Abstract
Red blood cells (RBC) have the special challenge of a large amount of reactive oxygen species (from their substantial iron load and Fenton reactions) combined with the inability to synthesize new gene products. Considerable progress has been made in elucidating the multiple pathways by which RBC neutralize reactive oxygen species via NADPH driven redox reactions. However, far less is known about how RBC repair the inevitable damage that does occur when reactive oxygen species break through anti-oxidant defenses. When structural and functional proteins become oxidized, the only remedy available to RBC is direct repair of the damaged molecules, as RBC cannot synthesize new proteins. Amongst the most common amino acid targets of oxidative damage is the conversion of asparagine and aspartate side chains into a succinimidyl group through deamidation or dehydration, respectively. RBC express an L-isoaspartyl methyltransferase (PIMT, gene name PCMT1) that can convert succinimidyl groups back to an aspartate. Herein, we report that deletion of PCMT1 significantly alters RBC metabolism in a healthy state, but does not impair the circulatory lifespan of RBC. Through a combination of genetic ablation, bone marrow transplantation and oxidant stimulation with phenylhydrazine in vivo or blood storage ex vivo, we use omics approaches to show that, when animals are exposed to oxidative stress, RBC from PCMT1 knockout undergo significant metabolic reprogramming and increased hemolysis. This is the first report of an essential role of PCMT1 for normal RBC circulation during oxidative stress.
Introduction
Dysfunction of red blood cells (RBC) is a component of numerous diseases and oxidant stress is a key component to both normal RBC aging and also pathological dysfunction.1 In addition, approximately one of every 70 Americans is transfused with packed RBC each year.2 RBC are stored for up to 42 days prior to transfusion out of logistical necessity, resulting in oxidative damage that affects their ability to function upon transfusion,3 thus impacting a wide variety of diseases for which transfusion is essential, including trauma, issues of chronic hemostasis, cancer, and chronic bone marrow disorders.2 Thus, damaged or dysfunctional RBC are a widespread factor in human disease and therapy.4
RBC face particular metabolic challenges regarding generation and mitigation of oxidative damage. First, RBC have a unique source of oxidative stress because of the high load of iron associated with hemoglobin (RBC account for appriximately 66% of bodily iron),4 which drives radical-generating Fenton reactions. Second, unlike most other tissues, repair of oxidant stress in RBC cannot involve new protein synthesis owing to the lack of nuclei or ribosomes. As such, elucidating and understanding the nonsynthetic pathways by which RBC manage oxidative stress can provide unique clues on cellular responses to oxidant injury. Although reactive oxygen species (ROS) are highly reactive, they preferentially attack particular chemical groups. When damaging proteins, the side chains of asparagine and aspartate are particularly susceptible to oxidative conversion into a succinimidyl group through deamidation or dehydration, respectively.5 The succinimidyl group is unstable, and decays approximately 70-85% of the time into an L-isoaspartyl residue, impacting the protein structure (backbone orientation) and function. Clarke, Ingrosso and colleagues identified L-isoaspartyl groups on RBC membrane proteins in response to aging or pathological oxidative stress, such as in glucose 6-phosphate dehydrogenase deficiency or Down syndrome.6–12 Thus, accumulation of L-isoaspartyl groups in essential RBC proteins correlates with RBC dysfunction and pathology.
A specific repair pathway exists for asparagines and aspartates that have been oxidized into L-isoaspartyl groups. The enzyme L-isoaspartyl methyltransferase (PIMT, gene name PCMT1)13 methylates L-isoaspartyl groups to form an isoaspartyl methyl ester, which can then spontaneously decompose into a normal aspartyl group (returning aspartate to its original stucture and converting asparagine into aspartate). However, the isoaspartyl methyl ester can also decompose back into the “damaged” L-isoaspartyl group, which can again be methylated by PIMT. Due to the stochastic nature of isoaspartyl methyl ester decomposition (as well as ongoing oxidation of aspartates and asparagines by ROS), multiple and ongoing cycles of PIMT-dependent methylation are required – a process that is constrained both by sufficient PIMT activity and a sufficient pool of methyl donors as co-factors (e.g., of S-Adenosyl-methionine [SAM]).
Recently, we have observed that oxidant stressors from either environmental insults or genetic disease results in the extensive methylation of at least 116 RBC proteins in proximity to their active sites.14 Moreover, tracing experiments with 13C15N-methionine (substrate for the synthesis of SAM – the methyl group donor in PIMT-dependent reactions) revealed that the formation of isoaspartyl 13Cmethyl- esters was increased in response to oxidative insults. Thus, a model has emerged in which oxidative damage induces formation of L-isoaspartyl groups, which are then methylated by PIMT, resulting in repair of oxidized protiens without the need for resynthesis. However, each of the observations that support his model are correlative. The goal of this paper is to test the causal role of the PIMT pathway in RBC oxidant-damage repair through genetic ablation of PCMT1.
Methods
Animal studies with mice
All the animal studies described in this manuscript were reviewed and approved by either the BloodworksNW Research Institute IACUC or the University of Virginia Institutional Animal Care and Use Committee. PCMT1+/- founders15 were acquired from the National Institutes of Health mouse embryo repository and were bred with C57BL/6 females. Interbreeding of PCMT1+/- (henceforth PCMT1 heterozygous) mice was designed to generate PCMT1-/- (PCMT1 KO), as confirmed by genotyping. The use of Ubi-GFP and HOD mice have been previously described in prior work from our group.16 Whole blood was drawn by cardiac puncture as a terminal procedure for PCMT1 KO or WT mice, followed by harvesting of organs (10 mg of tissues from brain, heart, kidney, liver and spleen). Tissues and blood were snap frozen in liquid nitrogen and stored at -80ºC until subsequent analysis. For transfusion studies, fresh RBC (never frozen) were used.
Diamide treatments
RBC from WT and PCMT1 KO mice were incubated with diamide (0.5 mM, Sigma Aldrich) at 37°C for 0, 3 and 6 hours (h), as previously described,17 prior to metabolomics analyses.
Bone marrow transplantation and phenylhydrazine treatment
BMT was performed as previously described, but with WT or PCMT1 KO donors and green fluorescence protein (GFP) recipients. 18 Engraftment was monitored by the appearance of GFP-negative RBC and the dissappearance of GFP-positive RBC. On average, GFP-positive RBC were undetectable after approximatley 56 days, consistent with the known RBC lifespan of mice.
Prior to treatment with phenylhydrazine (PHZ) [Sigma Aldrich, USA], mice were injected daily for 3 consecutive days with biotin- XSE (ThermoFisher, Waltham, MA, USA Cat# B1582) until 100% biotinylation of RBC was achieved. Each injection consisted of 1 mg biotin-XSE in a 8% solution of dimethyl sulfoxide (DMSO) in phosphate buffered saline. Mice were then given two intraperitoneal injections (6 h apart) of PHZ to achieve a final dose of 0.01 mg/g body weight. Perpheral blood was then harvested longitudinally and RBC stained with avidin-APC to allow enumeration of RBC that had been present at time of PHZ treatment and to distinguish them from RBC generated by hematopoeisis after PHZ injection.
Storage under blood bank conditions
RBC were collected, processed, stored, transfused and posttransfusion recovery was determined as previously described.19
Tracing experiments with labeled glucose and methionine
RBC from WT and PCMT1 KO mice (100 mL) were incubated at 37ºC for 1 h in the presence of 1,2,3-13C3-glucose or 13C-methionine, prior to determination of lactate isotopologues +2/+3 (as markers of pentose phosphate pathway to glycolysis fluxes) and 13C-SAM (as marker of methyltransferase activity14), as previously described.20
Sample processing and metabolite extraction
A volume of 50 mL of frozen RBC aliquots was extracted in 450 mL of methanol:acetonitrile:water (5:3:2, v/v/v). After vortexing at 4°C for 30 minutes (min), extracts were separated from the protein pellet by centrifugation for 10 min at 10,000 RPM at 4°C and stored at −80°C until analysis.
Ultra-high-pressure liquid chromatography-mass spectrometry metabolomics
Analyses were performed using a Vanquish ultra-high-pressure liquid chromatography-mass spectrometry (UHPLC) coupled online to a Q Exactive mass spectrometer (MS) (Thermo Fisher, Bremen, Germany). Samples were analyzed using a 3 min isocratic condition or a 5, 9 and 17 min gradient as described.21 Solvents were supplemented with 0.1% formic acid for positive mode runs and 1 mM ammonium acetate for negative mode runs. MS acquisition, data analysis and elaboration was performed as previously described.21–23
Proteomics
Proteomics analyses were performed via filter aided sample preparation (FASP) digestion and nano UHPLC-MS/MS identification (nanoEasy LC 1000 coupled to a QExactive HF, Thermo Fisher), as previously described.24
Statistical analyses
Graphs and statistical analyses (either t-test or repeated measures ANOVA) were prepared with GraphPad Prism 5.0 (GraphPad Software, Inc, La Jolla, CA) and MetaboAnalyst 4.0, which allows to calculate false discovery rate (FDR)-corrected P-values from ttest and ANOVA analyses.25 Heat maps include color are graphed as Z-score normalized values for each metabolite/protein (ranges included as legends within each panel containing a heat map); each single box in the heat maps indicates a distinct biological replicate per each group. All the raw data from the omics analyses reported in this study are included in the Online Supplementary Table S1.
Results
Lack of a single copy of PCMT1 results in significant alterations in the metabolome and proteome of red blood cells compared to wild type red blood cells
PCMT1+/- RBC were compared to WT controls (Figure 1A). As predicted, phenotypes between the two mouse strains differed significantly both at the protein and metabolite level (Figure 1B and C report the top 50 significant metabolites and proteins by t-test, respectively). Unsupervised analyses revealed a significant impact of PCMT1 heterozygosity on the omics phenotypes of RBC, with approximately 33% of the total variance between the two groups explained by a single principal component (PC1 in Figure 1D). Metabolites with the highest loading weights are highlighted in Figure 1E, showing a significant impact on pathways relevant to redox/NADPH homeostasis (glutathione, ribose phosphate, biliverdin, formyltetrahydrofolate, cysteinyl-glycine, gamma-glutamyl-cysteine) and amino acids (glutamate, tyrosine, glycine) in PCMT1+/- RBC. Of note, lower levels of enzymes involved in these pathways were detected in PCMT1+/- RBC, including 6-phosphogluconolactonase (6PGL), transaldolase (TALDO), glutathione peroxidase 1 (GPX1) (Figure 1C). Conversely, the levels of other enzymes involved in alternative redox regulation pathways were increased in PCMT1+/- RBC, including peroxiredoxin 2 (PRDX2), thioredoxin reductase 2 (TRXR2), hydroxyacyl glutathione hydrolase (GLO2), glucose 6-phosphate isomerase (G6PI), lactate dehydrogenase B (LDHB) (Figure 1C). Accordingly, we focused on these pathways to provide a more detailed overview of the metabolic changes observed in RBC from these mice (Figure 1F). Specifically, we noted an accumulation of methionine (the main methyl group donor precursor - Figure 1F), consistent with a decreased PIMT levels and activity. Likewise, differential D-methylation patterns were observed between WT and PCMT1+/- RBC, especially in regions in which aspartyl groups are flanked by other negatively charged residues (D, E, S) in proximity of deamidation susceptible asparaginyl groups (N – representative methylation of deamidated N is shown for the most abundant protein, hemoglobin - Figure 1G, the full list is stated in the Online Supplementary Table S1). Notably, significantly higher levels of band 3 methylation of residues D329 and 368 was observed in WT mice compared to heterozygous mice (Online Supplementary Table S1).
Lack of both copies of PCMT1 exacerbates the phenotypes observed in heterozygous mouse red blood cells
PCMT1 KO mice die at approximately 6-8 weeks from seizures due to accumulated protein damage in brain tissue.15,26 PCMT1 KO mice were obtained by crossing of PCMT1 heterozygous mice and were characterized at 5 weeks of age (prior to the onset of seizures) compared to age-matched WT mice; to ensure that the same background genetics were present in each group, PCMT1 KO, WT and heterozygous mice were all obtained from the same colony of interbreeding heterozygous mice (Figure 2A). Unsupervised principal component analysis (PCA) revealed a progressive alteration of the metabolic phenotype of RBC from WT to heterozygous and KO mice (Figure 2B), with the main metabolic discriminants across groups being pentose phosphate pathway (PPP) metabolites (6-phosphogluconate, sedoheptulose phosphate, ribose phosphate), polyunsaturated fatty acids (eicosa- and docosahexaenoic acid), oxylipins (15-HETE), tryptophan and tyrosine metabolites (including several indoles, picolinic acid, dopamine – Figure 2C). Changes in the metabolome and proteome (Figure 2D to E) extended the observations from the heterozygous mouse RBC, showing an apparent gene dose effect for many analytes. In particular, methionine consumption and utilization of S-adenosylmethionine were progressively impaired in the heterozygous and KO mice (Figure 2E), a finding that we confirmed with tracing of stable isotope labeled 13C-methionine (13C-SAM/13C-SAH levels increase almost 3-log fold in KO mice compared to WT - Figure 2G). However, in some cases, loss of a single copy of PCMT1 resulted in a change that was not further exacerbated by deletion of both PCMT1 copies (e.g., decreased GSH, G3P, and 5OXO as well as increased Glucose, DPG, PEP). These findings suggest a threshold effect for some metabolites in response to loss of a single copy of PCMT1. Overall, altered glutathione homeostasis (decreased levels of reducing equivalent pools) was accompanied by an apparent compensatory decrease in glycolysis (steady state levels of glycolytic intermediates in Figure 2E) and activation of the PPP (Figure 2E), which we confirmed with tracing of lactate isotopologues27 +2/+3 upon incubation with 1,2,3-13C3-glucose (Figure 2F).
Unexpectedly, genetic ablation of PCMT1 had a dose response effect on tryptophan and tyrosine metabolism (Online Supplementary Figure S1). Specifically, accumulation of tryptophan and the products of its breakdown (indole) and oxidation (picolinic acid and quinolinic acid, but not anthranilate) were observed in RBC of heterozygous and KO mice (Online Supplementary Figure S1). Similarly, accumulation of tyrosine and dopamine was observed in heterozygous and KO mice, in which the levels of taurine, hypotaurine and acetylcholine were decreased (Online Supplementary Figure S1).
Metabolic characterization of organs from wild-type and PCMT1 knockout mice
Since picolinic and quinolinic acid are neurotoxicants28 and PCMT1 KO mice are characterized by early mortality due to seizures, the metabolomes of different organs from WT and KO mice were characterized, namely brain, heart, kidney, liver and spleen (Figure 3A). PCA and heat maps (top 25 significant metabolites by t-test) for each organ are provided in alphabetical order in Figure 3B to F. Of note, differences in brain phenotypes were comparable to those observed in RBC, including decreases in polyunsaturated fatty acids, hypotaurine and increases in PPP metabolites and SAM in the KO mice (Figure 3B). However, PCMT1 KO mouse brains were characterized by lower tryptophan and indole. In contrast, all the other organs showed metabolic phenotypes clearly distinct from RBC and brains. In particular, hearts from KO mice were characterized by significant increases in the levels of several saturated fatty acids (7:0, 8:0, 10:0, 14:0, 16:0, 18:0) and decreased levels of purines (guanine, guanosine, hypoxanthine) and Krebs cycle metabolites (succinate – Figure 3C). Kidneys had similar trends with respect to purines (decreases in hypoxanthine, xanthine and urate), but an opposite phenotype with respect to several fatty acids and related oxidation products, which decreased in KO mice, whereas carboxylic acids increased (2-oxoglutarate, itaconate - Figure 3D). Comparably to kidneys, similar increases in several fatty acids were noted in livers from KO mice (myristoleic, palmitoleic, dodecanedioic and tetradecenoic acids), though PCMT1 KO livers were also characterized by significant decreases in the levels of several amino acids (leucine, lysine, cysteine, phenylalanine, threonine – Figure 3E). Spleens from KO mice had lower levels of several 18 carbon-chained fatty acids (similarly to brains and kidneys, opposite to RBC, heart and liver – Figure 3F). In summary, while different organs each had distinct metabolic phenotypes, all organs from KO mice were characterized by (i) lower levels of methionine utilization, (ii) decreased pools of glutathione and glutathione precursors, and (iii) compensatory activation of the PPP (as inferred from steady state levels of intermediates, especially in brain and spleen), suggestive of an altered capacity to cope with oxidant stress in KO mice.
Exacerbation of metabolic differences in red blood cells from wild-type and PCMT1 knockout mice by oxidant stress with diamide in vitro
While the above analysis showed a clear effect of PCMT1 on the metabolome, it was performed in a baseline state of oxidative stress. In order to test the hypothesis that PCMT1 KO RBC would be impaired in their response to an oxidative insult, RBC were incubated with diamide in vitro for up to 6 h (Figure 4A). While the overall metabolic impact of deleting PCMT1 was greater in the RBC metabolome than the effect of diamide treatment (Figure 4B), we identified a subset of metabolites that had a higher baseline level in WT RBC and which increased further following diamide treatment only in WT but not in KO mice (Figure 4C – blue box). Highlights of metabolites in this group are presented in Figure 4D including several tryptophan metabolites (kynurenine, nicotinamide, hydroxyindoleacetate), PPP and glutathione homeostasis metabolites (ribose phosphate, glutathionyl-cysteine, ergothioneine). The increase of these metabolites in WT but not PCMT1 KO mice demonstrate a defect in the ability of PCMT1 KO RBC to further activate the PPP following additional exogenous insults. Thus, in addition to impairing protein repair, lack of PCMT1 prevents compensatory metabolic shifts to environmental insults.
Increased in vivo hemolysis in response to systemic oxidant stress
In order to allow the study of PCMT1 KO RBC in adult mice in vivo, bone marrow transplant (BMT) was carried out with donors being PCMT1 KO or WT controls (5 weeks of age prior to seizure onset in PCMT1 KO) and recipients being Ubi-GFP recipient mice that express GFP in all blood lineages. This allowed monitoring of engraftment by evaluating the progressive decline in GFP positivity in circulating blood cells via flow cytometry (Figure 5A). No significant difference was seen in the rate of RBC engraftment from PCMT1 KO versus WT donors (approximately 56 days) (Figure 5B). Consistent with what was observed for RBC from 5-week-old PCMT1 KO mice compared to WT controls, the metabolic phenotypes of RBC from PCMT1 KO→WT mice showed significant accumulation of SAM compared to WT→WT RBC, as well as decreases in glutathione (total pool – including reduced and oxidized) and increases in dopamine and tryptophan metabolites (indole, indole acetate, 3-methyldioxyindol) (Figure 5C). Conversely, BMT PCMT1 KO→WT RBC showed higher levels of glycolytic intermediates (glucose 6-phosphate, glyceraldehyde 3-phosphate, 2,3-diphosphoglycerate and isobaric isomer, phosphoenolpyruvate) and lower steady state levels of PPP metabolites (6-phosphogluconate and ribose phosphate) or other pathways involved in NADPH homeostasis (folate, pyruvoyl-THF – Figure 5C). Together, these results suggested a lower antioxidant capacity of RBC from PCMT1 KO donors compared to WT donors; however, this difference did not affect the ability of hematopoietic stem cells to generate RBC or RBC to circulate in an otherwise healthy state.
In order to test the effects of oxidant stress that was increased from normal healthy conditions, mice were treated intravenously with a biotinylating reagent such that 100% of RBC were biotinylated. Mice were then injected with PHZ, which induces oxidant stress and is known to be toxic to RBC (Figure 6A). Clearance of RBC was then monitored by assessing the percentage of Biotin + RBC over time in peripheral blood. Significantly higher clearance of PCMT1 KO RBC was observed compared to WT RBC following two consecutive (6 h apart) injections of PHZ in vivo (Figure 6B), resulting in double the clearance of the PCMT1 KO RBC within the first 30 h from PHZ injection. Of note, the metabolic phenotypes of PCMT1 KO→WT RBC upon PHZ-induced stress showed significant decreases in the levels of glutathione pools (both reduced and oxidized) and accumulation of several short chain fatty acids (e.g., 5:0, 6:0, 8:0, 9:0-OH, 12:0) markers of fatty acid breakdown and oxidation in the mitochondria-devoid mature RBC (Figure 6C).
Increased oxidant stress but normal circulation in red blood cells from PCMT1 mice following blood storage
We hypothesized that RBC from PCMT1 KO mice would be more susceptible to storage-induced damage (Figure 7A). RBC were collected from WT and PCMT1 KO mice (at 5 weeks of age prior to seizures) and stored at 2-6 ºC for 12 days. Stored RBC were mixed with fresh tracer cells (HOD RBC that express a trackable transgene) in order to control for differences in injection or blood volume. The mixture was transfused into Ubi-GFP+ recipient mice. This approach allows us to determine the percentage of test RBC (GFP- HOD-) as a function of control RBC (GFP-HOD+) that are circulating at 24 h after transfusion, a parameter referred to as post-transfusion recovery (PTR) and one of the Food and Drug Administration-mandated gold standards for RBC storage quality in humans.29
Contrary to our prediction, both fresh and stored WT and PCMT1 KO RBC showed comparable PTR (Figure 7B). Despite the lack of a phenotype with respect to PTR, PCMT1 KO RBC were characterized by (i) a decrease in methionine consumption and accumulation of SAM, but not SAH (consistent with the lack of PCMT1 activity); (ii) significantly lower glutathione pools (especially GSSG) and increased oxidation of methionine thiols to sulfoxide; (iii) increased levels of purine and lipid oxidation markers (hypoxanthine; 15-HETE, 12,13-diHOME, 9-oxononanoic acid) (Figure 7C), previously identified as predictors of poor post-transfusion recoveries in mice and humans.16,30,31 On the other hand, stored PCMT1 KO mouse RBC showed significantly higher levels of PPP metabolites (gluconolactone phosphate), the CoA precursor pantotheine (indirectly involved in the Lands cycle for the recycling of oxidized lipids) and sphingosine 1-phosphate, at least at baseline (Figure 7C and D).
Discussion
Herein, we report that while PCMT1 is not required for normal erythropoiesis or RBC circulatory lifespan, upon exposure to oxidative stress, RBC lacking PCMT1 are more rapidly cleared from circulation. Metabolically, the absence of PIMT results in a severe depletion of the glutathione pools and a compensatory activation of the PPP to generate the reducing equivalents required for the recycling of oxidized thiols. This occurs in both RBC and also peripheral organs. Additional metabolic perturbations are also observed, including increased levels of tryptophan oxidation products (especially metabolites in the kynurenine pathway) and dopamine in both RBC and various organs, including the brain. Since metabolites in the kynurenine pathway can be neurotoxicant,28 it is interesting to note how our metabolomics analysis could provide a potential clue as to the mechanisms that drive early onset of seizures in mice lacking both copies of PCMT1. To our knowledge, this is the first report showing an essential role of PIMT in maintenance of RBC circulatory capability during oxidative stress.
The potential role of PIMT in RBC blood storage remains unclear. As with any negative finding, the lack of change in PTR for PCMT1 KO RBC does not formally rule out that PIMT may be involved in the biology of stored RBC, if one evokes the possibility of redundant pathways. However, given the metabolic changes seen in PCMT1 KO RBC, we do not believe redundant pathways are likely. More importantly, a number of the metabolites that are increased in PCMT1 KO RBC have shown a strong inverse correlation to PTR in both human and murine RBC, in particular increased hypoxanthine30 and oxidized lipid species.16 This correlation has held under multiple experimental conditions and contexts; as such, oxidized lipids have been posited to play a causal role in post-transfusion clearance of stored RBC. However, the current report instantiates a condition in which there is an increase in hypoxanthine and certain oxidized lipids, but normal PTR. Thus, this counts as importance evidence against the lipid oxidation hypothesis of RBC clearance after storage. On the other hand, it is worth noting that PCMT1 KO RBC have a compensatory activation of the PPP, which is relevant in the light of the recently reported decreases in PTR in blood donors with G6PD deficiency,29,32 the most common enzymopathy in humans and thus clinically relevant for ~400 million individuals worldwide. It remains to be assessed whether ablated PIMT activity may result in poor post-transfusion performances of the stored RBC in the context of G6PD deficiency or other genetic factors that have been shown to negatively impact post-transfusion recoveries in mice (e.g., STEAP316). In addition, the murine recipients here were all healthy, which is not consistent with the pro-oxidant environment the stored RBC faces upon transfusion in the critically ill or chronically hypoxic recipient (e.g., trauma or sickle cell patient, respectively).33,34
Several unanticipated alterations in metabolism were observed in PCMT1 KO mice, which serve as grounds for reasonable hypothesis driven speculation in the context of broader pathological consequences. Increased kynurenine pathway metabolites have been suggested to be responsible for neurotoxicity, as reported in Down syndrome.28 Prior reports had also suggested an increase in isoaspartyl damage and PIMT activation in Down syndrome.7 Dysregulation of methionine metabolism has also been reported in RBC from individuals with Down syndrome;35 this has been partially explained by the localization on chromosome 21 of genes coding for enzymes involved in homocysteine metabolism. Of note, alterations of RBC kynurenines and indoles (breakdown product of tryptophan metabolism) have been reported in RBC from aging mice and mouse models of parabiosis, raising the possibility of a role for PIMT-associated dysregulation of tryptophan metabolism in the aging process. On the other hand, dopamine – whose metabolism is dependent on NADPH – has been shown to accumulate in RBC from individuals with G6PD deficiency,32 as well as in response to exerciseinduced oxidant stress.36 These hypotheses will need to be rigorously assessed by subsequent studies, but the pattern that emerges from the current findings are consistent with a possible role of PIMT in these processes.
An important implication of the findings in this report is that there are widespread metabolic effects of deleting PCMT1, far beyond the predicted decrease in methionine consumption as a PIMT substrate. Rather, general shifts were seen in PPP activity, both at baseline as well as in response to oxidative stress. Moreover, as above, alterations in other metabolic pathways were likewise observed. This raises the important question of how PIMT contributes to widespread metabolic effects, since the known activity of PIMT is with regards to repairing oxidative damage once it has occurred and not altering the metabolism of oxidative pathways. The most likely hypothesis is that aspartate and asparagine residues are critically important in a wide range of metabolic enzymes and failure of PIMT to repair them when they are oxidized, results in alterations of these pathways through alteration of enzymatic activity or even inactivation. As isoaspartyl damage would impact protein backbone orientation, while methylation of isoaspartyl groups would affect the charge of side chains, one could speculate that PIMT activity (or lack thereof) could play a role in enzymatic function37 and RBC structural/function homeostasis through regulation of protein-protein interactions, in like fashion to phosphorylation. 38 A related, but distinct hypothesis is that lack of PCMT1 results in functional alteration of proteins involved in gene expression (e.g., transcription factors and/or enzymes involved in post-translational modification) as a result of unrepaired oxidation of aspartate and/or asparagine, which then changes the baseline genetic expression of enzymes involved in metabolic pathways. However, this explanation would only apply to nucleated cells, and not mature RBC. Given that the same general changes were seen in both RBC and other organs, this hypothesis is not favored as a general mechanism. A third and even less likely hypothesis, but one we cannot formally rule out based upon existing data, is that PIMT has additional enzymatic activities that have not yet been identified. The exact mechanism(s) by which PIMT activity broadly affects metabolism will require further investigation to elucidate; however, the findings in this manuscript demonstrate that PCMT1 is required for a number of distinct metabolic pathways and is essential for RBC to survive oxidative stress.
Footnotes
- Received July 13, 2020
- Accepted September 2, 2020
Correspondence
Disclosures
The authors declare that AD and KCH are founders of Omix Technologies Inc and AD of Altis Biosciencens LLC; AD and JCZ are a consultant for Rubius Therapeutics; AD is an advisory board member for Hemanext Inc and FORMA Therapeutics Inc. All the other authors disclose no conflicts of interest relevant to this study.
Contributions
AD and JCZ designed the study; AH, JCZ performed mouse studies. AD, BB, EJM performed metabolomics analyses; MD, KCH performed proteomics analyses; AD and JCZ analyzed data, prepared figures and wrote the first draft of the manuscript; AD and JCZ revised the manuscript. All the authors contributed to the finalization of the manuscript.
References
- Kaestner L, Minetti G. The potential of erythrocytes as cellular aging models. Cell Death Differ. 2017; 24(9):1475-1477. https://doi.org/10.1038/cdd.2017.100PubMedPubMed CentralGoogle Scholar
- Whitaker B, Rajbhandary S, Kleinman S, Harris A, Kamani N. Trends in United States blood collection and transfusion: results from the 2013 AABB Blood Collection, Utilization, and Patient Blood Management Survey. Transfusion. 2016; 56(9):2173-2183. https://doi.org/10.1111/trf.13676PubMedGoogle Scholar
- D’Alessandro A, Kriebardis AG, Rinalducci S. An update on red blood cell storage lesions, as gleaned through biochemistry and omics technologies. Transfusion. 2015; 55(1):205-219. https://doi.org/10.1111/trf.12804PubMedGoogle Scholar
- Nemkov T, Reisz JA, Xia Y, Zimring JC, D’Alessandro A. Red blood cells as an organ? How deep omics characterization of the most abundant cell in the human body highlights other systemic metabolic functions beyond oxygen transport. Expert Rev Proteomics. 2018; 15(11):855-864. https://doi.org/10.1080/14789450.2018.1531710PubMedGoogle Scholar
- Yang H, Zubarev RA. Mass spectrometric analysis of asparagine deamidation and aspartate isomerization in polypeptides. Electrophoresis. 2010; 31(11):1764-1772. https://doi.org/10.1002/elps.201000027PubMedPubMed CentralGoogle Scholar
- Galletti P, Manna C, Ingrosso D, Iardino P, Zappia V. Hypotheses on the physiological role of enzymatic protein methyl esterification using human erythrocytes as a model system. Adv Exp Med Biol. 1991; 307:149-160. https://doi.org/10.1007/978-1-4684-5985-2_14PubMedGoogle Scholar
- Galletti P, De Bonis ML, Sorrentino A. Accumulation of altered aspartyl residues in erythrocyte proteins from patients with Down’s syndrome. FEBS J. 2007; 274(20):5263-5277. https://doi.org/10.1111/j.1742-4658.2007.06048.xPubMedGoogle Scholar
- Ingrosso D, D’angelo S, di Carlo E. Increased methyl esterification of altered aspartyl residues in erythrocyte membrane proteins in response to oxidative stress. Eur J Biochem. 2000; 267(14):4397-4405. https://doi.org/10.1046/j.1432-1327.2000.01485.xPubMedGoogle Scholar
- Ingrosso D, Cimmino A, D’Angelo S. Protein methylation as a marker of aspartate damage in glucose-6-phosphate dehydrogenase- deficient erythrocytes: role of oxidative stress. Eur J Biochem. 2002; 269(8):2032-2039. https://doi.org/10.1046/j.1432-1033.2002.02838.xPubMedGoogle Scholar
- Janson CA, Clarke S. Identification of aspartic acid as a site of methylation in human erythrocyte membrane proteins. J Biol Chem. 1980; 255(24):11640-11643. https://doi.org/10.1016/S0021-9258(19)70177-XGoogle Scholar
- McFadden PN, Clarke S. Methylation at Daspartyl residues in erythrocytes: possible step in the repair of aged membrane proteins. Proc Natl Acad Sci U S A. 1982; 79(8):2460-2464. https://doi.org/10.1073/pnas.79.8.2460PubMedPubMed CentralGoogle Scholar
- O’Connor CM, Clarke S. Methylation of erythrocyte membrane proteins at extracellular and intracellular D-aspartyl sites in vitro. Saturation of intracellular sites in vivo. J Biol Chem. 1983; 258(13):8485-8492. https://doi.org/10.1016/S0021-9258(20)82090-0Google Scholar
- Desrosiers RR, Fanélus I. Damaged proteins bearing L-isoaspartyl residues and aging: a dynamic equilibrium between generation of isomerized forms and repair by PIMT. Curr Aging Sci. 2011; 4(1):8-18. https://doi.org/10.2174/1874609811104010008Google Scholar
- Reisz JA, Nemkov T, Dzieciatkowska M. Methylation of protein aspartates and deamidated asparagines as a function of blood bank storage and oxidative stress in human red blood cells. Transfusion. 2018; 58(12):2978-2991. https://doi.org/10.1111/trf.14936PubMedPubMed CentralGoogle Scholar
- Kim E, Lowenson JD, MacLaren DC, Clarke S, Young SG. Deficiency of a protein-repair enzyme results in the accumulation of altered proteins, retardation of growth, and fatal seizures in mice. Proc Natl Acad Sci U S A. 1997; 94(12):6132-6137. https://doi.org/10.1073/pnas.94.12.6132PubMedPubMed CentralGoogle Scholar
- Howie HL, Hay AM, de Wolski K. Differences in Steap3 expression are a mechanism of genetic variation of RBC storage and oxidative damage in mice. Blood Adv. 2019; 3(15):2272-2285. https://doi.org/10.1182/bloodadvances.2019000605PubMedPubMed CentralGoogle Scholar
- Kanias T, Sinchar D, Osei-Hwedieh D. Testosterone-dependent sex differences in red blood cell hemolysis in storage, stress, and disease. Transfusion. 2016; 56(10):2571-2583. https://doi.org/10.1111/trf.13745PubMedPubMed CentralGoogle Scholar
- Desmarets M, Cadwell CM, Peterson KR, Neades R, Zimring JC. Minor histocompatibility antigens on transfused leukoreduced units of red blood cells induce bone marrow transplant rejection in a mouse model. Blood. 2009; 114(11):2315-2322. https://doi.org/10.1182/blood-2009-04-214387PubMedPubMed CentralGoogle Scholar
- Zimring JC, Smith N, Stowell SR. Strain-specific red blood cell storage, metabolism, and eicosanoid generation in a mouse model. Transfusion. 2014; 54(1):137-148. https://doi.org/10.1111/trf.12264PubMedPubMed CentralGoogle Scholar
- D’Alessandro A, Nemkov T, Yoshida T. Citrate metabolism in red blood cells stored in additive solution-3. Transfusion. 2017; 57(2):325-336. https://doi.org/10.1111/trf.13892PubMedGoogle Scholar
- Nemkov T, Reisz JA, Gehrke S, Hansen KC, D’Alessandro A. High-throughput metabolomics: isocratic and gradient mass spectrometry-based methods. Methods Mol Biol. 2019; 1978:13-26. https://doi.org/10.1007/978-1-4939-9236-2_2PubMedGoogle Scholar
- Reisz JA, Zheng C, D’Alessandro A, Nemkov T. Untargeted and semi-targeted lipid analysis of biological samples using mass spectrometry-based metabolomics. Methods Mol Biol. 2019; 1978:121-135. https://doi.org/10.1007/978-1-4939-9236-2_8PubMedGoogle Scholar
- Stefanoni D, Shin HKH, Baek JH. Red blood cell metabolism in Rhesus macaques and humans: comparative biology of blood storage. Haematologica. 2020; 105(8):2174-2186. https://doi.org/10.3324/haematol.2019.229930PubMedPubMed CentralGoogle Scholar
- D’Alessandro A, Dzieciatkowska M, Nemkov T, Hansen KC. Red blood cell proteomics update: is there more to discover?. Blood Transfus. 2017; 15(2):182-187. Google Scholar
- Chong J, Wishart DS, Xia J. Using MetaboAnalyst 4.0 for comprehensive and integrative metabolomics data analysis. Curr Protoc Bioinformatics. 2019; 68(1):e86. https://doi.org/10.1002/cpbi.86PubMedGoogle Scholar
- Yamamoto A, Takagi H, Kitamura D. Deficiency in protein L-isoaspartyl methyltransferase results in a fatal progressive epilepsy. J Neurosci. 1998; 18(6):2063-2074. https://doi.org/10.1523/JNEUROSCI.18-06-02063.1998PubMedPubMed CentralGoogle Scholar
- Reisz JA, Wither MJ, Dzieciatkowska M. Oxidative modifications of glyceraldehyde 3-phosphate dehydrogenase regulate metabolic reprogramming of stored red blood cells. Blood. 2016; 128(12):e32-42. https://doi.org/10.1182/blood-2016-05-714816PubMedGoogle Scholar
- Powers RK, Culp-Hill R, Ludwig MP. Trisomy 21 activates the kynurenine pathway via increased dosage of interferon receptors. Nat Commun. 2019; 10(1):4766. https://doi.org/10.1038/s41467-019-12739-9PubMedPubMed CentralGoogle Scholar
- Francis RO, D’Alessandro A, Eisenberger A. Donor glucose-6-phosphate dehydrogenase deficiency decreases blood quality for transfusion. J Clin Invest. 2020; 130(5):2270-2285. https://doi.org/10.1172/JCI133530PubMedPubMed CentralGoogle Scholar
- Nemkov T, Sun K, Reisz JA. Hypoxia modulates the purine salvage pathway and decreases red blood cell and supernatant levels of hypoxanthine during refrigerated storage. Haematologica. 2018; 103(2):361-372. https://doi.org/10.3324/haematol.2017.178608PubMedPubMed CentralGoogle Scholar
- D’Alessandro A, Yoshida T, Nestheide S. Hypoxic storage of red blood cells improves metabolism and post-transfusion recovery. Transfusion. 2020; 60(4):786-798. https://doi.org/10.1111/trf.15730PubMedPubMed CentralGoogle Scholar
- D’Alessandro A, Fu X, Kanias T. Donor sex, age and ethnicity impact stored red blood cell antioxidant metabolism through mechanisms in part explained by glucose 6-phosphate dehydrogenase levels and activity. Haematologica. 2021; 106(5):1290-1302. https://doi.org/10.3324/haematol.2020.246603PubMedPubMed CentralGoogle Scholar
- Culp-Hill R, Srinivasan AJ, Gehrke S. Effects of red blood cell (RBC) transfusion on sickle cell disease recipient plasma and RBC metabolism. Transfusion. 2018; 58(12):2797-2806. https://doi.org/10.1111/trf.14931PubMedPubMed CentralGoogle Scholar
- Peltz ED, D’Alessandro A, Moore EE. Pathologic metabolism: an exploratory study of the plasma metabolome of critical injury. J Trauma Acute Care Surg. 2015; 78(4):742-751. https://doi.org/10.1097/TA.0000000000000589PubMedPubMed CentralGoogle Scholar
- Culp-Hill R, Zheng C, Reisz JA. Red blood cell metabolism in Down syndrome: hints on metabolic derangements in aging. Blood Adv. 2017; 1(27):2776-2780. https://doi.org/10.1182/bloodadvances.2017011957PubMedPubMed CentralGoogle Scholar
- San-Millán I, Stefanoni D, Martinez JL. Metabolomics of endurance capacity in World Tour professional cyclists. Front Physiol. 2020; 11:578. https://doi.org/10.3389/fphys.2020.00578PubMedPubMed CentralGoogle Scholar
- Pallotta V, D’Alessandro A, Rinalducci S, Zolla L. Native protein complexes in the cytoplasm of red blood cells. J Proteome Res. 2013; 12(7):3529-3546. https://doi.org/10.1021/pr400431bPubMedGoogle Scholar
- Longo V, Marrocco C, Zolla L, Rinalducci S. Label-free quantitation of phosphopeptide changes in erythrocyte membranes: towards molecular mechanisms underlying deformability alterations in stored red blood cells. Haematologica. 2014; 99(7):e122-e125. https://doi.org/10.3324/haematol.2013.103333PubMedPubMed CentralGoogle Scholar
Data Supplements
Figures & Tables
Article Information
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.