Академический Документы
Профессиональный Документы
Культура Документы
(Thymallus thymallus)
Spiros Papakostas,† L. Asbjørn Vøllestad,‡ Craig R. Primmer,† and Erica H. Leder*,†
Division of Genetics and Physiology, Department of Biology, University of Turku, Finland, and Centre for
Ecological and Evolutionary Synthesis, Department of Biology, University of Oslo, Norway
Salmonids are teleost fish of profound research and economic interest. Embryonic development is a
key aspect of salmonid biology that can be critically affected by environmental parameters. Still, their
proteome during embryogenesis remains largely unexplored. This study investigates the proteome of
the eyed-egg and hatching stages of embryonic development of a salmonid species, European grayling
(Thymallus thymallus), using a shotgun proteomic approach. To deal with limited grayling protein
resources, the generated spectra were compared against an all-salmonid database using search and
multiple protein grouping algorithms to infer identifications. Functional enrichment analysis was carried
out at different levels (gene ontologies, pathways, networks) using zebrafish as a reference genome.
A total of 213 and 239 proteins were confidently detected in eyed and hatching stages, respectively.
Cell cycle, energy, and protein metabolism were the major processes common to both stages. Nuclear
activity and brain and eye development were the predominant functions in the eyed-stage proteome,
while central nervous system, skeletal muscle, and heart development prevailed in the hatching stage.
Overall, this research constitutes the first effort to describe the proteome during embryogenesis in
grayling or any salmonid species. It also presents a systematic approach by which existing resources
can enable proteome research in salmonids.
4790 Journal of Proteome Research 2010, 9, 4790–4800 10.1021/pr100507s 2010 American Chemical Society
Published on Web 07/05/2010
Proteomic Profiling of Early Life Stages of European Grayling research articles
been studied in any salmonid fish. In European grayling drated for 12 h and focused on an IPGphor (Amersham
(Thymallus thymallus), the evolutionary importance of embry- Pharmacia Biotech, Piscataway, NJ) using the following pro-
onic development rate and hatching time has been demon- gram: hold at 500 V for 1 h, linear gradient from 500 to 1000 V
strated in recently introduced populations inhabiting the for 15 min, hold at 1000 V for 1 h, linear gradient from 1000 to
Lesjaskogsvatnet lake system in southern Norway.24,25 There- 8000 V for 30 min, hold at 8000 V for 3.5 h. The gel strip was
fore, grayling individuals from this system can be useful cut into six fractions and peptides were extracted from the gel
candidates for investigating early development at the protein strip by washing three times: first in 0.1% TFA, second in 0.1%
expression level in the context of evolutionary and ecological TFA, 50% acetonitrile (ACN), and third in 0.1% TFA, 100% ACN,
adaptation. In this study, a MS-based approach was employed all for 15 min at 37 °C. The solution from the three washes
to identify similarities and differences of the proteomes ex- was combined and evaporated. The peptides were dissolved
pressed at the eyed-egg and hatching larvae stages in grayling in 0.1% TFA, 5% ACN and desalted using UltraMicroSpin C18
from Lake Lesjaskogsvatnet. Overall, this work provides the first columns (The Nest Group) according to the manufacturer’s
systematic investigation of the early development proteome in directions.
salmonids. Due to the limited number of available protein Mass Spectrometry Analysis. LC-MS/MS analyses were
sequences in grayling, this study also demonstrates a way by made using a system consisting of a Waters Cap-LC (Waters,
which existing sequence resources can be used for the MS- Milford, MA) coupled to a QSTAR Pulsar ESI-hybrid quadrupole
based protein detection in any salmonid species. time-of-flight instrument (Applied Biosystems, Foster City, CA).
A 0.3 mm ×5 mm PepMap C18 µ-precolumn (LC Packings,
Materials and Methods Dionex, Sunnyvale, CA) was used for sample loading and was
Sample Collection and Experimental Design. The samples coupled to a 15 cm × 75 µm i.d. fused silica capillary column
derived from two grayling spawning sites (Sandbekken and packed with 5 µm Magic C18 (Michrom BioResources, Inc.,
Valåe) of the Lesjaskogsvatn lake system. They were the product Auburn, CA) for separation. A 90 min binary gradient from 2%
of the artificial fertilization between the pooled gametes of 30 to 30% phase A was used at a flow rate of 200 µL/min (phase
individuals (15 sires, 15 dams) per spawning population. Eggs A ) 5% ACN, 0.1% HCOOH; phase B ) 95% ACN, 0.1%
from each population were reared in common garden condi- HCOOH). Data acquisition and instrument control was per-
tions at 6 °C. Embryos were sampled daily from fertilization to formed using Analyst QS 1.1 software (Applied Biosystems,
when the yolk sac was absorbed. Individuals were selected from Foster City, CA), with the mass spectrometer set to perform
a time point corresponding to the appearance of eye pigmenta- 1 s survey scans followed by two 2 s MS/MS scans of the two
tion (98-100 degree days, eyed-egg stage) and from when more most intense peaks (10 cps), with dynamic exclusion for 5 min.
than 50% of the individuals are hatched (188 degree days, Database Search and Protein Detection Parameters. Protein-
hatching stage). Four individuals from each population for each Pilot software (v.3, Applied Biosystems, ABI) was employed to
time point were used for protein analysis. All fish embryos used infer protein identifications. The detected protein threshold in
in this study were collected according to animal experimenta- the Paragon algorithm was set to 1.3 (95% confidence level).
tion guidelines at the sampling location in Norway. The search database was comprised of all the salmonid proteins
Sample Preparation. Proteins were isolated from whole (14 885 in total) submitted in UniProt database (15.11 release).
embryos using Tri Reagent (Sigma), following the manufacturer The common contaminants file provided by ABI (ABSciex-
instructions. Protein pellets resulting from the extraction _ContaminantDB file) was also appended to these protein
protocol were dissolved in 1% SDS and quantified using a DC sequences. The search effort was set to thorough ID, and both
protein assay (Bio-Rad). Because within-stage quantification biological modifications and amino acid substitutions were
will be subsequently investigated, samples that were used in included. The quality of the Paragon algorithm detection results
this study for the MS analysis were labeled with iTRAQ reagents was assessed by performing a false discovery rate (FDR) analysis
following the manufacturer’s protocol (Applied Biosystems). at 1% global error rate using a decoy database as implemented
Briefly, 100 µg of each sample was acetone precipitated and within the proteomics system performance evaluation pipeline
resuspended in 20 µL of 0.5 M triethyl ammonium bicarbonate (PSEP) of the ProteinPilot software. The protein groups were
(TEAB) and 1 µL of 2% SDS (supplied in the iTRAQ kit). Samples created by the Pro Group algorithm of the ProteinPilot software.
were reduced by the addition of 2 µL of 50 mM tris-carboxy- This algorithm helped reduce the protein redundancy of the
ethyl phosphine hydrochloride (TCEP) and incubation for 1 h search database since it groups together proteins with the same
at 60 °C, and then alkylated by the addition of 1 µL of 200 mM spectral evidence. Yet, these groups were further inspected on
methyl methane thiosulfonate (MMTS) and incubation for 10 the basis of the multidetected protein groups suggested by the
min at room temperature (RT). Following this step, 10 µL of same algorithm. Multidetected groups that involved proteins
trypsin (1 µg/µL) (Promega, Madison, WI) was added, and annotated under the same gene name or name entries in
samples were digested overnight at 37 °C. Ethanol was added UniProt were manually grouped together. For each protein, the
to the iTRAQ labels, and the reconstituted labels were added number of credible spectra (g95% confidence) uniquely as-
to their respective samples. Samples were incubated for 1 h, signed to them was counted, and for proteins detected only in
combined into their respective multiplexes, and then evapo- one of the stages, a one-tailed χ-square test with the Yates’
rated. Samples were dissolved in 50 µL of 0.1% trifluoroacetic continuity correction was performed to examine whether the
acid (TFA) and desalted using MacroSpin C18 columns (The spectra present were significantly greater than zero (p e 0.05).
Nest Group, Southborough, MA) according to the manufac- Gene Ontology (GO) Enrichment Analysis. Hierarchical GO
turer’s directions. Samples were dissolved in isoelectric focusing overrepresentation tests were performed in Cytoscape 2.6.326
buffer consisting of 2 M urea and 2% IPG buffer, pH 3-10 (GE using the BiNGO 2.3 plugin.27 Four gene lists were investigated.
Healthcare), and fractionated before mass spectrometry analy- These involved all proteins that were detected in the eyed stage
sis using isoelectric focusing on a 13 cm, pH 3-10 linear, (gene set a), all those detected in the hatching stage (set b),
Immobiline DryStrip (GE Health Care). The strips were rehy- those detected only in the eyed stage (set c), and those detected
Figure 1. Combined network view of the GO-based functional groups in the grayling eyed and hatching stages. A total of 20 groups
were detected. Dashed lines distinguish the generated clusters to the biological process, cellular component, or molecular functions
GO categories. For each category, every group has its own distinctive color. Multicolored nodes represent GO terms that belong to
more than one group. A description of each group on the basis of the high- (more general) and low-level significant GO terms is given
in Table 1. Since this is a combined view of the eyed and hatching stage networks, nodes were set to the same size, while node
abundances and connectivity are only indicative of the general network organization. More details can be found in Supplementary
Tables 3 and 4, Supporting Information.
Table 1. Description of the GO-Based Functional Groups Identified by ClueGO in the Eyed (eye) and Hatching (hatch) Stagesa
functional stage
group high-level GO detected terminal GO
Biological Process
G1 cellular homeostasis eye cellular iron-ion homeostasis
G2 cellular macromolecule metabolic process both regulation of cell cycle
G3 macromolecule biosynthetic process both translational elongation
G4 DNA metabolic process eye DNA replication initiation
G5 cellular macromolecular complex subunit organization both nucleosome assembly
G6 protein metabolic process both protein folding
G7 carbohydrate catabolic process both glycolysis
G8 lipid localization both lipid transport
G9 ns ns
G10 nucleobase, nucleoside, and nucleotide metabolic process eye nucleotide metabolic process
Cellular Component
G11 intracellular non-membrane-bound organelle both keratin filament
G12 mitochondrial membrane part both proton-transporting ATP synthase complex
G13 chromosomal part both nucleosome
G14 ribonucleoprotein complex both large/small ribosomal subunit
G15 proteasome complex hatch proteasome core complex
Molecular Function
G16 ns ns
G17 hydrolase activity, acting on acid anhydrites eye nucleoside triphosphatase activity
G18 ns ns
G19 oxidoreductase activity, acting on CH-OH group of donors both L-malate dehydrogenase activity
G20 translation factor activity, nucleic acid binding both translation elongation factor activity
a
The most significant high-level terms derived from ClueGO, whereas the low-level, terminal ones derived from BiNGO. Three of the groups (G9, G16,
and G18) had no significant terms (ns) at p e 0.05. All the significant GO terms for the BiNGO and ClueGO analyses are given in Supplementary Tables 3
and 4, Supporting Information, respectively.
through activation of Rho pathways.39 Lastly, the granzyme functions. Previous studies have reported a number of
A pathway is involved in apoptosis,40 while the protein proteins associated with integrin and apoptosis signaling in
ubiquitination pathway is part of the protein catabolic hatching stage zebrafish embryos.16 According to the results
mechanism. The presence of these pathways in both pro- of this work, the functioning of these pathways is likely to
teomes underlines the organogenesis-related processes of be present earlier in development and presumably through-
cell proliferation, migration, and death among their common out teleost embryogenesis.
Figure 2. (A) Detailed view of the molecule interactions for the network NE3 identified in the eyed stage. Its functions include
“cardiovascular and visual system development”, which to some extent characterizes the specific functioning of the eyed-stage proteome.
(B) Detailed view of the hatching-stage network NH3, which is involved in “skeletal and muscular system development and function”.
More information about the interacting molecules is given in Supplementary Table 7, Supporting Information. Both networks were
drawn in Cytoscape 2.6.3 according to the results generated by the IPA software.
including lamination abnormalities of the cortex and the eye 2). Again, the RAN pathway is involved in neural, brain and
retina.43 Similarly, the 14-3-3 protein has been implicated in eye development. During mouse embryogenesis Ran protein
the eye and the early neural development in Drosophila and expression was detected initially in neuroectoderm, neural
mice.44,45 On a different functional level, five of the detected groove, neural folds, neural tube, and later within the encepha-
pathways were also significant only in the eyed stage (Table lic vesicle wall and spinal cord, especially during eye develop-
Figure 3. The result from the direct comparison of the eyed and hatching stages with ClueGO. Color gradient shows the proportion of
genes in the GO term associated with each of the two stages. Red coloring characterizes eyed-stage related GOs, while green the
hatching-stage ones. Equal proportions are shown in white. Node size is related to the statistical significance of the GO. Only terms
with a corrected p-value g 0.05 are effectively visible. Edge line width is proportional to the kappa scores, while node connectivity was
inferred using kappa statistics.22 Seven overrepresented GO terms and term clusters were related to the eyed stage (E1-E7) and eight
to the hatching stage (H1-H8). The most significant GOs for the clusters or terms are nucleotide biosynthetic process (E1), iron-ion
transport (E2), nuclear transport (E3), DNA replication initiation (E4), mitochondrial membrane part (E5), mitochondrial matrix (E6),
ribose phosphate diphosphokinase activity (E7), NADPH regeneration (H1), regulation of actin filament polymerization (H2), proton-
transporting ATP synthase complex (H3), proteasome core complex, alpha subunit complex (H4), ribosome (H5), inorganic cation
transmembrane transporter activity (H6), translation factor activity, nucleic acid binding (H7), threonine-peptidase activity (H8). More
details can be found in Supplementary Table 4, Supporting Information.
Figure 4. Scatter plot of the average number of unique credible (g95 confidence) spectra per technical replicate assigned to each of
the 313 detected proteins in the eyed and hatching stages. The names of those proteins with the most assigned spectra are noted on
the graph. A detailed view of these data can be found in Supplementary Table 8, Supporting Information.
Hatching-Specific Functions. A total of 19 proteins and five analysis. The GO terms of actin filament polymerization,
pathways were significantly detected only in the hatching stage creatine kinase activity, and calcium ion binding were signifi-
(Table 2, Supplementary Table 8, Supporting Information). cant only in the hatching stage (Figure 3, Table 4). The RhoA
Skeletal muscle development was found to be a major hatch- and actin cytoskeleton signaling pathways, both involved in
ing-specific function. Creatine kinases 1 and 2, both among muscle tissue development,60 were detected only in the hatch-
the significant proteins, are involved in muscle-specific me- ing stage (Table 2). Lastly, one of the most significant functions
tabolism.56 Troponin T isoform 1, another significant protein, of the network NH3 was “skeletal and muscular system
is involved in muscle maturation.57 In addition to the specific development and function” (Table 3). A number of genes,
presence of particular proteins, their relative abundance, as can whose proteins were detected only in the hatching stage,
be inferred by the spectral evidence,58,59 also highlights the participate in this network (Figure 2B).
muscle functioning of the hatching-stage detected proteins. A Other hatching-specific functions involve central nervous
number of skeletal muscle proteins, namely the myosin heavy system and heart development. In particular, the fatty acid
chain, actin alpha, and actin beta, had considerably higher binding protein (FABP) brain type and heart type are two
spectral counts in the hatching than in the eyed stage (101, proteins detected only in the hatching stage (Supplementary
80, and 103 versus 2.6, 23, and 8, respectively; Figure 4, Table 8, Supporting Information). The FABP brain type is an
Supplementary Table 8, Supporting Information). Evidence for intracellular protein expressed particularly in the central
hatching-specific myoskeletal development functioning can nervous system with an important role during neurogenesis
also be found in the GO, pathway, and network levels of in vertebrates.61,62 During zebrafish late embryogenesis, FABP