|Home | About | Journals | Submit | Contact Us | Français|
Macaques are the closest evolutionary relatives of humans routinely used in basic and applied biomedical research. Their genetic, physiological, immunological and metabolic similarity to humans, second only to that of the great apes, makes them invaluable models of human disease. These similarities also mean that macaques are often the only experimental models available for evaluating increasingly specific drugs in development, and as a proof-of-concept bridge can help reduce the numbers of compounds that fail in clinical pharmaceutical research. In vertebrates, red blood cells (RBCs) diseases are frequently severe as their role as sole gas transporter makes them indispensable to survival; much research has therefore focused on an in-depth understanding of the functioning of the RBC. RBCs also host malaria, babesia and other parasites. Recently, we presented an in-depth proteome for the human RBC and a comparative human/mouse RBC proteome.
Here, we present directly comparable data for the human, mouse and rhesus RBC proteomes. All proteins were identified, validated and categorized in terms of sub-cellular localization, protein family and function and, in comparison with the human and mouse RBC, were classified as orthologues, family-related or unique. Splice isoforms were identified and polypeptides migrating with anomalous apparent molecular weights were grouped into putatively ubiquitinylated or partially degraded complexes.
Overall there was close concordance between mouse, human and rhesus proteomes, confirming the unexpected RBC complexity. Several novel findings in the human and mouse proteomes have been confirmed here. This comparison sheds light on several open issues in RBC biology and provides a departure point for more comprehensive understanding of RBC function.
Animal research has been indispensable in understanding human biology and disease. As a readily bred and maintained primate species, rhesus monkeys (Macaca mulatta) offer invaluable models for disease-related research because of their genetic, physiological, immunological and metabolic similarity to humans, and are thus the most widely used non-human primate model in biomedical research. Research in macaques continues to contribute to fields of applied and basic biomedical research including infectious diseases (e.g. HIV1,2, malaria3,4, varicella5, tuberculosis6), pharmacology (drug development and testing7; vaccine trials8,9), immunity10,11 and autoimmunity (e.g. arthritis12, multiple sclerosis13), neuroscience14,15, behavioural biology16 (social adjustment, learning, drug addiction, alcoholism and a range of behavioural disorders), reproductive physiology17, transplantation18,19, endocrinology20; cardiovascular, diabetes and obesity studies21,22; well-controlled studies of diet (e.g. calorie restriction)23, exposure to chemical/biological agents24, long-term studies of the physiology of aging25. With respect to the major infectious diseases of developing countries, macaque models are particularly suited for studies of malaria physio-pathology and drug/vaccine testing3,4, studies of vaccine efficacy for TB6 and studies of disease progression and pathogenesis in AIDS even allowing for studies of infection, pathogenesis and treatment of animals infected in utero26.
The red blood cell (RBC) is one of the more important and readily accessible cell populations in any mammalian organism and has therefore been an object of intensive investigation for many years. The human blood factor, Rh, is named after the rhesus monkey, because studies in these animals contributed to a definitive understanding of blood antigens. Recently, the use of a rhesus macaque animal model was proposed for the testing of Haemophilia B gene therapy27.
The availability of high accuracy and high sensitivity mass-spectrometry techniques and publication of the first macaque genome28 provides an opportunity to derive in depth proteomes for critical macaque cell types. Because of the experimental importance of the rhesus macaque we have derived a deep coverage rhesus RBC proteome and compared it with our recently derived deep coverage human29 and mouse30 RBC proteomes. To our knowledge this is the first global analysis of rhesus membrane and soluble RBC proteins, but also the first three-way comparative RBC proteomic analysis in which samples were prepared and analyzed under identical conditions and data validated according to consistent criteria.
Whole blood of adult Rhesus macaques was collected in citrate by saphenous vein puncture under ketamine anaesthesia. All aspects of the work had received prior approval from the BPRC animal experimental committee (DEC). Saphenous samples were processed at 4 °C. White cells were removed (Plasmodipur®, Euro-diagnostica B.V., Arnhem) according to manufacturer instructions before RBC were pelleted (1,700 x g, 5 min), resuspended and layered on 30 mL of Lympholyte-H, Cell Separation Media (CL5020, Cederlane Laboratories Limited, Ontario, Canada) (1,700 x g, 5 min) to remove granulocytes and residual lymphocytes; at each step, along with the supernatant, the upper RBC layer was removed. As for our previous study of human/mouse RBC29,30 samples were stored before analysis (4 °C, 72–96 hours) to allow maturation of reticulocytes to RBC. Finally, RBCs were washed (5 times, cold RPMI-1640 (Sigma-Aldrich, Zwijndrecht, The Nederlands) before preparation of membrane and cytoplasmic fractions.
RBC from five separate purifications were diluted (RPMI-1640) for counting (haematocytometer, taking the mean of five fields, repeated three times).
Counts consistent with reference data for rhesus RBC were obtained (4.63*109 cell/mL - 5.97*109 cell/mL). Conventional slides of processed (5 batches) and unprocessed blood, stained with Azure B31 or May-Grünwald Giemsa32, were counted for white blood cell (WBC), granulocyte, monocyte, reticulocytes and platelets.
Preparation and analyses were essentially as previously described for human/mouse RBC29,30. Briefly, membrane fractions were obtained by hypotonic phosphate buffer lyses and soluble fractions were obtained through repeated freeze/thaw cycling. Proteins in both membrane and soluble fractions were further separated by SDS-PAGE; membrane samples were also extracted using combinations of sodium carbonate, ethanol and EDTA.
Samples were enzymatically digested either in solution or in gel33 prior to LC-MS/MS using a Q-STAR (Q-STAR Pulsar; PE Sciex, Canada, 17 runs) or LTQ-FT (Hybrid-2D-Linear Quadrupole Ion Trap - Fourier Transform Ion Cyclotron Resonance (FTICR) Mass Spectrometer; Thermo Electron, Germany, 3 runs). MS/MS spectra were searched against the non-redundant International Protein Index (IPI) rhesus sequence database (www.ensembl.org, 43339 sequences; release 07–2008)34 using Mascot software35 (version 2.0) using the same parameters as used for the human/mouse RBC proteome29,30 (Addendum 1). Spectra were also searched using the corresponding reverse database to estimate false positive peptide identifications (0.5%). Because annotation databases are by nature incomplete and changing it is imperative that outputs are subject to critical evaluation, to ascribe for example the most likely function(s) when several are available. Annotation data presented here result from assembly of all information available via Uniprot (Swiss-Prot and TrEMBL)36 and EBI databases34 (www.ebi.ac.uk) using protein accession numbers. As such databases are incomplete and dynamic such analyses cannot be absolute. When in doubt, a further crosscheck was provided by submitting the protein description to PubMed37 and evaluating the resulting literature for relevance to RBC. Practical limits preclude citation of all relevant papers; as might be expected findings were sometimes contradictory.
Validation by MSQuant (version 1.4.3, official release 11-02-2008), an open source software developed by our laboratory (http://msquant.sourceforge.net), provided a manual score and spectrum evaluation for each of the peptides that had led to the identification of a given protein. The same stringent criteria were applied as for the human/mouse RBC proteome29,30. Proteins were then blasted, all versus all (cutoff 95% to remove redundancy). Swiss-Prot/TrEMBL (http://www.expasy.org)36, Ensembl (http://www.ensembl.org)34 and Gene Ontology databases38 were used for annotation. Unique Swiss-Prot/TrEMBL/Ensembl numbers provided access to sequence, isoform, family, localization and function data for identified proteins; in parallel IPI numbers were queried against the Gene Ontology database38, enabling grouping by class, function or localization and quantitation within grouping (for example within the different signal transduction pathways).
An all versus all blast of validated proteins was done to eliminate redundancy, providing final membrane, soluble and mixed fraction protein lists (Supplementary material).
In particular, where proteins were identified only with only one method (output for each method is typically the result of 3 runs of pooled RBCs prepared and analyzed using the same conditions), annotation databases and the literature were extensively used to define proteins as genuine red cell components or as probable contaminants from other blood sources.
RBC preparations for mass spectrometry were analyzed for cell purity using stained slides. Automated counts were performed with a Sysmex XT-2000i instrument (Sysmex Nederland B.V., Etten-Leur), which combines automated and sophisticated reticulocyte analysis haematological determinations with mammal-specific settings. In one of five samples we observed a maximum of 1 WBC in one million RBCs; no granulocytes, monocytes, platelets or reticulocytes were detected. This low degree of contamination was confirmed by the fact that high copy number molecules from other circulating cell types, namely CD45 (leucocyte-specific)39, transferrin receptor (reticulocytes, lost in exosomes during maturation)40 and ferritin receptor41, were not in our final protein list. A comparison of the proteins found in our study and those found in platelets (“secretomes”)42–44 shows that only proteins that are known to belong to both platelets and RBCs or to bind to both cell types are found in both proteomes, while platelet-specific proteins such as coagulation factors, fibrinogen and different platelet glycoproteins are not found in our RBC dataset. Worthy of mention is thrombospondin 1, which is known to bind to different blood cells45 and was identified with a high number of peptides in platelets proteomes. This protein was found associated to both human and rhesus membranes, but not in the mouse proteome. Interestingly, the 14-3-3 zeta-delta protein, which was until now attributed solely to platelets by the Uniprot database46, appears to be present also in RBCs, since it was found consistently in all the three proteomes. A special case is that of the Amyloid A4 protein, which does not appear to be present in RBC, but was identified in the membrane fraction all three proteomes. Literature evidence suggests, that this protein may bind to the RBC membrane in the central nervous system47 and in the plasma48.
Although these results suggest that the extensive RBC purification and stringent protein inclusion criteria yielded an essentially pure RBC proteome, due to the intrinsic need to balance sensitivity and specificity, we cannot rule out inclusion of a small number of false positive identifications nor the occasional presence of contaminating proteins wrongly assigned to the RBC.
The study goal was to obtain the most complete possible proteome; using the same strategies (extensive purification of the starting material, stringent identification parameters), and MS-machines previously used to determine the human and the mouse proteome29,30. To address the dynamic range challenge arising from the wide difference in copy numbers between abundant and rare proteins typical of the RBC (e.g. Band 3 1 million copies/RBC versus CR1 100 copies/RBC), we used parallel approaches, employing both LTQ-FT and QSTAR for MS. The LTQ-FT provides rapid cycling, high mass accuracy (FT), high fragmentation speed (LTQ) and significantly enhances the dynamic range allowing for the identification of extremely low abundance extra-cellular RBC-binding proteins (Complement C1q subcomponent subunit A and C, Complement C1s and C1r subcomponent, Complement C3-like, C2 and factor B). To maximize protein detection using the QSTAR, exclusion lists were created.
As RBC membrane proteins have different biochemical characteristics and are partially shielded from digestion by the lipid membrane core; different extraction methodologies (e.g. for membranes using various solvents, detergents, ion-chelators and ionic solutions) were used to ensure an in-depth analysis of the human, mouse and rhesus ghost samples. These methodologies can be combined (e.g. NaCO3 and EtOH) or applied independently to aliquots of the same sample prior to MS-analysis, furnishing a very broad picture of the sample at hand. RBC membrane samples were first prepared for extraction by hypotonic buffer lysis - a procedure ensuring the preservation of RBC membrane characteristics. Extraction procedures of choice included delipidation (e.g. EtOH), differential disruption of ionic bonds (using calcium carbonate) and disanchoring of the cytoskeleton (e.g, EDTA) (Table I). The evidence summarized in Table I suggests that the Rhesus membrane behaviour vis-à-vis of the different extraction procedures most closely resembles that of the mouse membrane. Results from repeated (twice) extraction with calcium carbonate suggest that the rhesus and mouse RBC membrane contain less integral proteins, than the human RBC membrane. The trend in the differential extraction procedures and a visual inspection of the gel (10μL ghost loaded Figure 1) suggest that the amount of overall membrane proteins may be decreasing in the order human (341), rhesus (328) and mouse (247/262 considering mixed fraction proteins orthologous to human membrane). These results are in agreement with Bicinchoninic Acid Protein Assay (BCA) determinations on ghost RBC membrane protein’s content (human: 2g protein/10mL blood; rhesus: 1.8g protein/10mL blood; mouse: 1.6g protein/10mL blood).
It is interesting to notice, that with eight more proteins than the human RBC proteome, the rhesus proteome features the highest amounts of extracellular proteins (28). Aside from a few more apolipoproteins (A-IV49; B50; H51), which may not have been detected in the human proteome because of the low abundance and the transient nature of their binding to the RBC surface, a number of complement-related proteins were found.
Among the apolipoprotein, particulatly interesting is ApoH (beta2-glycoprotein I), as it seems to exert a number of roles across the RBc life span: it is involed in the inhibition of thymidine incorporation in fetal calf erythroid cells52; it may induce RBC discocyte-echinocyte-spherocyte shape transformation and subsequent agglutination of RBCs53 and it is proposed to be among the non-immune mediators of RBC clearance54,55.
In the context complement-related proteins it is interesting to notice that the analysis of the different mammalian RBC proteomes gave very varied results.
Common to the human and rhesus RBC membrane proteomes were the complement receptor 1 (CR1), complement factor C3b and IgG proteins, while C1Q and C4 were common to the rhesus and mouse RBC membrane proteomes. The complement receptor is notoriously absent from the mouse RBC, but the complement receptor related protein (Crry) was found. In the rhesus RBC proteomes C1s, C1r, C2 and the C3/C5 convertase were additionally found.
In the human proteome, we were able to highlight the presence the high MW C3b-IgG complexes first discovered by Lutz and colleagues56,57, while in the mouse these complexes were absent. In the rhesus, most of the IgG were found in the gel either at their MW or at a lower MW indicating degradation. However, in 1 rhesus membrane gel, high MW C3b-IgG complexes were found. These findings, should be considered in the context of literature data58–63 suggesting that the lower the organism the more RBCs are disappearing at random. Hence, the absence of C3b2-IgG complexes within a lower mammalian species, may illustrate that this species has lost the majority of RBC before they reached old age (the half-life of RBCs decreases progressively from human (120 days) to rhesus (90 days) to mouse (60 days) or may indicate, that while the elimination of senescent RBCs in humans happens with prevalence through the non-classical complement pathway, in the mouse it happens through the classical complement pathway and that in the rhesus macaque both pathways play a role in this process. It is interesting to notice that while IgM NAbs are the primary NAbs in lower vertebrates (eg mouse), IgG natural Abs (NAbs) are rare in mouse, but common in human and increasingly common in monkeys (finding confirmed by the comparison of the three mammalian RBC proteomes). This can be explained by the fact that IgM NAbs are primarily directed to neo-antigens, while IgG NAbs can be directed to a pre-existing antigen, if its extent of oligomerization changes. At the moment, hard conclusions stemming from the interpretation of the above findings are difficult as very little is known about species-differences with regard to mechanisms of RBC clearance.
Multiple members of the same protein family and protein isoform where critically evaluated and evidence of unique peptides with spectrum scores over 30 and correct amino acid sequence attribution was searched for in the proteomics data. The unequivocal attribution of membrane fraction splice isoforms was possible for 7 of 20 proteins in the mouse RBC, 22 of 25 in the Rhesus RBC and 48 of 54 in the Human RBC. In the analysis of the soluble fraction, 6 of 17 splice isoform for the mouse, 6 of 6 for the Rhesus and 28 of 28 in the human could be unequivocally assigned (Mouse mixed fraction 4 of 11 splice isoforms assigned).
These data suggest that in the human RBC proteome a much higher amount of splice isoforms (more than double) was found and in most cases unequivocally assigned when compared to the other two mammalian RBC proteomes. This could be interpreted either as a consequence of the annotation and the significantly higher amount of knowledge we have about human isoforms, or as an objective finding suggesting that human proteins give rise to more tissue and function specific protein variants than proteins in other -even closely related- mammals.
A closer look at the different families of cytoskeletal RBC proteins (Table II) identified in the three proteomes, demonstrated the overall similarity in the sub-cellular network. However, there are a few interesting differences: adducin-g is expressed only in mouse and rhesus erythrocytes and its expression in human RBCs is associated to spherocytic hereditary elliptocytosis64; Tubulin α-3 is not expressed by rhesus RBCs and its expression thus appears to be confined to the human RBC. It is interesting to notice that a new member of the spectrin family has been identified in rhesus RBCs, namely Spectrin β-chain, brain 2 (Beta-III spectrin). While Beta-III spectrin has never been described as a part of the RBC cytoskeleton before, mutations seem to be linked to a form of neuro-degeneration know as spinocerebellar ataxia65,66. Links have been found between forms of this disease and the Creutzfeld-Jacob disease67, which can itself be associated with a cerebellar ataxia onset68,69. This observation is interesting, as the pooled rhesus RBC samples reveled a number of Creutzfeld-Jacob disease-related proteins (Major prion protein PrP (CD230), Cystatin C, prion inteacting protein-1) most of which are characterized in the databases as either secreted or GPI-anchored.
In the human RBC a number of glycolytic enzymes were found to be membrane/cytoskeleton-bound29, while in the mouse most glycolytic enzymes were found exclusively in the soluble RBC fraction30. In this context, the Rhesus RBC offers a more variegated picture (Table III), which appears to be consistent with the supra-molecular organization of glycolytic enzymes in other macaque tissues70. It is noteworthy that just one enzyme variant (6-phosphofructokinase muscle and liver type) was found in the Rhesus RBC, while a few glycolitic enzyme variants were found in mouse and human RBCs.
This observation is in line with reports by Buettner-Janusch and colleagues who hardly found any variants of glycolytic enzymes in the RBCs of the semi-free-ranging population of rhesus macaques that inhabit Cayo Santiago71. A detailed investigation by Kurganov and others found that with two independent binding sites 6-phosphofructokinase has a key role in the formation of the multi-enzyme complex, which leads to the compartmentalization of the glycolytic process72, although the bulk of the associations remain transient in nature73.
Researchers have argued on whether isozymes are central to the differential propensity of a same enzyme for adhering to erythrocyte membranes74 often supporting their arguments with different findings; evidence from the mammalian RBC proteomes at hand suggest that isozymes may be important (e.g. 6-phosphofructokinase, GAPD) but also that differential compartmentalization of glycolysis may be brought about by different enzymes in different mammals.
Having eliminated redundancy, final human, mouse and rhesus RBC protein lists of both the membrane (Figure 2A) and soluble fractions(Figure 2B), were compared to identify orthologs and family related proteins/paralogs. To increase prediction reliability, a pipeline of three independent algorithms (Blast homology algorithm based on NCBI Blast.exe, Peer Bork algorithms (STRING algoritms75, Blast2e and related algorithms76, http://smart.embl.de77) and EBI PHYML) was used in the determination of commonality/uniqueness and evolutionary relationships among proteins. The PHYML algorithm created by the European Bioinformatic Institute offers an advantage over most other algorithms as it enables not only to find simple unique orthologs characterized by clear concordance between reciprocal best approaches, but also more complex one-to-many and many-to-many relationships.
Central to PHYML78–80 is the generation of maximum likelihood phylogenetic gene trees, which are reconciled with species trees (using RAP), annotating internal nodes distinguishing duplication/speciation events, to represent an evolutionary history for gene families. In this study, proteins were defined as orthologs when homology Blast returned > 80% identity and the Bork algorithm tree showed a common precursor, or when EBI predictions indicated that two proteins were orthologs. In the Venn-Diagrams (Figure 2A and andB),B), the tri-way orthology/paralogy is represented in a simplified manner: proteins pertaining to the mixed fraction of the mouse proteome is either been attributed to the soluble or to the membrane fraction depending on the sub-cellular fraction in which the rhesus orthologs were found
These combined approaches returned orthologs as follows: 164 MM/HM/RM, 60 HM/RM, 36 MM/RM, 54 HM/MM, 49 unique to HM, 46 unique to RM and 37 unique to MM. These and family-related proteins 27 (MM/HM/RM), 7 (HM/RM), 9 (MM/RM), 5 (MM/HM) are detailed in Fig 2A. The analysis of the soluble fractions yielded the following results: 114 MS/HS/RS, 28 HS/RS, 20 MS/RS, 11 HS/MS, 88 unique to HS, 25 unique to RS and 137 unique to MS. These and family-related proteins 15 (MS/HS/RS), 6 (HS/RS), 5 (MS/RS), 7 (MS/HS) are detailed Figure 2B.
The membrane protein dataset was analysed for sub-localization (Figure 3). For 296 of the total 328 membrane proteins a sub-localization was available in the annotation database: 79 proteins were integral to the membrane; 3 GPI-anchored; 58 membrane associated/bound (29 are cytoplasmic proteins that upon activation transfer to the membrane); 40 cytoskeletal (of which 10 cytoskeleton associated); 39 organellar (of which 13 are involved in cytoplasm vesicle trafficking); 49 cytoplasmic, 28 extracellular (of which several are described to bind to the RBC surface e.g. galectin-3, SDR-1, SPARC, apolipoproteins, complement factors etc.).
As is the case for human and mouse, most of the proteins found are involved in binding, transport, inter-(blood groups) and intra-cellular signaling or have structural activity.
Transport processes in both the membrane and soluble fractions mostly involved ions and proteins. Redox and cellular water homeostasis, the control of cell shape/volume, vascular processes and salt levels are predominant in terms of membrane cellular functions. In the context of vascular process, we have consistently with our previous reports, once again found evidence of proteins involved in nitric oxide modulation81,82 (guanine nucleotide binding [Gi]83 and Hsp-90) and of the low abundance receptor-type tyrosine-protein phosphatase alpha consistently with the theory of Battacharya84.
The rhesus membrane proteome is composed of 328 proteins, of which 46 appear to be uniquely expressed. Eight of these proteins are the products of novel genes that are so far only ascribed to the rhesus macaque. The GO annotation for these proteins obtained from the Embl Database34 shows 2 proteins likely to belong to the IgG family, 2 proteins likely involved in the intracellular vesicle mediate transport, 1 protein member of the fatty acid binding family, 1 of the iron/heme binding family, 1 of the phenol-pyruvate kinase family and a protein for which no go annotation could be retrieved.
Interesting is the detection of some proteins associated with adhesion, rosetting and interaction between RBCs and the neutrophils (cytohesin-1, vanin-2 and Dynein), which consistently with their role in the formation of stable complexes all appear at a higher molecular weight; the detection of proteins known from literature to be associated with the Kreuzfeld-Jacob syndrome (see detection and validation of isoforms) and the detection of a few contaminants from the skeletal muscle (RB1-inducible coiled-coil protein 1, Creatine kinase, sarcomeric mitochondrial precursor) and the endothelial cell layer lining the blood vessels (LIN-7 homolog A). A few secreted proteins are also found (GLPG464, Histidine-rich glycoprotein and Tumor necrosis factor receptor superfamily member 6). According to literature evidence, nearly all other proteins found are associated either to the mature RBCs or to reticulocytes, where the latter are generally not found at their molecular weight. Overall, of the 46 unique rhesus membrane proteins, 20 are found in bands corresponding to their MW, 16 both at their MW and at a higher MW (> 188 kDa), 9 are found exclusively at higher MW and 1 exclusively at lower MW.
The soluble protein dataset was analysed for sub-localization (Figure 4). For 181 of the total 202 soluble proteins a sub-localization was available in the annotation database: 144 proteins were cytoplasmic; 13 cytoplasmic and nuclear, 7 nuclear (of which 3 membrane associated), 9 Golgi and/or endoplasmic reticulum (of which 4 membrane associated), 8 mitochondrial and/or cytoplasmic (of which 2 membrane associated). Consistently with the continuous in-development nature of the RBC most proteins identified were involved in the catabolic protein metabolism (e.g. proteasome family members (21), ubiquitin-related metabolic proteins ) and only a few (10) in protein synthesis. Reticulocyte-legacy proteins found included the members of a number of macromolecular complexes (e.g. eukaryotic pre-initiation and initiation complexes, the nucleosome complex) probably representing the results of incomplete degradation (did not show their expected MW on gel and were often found at lower MW).
The rhesus soluble RBC proteome contains a total 202 proteins, of which 24 are unique. According to literature these 14 of these proteins have either been attributed to the RBC before (9) or are reticulocyte legacy proteins (5). Two of these proteins are found at a different MW from the expected one, while further two are found both at the expected MW and either at a higher/lower MW (2). Only one of these proteins is found consistently at its molecular weight. Of the other proteins, 4 are attributed by the databases to the bone marrow, 3 are ubiquitous, 1 is involved in the metabolism of nitrogen compounds and for 2 proteins no relevant information could be found (1 appears to be a novel rhesus specific product). Interestingly there are a few proteins that in nucleated cells are associated with the membrane of internal organelles, which in the RBC have been described to associate with the plasma membrane, but are found in the soluble rather than in the membrane fraction.
Although RBC’s are devoid of nuclei, a few proteins annotated as organellar were identified in all three proteomes. Among others, these include transcription and translation regulators, the 78kDa glucose regulated protein and some members of the 4-3-3 protein family. Overall, closely similar sets of organellar proteins were found in all three mammals to migrate at anomalously high (elongation factor A, eukaryotic translation initiation factor 2) or low MW. Since mass-spectrometry reveals tryptic peptides it is not unthinkable that partially degraded and/or ubiquitinilated proteins may be detected, which are simply reticulocyte legacy proteins. These “left over proteins”, a legacy of RBC development are generally expected to be inactive, but some may have yet unknown roles in mature RBCs.
Ribosomal proteins S19 (MM), S27 (HM) and S27a (RS) and 40S ribosomal protein S3 (MS/RS) and S6 (HS/RS) may provide one example. rpS3 and rpS6 have been found to interact with heat shock protein 90 (Hsp-90) thereby preventing ubiquitination and proteasome-dependent degradation85. A further regulatory mechanism involves Hsp-70, which associates with free rpS3 promoting its degradation85. Both chaperones occur in mature red blood cells86 and were found in our three mammalian proteome studies (Figure 5).
These findings suggest a common maturation schedule in terms of protein degradation for the RBCs of different mammals even though their half time differs sometimes quite significantly (human: 120 days, macaque: ~ 80 days and mouse: ~ 30 days).
Trypsin-digested samples were analyzed by capillary liquid chromatography coupled on-line with tandem mass-spectrometry (LC-MS/MS) using an Agilent 1100 series system and a Q-STAR (Q-STAR pulsar from PE Sciex, Canada, 17 runs) or LTQ-FT (Hybrid-2D-Linear Quadrupole Ion Trap - Fourier Transform Ion Cyclotron Resonance (FTICR) Mass Spectrometer, Thermo Electron, Germany, 3 runs). Samples from 3 μg protein were separated by reverse-phase chromatography, (3 mm Reposil C18, 75mm x 12 cm column) using a gradient from 98 % MS Buffer A and 2 % MS Buffer B solution at 0.5 mL/min flow rate. MS Buffer A was 0.5 % glacial acidic acid v/v; MS Buffer B, was 80% acetonitrile, 0.5 % glacial acidic acid v/v. At 24 min the flow was decreased to 0.25 mL/min and the amount of buffer B was increased to 7% (27 min), 13% (35 min), 33% (95 min), 50% (112 min), 60% (117 min) and finally to 80% (123 min). Eluted peptides were ionized to charge states 1+, 2+ or higher by the electrospray source and peptides that were at least doubly charged were analyzed in data-dependent MS experiments with dynamic exclusion. Two FT methods were used, the 3 most intense ions method43 and the 5 mass ranges method, the methods differing in the mode in which the ion trap is filled. For the Q-star, raw spectra were converted to a detected peak list of 1+, 2+ and 3+ ions, providing an exclusion list that was iteratively applied, thereby progressively excluding the most abundant ions from sequencing and enabling progressively less abundant peptides to be sequenced.
This work was supported by the ZonNW-NGI HORIZON doorbraakproject (dossiernummer 93519014), the NWO-CLS malaria (dossiernummer 635100026) and the European Network of Excellence EviMalaR (grant number 242095).