Absolute proteome quantification of highly purified populations of circulating reticulocytes and mature erythrocytes

Abstract

Reticulocytes produced in the bone marrow undergo maturation in the bloodstream to give rise to erythrocytes. Although the proteome of circulating red cells has been the subject of several reports, the cellular populations used for these studies were never completely devoid of reticulocytes. In our current study, we used highly purified erythrocyte and reticulocyte populations to quantify the absolute expression levels of the proteins in each cell population. Erythrocytes and reticulocytes were purified in a multistep process involving cellulose chromatography, Percoll gradient centrifugation, and fluorescence cell sorting after thiazole orange labeling. Proteins were analyzed by mass spectrometry from whole cells and erythrocyte plasma membrane (ghosts), leading to the identification and quantification of 2077 proteins, including 654 that were reticulocyte-specific. Absolute quantifications of these proteins were made using the mean corpuscular hemoglobin content of the cells as a standard. For each protein, we calculated the percentage loss during the terminal stages of reticulocyte maturation and the percentage of association with the plasma membrane. In addition, we used modified adenosine triphosphate and adenosine diphosphate molecules that enable the transfer of a biotin molecule to the catalytic sites of kinases to isolate active kinases in the erythrocytes and determined the absolute expression of 75 protein kinases and the modification of their expression during reticulocyte maturation. Our findings represent the first absolute quantification of proteins that are specifically expressed in normal erythrocytes with no detectable contamination by reticulocytes. Our findings thus represent a reference database for the future proteomic analysis of pathological erythrocytes.

Graphical abstract

Figure 1.

Peripheral blood cell purification flowchart. Reticulocytes and erythrocytes were purified from the peripheral blood of healthy donors. Blood was centrifuged for 10 minutes at 150g to remove plasma and a portion of the leukocytes and platelets. During this step, crude cell populations were sampled for MS analysis. The remaining leukocytes and platelets were removed by passing the cells through a cellulose column packed in 2-mL plastic syringes. Cells not retained in the column were separated into 2 fractions that were centrifuged through Percoll layers of different concentrations to obtain fractions depleted (P3) or enriched (P4) in reticulocytes. A cell sample from P3 (EEP) was used for MS analysis. In the experiment reported, the EEP cell population contained 0.62% of reticulocytes. Cells were labeled with thiazole orange (TO), and erythrocytes (EPP; TO⁻ cells) and reticulocytes (Retic; TO⁺) were purified by FACS from P3 and P4 fractions, respectively, using a FACSJazz cell sorter (BD Biosciences). Between 750 000 and 2 million purified reticulocytes and between 6 and 15 million erythrocytes were obtained from each purification procedure and used for subsequent analyses. The purity of isolated cell populations was controlled by cytocentrifuged cells stained with New Methylene Blue (reticulocyte stain; Sigma-Aldrich).

Figure 2.

Overall analysis of the proteomic data for whole-cell preparations. (A) Contribution of globin MS signal to the MS signal of all identified proteins. Identified peptides were ranked according to their intensity and the contribution of globin peptides to the sum of the intensities of all identified peptides was calculated at each point. (B) Number of proteins quantified in the EEP, EPP, and Retic populations. (C) Number of proteins quantified in single-cell, 2-cell, or 3-cell populations. (D) Number of proteins only quantified in the EEP, Retic, or EPP cell population. (E) Dynamic range of protein quantification in the EEP, EPP, and Retic cell populations. Proteins were ranked according to their expression level. (F-H) Reproducibility of the quantifications in the different experiments. The numbers in the bottom panels indicate Pearson correlation coefficients. (I) Relationship between protein expression levels in the EEP, EPP, and Retic cell populations. Numbers are Pearson correlation coefficients. (J) The absolute quantification is not affected by dilution of the analyzed peptides. SCX fractions were analyzed directly and after a twofold dilution.

Figure 3.

Absolute protein quantification in whole-cell and in ghost preparations. (A) Comparison of the quantification values for transmembrane proteins. Proteins with transmembrane domains were selected using the keyword annotations in Uniprot. Proteins quantified in both whole-cell and ghost preparations were selected and the quantification values were compared. (B) Comparison of the quantification values for all proteins quantified in the whole-cell preparations. (C) Comparison of the quantification values obtained in the present study with those reported previously by Bryk and Wiśniewski. (D) Comparison of the present study quantification values with those obtained by biochemical methods and compiled by Burton and Bruce except for HBA1, which was calculated starting from a hemoglobin content of 30 pg per erythrocyte and CA2 that was obtained from Sterling et al.

Figure 4.

Protein kinases and phosphatases of the red cell proteome. Protein kinases (A) or phosphatases (B) quantified from whole-cell and/or ghost preparations were ranked according to their expression levels in erythrocytes. Absolute quantifications in erythrocytes (EPP) and in reticulocytes (Retic) are indicated (right panel). Left panels, The expression ratio of these proteins in erythrocytes vs reticulocytes. (C) Protein kinases identified by the kinome-targeted approach using modified ATP or ADP probes. The pie chart is in proportion to the LFQ value of each identified kinase.

Figure 5.

Modifications of the red cell proteome during reticulocyte maturation. (A) Comparison of the expression level of each quantified protein in reticulocytes and erythrocytes. (B-D) Proteins whose expression decreased by at least 90% during reticulocyte maturation, including proteins that were no longer detected in erythrocytes, were compared with the proteins quantified in erythrocytes that did not present such a decrease during reticulocyte maturation. Term enrichment analysis in these 2 sets of proteins was done with Perseus software using annotations from the Kyoto Encyclopedia of Genes and Genomes (KEGG), from the Gene Ontology Biological Process (GOBP), or from the Gene Ontology Molecular Functions (GOMF) databases. (E) Modification of the expression level of glycolysis pathway proteins during reticulocyte maturation. Expression level (as copy number per cell) of each protein in the reticulocytes (open circles) and in the erythrocytes (closed circles) is indicated (right panel), the expression ratio in erythrocyte vs reticulocyte is indicated (left panel).

Figure 6.

Expression of specific proteins on the red cell membrane during erythrocyte maturation. Proteins known to play a specific role in erythrocyte membrane organization and/or function were selected. Their expression levels in reticulocytes (open circles) and in erythrocytes (filled circles) are indicated.