Global mass spectrometry based metabolomics profiling of erythrocytes infected with Plasmodium falciparum

Abstract

Malaria is a global infectious disease that threatens the lives of millions of people. Transcriptomics, proteomics and functional genomics studies, as well as sequencing of the Plasmodium falciparum and Homo sapiens genomes, have shed new light on this host-parasite relationship. Recent advances in accurate mass measurement mass spectrometry, sophisticated data analysis software, and availability of biological pathway databases, have converged to facilitate our global, untargeted biochemical profiling study of in vitro P. falciparum-infected (IRBC) and uninfected (NRBC) erythrocytes. In order to expand the number of detectable metabolites, several key analytical steps in our workflows were optimized. Untargeted and targeted data mining resulted in detection of over one thousand features or chemical entities. Untargeted features were annotated via matching to the METLIN metabolite database. For targeted data mining, we queried the data using a compound database derived from a metabolic reconstruction of the P. falciparum genome. In total, over one hundred and fifty differential annotated metabolites were observed. To corroborate the representation of known biochemical pathways from our data, an inferential pathway analysis strategy was used to map annotated metabolites onto the BioCyc pathway collection. This hypothesis-generating approach resulted in over-representation of many metabolites onto several IRBC pathways, most prominently glycolysis. In addition, components of the "branched" TCA cycle, partial urea cycle, and nucleotide, amino acid, chorismate, sphingolipid and fatty acid metabolism were found to be altered in IRBCs. Interestingly, we detected and confirmed elevated levels for cyclic ADP ribose and phosphoribosyl AMP in IRBCs, a novel observation. These metabolites may play a role in regulating the release of intracellular Ca(2+) during P. falciparum infection. Our results support a strategy of global metabolite profiling by untargeted data acquisition. Untargeted and targeted data mining workflows, when used together to perform pathway-inferred metabolomics, have the benefit of obviating MS/MS confirmation for every detected compound.

Figure 1. Features detected by global untargeted analysis and inferred from METLIN database searches having an absolute fold change |[IRBC/NRBC] Log2| ≥1.0.

A complete listing of all features detected by global untargeted analysis is presented in Table ST1. Compounds confirmed by MS/MS and/or RT matching are indicated with a “§” symbol.

Figure 2. A partial list of identified compounds resulting from “targeted” analysis, showing differential abundances between IRBC and NRBC.

The relative abundances are based on IRBC/NRBC Log₂ ratios for each entity that were significantly (P<0.05) different. Also indicated are the biological pathways associated with some of these compounds.

Figure 3. Pie Chart summarizing the distribution and super class (from HMDB) information for IRBC and NRBC metabolites, detected by “untargeted” and “targeted” data mining analyses.

Figure 4. Extracted Ion Chromatograms (EICs) for a derivitized standard mixture containing both (A) cADPR and (B) pRib-AMP, analyzed by LC/MS ESI (−) mode.

Figure 5. Extracted MS/MS spectra for m/z 558.0644, detected in ESI (−) mode at three collision energies: 10, 20, 40 eV.

The empirical formula corresponding to the neutral mass, C₁₅H₂₃N₅O₁₄P₂ corresponds to a compound with a single database entry in METLIN and annotated as N1-(5-phospho-D-ribosyl)-AMP (pRib-AMP). MS/MS spectra were collected for this compound from: (A) the chemically derivitized cADPR product, pRib-AMP, and, (B) an IRBC sample.

Figure 6. A proposed enzymatic pathway for generating pRib-AMP from cADPR in P. falciparum.

Figure 7. Untargeted, global LC/MS metabolomics analysis for P. falciparum infected RBC cultures reveal metabolites associated with the mitochondrial TCA cycle.

Differential and quantified metabolites are shaded in red.