Proteomic study of the brackish water mussel Mytilopsis leucophaeata

Background

We encountered the opportunity to study proteochemically a brackish water invertebrate animal, Mytilopsis leucophaeata, belonging to the bivalves which stem from the second half of the Cambrian Period (about 510 million years ago). This way, we were able to compare it with the vertebrate animal, the frilled shark (Chlamydoselachus anguineus) that stems from a much later period of geologic time (Permian: 245–286 MYA).

Results

The mussel contains a well-adapted system of protein synthesis on the ER, protein folding on the ER, protein trafficking via COPI or clathrin-coated vesicles from endoplasmic reticulum (ER) to Golgi and plasmalemma, an equally well-developed system of actin filaments that with myosin forms the transport system for vesicular proteins and tubulin, which is also involved in ATP-driven vesicular protein transport via microtubules or transport of chromosomes in mitosis and meiosis. A few of the systems that we could not detect in M. leucophaeata in comparison with C. anguineus are the synaptic vesicle cycle components as synaptobrevin, cellubrevin (v-snare) and synaptosomal associated protein 25-A (t-snare), although one component: Ras-related protein (O-Rab1) could be involved in synaptic vesicle traffic. Another component that we did not find in M. leucophaeata was Rab11 that is involved in the tubulovesicular recycling process of H⁺/K⁺-ATPase in C. anguineus. We have not been able to trace the H⁺/K⁺-ATPase of M. leucophaeata, but Na⁺/K⁺-ATPase was present. Furthermore, we have studied the increase of percent protein expression between 1,070 MYA (the generation of the Amoeba Dictyostelium discoideum) and present (the generation of the mammal Sus scrofa = wild boar). In this time span, three proteomic uprises did occur: 600 to 500 MYA, 47.5 to 4.75 MYA, and 1.4 to 0 MYA. The first uprise covers the generation of bivalves, the second covers gold fish, chicken, brine shrimp, house mouse, rabbit, Japanese medaka and Rattus norvegicus, and the third covers cow, chimpanzee, Homo sapiens, dog, goat, Puccinia graminis and wild boar. We hypothesise that the latter two uprises are related to geological and climate changes and their compensation in protein function expression.

Conclusions

The proteomic and evolutionary data demonstrate that M. leucophaeata is a highly educatioanal animal to study.

Mytilopsis leucophaeata or the brackish water mussel, belonging to the Dreissenidae or bivalve mussels, originated from Europe more than 60 million years ago (Paleocene, Verween et al. 2010). Subsequently, it disappeared to Central America and returned to Europe (harbour of Antwerp) in 1835. Since then, it is a stable inhabitant of European brackish waters.

In the period of September 2009 to September 2010, two of my colleagues studied the influence of a number of parameters (depth, temperature, salinity and illumination) on size, growth condition, diet and attachment via development of byssal threads (Grutters and Verhofstad 2010). Bivalves have survived a long history from about 510 million years ago to present. This means that it has been adapted to geologically and climate-changing conditions, which might be reflected in the evolution of their proteome. For this reason, the present study was started in order to see whether the presence of certain proteins might unveil certain metabolic systems in this aquatic animal. Almost simultaneously, an article from Riva et al. (2012) did so in relation to the effect of a pollutant (triclosan) on the metabolism of Dreissena polymorpha with emphasis on gills. In the same year, Fields et al. (2012) and Tomanek et al. (2012) published a study on the effect of temperature and hyposalinity on protein expression in the gills of the Mytilidae Mytilus galloprovincialis and Mytilus trossulus. In addition, we were interested in comparing amino acid sequences of our mussel with animals stemming from later periods of life in order to cheque the phylogenetic developments that had taken place meanwhile.

Forty individuals of the brackish water mussel, caught from a branch of the North Sea Channel to Amsterdam harbour, were taken by scalpel knives and tweezers from their shells, yielding a total wet weight of 1.8 g, sufficient for further analysis. The body parts were taken up in 5 ml triethanolamine HCl, pH 7.0 in 25% glycerol in the presence of 0.5 mM phenylmethylsulfonyl fluoride (PMSF) to prevent autolysis (Schuurmans Stekhoven et al. 2003). Further procedures, such as Potter-Elvehjem homogenisation; fractionated centrifugation to fractions F1, F2 and F3; delipidation of fractions prior to electrophoresis; electrophoretic separation of proteins in the fractions; staining and destaining of the gels; determination of the apparent molecular weights of the protein bands on gel; excision of the protein bands; transport to the mass spectrometric analysis laboratory in Leicester; as well as the mass spectrometric analysis itself, is given in full detail in our previous publication (Schuurmans Stekhoven et al. 2010).

Information as to the cellular localisation and function of the analysed proteins stem from handbooks like Biochemistry of Hubert Stryer, Google (Scholar), Pubmed.com, BLAST and UniProtKB/Swiss-Prot Protein Knowledgebase and literature referred to therein. The absorption spectrum of the brownish coloured F1 fraction (325 to 750 nm) was made with a Zeiss M4QIII spectrophotometer at 20- to 50-nm intervals. A 100 μl of the F1 fraction was dissolved in 1 ml 2% SDS, subsequently centrifuged for 5 min at 5,000 rpm in a table top centrifuge, and the supernatant scanned.

Homogenisation and fractional centrifugation

Potter-Elvehjem homogenisation of the mussels required very harsh and frequent pottering, yielding a brownish homogenate. Subsequent centrifugation at 1,200, 9,000, and 100,000 g yielded the F1 to F3 fractions. Total protein (mg) of the fractions amounted to 78.3 for F1, 7.03 for F2 and 6.2 for F3, hence ratio F1:F2; F3 = 12.6:1.13:1.0. This ratio brought about association with the kidney (21.5:11.2:1.0) and colon (19.4:3.9:1.0) of the frilled shark Chlamydoselachus anguineus (Schuurmans Stekhoven et al. 2012) in which particular proteins (L-plastin, moesin, filamin A and α-actinin) are serving as additional construct in linking filaments (microtubules) to the plasma membrane. However, in the mussel case, in particular in relation to the brown colour of F1, and less so of F2, we had to think more in terms of byssal threads, the biopolymers by which mussels attach themselves to their substrate like rocks or even ship walls. The brown colour is based on an aqueous solution of pheomelanin (Napolitano et al. 2008) of which the almost exponential absorption curve (down to 325 nm) fits to our curve of M. leucophaeata F1 (Figure 1). Byssal threads apparently are high MW biopolymers as F1 did not demonstrate any entrance of protein into the gel. This started only in the lightly brown F2 and came to full expression by the light yellow F3, which demonstrated proteins in the apparent molecular weight range of 14.1 to 240 kDa (Tables 1, 2, 3 and 4). The tables are subdivided into prokaryotic and eucaryotic ribosomal subunits (Table 1), proteins from the ER, Golgi network and plasma membrane (Table 2), proteins of the cytoskeleton and muscle (Table 3) and cellular vacuoles, vaults, nuclei and mitochondria (Table 4). From all these proteins, the prokaryotic or eucaryotic origin is mentioned as well as the function and cellular localisation as could be found in literature, including data banks. All proteins are accompanied by their accession numbers from [UniProtKB/SwissProt] between square brackets. Confusion between capital O and the number zero is excluded since capital O is only present at the first position and number zero in any position from 2 to 6 of the accession series. Translation of the accession data to protein easily occurs by using the programme PubMed (www.ncbi.nlm.nih.gov/pubmed) by choosing the term protein.

Absorption spectrum of SDS-solubelised F1 fraction from M. leucophaeata from 325–750 nm. Data points are indicated by black spots that have been line connected.

Full size image

Table 1 shows that 52% of the ribosomal subunits is of bacterial origin with a decreasing order in percentage for Pseudomonas fluorescens (32.8%), Pseudomonas stutzeri (7.5%), Pseudomonas mendocina (3%), Pseudomonas syringae (3%), Escherichia coli (3%), Pseudomonas putida (1.5%) and Pseudomonas aeruginosa (1.5%). The other half of the components is occupied by eucaryotes, ranging from pearl oyster to Norway rat. Possible causes and impacts of the bacterial contamination will be handled under ‘Discussion’ section.

Outside Table 1, only very few bacterial proteins have been identified, except peptidoglycan-associated lipoprotein in the plasma membrane of P. putida (Table 2), outer membrane porin F from P. fluorescens (Table 2) and 2,3-bisphosphoglycerate-dependent phosphoglycerate mutase from Rhodospirillum centenum SW (Table 2). The latter photosynthetic bacterium is housing in marine and brackish water and so can be easily caught by the mussel valves. Still another intruder in the list of proteomics is ribulose bisphosphate carboxylase from Agrostis stolonifera (creeping bent grass) as this reaction takes place in chloroplasts. The habitat of creeping bent grass is on wetlands with tolerance to flooding (Garry Oak Ecosystems Recovery team: www.goert.ca/documents/A.stolonifera.pdf) or inundation of riparian zones which may have brought the plants in contact with the mussels.

Major intracellular activities, presented in Table 2, are protein folding on endoplasmic reticulum (ER) (endoplasmin, peptidyl-prolyl cis-trans isomerase C = cyclophilin C), assembly of multimeric protein complexes inside the ER (heat-shock 70 kDa protein cognate 3, 78 kDa glucose-regulated protein) and protein translocon formation across the ER (dolichyl-diphosphooligosaccharide protein glycosyltransferase). In addition, we found a number of transport processes, such as cargo transport from trans-Golgi to plasma membrane (guanine nucleotide-binding protein), protein sorting at trans-Golgi network and recruitment of clathrin to the membrane (AP-1 complex, clathrin heavy chain 1), protein trafficking via COPI or clathrin-coated vesicles from ER to Golgi and plasmalemma (ADP-ribosilation factor 1 = Arf 1), tubulovesicular recycling of protein from Golgi → ER (glyceraldehyde-3-phosphate dehydrogenase 2) and synaptic vesicle traffic (O-Rab1). Further, a few constructional processes are involved like fragmentation and reassembly of Golgi stacks during and after mitosis + formation of tER (transitional endoplasmic reticulum ATPase), contractile ring formation (actin, cytoplasmic 2) and regulation of actin filament assembly + disassembly (gelsolin-like protein 1). Subsequently, we record a number of cellular signalling components like neural signalling GTPase (guanine nucleotide-binding protein G(o) or G(q), subunit α), integrin signalling at adhesions (Rack 1) and Gαβγ activation of phospholipase c in a cellular signalling process (guanine nucleotide-binding protein subunit β-1). Another constructional component is heat-shock 70 kDa protein, yielding conservation of protein shape and protection against stress. In addition, we noticed a number of metabolic enzymes (α-amylase, glycogen phosphorylase, trypsin, enolase) and cation-activated enzymes (sarcoplasmic/endoplasmic reticulum calcium ATPase as modulated by Rap-1b, calcium-transporting ATPase, Na⁺/K⁺-ATPase) and the selectivity conferring protein in renal brush border nucleic acid conducting channel (cytoplasmic malate dehydrogenase).

The contribution of cellular signalling components appears to be modest in number but has been encountered before in a proteomic analysis of Mytilus galloprovincialis and Mytilus trossulus: three to four signalling components in a total of 47 to 61 proteins, i.e. 6.4% to 6.5% (Tomanek and Zuzow 2010).

The cytoskeletal and muscle components (Table 3) with their numbers between parentheses can be summarised as follows: actin non-muscle (11), cytoskeleton (microfilaments) = transport track for myosin; linkers of actin to plasma membrane: plastin, radixin, spectrin α and β chains (4); myosin-9, involved in vesicle transport via actin filaments (1); tubulin α and β chains (microtubules, involved in ATP-driven vesicle transport or transport of chromosomes in mitosis and meiosis) (4), myosin muscle, involved in contraction: myosin-11 (smooth muscle) (1), myosin LC-1 + heavy chain + adductor muscle light chain (3), paramyosin (byssus retractor muscle) (1), adductor muscle actin (precursor) → contraction (1), actin, muscle precursor, tropomyosin, α-actinin, actin larval muscle (4).

Additional cytoskeletal organised components are: the chaperonin TCP-1, heat-shock cognate protein HSP90-β (with a plurality of functions), ubiquitin (involved in proteolysis), elongation factor 1α (involved in protein synthesis) and eucaryotic initiation factor 4A-I (involved in mRNA binding to the ribosome).

After corrections for bacterial and plant contaminations in Tables 1 and 2, we come to a total of M. leucophaeata-related analyses of 112. The analyses, related to the cytoskeletal and muscle components (1st paragraph of Table 3) amount to 30, i.e. 26.8% of all analyses. Similar results have been scored by Tomanek and Zuzow (2010) for M. trossulus and M. galloprovincialis: 16.4% and 25.5%, respectively. In a later study (Fields et al. 2012), they even scored 39.4% and 52.8%, respectively. These numbers underline the importance for these water-bound animals of a sturdy built body with solid protection against predators. Intracellular stability by the cytoskeleton via linking of actin filaments to the cell membrane, presence of adductor muscles for closure of the shell halves provide additional support for the above ideas, and of course, the presence of paramyosin is essential for binding of the animal to the substrate, including stones and ship walls that bring them to the harbours.

In this last part of our analyses (Table 4), we find only transport vesicles (V-type H⁺-ATPase, major vault protein); histones H2A, H2AV, H2B, H3 and H4; proteasome subunit α type-2, −5A and −7; the 14-3-3 protein ε; ATP synthase subunit α- + β-precursor; the citric acid cycle enzymes succinate dehydrogenase; probable malate dehydrogenase 3; and phosphoenolopyruvate carboxykinase that is responsible for the conversion of oxaloacetate to phosphoenolpyruvate in the gluconeogenesis pathway (Stryer 1995b); and the mitochondrial ADP/ATP exchanger.

This gives a fair picture of what is going on in the subcellular compartments indicated, but some of the animals (or plants) of comparison (second column) require some criticism with regard to their comparability as will be brought forward in the discussion of Tables 3 and 4.

As mentioned in the ‘Introduction’ section, we were also interested in the protein chemical developments that had taken place in the mussel's long history. For that reason, we have registered the generation time for all vertebrates and invertebrates that had provided sequences that led to the identification of proteins that have been presented in Tables 1, 2, 3 and 4. Proteins that could be considered as intestinal contaminants: bacterial proteins, but also a single plant, living in flooded wetlands, have been omitted. Figure 2A,B provides the graphical result of MYA from 1,100 to zero vs. the sum of the vertebrate + invertebrate contribution. It can be seen that the graph curves upward from 600 to 500 MYA, thereby covering the generation of bivalves. From 320 to 50 MYA, the line is linear, covering the insect Drosophila melanogaster (321 MYA) teleosts Takifugu rubripes and goldfish (70 to 48 MYA). Thereupon follows a second uprise until 4.75 MYA in which period Birds (Phasianidae), Decapoda (brine shrimp), Muridae (rat and house mouse), rabbit and the Japanese medaka participate. After this period, there is some levelling off, but is followed by a third uprise, lasting from 1.4 to 0 MYA (Figure 2B). Animals and fungi, involved in this period, are: cow (5), chimpanzee (3), Homo sapiens (7), dog (1), goat (1), stemrust (1) and wild boar (2) (Table 5). From the total of 20 items, 10 (50%) belong to Table 1 and are part of ribosomes that are involved in the protein synthesis machinery. The only cow that was missing in Table 1 can be found in Table 2 as a provider of cyclophilin C, a protein folding catalyst on the ER. Chimpanzee can be found in Table 2 at Ras-related protein Rap-1b and in Table 3 at tubulin α-1 A chain and eucaryotic initiation factor 4A-I. Rap-1b is a cellular signalling component in neuronal cells (Sahyoun et al. 1991) and lymphocytes (Awasthi et al. 2010). Tubulin α-1 A chain is part of the cytoskeleton (microtubules), which supports the cell shape and serves as transport track for vesicles. Eucaryotic initiation factor 4A-I is involved in binding of the messenger RNA to the ribosome. H. sapiens (human) is for about half categorised under the ribosomes at Table 1. Further, humans can be found in Table 2 at α-amylase, glycogen phosphorylase, and dolichyl-diphosphooligosaccharide protein glycosyltransferase and in Table 3 at myosin-9. Functions are: provision of glucose from its polymers, protein transport across the ER and transport of vesicles along actin filaments. The only dog is related to protein synthesis (Table 1); stemrust (Puccinia graminis) is also related to the cytoskeletal actin (Table 3) and wild boar to transitional ER ATPase (TERA) and the serine protease trypsin (Table 2). The first is involved in breakdown and repair of Golgi stacks during and after mitosis, respectively, plus formation of the transitional ER. The second is involved in proteolysis.

A + B. Graphical expression of pro- and eucaryotes from which proteins were found in the mussel's proteome. The x-axis indicates the time of generation in MYA (million years ago), and the y-axis indicates the accumulated expression in percent of total (100%). Point for point has been connected graphically for the period from 1,070 to 6.5 MYA in part A and from 6.5 to 0 MYA in part B. A complete survey of the data is shown in Table 5.

Full size image

Summarising all processes involved in the third proteochemical uprise (covering the Pleistocene and Holocene) shows a composition of protein synthesis, breakdown, folding and transportation across the ER. In addition, binding of messenger RNA to the ribosome, cellular signalling, microtubular vesicle transport, breakdown and repair of Golgi stacks during and after mitosis and formation of the transitional ER are also involved. Glucose supply is provided by α-amylase and glycogen phosphorylase; in the latter case yielding glucose-1P. All these processes are essential for survival of the animal, and during its evolution, it may have been necessary to adapt to the changing environmental conditions.

In addition, we liked to analyse also the second proteochemical uprise (48 to 4.75 MYA, corresponding with half Eocene to second half of Pliocene). Animals belonging to this time range are chicken (9), brine shrimp (2), house mouse (4), rabbit (1), Japanese medaka (1), and Rattus norvegicus (6), with the number of recorded protein components between parentheses. Summation yields a number of 23, slightly more than the third uprise with 20 components, but taking an appreciably longer time span: 43 vs. 1.4 million years.

Processes involved in this second period of development of the mussel are: transcription regulation, DNA repair and replication + chromosome stabilisation (chicken histone H2A. V: 1 item), protein synthesis at the ER by ribosomes from Muridae (Rattus and Mus musculus: 5 items), binding of aminoacyl-tRNA to the ribosomes in protein synthesis (elongation factor 1α, brine shrimp: 1 item), protein folding at the ER (endoplasmin, rabbit: 1 item), cargo sorting at the plasma membrane and trans-Golgi network (clathrin heavy chain 1, R. norvegicus: 1 item, and AP-1 complex, subunit β-1, M. musculus: 1 item), also including endocytosis.

Besides protein synthesis, folding and sorting, there is also the recycling from Golgi to ER (glyceraldehyde-3 phosphate dehydrogenase 2, R. norvegicus: 1 item). Not only protein synthesis is involved, but also protein cleavage (proteasome subunit α type-7, chicken: 1 item). Protein is transcellularly transported via phosphorylation/dephosphorylation of myosin on the cytoskeletal actin tracks. Components of this system are provided by brine shrimp (actin, clone 205: 1 item) and Japanese medaka (cytoplasmic actin: 1 item). Intracellular stability is given by linking of actin filaments to each other (plastin-1, chicken: 1 item) or to the plasma membrane (spectrin α chain, chicken: 1 item, and radixin, house mouse: 1 item). Intracellular vesicle transport is provided by tubulin α-3 chain, which is also involved in mitosis (house mouse: 1 item). Heat-shock proteins: 70 kDa and cognate protein HSP 90-β are involved in conservation of protein shape (anti-stress protectant) and cytoskeletal stabilisation + signal transduction (chicken: 2 items). Last, but not least, muscle contraction for closing and opening the valves is effected by sarcoplasmic/endoplasmic reticulum calcium ATPase 3 and myosin-11 (chicken: 2 items). Not to forget is ATP synthase (chicken: 1 item), which will make transportation and muscle contraction, besides many other processes, possible via its formation of ATP.

Comparison of the processes, represented by the proteins of the second and third uprise, shows similarities and differences. For instance: protein synthesis via ribosomes at the ER counts for 50% in uprise 3, but only for half as much (26%) in uprise 2. Uprise 2 covers transcription regulation, DNA repair and replication + chromosome stabilisation (1 item), whereas uprise 3 does not. Protein folding at ER (1 item) occurs in both periods, but cargo sorting at the plasma membrane and trans-Golgi network (2 items) is only in period 2. The same holds for recycling from Golgi → ER (1 item), whereas protein transport across the ER (1 item) belongs only to period 3. On the other hand, protein cleavage (either by proteasome subunit α type-7 or trypsin) occurs in both periods. The same holds true for actin tracks (2 and 1 item, respectively), but not for actin linkers (3 in period 2 only). Both periods contain vesicle transport (1 and 2 items in periods 2 and 3, respectively). Further activities that relate only to period 2 are protein shape conservation + cytoskeletal stabilisation + signal transduction by heat-shock proteins (2 items). Activities that are related to closing and opening the valves, plus the enzyme that this facilitates (ATP synthase) (3 items), only occur in the second uprise, but cellular signalling by Rab-1b and enzymes involved in the hydrolysis of polysaccharides are confined to uprise 3 (4 items).

Table 1 showed an abundance of 52% in bacterial ribosomal subunits. Although the presence of bacteria in the intestine of a eucaryote is a common phenomenon, the capacity of the present bacteria to break down and thereby detoxicate organic pollutants raises the possibility that these bacteria have been added on purpose to the canal inhabited by the mussel. The following detoxifying properties have been ascribed to some of the indicated strains: P. fluorescens is beneficial for plants in terms of suppressing pathogens, aiding nutrient absorption and degrading environmental pollutants (www.buzzle.com/articles/pseudomonas-fluorescens.html). P. putida is a versatile environmental isolate that is capable of growth on several aromatic hydrocarbons, including benzene, toluene, ethylbenzene and p-cymene. Its broad substrate toluene dioxygenase has been widely utilised in biocatalytic synthesis of chiral chemicals, as well as in the metabolism and detoxification of trichloroethylene (TCE). P. putida F1 is known to be chemotactic to aromatic hydrocarbons and chlorinated aliphatic compounds and has the potential for use in biomediation applications (genome.jgi-psf.org/psepu/psepu.home.html) (site of DOE Joint Genome Institute, University of California). On the other hand, the strain W619 that showed up in our analyses is more competent with regard to heavy metal resistances and beneficial effects on plants (Wu et al. 2011).

P. mendocina DSWY0601 and ymp extrude a polyhydroxybutyrate (PHB) depolymerase that can degrade PHB plastic (Yan et al. 2012). P. stutzeri strain A1501 is equally beneficial to plants by denitrification of NO₃ ⁻, converting it to N₂ and fixation of N₂ → 2 NH₃. Subsequently, NH₄ ⁺ is coupled to α-ketoglutarate under formation of glutamate (Stryer 1995c; Lalucat et al. 2006).

In contrast to the above positive descriptions, the list of bacteria also contains some negatively acting contributors: P. aeruginosa, P. syringae and E. coli. P. aeruginosa, despite its positive contribution in oil degradation in the presence of glycerol or the biosurfactant rhamnolipid (Zhang et al. 2005b), also excretes toxins that are deleterious for the pulmonary system (Roy-Burman et al. 2001). P. syringae is a plant-pathogenic bacterium, infecting bean to tomato, causing bacterial speck to bacterial cancer (P. syringae Genome Resources home page: Pseudomonas-Plant Interaction (PPI) from Cornell University: Department of Plant Pathology: www.pseudomonas-syringae.org). E. coli APEC01 is a deleterious avian pathogenic bacterium causing epidemic colibacillosis in the poultry industry (Kabir 2010).

Some data of similarity with C. anguineus (Schuurmans Stekhoven et al. 2012) are: synaptic vesicle traffic (Ngsee et al. 1991) and enolase in the plasma membrane of synaptosomes (Ueta et al. 2004). The first reminded us of the neurotransmitter cycle that we found in the brain of C. anguineus via its modulator α-synuclein and v- and t-snares VAMP1/2 and SNAP-25 + syntaxin 1. However, although bivalves contain a nervous system (Encyclopaedia Brittanica: www.brittannica.com/EBchecked/topic/67293/bivalve/35745/The-shell), we have not been able to find the abovementioned v- or t-snares for bivalves. In partial contrast to this are the results obtained for M. galloprovincialis (Venier et al. 2009) in which results for three t-snares in the Mediterranean mussel have been obtained via transcribed sequences: SNAP-25A [Accession No. Q5TZ66], SNAP-type protein [Accession No. Q25391] and SNAP-47 [Accession No. Q0P4A7]. Yet, v-snares have not been detected either in this case. Presence of enolase in the plasma membrane appears to have an endangering effect via its complex formation with plasminogen that by subsequent activation to plasmin can break down the extracellular matrix and so can allow invasion of pathogens, viruses and metastatic cancer cells (Liu and Shih 2007; Díaz-Ramos et al. 2012). Normally, plasmin is used to dissolve fibrin blood clots but upon generation on the cell surface might cause the above effects. However, in our analyses, neither plasminogen nor plasmin (MW 81 and 75.4 kDa, Barlow et al. 1969) or plasminogen activator (tPA, MW 72 kDa, Manosroi et al. 2001) has been traced. On the other hand, in the transcribed sequences of M. galloprovincialis, two sequences were found that matched plasminogen [Accession No's. Q01177 and Q6PBA6] (Venier et al. 2009). Furthermore, despite the clear presence of Na⁺/K⁺-ATPase in our analyses, we have been unable to find the presence of phospholemman (FXYD1), known as a modulator of Na⁺/K⁺-ATPase (Mahmmoud et al. 2000), even though our analyses covered a wide range of molecular weights (14.1 to 240 kDa). Since phosphorylation of Na⁺/K⁺-ATPase causes dissociation of phospholemman, this may have led to its absence in the analyses. In addition, salinity may also decrease the FXYD content relative to the Na⁺/K⁺-ATPase content (Wang et al. 2008). In another report (Horisberger 2006), it has been indicated that no FXYD protein can be found in arthropods or any nonvertebrate animals. We think that the only way that is left to trace the absence or presence of FXYD in bivalves is to analyse their DNA.

One of the fungi that have entered the list of comparative sequences is P. graminis at actin in Table 3. This mould spreads its occurrence by spore formation via two different hosts, thereby causing the so-called stemrust, especially in wheat and barley (Schumann and Leonard 2000). Although contamination of the brackish water mussel with infected wheat and/or barley from freight ships in the harbour cannot be excluded, another possibility is indistinguishable peptides formed by trypsin treatment (cf. Schuurmans Stekhoven et al. 2010) as used in the analysis of actin from Crassostrea gigas [Accession No. O17320] or Puccinia graminis [Accession No. P50138]. A few of the possibilities are a²⁰gfagddapr²⁹, h⁴¹qgvmvgmgqk⁵¹ and y⁷⁰piehgivtnwddmek⁸⁵ for Crassostrea gigas and the same sequences for Puccinia graminis, but with a numbering of a¹⁹-r²⁸, h⁴⁰-k⁵⁰ and y⁶⁹-k⁸⁴. It appears that the sequences are quite conserved since they are also found in β-actin of the mammal M. musculus [Accession No. ABL01512].

Another subject of criticism is the possibly hereditary plant sequences in the genome of the mussel. An example could be the occurrence of Proteasome subunit α type-5A and ADP/ATP carrier protein from Arabidopsis thaliana (thale cress). Thale cress grows on edges of agricultural fields, stone walls alongside tracks and roads and Mediterranean scrublands with scattered holm oaks but is no inhabitant of wet lands (Picó et al. 2008). Therefore, we have looked for comparable sequences in MytiBase: a knowledgebase of mussel (M. galloprovincialis) with 3,275 transcribed sequences (Venier et al. 2009). To our surprise: also in this large list of transcribed proteins, a few examples of plant heritage were met: first the occurrence of 14-3-3-like protein b of Oryza sativa (India Group) [Accession No. ABR25888] together with 14-3-3 C1 protein from Oncorhynchus mykiss [Accession No. Q6UFZ7]. Identical sequences, found for Oryza sativa with those for O. mykiss (the latter between parentheses), are e¹⁶-e³⁴ (e¹¹³-e¹³¹), p⁶⁵-f⁸⁰ (p¹⁶²-f¹⁷⁷), l⁹⁴-d¹⁰⁰ (l¹⁹¹-d¹⁹⁷) and s¹¹³-d¹³⁴ (s²¹⁰-d²³¹). It is evident that the sequence for Oryza sativa is 97 amino acids less than that of O. mykiss, due to incomplete DNA. Yet, we can calculate an identity that must be minimally 25% the same. Since the sequences in case of M. galloprovincialis have been determined via DNA, we have to accept a genetic link between animals and plants, and so the link with proteasome subunit α type 5A, ADP/ATP carrier protein 3 and Arabidopsis thaliana may be genuine and not artificial.

A second example of plant heritage by M. galloprovincialis is found in the presence of probable ATPase from the chloroplast of Oenothera organensis (organ Mountains evening primrose) [Accession No. Q0H0T1] which grows in the mountains of New Mexico, far away from the mussel of the Mediterranean Sea. Hence, there has been a time that they were neighbours. This hypothesis is built on three assumptions:

1. 1.

   Some DNA of consumed food can be taken up in cells and incorporated into the DNA of the consumer if it displays some similarity with DNA of that consumer.

2. 2.

   Since bivalves do not inhabit the mainland, the plant consumers could be snails (Gastropods), which form a sister clade with bivalves, forming the Pleistomollusca (Kokot et al. 2011).

3. 3.

   If geological conditions, like flooding, would force gastropods to evolve to bivalves, it is not unthinkable that bivalves would contain land plant sequences in their genome as has been shown by Venier et al. (2009). Before accepting this hypothesis, it will be necessary to trace the snail's genome or RNA for plant resemblances.

In the generation of the brackish water mussel M. leucophaeata in the period of 1,070 to 0 MYA, two additional genetic uprises occurred beyond the uprise caused by the generation of the bivalves per se (approximately 510 MYA). The question arises why this has to be achieved by acceleration in a developmental uprise: the second in the period of 48 to 4.75 MYA and the third in the period of 1.4 to 0 MYA. In the 48 to 4.75 MYA period (Eocene-Pliocene), earth was in motion with formation of mountains and separation or collision of geological plates, volcano formation, followed by climate cooling (Pidwirny 2012). In the later period (1.4 to 0 MYA: Pleistocene + Holocene), earth was subject to freezing (Pleistocene Ice Age) with extinction of many species (Pidwirny 2012), which may have forced the mussel to a counterreaction by speeding up its adaptation of DNA to that of modern species (cow, chimpanzee, human, dog, stemrust and wild boar, Table 5).

Evidence for climate change-induced effects on adaptation of the genome of animals and plants has been recently published (Reusch and Wood 2007; Buckley et al. 2012; Franks and Hoffmann 2012). Failure to adapt to the changes may eventually lead to extinction. In this respect, M. leucophaeata did not fail, otherwise it could not have been able to survive for 500 million years until present (Figure 2). Therefore, it is remarkable that it has only a limited range of salt concentration: 6.7 to 7.4 ppt (⁰/₀₀) to provide for an optimal condition. At higher salinities, the condition index is reduced to 50 at a salinity of 11 ppt (Grutters and Verhofstad 2010). Seawater usually has a salinity of 35 ppt (Office of Naval Research: www.onr.navy.mil/focus/ocean/water/salinity 1.htm). Therefore, one may question how M. leucophaeata can survive the trip via the ocean to the brackish North Sea Channel. Transportation of larvae and postlarvae with tolerance to salinity of 32 ppt in ballast water makes the trip possible (Verween et al. 2010). The salinity of the North Sea Channel varies from 1.7 to 9.2 ppt (Van der Velde et al. 1998), which M. leucophaeata can enter with confidence.

We like to thank M.Sc. Michiel Verhofstad and M.Sc. Bart Grutters for their gift of M. leucophaeata, which is at the basis of this study. As in my previous publication (Schuurmans Stekhoven et al. 2012), M.Sc. Ruud van Hintum helped me by photography of the gels after electrophoresis of the F1 to F3 fractions. Also, K. van Benthem is acknowledged for the construction of Figure 1 of this paper and my son M.G.H.P. Schuurmans Stekhoven for the construction of Figure 2, Lady S. Ibrahim (Protein and Nucleic Acid Chemistry Laboratory, Proteomics Facility), University of Leicester for her contribution in the protein analyses of this paper and of course Kiki Wu for her textual settings.

Affiliations

1. Department of Animal Ecology and Ecophysiology, Faculty of Science, Radboud University Nijmegen, Heyendaalseweg 135, 6525 AJ, Nijmegen, The Netherlands
   • Feico MAH Schuurmans Stekhoven
   •  & Gerard van der Velde
2. Department of Life Sciences, National Chung-Hsing University, Taichung, 402, Taiwan
   • Tsung-Han Lee
3. Protein and Nucleic Acid Chemistry Laboratory, Proteomics Facility, University of Leicester, Lancaster Road, Leicester, LE1 9HN, UK
   • Andrew R Bottrill
4. Naturalis Biodiversity Center, P.O. Box 9517, 2300RA, Leiden, The Netherlands
   • Gerard van der Velde

Corresponding author

Correspondence to Feico MAH Schuurmans Stekhoven.

Competing interests

The authors declare that they have no competing interests.

Authors’ contributions

FSS carried out the homogenisation and fractional centrifugation of the bivalve homogenate, electrophoresis of the fractions, staining and destaining of the gels and conveying the gel strips to the analytical laboratory of ARB in Leicester. Further, he interpreted the data and wrote this article. THL offered his contribution in formatting of the text and provided his data (included in this paper) on the effect of salinity change on the expression of FXYD, a modulator of Na⁺/K⁺-ATPase. GvdV, the malacologist of our department, provided essential information on the anatomy and physiology of mollusks, and ARB, together with Lady S. Ibrahim, analysed the gel strips and sent me the proteomic data together with their accession numbers in the data banks. All authors read and approved the final manuscript.

Reprints and Permissions