a p.G849D variant in the
neurexin 2 gene (NRXN2) in a SouthAfrican multiplex PD family
(Sebate etal., 2021). e translated protein, NRXN2α, is a synaptic
regulation protein involved in processes such as calcium channel
regulation, neuronal cell adhesion, and transmembrane signaling
(Craig and Kang, 2007). ere have been a limited number of
studies on NRXN2α in disease, but a few have implicated the
protein in neuronal and synaptic disorders (Missler etal., 2003;
Dachtler etal., 2015). In addition, it has been shown that neurexins
and their common binding partners, neuroligins, link synaptic
dysfunction to cognitive disease (Südhof, 2008).
Here, weaimed to further investigate the eect of the p.G849D
variant on biological pathways by using a proteomics approach in
an SH-SY5Y cellular model of PD, transfected with wild-type and
mutant NRXN2α plasmids. To this end, weexamined the total
proteome of the dierent treatment groups in an attempt to
understand the changes in biological pathways. Wehypothesized
that overexpression of the wild-type NRXN2α would provide an
indication of the pathways related to NRXN2α’s function, while
overexpression of the mutant NRXN2α could give an indication
of the method of action by which it potentially leads to
neurodegeneration. Together, these ndings may provide a better
understanding of the function of both the wild-type and mutant
NRXN2α, and its possible involvement in PD.
Materials and methods
Ethical considerations
Ethical approval was obtained from the Health Research Ethics
Committee (Protocol numbers 2002/C059 and S20/01/005 PhD)
and the Research Ethics Committee: Biological and Environmental
Safety (Protocol number BEE-2021-13149). Both committees are
located at Stellenbosch University, Cape Town, SouthAfrica.
Cell culture
SH-SY5Y cells were cultured in DMEM with high glucose
(4.5 g/l) and 4 mM L-Glutamine (Lonza). In addition, the
media was supplemented with 15% FBS (Gibco) and 1%
penicillin/streptomycin (Sigma Aldrich). Cells were
maintained at 37°C and 5% CO2 in a humidified incubator
(ESCO Technologies).
Plasmids
NRXN2α wild-type
e NRXN2α-ECFP-N1 plasmid is a kind gi from Prof.
Ann Marie Craig (University of British Columbia, Canada). is
plasmid expresses wild-type mouse NRXN2α-CFP and was
generated as per Kang etal. (2008). e pECFP-N1 plasmid
without an insert (empty vector) was a kind gi from Prof.
Harald Sitte (Medical University of Vienna, Austria).
Site-directed mutagenesis
In order to generate the p.G849D mutant plasmid [p.G882D
in our mouse model, mouse genomic position: 540693_chr19
(GRCm39)], site-directed mutagenesis was performed on the
wild-type NRXN2α-ECFP-N1 plasmid using the Q5 Site-
Directed Mutagenesis kit (New England Biolabs), as per the
manufacturer’s instructions. More information on the primers
used and PCR conditions can befound in the Supplementary
material.
Treatment groups
A total of four treatment groups were used for the analysis: (1)
non-transfected cells (NT), (2) cells transfected with the wild-type
plasmid (WT), (3) cells transfected with the mutant plasmid
(MUT), and (4) cells transfected with the empty vector (EV). All
treatments were performed in triplicate.
Transfection
SH-SY5Y cells were grown in sterile 25 cm
3
asks until 70%
conuent and transfected using Lipofectamine3000 (Invitrogen)
as per the manufacturer’s instructions. e transfection
eciency was determined by examining the cells under an
Oxion Inverso Fluo E4 uorescent microscope (Euromex) at
100x magnication for the presence of cyan uorescent
protein (CFP).
NRXN2α levels
Prior to proteomics analysis, NRXN2α protein levels were
determined using immunouorescent ow cytometry and
measured with the Guava® Muse® Cell Analyzer (Luminex) to
conrm overexpression of NRXN2α. Please see
Supplementary material for more details.

Cuttler et al. 10.3389/fnagi.2022.1002777
Frontiers in Aging Neuroscience 03 frontiersin.org
Proteomics analysis
Protein extraction and clean-up
Cells were detached using Trypsin–EDTA and centrifuged at
2739 ×g for 5 min to collect cell pellets. Cell pellets were stored at
−80°C until required. e pellets were then thawed in 100 mM
Tris buer pH 8 containing 0.5% sodium dodecyl sulfate (SDS,
Sigma), 100 mM NaCl (Sigma), 5 mM triscarboxyethyl phosphine
(TCEP, Sigma), protease inhibitor cocktail (ermo Fisher), and
2 mM EDTA (ermo Fisher). Once thawed, the pellets were
submerged in an ice-cold sonic bath for 30 s prior to vortexing for
30 s. is cycle was repeated three times and the pellets were
completely dissolved. Extraction reagents were then removed
using a chloroform-methanol–water liquid–liquid extraction
method. More information on the protein extraction, on-bead
digest and liquid chromatography performed in preparation for
mass spectrometry, can befound in the Supplementary material.
Mass spectrometry
Mass spectrometry was performed by Stellenbosch
University’s Central Analytical Facilities (CAF) using a ermo
Scientic Fusion mass spectrometer equipped with a Nanospray
Flex ionization source. e sample was introduced through a
stainless-steel emitter. Data were collected in positive mode with
spray voltage set to 1.8 kV and ion transfer capillary set to
280°C. Spectra were internally calibrated using polysiloxane ions
at m/z = 445.12003 and 371.10024. MS1 scans were performed
using the Orbitrap detector set at 120,000 resolution over the scan
range 350–1,650 with automatic gain control (AGC) target at 3 E5
and maximum injection time of 40 milliseconds. Data were
acquired in prole mode.
MS2 acquisitions were performed using monoisotopic
precursor selection for ion with charges +2 − +7 with error
tolerance set to ± 10 ppm. Precursor ions were excluded from
fragmentation once for a period of 60 s. Precursor ions were
selected for fragmentation in higher-energy C-trap dissociation
(HCD) mode using the quadrupole mass analyzer with HCD
energy set to 32.5%. Fragment ions were detected in the Orbitrap
mass analyzer set to 30,000 resolution. e AGC target was set to
5E4 and the maximum injection time to 80 milliseconds. e data
were acquired in centroid mode.
Data analysis
e raw les generated by the mass spectrometer were
imported into Proteome Discoverer v1.4 (ermo Fisher) and
processed using the SequestHT algorithm. Database interrogation
was performed against the UniProt H. Sapiens database
concatenated with the cRAP contaminant protein database.
1
Semi-
tryptic cleavage with 2 missed cleavages was allowed for. Precursor
mass tolerance was set to 10 ppm and fragment mass tolerance set
to 0.02 Da. Deamidation (NQ) and oxidation (M) were allowed as
1 https://www.thegpm.org/crap
dynamic modications and thiomethyl of C as static modication.
Peptide validation was performed using the Target-Decoy PSM
validator node. e results les were imported into Scaold 1.4.4
(Searle, 2010) and identied peptides validated with X!Tandem
and the Peptide and Protein Prophet algorithms included in
Scaold. Quantitation was performed by Scaold aer one-way
ANOVA and Student’s t-test were performed.
Pathway analysis and enrichment
analysis
First, the data for the separate treatment groups were
combined into a Venn diagram using Venny 2.12 (Oliveros, 2015)
to identify proteins unique to each treatment group. ereaer, in
order to identify dierentially abundant proteins, the data were
compared as follows: empty vector transfected cells vs
non-transfected cells (EV vs NT), wild-type transfected cells vs
non-transfected cells (WT vs NT), mutant transfected cells vs
non-transfected cells (MUT vs NT), and mutant transfected cells
vs wild-type transfected cells (MUT vs WT). Functional
information for unique and dierentially abundant proteins was
obtained from UniProt3 (e UniProt Consortium, 2021).
Pathway analysis was conducted using the KEGG4 (Kanehisa etal.,
2021) and STRING5 (Szklarczyk etal., 2019) databases. Each
protein set was then uploaded to WebGestalt
6
(Liao etal., 2019)
for enrichment analysis as per the default parameters.
Results
Overexpression of NRXN2α
Transfection eciency, determined by evaluating CFP
microscopically, was 68% for the wild-type construct, 65% for the
mutant construct, and 67% for the empty vector construct. In
addition, overexpression of NRXN2α was conrmed using the
Guava
®
Muse
®
Cell Analyzer (Luminex). ere was a 26 and 21%
increase in NRXN2α levels in the wild-type and mutant samples,
respectively, with no change in the empty vector sample when
compared to non-transfected cells (Supplementary Figure S1).
Total proteins identified
e total ion chromatograms in Supplementary Figure S2
show the successful digestion of peptides in each sample as well
2 https://bioinfogp.cnb.csic.es/tools/venny
3 https://www.uniprot.org
4 https://www.kegg.jp
5 https://string-db.org
6 http://www.webgestalt.org

Cuttler et al. 10.3389/fnagi.2022.1002777
Frontiers in Aging Neuroscience 04 frontiersin.org
as their protein prole. Quantitation and regression analysis
performed in Scaold showed that most data points clustered
within one standard deviation from the mean and that all the
sample sets showed the same grouping (Figure1A), indicating a
successful experiment. e total number of proteins detected was
2,667, 2,630, 2,691, and 2,646 for the non-transfected cells (NT),
wild-type transfected cells (WT), mutant transfected cells
(MUT), and empty vector transfected cells (EV), respectively.
Since all treatments were performed in triplicate, a protein had to
bepresent in a minimum of 2 replicates in order to beconsidered
an identied protein.
Unique proteins in each group
e Venn diagram in Figure 1B shows that 1822 proteins
were shared by all four treatment groups. NT cells had 31 unique
proteins, while WT, MUT, and EV cells had 22, 44, and 28 unique
proteins, respectively. Functional information for all of these
proteins can be found in Supplementary Tables S1–S4.
Each protein set was then uploaded to STRING7 (Szklarczyk
etal., 2019) for pathway analysis.
Enrichment terms obtained from STRING for the unique
proteins in each treatment group are shown in Table1. ere
were no enriched terms for the EV cells, so they were excluded
from the table. e NT cells showed enrichment terms for
“compound binding.” “Acetylation” was the only term enriched
for in the WT cells. STRING analysis of the proteins involved in
acetylation shows that they are not predicted to interact with each
other and do not form part of the same networks. erefore,
enrichment of “acetylation” in these cells is likely to bea chance
nding, showing that the cells are undergoing modication upon
transfection with the WT plasmid. However, since acetylation of
proteins is also potentially implicated in neurodegenerative
disorders, such as PD (Yakhine-Diop etal., 2019), it could also
be an important mechanism of action for the WT NRXN2α.
“Metabolic processes” were enriched in the MUT cells, which is
also possibly a result of introducing the MUT NRXN2α into the
7 https://string-db.org
AB
CD EF
FIGURE1
Unique and dierentially abundant proteins identified in this experiment. (A) Scatterplot of the standard deviation (log10) vs mean (log10)
shows good clustering of all samples along the linear regression line. Graph generated by Scaold 1.4.4. (B) Venn diagram of the total
proteins identified in each treatment group shows the distribution of shared proteins and numbers of proteins unique to each group.
Diagram generated with Venny 2.1 (https://bioinfogp.cnb.csic.es/tools/venny; Oliveros, 2015). Volcano plots showing dierentially
abundant proteins when comparing EV vs NT (C), WT vs NT (D), MUT vs NT (E), and MUT vs WT (F). The gray line indicates the
significance threshold. Significant proteins (p ≤ 0.05, Students t-test) are colored purple. Where possible, individual proteins have been
labelled. Proteins found within the red funnel may become statistically insignificant with increased sample sizes. Graphs generated with
GraphPad Prism® 5.02. Abbreviations: EV: empty vector transfected cells, MUT: mutant transfected cells; NT: non-transfected cells; WT:
wild-type transfected cells.

Cuttler et al. 10.3389/fnagi.2022.1002777
Frontiers in Aging Neuroscience 05 frontiersin.org
cells. Interestingly, terms related to RNA processes were also
enriched in the MUT cells. ese cells contain several unique
proteins which are involved in RNA metabolism and processing,
thus showing that there may bechanges in transcription in these
cells. erefore, it is possible that the MUT protein is somehow
disrupting RNA processing.
Dierentially abundant proteins between
groups
For downstream analyses, all proteins identied in the EV
cells were then compared to those in the NT cells as a control
since wespeculate that the vector backbone should not cause
signicant changes to the cellular proteome. e proteins in the
WT cells and MUT cells were then each compared to the NT to
better understand their individual contributions to the proteome.
Finally, the main analysis involved comparison of the MUT cells
to the WT cells in an attempt to understand the eect of the
p.G849D variant. Table 2 shows the number of dierentially
abundant proteins found for each comparison group, divided into
those that are less abundant and those that are more abundant.
Volcano plots representing the dierentially abundant proteins in
each analysis are shown in Figures1C–F. Functional information
for all of these proteins can be found in Supplementary
Tables S5–S8.
Each protein set was uploaded to both KEGG8 (Kanehisa
etal., 2021) and STRING9 (Szklarczyk etal., 2019) for pathway
analysis. KEGG examines which pathways the proteins in each set
are involved in, whereas STRING identies the pathways that both
the proteins and their immediate interactors are involved in.
STRING interaction diagrams are shown in Figure2. As can
beseen in the EV vs NT analysis (Figure2A), there are not many
interactions and most consist of only a few nodes. is shows that
these protein networks are likely enriched by chance. e MUT vs
NT analysis (Figure2C) has the largest number of interactions,
showing the potential of the MUT to inuence cellular pathways.
In the main analysis (MUT vs WT; Figure2D), multiple proteins
are shown to interact in succession, hinting that there is a single
mode of action for the MUT NRXN2α. Each analysis using
STRING also provided a set of enrichment terms which show
which molecular mechanisms the generated protein network is
involved in.
Enrichment terms from STRING were combined with those
from KEGG and are shown in Table3. “Metabolic pathways” are
enriched across all analyses, suggesting that any treatment could
have an eect on the general cellular metabolic pathways. In the
EV vs NT analysis, “Alzheimer disease,” “Huntington disease,” and
“Parkinson disease” were enriched. However, the same three
proteins (QCR8, ATPD, and SNCA) were present in each group,
showing that this may bea chance nding. “Alzheimer disease,”
“Amyotrophic lateral sclerosis,” “Huntington disease,” “prion
disease,” and “Parkinson disease” were enriched in both the WT
vs NT and MUT vs NT analyses. When examining the MUT vs
WT, the only enrichment term related to neurodegenerative
disorders was “Huntington disease.” However, “ribosome,”
“spliceosome,” “mRNA surveillance pathway,” and
“nucleocytoplasmic transport” were also enriched in the MUT vs
WT analysis. is may hint toward a mode of action of the mutant
protein whereby protein translation and transport are aected by
its overexpression.
In the MUT vs WT analysis, the only enrichment term related
to neurodegenerative disorders was “Huntington disease.” e
three proteins involved in this pathway were AP2A1, RPB2, and
SDHB. STRING analysis shows that these proteins are not
predicted to interact with each other and do not form part of the
same networks. erefore, this enrichment is more likely to bea
random result of introducing cDNA to the cells. However,
“ribosome,” “spliceosome,” “mRNA surveillance pathway,” and
“nucleocytoplasmic transport” were also enriched in this analysis.
RPL8, RPS6, RPS21, and RPS25 are shown to bepart of the
“ribosome.” XPO2, PHAX, and PNN are involved in
“nucleocytoplasmic transport.” FUS, PNN, and PP1B are involved
in the “mRNA surveillance pathway.” FUS, RBMX, and PRP4 are
involved in the “spliceosome.” RPL8, RPS6, RPS21, and RPS25 are
8 https://www.kegg.jp
9 https://string-db.org
TABLE1 Enrichment terms obtained from the STRING online tool for
the unique proteins in each treatment group.
NT (no. of
proteins)
WT (no. of
proteins)
MUT (no. of
proteins)
Heterocyclic compound
binding (24)
Acetylation (12) Cellular metabolic process
(31)
Organic compound
binding (24)
Macromolecule metabolic
process (28)
RNA processing (12)
Metabolism of RNA (9)
STRING: https://string-db.org (Szklarczyk etal., 2019). NT: non-transfected cells; MUT:
mutant transfected cells; and WT: wild-type transfected cells.
TABLE2 The number of dierentially abundant proteins found for
each comparison.
Treatment
Comparison
Total
dierentially
abundant
proteins (No.)
More
abundant
proteins
(No.)
Less
abundant
proteins
(No.)
EV vs NT 28 8 20
WT vs NT 52 11 41
MUT vs NT 61 31 30
MUT vs WT 37 16 21
EV: empty vector transfected cells; NT: non-transfected cells; MUT: mutant transfected
cells; WT: wild-type transfected cells.

Cuttler et al. 10.3389/fnagi.2022.1002777
Frontiers in Aging Neuroscience 06 frontiersin.org
all ribosomal subunits, involved in ribosomal and
mRNA pathways.
STRING analysis shows that PHAX, XPO2, FUS, RBMX,
and PNN potentially interact in sequence. PHAX is a
phosphoprotein adaptor involved in RNA export and XPO2
plays a role in protein import/export within the nucleus. FUS
and RBMX are both RNA-binding proteins. FUS plays a role in
processes such as transcription regulation, RNA splicing, RNA
transport, DNA repair, and damage responses, while RBMX
plays several roles in the regulation of pre- and post-
transcriptional processes. PNN is a transcriptional activator for
the E-cadherin promoter gene, but it is also involved in RNA
binding and mRNA splicing via the spliceosome. PRP4 is a U4/
U6 small nuclear ribonucleoprotein which participates in
pre-mRNA splicing. PP1B is a serine/threonine-protein
phosphatase. Protein phosphatases are essential for cell division,
and PP1B participates in the regulation of glycogen metabolism,
muscle contractility, and protein synthesis. It has also been
shown to beinvolved in the mRNA surveillance pathway. Taken
together, the pathways of these proteins all relate back to the
regulation of transcription, strongly suggesting that this process
is being aected in the MUT cells.
AB
CD
FIGURE2
STRING protein–protein interaction diagrams of the dierentially abundant proteins. (A) EV vs NT, (B) WT vs NT, (C) MUT vs NT, (D) MUT vs WT.
Proteins which do not form part of an interaction network have been excluded. STRING: https://string-db.org (Szklarczyk etal., 2019). EV: empty
vector transfected cells, MUT: mutant transfected cells; NT: non-transfected cells; WT: wild-type transfected cells.

Cuttler et al. 10.3389/fnagi.2022.1002777
Frontiers in Aging Neuroscience 07 frontiersin.org
Enrichment analysis
WebGestalt facilitates the uploading of a gene or protein set
and the corresponding fold changes and then provides enriched
Gene Ontology (GO) terms. e graphical results of this analysis
are shown in Supplementary Figure S3. “Negative enrichment”
scores refer to GO terms that are downregulated or suppressed,
while “positive enrichment” scores refer to GO terms that are
upregulated or promoted. No specic enrichment was observed
for the EV vs NT analysis, suggesting that transfecting the cells
with the empty vector had a minimal eect.
In the WT vs NT analysis, the highest positive enrichment
was for “aminoacyl-tRNA biosynthesis.” Interestingly, “oxidative
phosphorylation,” “Alzheimer disease” and “Parkinson disease”
also showed positive enrichment scores. e mitochondrial
proteins QCR1, QCR2, and SDHB were the ones enriched in all
these pathways, again showing that the mitochondrial dysfunction
may beaected in the WT cells.
In the MUT vs NT analysis, “metabolic pathways” was
positively enriched with the increased abundance of NADC, a
protein involved in the catabolism of quinolinic acid. In the MUT
vs WT analysis “Huntington disease” and “metabolic pathways”
were negatively enriched and had false discovery rates (FDRs)
lower than 0.05. RPB2, SPSY, SDHB, PGM1, PUR9, and ODPB
were negatively enriched for “metabolic pathways.” All these
proteins, except RPB2, are involved in carboxylic acid metabolism.
AP2A1 and RPB2 were negatively enriched in “Huntington
disease.” AP2A1 is part of the adaptor protein complex 2 which
functions in protein transport via transport vesicles. It is also
involved in endolysosomal tracking and is thus implicated in
several neurodegenerative disorders (Müller, 2014; Heaton etal.,
2020; Srinivasan etal., 2022). RPB2 is a subunit of DNA-dependent
RNA polymerase II, and therefore, its main function is in
RNA transcription.
Findings from our study suggest that transfecting cells with
the plasmids is having an eect on ribosomal processes. erefore,
enrichment scores for the GO term “ribosome” across each
analysis have been summarized in Supplementary Figure S4. In
the WT vs NT analysis, only this GO term had an FDR ≤ 0.05,
showing that it highly likely that it has been negatively enriched in
this protein set. In the MUT vs NT analysis, “ribosome” was again
negatively enriched and had an FDR ≤ 0.05. However, in the MUT
vs WT analysis, “ribosome” was positively enriched, but its FDR
was above 0.05. Still, dysregulated ribosomal functioning seems to
be common among the analyses and may be an important
biological process related to the NRXN2α protein.
Discussion
is exploratory analysis has revealed that wild-type NRXN2α
may play a role in pathways related to neurodegenerative
disorders. Since the transfection eciency and NRXN2α levels
between the WT and MUT were similar, wecan be relatively
condent that the proteomics analysis showed dierences caused
by the overexpressed proteins and not by other technical
dierences between the two groups. In addition, while
overexpression of the empty vector plasmid did show similar
enrichment terms to the other analyses (Table 2), when
performing enrichment analysis for the EV transfected cells vs NT
TABLE3 Enrichment terms obtained from the KEGG and STRING online tools for each protein set.
EV vs NT (no. of proteins) WT vs NT (no. of
proteins)
MUT vs NT (no. of proteins) MUT vs WT (no. of proteins)
Metabolic pathways (7) Metabolic pathways (15) Metabolic pathways (12) Metabolic pathways (5)
Alzheimer’s disease (4) Alzheimer’s disease (7) Ribosome (8) Coronavirus disease—COVID-19 (4)
Pathways of neurodegeneration—multiple
diseases (3)
Amyotrophic lateral sclerosis (7) Amyotrophic lateral sclerosis (5) Ribosome (4)
Huntington’s disease (3) Huntington’s disease (7) Alzheimer’s disease (5) Alcoholism (3)
Oxidative phosphorylation (3) Parkinson’s disease (6) Biosynthesis of amino acids (5) Diabetic cardiomyopathy (3)
Parkinson’s disease (3) Prion disease (6) Carbon metabolism (5) Huntington’s disease (3)
Alanine, aspartate and glutamate
metabolism (4)
Pathways of neurodegeneration— multiple
diseases (5)
mRNA Surveillance Pathway (3)
Aminoacyl-tRNA biosynthesis (3) Salmonella infection (5) Nucleocytoplasmic transport (3)
Cysteine and methionine
metabolism (3)
Diabetic cardiomyopathy (4) Spliceosome (3)
Carbon metabolism (3) DNA replication (4)
Huntington’s disease (4)
Parkinson’s disease (4)
Prion disease (4)
KEGG: https://www.kegg.jp (Kanehisa etal., 2021), STRING: https://string-db.org (Szklarczyk etal., 2019). EV: empty vector transfected cells; NT: non-transfected cells, MUT: mutant
transfected cells; and WT: wild-type transfected cells.

Cuttler et al. 10.3389/fnagi.2022.1002777
Frontiers in Aging Neuroscience 08 frontiersin.org
cells, it can beseen that none of these terms were signicantly
enriched (Supplementary Figure S3A). erefore, wepostulate
that the EV only had a minimal eect on the cells and the majority
of changes in the other analyses are in fact due to the NRXN2α
cDNA insert (WT or MUT).
Overexpression of the WT protein in SH-SY5Y cells led to the
enrichment of proteins involved in neurodegenerative diseases,
such as Alzheimer’s disease, Amyotrophic lateral sclerosis, and
Parkinson’s disease. In particular, the enriched proteins were
involved in mitochondrial and lysosomal functioning, which are
known to be dysregulated in PD and other neurodegenerative
disorders (Rego and Oliveira, 2003; Wang etal., 2018). us, the
wild-type protein may be involved in pathways related to the
development of neurodegenerative disorders, such as PD. is
provides further evidence potentially implicating synaptic proteins
in the pathobiology of PD. Overexpression of the MUT
NRXN2α-CFP protein showed similar results. Proteins unique to
the MUT cells were enriched for terms related to “ribosome.” In
addition, when directly comparing the WT transfected cells with
the MUT transfected cells, terms related to “ribosome” were
enriched. is may thus hint at a mode of action for the p.G849D
mutant protein. Since the main function of the ribosome is
translation of mRNA into protein, dysregulated translation could
beimplicated as a biological process involved in neurodegeneration.
Furthermore, both cytoplasmic and mitochondrial ribosomal
proteins were enriched. Indeed, it has been shown that if synaptic
translation is dysregulated, mitochondrial physiology can bealtered
(Kuzniewska etal., 2020). In addition, EIF4G1, another protein
implicated in PD, is known to beinvolved in protein translation
processes (Chartier-Harlin etal., 2011). Furthermore, the DJ-1 and
SYNJ1 proteins implicated in PD (Bonifati etal., 2003; Krebs etal.,
2013) also have RNA-binding functions. DJ-1 acts to protect cells
from oxidative stress and cell death by acting as an oxidative stress
sensor and redox-sensitive chaperone and protease, while SYNJ1 is
a phosphatase involved in synaptic vesicle endocytosis and
neurotransmitter transport. A few studies have additionally
identied mitochondrial ribosomal proteins in PD. Gaare etal.
(2018) identied MRPL4, which encodes a component of the large
mitochondrial ribosome subunit, in an analysis of two PD cohorts,
while Billingsley etal. (2019) identied MRPL43 and MRPS34,
encoding components of the large and small mitochondrial
ribosome subunits, using data from a PD genome-wide association
study (GWAS). Both these studies thus link mtDNA translation to
PD risk. Dysregulated mRNA translation can therefore
beconsidered to play a role in PD pathogenesis (Martin, 2016). In
addition, a recent RNA-sequencing analysis showed that there was
dierential expression of ribosomal-related pathways in their PD
cohort (Hemmings etal., 2022). erefore, it is plausible that
synaptic translation could also beimportant in PD pathogenesis.
Here, changes in translation could aect oxidative stress and the
transport of neurotransmitters, thereby causing cells to bemore
susceptible to cell damage and death. In addition, a study on
lymphoblasts generated from PD patients showed an overall
downregulation of genes involved in protein synthesis (Annesley
etal., 2022). us, recent literature has shown that dysregulated
synaptic translation and mitochondrial dysfunction are linked
(Kuzniewska etal., 2020), mitochondrial ribosomal proteins have
been linked in a PD GWAS (Billingsley et al., 2019), pathways
related to ribosomes are enriched in an RNA-sequencing analysis
of a PD cohort (Hemmings etal., 2022), and that lymphoblasts
generated from PD patients have dysregulated expression of genes
involved in protein synthesis (Annesley etal., 2022). In addition,
some of the known PD-associated proteins are also shown to have
RNA or protein translation roles. is link between mRNA
translation is poorly understood, but a few reviews have highlighted
that restoring translation and proteostasis might bea useful target
for new therapeutics (Correddu and Leung, 2019; Zhou etal.,
2019). Impaired proteostasis at the synapse could also beimportant
for PD (Nachman and Verstreken, 2022) while reduced synaptic
activity and dysregulated extracellular matrix pathways have
recently been reported in midbrain neurons from PD patients,
providing evidence that synaptopathy is a general phenotype in PD
(Stern et al., 2022). us, biological processes related to the
ribosome, translation, and tRNA, specically at the synapse, could
possibly bean imp ortant molecular mechanism in PD pathobiology.
e strength of this study is that weexamined the eect of
overexpression of both the wild-type and mutant protein using
a hypothesis-free approach. In this way, wewere able to show
that potential mode of action of the mutant protein but were
also able to conclude that the wild-type protein is also involved
in pathways related to neurodegeneration. erefore, it is
possible that any dysregulation of NRXN2α could potentially
lead to neurodegeneration.
However, wealso acknowledge several limitations, including
the use of a commercial cell line for this study. While SH-SY5Y
cells are good in vitro model for PD as they display a
catecholaminergic phenotype, producing both dopamine and
noradrenaline (Xicoy etal., 2017), there are always limitations
when using cell lines to study a complex human disorder.
Unfortunately, we were not able to obtain dermal broblast
samples from the individuals harboring the NRXN2 variant as an
ex-vivo model for this study. Wealso acknowledge the limitations
of overexpressing a murine gene in a human cell line. erefore,
in the future it would beimportant to repeat these experiments
in broblasts from the patients or in animal models. Another
limitation is the use of shotgun proteomics. Since this study is
explorative, weinvestigated the total proteome to determine
which biological pathways were being aected. However, it may
beimportant to do more targeted proteomics work in future,
such as looking into post-translational modications as well as
investigating phospho-proteomics to determine signaling
changes. Indeed, several kinases and phosphatases were observed
in the dierent analyses, therefore phospho-proteomics would
be required to better understand the eect of these
protein changes.
In conclusion, ndings from this exploratory study possibly
implicate the NRXN2α protein in neurodegenerative processes
and show that synaptic ribosomal and translation processes may

Cuttler et al. 10.3389/fnagi.2022.1002777
Frontiers in Aging Neuroscience 09 frontiersin.org
beimportant in PD and/ or other neurodegenerative disorders.
However, further validation of NRXN2α and the proteins
implicated in synaptic ribosomal and translation processes in
other models of PD or neurodegenerative disorders would
berequired to prove or disprove this hypothesis.
Data availability statement
e mass spectrometry proteomics data have been deposited
to the ProteomeXchange Consortium via the PRIDE partner
repository with the dataset identier PXD036636 and 10.6019/
PXD036636.
Ethics statement
Ethical approval was obtained from the Health Research
Ethics Committee (Protocol numbers 2002/C059 and S20/01/005
PhD) and the Research Ethics Committee: Biological and
Environmental Safety (Protocol number BEE-2021-13149). Both
committees are located at Stellenbosch University, Cape Town,
South Africa.
Author contributions
KC conducted all experiments, performed all analyses, and
wrote the rst dra of the manuscript. SF assisted with analysis
of the mass spectrometry data. AM-N assisted with data
processing. MV performed the protein extraction and mass
spectrometry. RC assisted with writing and editing of the
manuscript. KC and SB conceptualized the study and acquired
funding. All authors contributed to the article and approved the
submitted version.
Funding
is work is based on the research supported in part by the
National Research Foundation of South Africa (NRF) (Grant
Numbers: 129249 and 146254); the SouthAfrican Medical Research
Council (SAMRC) (self-initiated research grant); the Harr y Crossley
Foundation and Stellenbosch University, SouthAfrica. SAMRC and
e Higher Education Department, Next Generation of Academic
Programme (nGAP), provided support for RC in the form of a
fulltime academic position and salary.
Acknowledgments
e authors thank the study participants for their participation
in and contribution to this study. e authors would also like to
thank Ann Marie Craig (University of British Columbia, Canada)
and Harald Sitte (Medical University of Vienna, Austria) for the
NRXN2α-ECFP-N1 plasmid and pECFP-N1 plasmid, respectively.
e authors additionally acknowledge the support of the DSI-NRF
Centre of Excellence for Biomedical Tuberculosis Research,
SouthAfrican Medical Research Council Centre for Tuberculosis
Research, Division of Molecular Biology and Human Genetics,
Faculty of Medicine and Health Sciences, Stellenbosch University,
Cape Town, SouthAfrica.
Conflict of interest
e authors declare that the research was conducted in the
absence of any commercial or nancial relationships that could
beconstrued as a potential conict of interest.
Publisher’s note
All claims expressed in this article are solely those of the
authors and do not necessarily represent those of their aliated
organizations, or those of the publisher, the editors and the
reviewers. Any product that may be evaluated in this article, or
claim that may be made by its manufacturer, is not guaranteed or
endorsed by the publisher.
Supplementary material
e Supplementary material for this article can be found
online at: https://www.frontiersin.org/articles/10.3389/fnagi.
2022.1002777/full#supplementary-material
References
Annesley, S. J., Allan, C. Y., Sanislav, O., Evans, A., and Fisher, P. R. (2022).
Dysregulated gene expression in Lymphoblasts from Parkinson’s disease. Proteomes
10:20. doi: 10.3390/proteomes10020020
Billingsley, K. J., Barbosa, I. A., Bandrés-Ciga, S., Quinn, J. P., Bubb, V. J.,
Deshpande, C., et al. (2019). Mitochondria function associated genes contribute to
Parkinson’s disease risk and later age at onset. NPJ Parkinsons Dis. 5, 1–9. doi:
10.1038/s41531-019-0080-x
Bonifati, V., Rizzu, P., Van Baren, M. J., Schaap, O., Breedveld, G. J., Krieger, E.,
et al. (2003). Mutations in the DJ-1 gene associated with autosomal recessive early-
onset Parkinsonism. Science 299, 256–259. doi: 10.1126/science.1077209
Chartier-Harlin, M. C., Dachsel, J. C., Vilariño-Güell, C., Lincoln, S. J.,
Leprêtre, F., Hulihan, M. M., et al. (2011). Translation initiator EIF4G1 mutations
in familial Parkinson disease. Am. J. Hum. Genet. 89, 398–406. doi: 10.1016/j.
ajhg.2011.08.009
Correddu, D., and Leung, I. K. H. (2019). Targeting mRNA translation in Parkinson’s
disease. Drug Discov. Today 24, 1295–1303. doi: 10.1016/j.drudis.2019.04.003
Craig, A. M., and Kang, Y. (2007). Neurexin-neuroligin signaling in synapse
development. Curr. Opin. Neurobiol. 17, 43–52. doi: 10.1016/j.conb.2007.01.011
Dachtler, J., Ivorra, J. L., Rowland, T. E., Lever, C., John Rodgers, R., and
Clapcote, S. J. (2015). Heterozygous deletion of α-neurexin Ior α-neurexin II results

Cuttler et al. 10.3389/fnagi.2022.1002777
Frontiers in Aging Neuroscience 10 frontiersin.org
in behaviors relevant to autism and schizophrenia. Behav. Neurosci. 129, 765–776.
doi: 10.1037/bne0000108
Dorsey, E. R., Elbaz, A., Nichols, E., Abd-Allah, F., Abdelalim, A., Adsuar, J. C.,
et al. (2018). Global, regional, and national burden of Parkinson’s disease,
1990–2016: a systematic analysis for the global burden of disease study 2016. Lancet
Neurol. 17, 939–953. doi: 10.1016/S1474-4422(18)30295-3
Feigin, V. L., Krishnamurthi, R. V., eadom, A. M., Abajobir, A. A., Mishra, S. R.,
Ahmed, M. B., et al. (2017). Global, regional, and national burden of neurological
disorders during 1990–2015: a systematic analysis for the global burden of disease
study 2015. Lancet Neurol. 16, 877–897. doi: 10.1016/S1474-4422(17)30299-5
Gaare, J. J., Nido, G. S., Sztromwasser, P., Knappskog, P. M., Dahl, O., Lund-Johansen, M.,
et al. (2018). Rare genetic variation in mitochondrial pathways inuences the risk for
Parkinson’s disease. Mov. Disord. 33, 1591–1600. doi: 10.1002/mds.64
Heaton, G. R., Landeck, N., Mamais, A., Nalls, M. A., Nixon-Abell, J., Kumaran, R.,
et al. (2020). Sequential screening nominates the Parkinson’s disease associated
kinase LRRK2 as a regulator of Clathrin-mediated endocytosis. Neurobiol. Dis.
141:104948. doi: 10.1016/j.nbd.2020.104948
Hemmings, S. M. J., Swart, P., Womersely, J. S., Ovenden, E. S., van den
Heuvel, L. L., McGregor, N. W., et al. (2022). RNA-seq analysis of gene expression
proles in posttraumatic stress disorder, Parkinson’s disease and schizophrenia
identies roles for common and distinct biological pathways. Discov. Ment. Heal. 2,
1–18. doi: 10.1007/s44192-022-00009-y
Kanehisa, M., Furumichi, M., Sato, Y., Ishiguro-Watanabe, M., and Tanabe, M.
(2021). KEGG: integrating viruses and cellular organisms. Nucleic Acids Res. 49,
D545–D551. doi: 10.1093/nar/gkaa970
Kang, Y., Zhang, X., Dobie, F., Wu, H., and Craig, A. M. (2008). Induction of
GABAergic postsynaptic dierentiation by α-neurexins. J. Biol. Chem. 283,
2323–2334. doi: 10.1074/jbc.M703957200
Krebs, C. E., Karkheiran, S., Powell, J. C., Cao, M., Makarov, V., Darvish, H., et al.
(2013). e sac1 domain of SYNJ1 identied mutated in a family with early-onset
progressive Parkinsonism with generalized seizures. Hum. Mutat. 34, 1200–1207.
doi: 10.1002/humu.22372
Kuzniewska, B., Cysewski, D., Wasilewski, M., Sakowska, P., Milek, J., Kulinski, T. M.,
et al. (2020). Mitochondrial protein biogenesis in the synapse is supported by local
translation. EMBO Rep. 21:e48882. doi: 10.15252/embr.201948882
Liao, Y., Wang, J., Jaehnig, E. J., Shi, Z., and Zhang, B. (2019). WebGestalt 2019:
gene set analysis toolkit with revamped UIs and APIs. Nucleic Acids Res. 47, W199–
W205. doi: 10.1093/nar/gkz401
Martin, I. (2016). Decoding Parkinson’s disease pathogenesis: the role of
deregulated mRNA translation. J. Parkinsons Dis. 6, 17–27. doi: 10.3233/JPD-150738
Missler, M., Zhang, W., Rohlmann, A., Kattenstroth, G., Hammer, R. E.,
Gottmann, K., et al. (2003). α-neurexins couple Ca 2+ channels to synaptic vesicle
exocytosis. Nature 423, 939–948. doi: 10.1038/nature01755
Müller, S. (2014). In silico analysis of regulatory networks underlines the role of
miR-10b-5p and its target BDNF in huntington’s disease. Transl. Neurodegener. 3,
1–5. doi: 10.1186/2047-9158-3-17
Nachman, E., and Verstreken, P. (2022). Synaptic proteostasis in Parkinson’s
disease. Curr. Opin. Neurobiol. 72, 72–79. doi: 10.1016/j.conb.2021.09.001
Oliveros, J. C. (2015). Venny. An interactive tool for comparing lists with Venn’s
diagrams. Available at: https://bioinfogp.cnb.csic.es/tools/venny/
Panicker, N., Ge, P., Dawson, V. L., and Dawson, T. M. (2021). e cell biology of
Parkinson’s disease. J. Cell Biol. 220:e202012095. doi: 10.1083/jcb.202012095
Rego, A. C., and Oliveira, C. R. (2003). Mitochondrial dysfunction and reactive
oxygen species in excitotoxicity and apoptosis: implications for the pathogenesis of
neurodegenerative diseases. Neurochem. Res. 28, 1563–1574. doi:
10.1023/A:1025682611389
Searle, B. C. (2010). Scaold: a bioinformatic tool for validating MS/MS-based
proteomic studies. Proteomics 10, 1265–1269. doi: 10.1002/pmic.200900437
Sebate, B., Cuttler, K., Cloete, R., Britz, M., Christoels, A., Williams, M., et al.
(2021). Prioritization of candidate genes for a south African family with Parkinson’s
disease using in-silico tools. PLoS One 16:e0249324. doi: 10.1371/journal.
pone.0249324
Srinivasan, S., Gal, J., Bachstetter, A., and Nelson, P. T. (2022). Alpha adaptins
show isoform-specic association with neurobrillary tangles in Alzheimer’s
disease. Neuropathol. Appl. Neurobiol. 48:e12776. doi: 10.1111/nan.12776
Stern, S., Lau, S., Manole, A., Rosh, I., Percia, M., Ezer, R. B., et al. (2022). Reduced
synaptic activity and dysregulated extracellular matrix pathways are common
phenotypes in midbrain neurons derived from sporadic and mutation-associated
Parkinson’s disease patients. NPJ Parkinsons Dis. 8:103. doi: 10.1038/
S41531-022-00366-Z
Südhof, T. C. (2008). Neuroligins and neurexins link synaptic function to
cognitive disease. Nature 455, 903–911. doi: 10.1038/nature07456
Szklarczyk, D., Gable, A. L., Lyon, D., Junge, A., Wyder, S., Huerta-Cepas, J., et al.
(2019). STRING v11: protein-protein association networks with increased coverage,
supporting functional discovery in genome-wide experimental datasets. Nucleic
Acids Res. 47, D607–D613. doi: 10.1093/nar/gky1131
e UniProt Consortium (2021). UniProt: the universal protein knowledgebase
in 2021. Nucleic Acids Res. 49, D480–D489. doi: 10.1093/nar/gkaa1100
Wang, C., Telpoukhovskaia, M. A., Bahr, B. A., Chen, X., and Gan, L. (2018).
Endo-lysosomal dysfunction: a converging mechanism in neurodegenerative
diseases. Curr. Opin. Neurobiol. 48, 52–58. doi: 10.1016/j.conb.2017.09.005
Xicoy, H., Wieringa, B., and Martens, G. J. M. (2017). e SH-SY5Y cell line in
Parkinson’s disease research: a systematic review. Mol. Neurodegener. 12, 10–11. doi:
10.1186/s13024-017-0149-0
Yakhine-Diop, S., Martínez-Chacón, G., Uribe-Carretero, E., Niso-Santano, M.,
González-Polo, R., and Fuentes, J. (2019). e paradigm of protein acetylation in
Parkinson’s disease. Neural Regen. Res. 14, 975–976. doi: 10.4103/1673-5374.
250575
Zhou, Z. D., Selvaratnam, T., Lee, J. C. T., Chao, Y. X., and Tan, E. K. (2019).
Molecular targets for modulating the protein translation vital to proteostasis and
neuron degeneration in Parkinson’s disease. Transl. Neurodegener. 8, 6–14. doi:
10.1186/s40035-019-0145-0