Skip to main content
NCBI home page
The Journal of Infectious Diseases logoLink to The Journal of Infectious Diseases
. 2012 May 7;206(2):238–248. doi: 10.1093/infdis/jis339

Metabolomic Analysis of Patient Plasma Yields Evidence of Plant-Like α-Linolenic Acid Metabolism in Plasmodium falciparum

Viswanathan Lakshmanan 1,2, Kyu Y Rhee 6,7, Wei Wang 3, Yiting Yu 4, Kamil Khafizov 5, Andras Fiser 5, Peng Wu 3, Omar Ndir 8, Souleymane Mboup 8, Daouda Ndiaye 8, Johanna P Daily 1,2
PMCID: PMC3490690  PMID: 22566569

Abstract

Metabolomics offers a powerful means to investigate human malaria parasite biology and host-parasite interactions at the biochemical level, and to discover novel therapeutic targets and biomarkers of infection. Here, we used an approach based on liquid chromatography and mass spectrometry to perform an untargeted metabolomic analysis of metabolite extracts from Plasmodium falciparum–infected and uninfected patient plasma samples, and from an enriched population of in vitro cultured P. falciparum-infected and uninfected erythrocytes. Statistical modeling robustly segregated infected and uninfected samples based on metabolite species with significantly different abundances. Metabolites of the α-linolenic acid (ALA) pathway, known to exist in plants but not known to exist in P. falciparum until now, were enriched in infected plasma and erythrocyte samples. In vitro labeling with 13C-ALA showed evidence of plant-like ALA pathway intermediates in P. falciparum. Ortholog searches using ALA pathway enzyme sequences from 8 available plant genomes identified several genes in the P. falciparum genome that were predicted to potentially encode the corresponding enzymes in the hitherto unannotated P. falciparum pathway. These data suggest that our approach can be used to discover novel facets of host/malaria parasite biology in a high-throughput manner.

Plasmodium falciparum is responsible for nearly 800 000 malaria deaths worldwide annually [1]. Only a handful of antimalarials are currently clinically effective against P. falciparum [2]. Moreover, there are no available vaccines against any of the malaria species [3, 4].

Clinical infection is associated with a complex interplay between the human host and malaria parasites [5, 6]. The dynamic nature of host-parasite responses during infection may impact the biochemical landscape of both the parasite and host. Parasite metabolomics, in combination with genomics and proteomics, is therefore likely to be a rich untapped source of novel biological insights.

Metabolites perform several critical functions, including energy storage, signal transduction, and developmental regulation. However, because predicting the metabolite repertoire through genomic and proteomic analyses may not be straightforward, direct biochemical measurement of metabolites is required. Some recent reviews provide an overview of metabolomics in investigating malaria biology [7–9].

Plasmodium species metabolomic studies have begun to uncover important/novel aspects of parasite biology. Using proton nuclear magnetic resonance analysis, Teng et al identified >50 parasite-specific metabolites in saponin-released P. falciparum [10]. Basant et al also used proton nuclear magnetic resonance analysis to study in vivo metabolic alterations in urine, serum, and brain samples during disease progression in Plasmodium berghei–infected mice [11]. Metabolic alterations were correlated with sexual dimorphism and were suggested to affect disease prognosis and treatment [11].

Olszewski et al quantified the levels of approximately 200 known metabolites in P. falciparum during the 48-hour intraerythrocytic stage, using liquid chromatography with tandem mass spectrometry [12]. Important discoveries were the conversion of arginine to ornithine by parasite arginase, as well as the potential link of parasite-induced hypoargininemia with the development of cerebral malaria [12].

More recently, Olszewski et al used 13C-labeled precursors to show that the amino acids glutamate and glutamine serve as the major carbon sources for the tricarboxylic acid pathway in P. falciparum [13]. Metabolomics also led to the identification in P. falciparum of isoprenoid [14] and carotenoid [15] biosynthesis pathways and to the unexpected finding that these parasites can produce glycerol and glycerol phosphate [16].

Even though P. falciparum is an obligate intracellular parasite, patient plasma offers an important window for understanding host/parasite biology during infection. Moreover, metabolomic analysis of plasma samples provides a powerful tool for phenotypic biology and clinical biomarker discovery, when compared to in vitro studies of parasite biology.

To explore the perturbations in the human/parasite metabolome associated with P. falciparum infection, we performed an untargeted liquid chromatography–mass spectrometry (LC-MS) analysis of metabolites in plasma samples collected from P. falciparum–infected and uninfected patients. We also used in vitro cultured P. falciparum–infected and uninfected erythrocytes to characterize parasite-specific metabolites.

We identified several metabolites whose abundances were significantly different between infected and uninfected samples. Metabolites of the plant ALA pathway were enriched in infected plasma and erythrocyte samples. Labeling with 13C-ALA showed evidence of a plant-like pathway in P. falciparum. Ortholog searches based on annotated ALA pathway enzymes in 8 plant genomes revealed several genes that potentially encode the corresponding enzymes in the hitherto unannotated P. falciparum pathway. Reverse transcription–polymerase chain reaction (RT-PCR) showed transcription of these genes in P. falciparum.

This study demonstrates the power of an untargeted approach to defining the metabolite repertoire during infection. This approach could be used to enhance our understanding of host/parasite biochemistry during infection and foster efforts toward identifying novel therapeutic targets and biomarkers of infection.

METHODS

Ethics Statement

Institutional review board approvals were obtained from the Ethics Committee of the Ministry of Health in Senegal, from the Human Subject Committee of the Harvard School of Public Health, and, for secondary analysis, from the Albert Einstein College of Medicine Committee for Clinical Investigations. Written informed consent from all patients enrolled was obtained at the Section de Lutte Antiparasitaire de Thies, Senegal (Table 1).

Table 1.

Demographic, Clinical, and Laboratory Parameters of Plasmodium falciparum–Infected and -Uninfected Patients Whose Plasma Samples Were Used in This Study

Parameter Smear Positive (n = 10) Smear Negative (n = 9) Pa
Age (y) 21.5 (15.8–26.0) 21 (15.0–36.5) .74
Height (m) 1.6 (1.6–1.8) 1.7 (1.5–1.9) .95
Weight (kg) 67 (57–80) 53 (34–66) .06
Days ill (no.) 4 (2.5–8.75) 4 (3–7) .83
Time since last fever (d) 0.46 (0.4–0.5) 0.5 (0.1–1.3) .89
Temperature (°C) 39.3 (37.8–42.1) 38.7 (38–39.7) .39
Hematocrit (%) 40.5 (35.3–43.8) 40 (35–44) 1
Parasitemia (%) 0.5525 (0.38–1.1) Not applicable
Open in a new tab

Data are median values (interquartile range).

a Determined using the Mann-Whitney U test (GraphPad Prism).

Total Metabolite Extraction

LC-MS–grade water, acetonitrile, methanol, and acetic acid and molecular biology–grade acetone and chloroform were purchased from Fisher Scientific. Total metabolites (polar and nonpolar) were extracted as previously described [17].

RESULTS

Metabolomic Analysis Robustly Segregated P. falciparum–Infected and -Uninfected Samples

Metabolite profiles of polar and nonpolar fractions from plasma and erythrocyte samples were obtained by electrospray ionization–based LC-MS in both positive and negative modes. Metabolite profiles were filtered using Agilent Mass Profiler Professional (Agilent) to contain only metabolite species whose abundances were above the quantitative detection threshold of 1000 ion counts (indicative of metabolite abundance) and were statistically significant (P < .05 by the Mann-Whitney U test and correction for multiple testing by the Benjamini-Hochberg approach) between infected and uninfected samples. After eliminating metabolite species that could not be nominated any identity by annotation, drugs, and dipeptides and tripeptides, we obtained 149 and 75 putatively annotated metabolites for plasma and erythrocyte samples, respectively (Supplementary Table 1). Both data sets contained groups of covariant ions that were nominated isomeric identities, with the covariant ions in each group possessing the same ion counts. Each of these groups represented a unique metabolite feature.

Two-dimensional partial least-squares discriminant analysis (PLSDA) score plots were generated for plasma and erythrocyte metabolite data sets, both of which showed very clear segregation of the infected group from the uninfected group (Figure 1A and 1B). Cross-validation of the PLSDA plots, which assessed the validity and robustness of PLSDA, excluded the possibility that this segregation was attributable to disparities in sample and metabolite numbers [18]. Accurate classification of the validation samples provided cross-validation of the PLSDA models. For the plasma PLSDA model, we obtained R2Y and Q2Y scores of 93%. For the erythrocyte PLSDA model, we obtained R2Y and Q2Y scores of 85% and 88%, respectively.

Figure 1.

Figure 1.

Open in a new tab

Two-dimensional partial least-squares discriminant analysis plots of plasma and erythrocyte metabolite data sets and identification of traumatin in infected and uninfected plasma samples. Score plots were generated using putatively annotated metabolites that had significantly different abundances (P < .05 by Mann-Whitney U test) between Plasmodium falciparum–infected and uninfected samples showing segregation of infected from uninfected plasma samples (A) and infected from uninfected erythrocytes (B). Each circle represents one sample. Open and dark circles represent infected and uninfected samples, respectively. C, Negative-mode extracted ion chromatogram (EIC) peak of pure traumatin at 1, 5, and 10 μM, and of the metabolite nominated as traumatin in plasma samples. The EIC peak in all samples was detected at a retention time (RT) of approximately 11.2 minutes. The y-axis represents ion counts (abundance), and the x-axis represents the RT. D, Extracted mass spectrometry (MS) spectrum peak (at a mass of 211.134) for the metabolite nominated as traumatin in plasma samples before and after spiking with 2.5 μM of pure traumatin, as well as the extracted MS spectrum peak for pure traumatin at 2 μM. The y-axis represents ion counts (abundance), and the x-axis represents the m/z value.

The root mean square error of prediction (RMSEP), which is the average error in cross-validation scores when using the training models for future predictions, similarly validated the predictive accuracy of this model. A smaller RMSEP value of a model (closer to zero) indicates a better ability to perform accurate future predictions. The RMSEP values for plasma and erythrocyte models were 0.1418 and 0.1947, respectively.

Putatively annotated metabolites with significantly abundant differences belonged to several biochemical classes, including amino acids and products of amino acid metabolism, glycolytic and Krebs cycle intermediates, and fatty acids and lipids.

Some of the metabolites that had been detected in the proton nuclear magnetic resonance study by Teng et al [10] were also detected in our infected erythrocyte samples. These included glycerophosphocholine, glycerophosphoethanolamine, and hypoxanthine (Supplementary Table 1). Interestingly, we detected formaldehyde in both our plasma (approximately 1.4-fold higher levels in infected plasma) and erythrocyte (approximately 4.5-fold higher levels in infected erythrocytes) data sets. In contrast, Teng et al had detected formate in their study [10]. In our study, taurine and hypoxanthine were detected at approximately 2.2- and 1.3-fold higher levels, respectively, in uninfected erythrocytes. These levels were comparable to the approximately 1.8- and 1.2-fold higher levels of taurine and hypoxanthine, respectively, in uninfected erythrocytes detected by Olszewski et al [12].

Plant-Like Metabolites of 13C–α-Linolenic Acid (ALA) Enriched During In Vitro P. falciparum Infection

Unexpectedly, some of the putatively annotated metabolites in our untargeted analysis belonged to a plant-like ALA metabolism pathway. Of these, the metabolite annotated as traumatin (12-oxo-10E-dodecenoic acid; expected 12C mass, 211.134) was one of the most differentially abundant between infected and uninfected plasma (approximately 8.5-fold greater average abundance in infected plasma; P < .0001) and infected and uninfected erythrocyte (approximately 1.5 fold higher in infected plasma; P < .0079) samples.

This metabolite eluted with a retention time of approximately 11.2 minutes in the infected plasma and erythrocyte (not shown) samples (Figure 1C). The identity of this putatively annotated metabolite in the infected plasma and erythrocyte (not shown) samples was confirmed using a pure 12C form of traumatin synthesized at a purity of >95% (Figure 1C, Supplementary Materials). The retention time of the latter (at concentrations of 1 μM, 5 μM, and 10 μM) matched the retention time of traumatin in the infected plasma sample with a mass difference of <2 ppm (Figure 1C). Furthermore, spiking infected and uninfected plasma samples with pure 12C traumatin increased the ions counts of the metabolite nominated as traumatin in the plasma samples in proportion to the concentration used (Figure 1D).

The detection of traumatin in patient plasma and erythrocyte samples led us to investigate the presence of additional intermediates of the plant ALA pathway in P. falciparum–infected samples using 13C tracing studies of ALA metabolism. Late-stage P. falciparum–infected erythrocytes enriched with Percoll approximately 30 hours after invasion were incubated with a 13C-ALA-ester mixture for 8 hours, with analyses conducted at 4 hours (approximately 34 hours after invasion) and 8 hours (approximately 38 hours after invasion). Total metabolites were extracted at each time point, and the uptake of 13C-ALA and the presence of 13C-labeled intermediates downstream of ALA within P. falciparum–infected erythrocytes were determined by LC-MS analysis. Uninfected erythrocytes were used as a control. The uptake of 13C-ALA within infected and uninfected erythrocytes was comparable and therefore not significantly different at both time points (Figure 2A). However, several putative metabolites of the ALA pathway were significantly enriched (P < .05) in the infected erythrocytes as compared to uninfected erythrocytes at both time points (Figure 2B–E).

Figure 2.

Figure 2.

Open in a new tab

Comparison of abundance levels of 13C-labeled metabolites of α-linolenic acid (ALA) in in vitro–cultured Percoll-enriched P. falciparum-infected erythrocyte versus uninfected erythrocyte pellets at 4 hours and 8 hours after incubation with 13C-labeled ALA. The y-axis shows ions counts (abundance) of 13C-labeled metabolites in infected samples (black bars) and uninfected samples (white bars). Data are representative of 3 repeats, and error bars reflect the standard error of the mean. All 13C-labeled ALA metabolites detected (except jasmonate at 4 hours) were significantly enriched in infected erythrocytes as compared to uninfected erythrocytes (P < .05, by the Mann-Whitney U test).

At the 4-hour time point, we detected 13C-labeled masses corresponding to aldehydes and oxoacids of the lipoxygenase-hydroperoxide lyase arm of the pathway. Specifically, 9-oxononadienal and 3-hexenal-like masses were present at 2- and 4-fold higher levels, respectively, in infected samples as compared to uninfected samples (Figure 2B). At 8 hours, the levels of these 2 metabolites were 5- and 8-fold higher, respectively, in infected samples (Figure 2B).

In the cyclopentenone-jasmonate branch, masses corresponding to 3-oxo-OPC8-CoA, trans-2-enoyl-OPC6-CoA, and methyl jasmonate were enriched in infected samples by 1.6-, 1.4-, 1.9-fold as compared to uninfected samples, whereas jasmonate levels were comparable (not significantly different) between infected and uninfected samples at 4 hours (Figure 2C). At 8 hours, the levels of masses corresponding to 3-oxo-OPC8-CoA, trans-2-enoyl-OPC6-CoA, jasmonate and methyl jasmonate were 3.5-, 4.5-, 4- and 2.8-fold higher, respectively, in infected samples (Figure 2C).

Masses corresponding to 3,6-nonadienal and OPC6-CoA were detected only at 4 hours and were present at 6- and 2-fold higher levels, respectively, in the infected samples (Figure 2D). At the 8-hour time point, values for these metabolites were below the quantitative detection threshold. In contrast, levels of heptadecatrienal, traumatin, traumatic acid, trans-2-enoyl-OPC4-CoA, and 3-oxo-OPC4-CoA-like masses were below the quantitative detection threshold at the 4-hour time point and were enriched in the infected samples by 3-, 5-, 4-, 8-, and 10-fold, respectively, only at the 8-hour time point (Figure 2E).

Tandem mass spectrometry (MS/MS) of any metabolite generates a signature fragmentation pattern that is a function of the metabolite's chemical structure. To further confirm the identity of the 13C-labeled forms of traumatin, traumatic acid, and methyl jasmonate, we carried out MS/MS of 13C-ALA-labeled parasite metabolite extracts. The MS/MS fragment masses of 13C-labeled traumatin were calculated from the MS/MS data for the pure 12C form of traumatin. The masses of MS/MS fragments of 13C-labeled traumatic acid and methyl jasmonate were calculated from MS/MS data of their 12C forms, available on the Metlin database (http://metlin.scripps.edu/). The MS/MS fragmentation was performed in the negative mode for traumatin and traumatic acid and in the positive mode for methyl jasmonate.

The expected masses of 13C-labeled MS/MS fragments of traumatin (193.1681, 177.1732, and 103.086) closely matched the observed masses of MS/MS fragments of traumatin and were detected at quantifiable levels (Figure 3A). Similarly, the expected masses of 13C-labeled MS/MS fragments of traumatic acid (176.1654, 116.0894, and 88.067) and methyl jasmonate (178.1161, 98.0777 and 85.0744) closely matched the masses of the MS/MS fragments obtained for each of these metabolites and were detected at quantifiable levels (Figure 3B and 3C, respectively).

Figure 3.

Figure 3.

Open in a new tab

Tandem mass spectrometry (MS/MS) data for 3 13C-labeled metabolites of α-linolenic acid detected in our labeling experiments. Chemical structures, structure of metabolite species expected upon fragmentation, expected mass, ionization mode, and the MS/MS peak for each fragment are depicted for traumatin (A), traumatic acid (B), and methyl jasmonate (C).

The levels of heptadecatrienoic acid and OPC4-CoA-like masses and metabolites with the redundant formulas [13C]18H30O3, [13C]18H30O4, and [13C]18H28O3 remained relatively constant (on average, 2-fold higher in infected samples) at both time points.

Identifying ALA Pathway Enzymes in P. falciparum and T. gondii by Using Orthologous Sequence Information From Plant Genomes

All enzymes in the ALA pathways of Arabidopsis thaliana, Arabidopsis lyrata, Populus trichocarpa, Ricinus communis, Vitis vinifera, all except 2 in Oryza sativa japonica, Sorghum bicolor, and all except 3 in Zea mays are annotated (http://www.kegg.com/; Kyoto Encyclopedia of Genes and Genomes pathway identification number map00592). In contrast, only 1 gene (PF14_0484) of this pathway is annotated in P. falciparum (http://www.plasmodb.org/). We thus used sequences of enzymes in the ALA pathway of these 8 plants to search for orthologs in the P. falciparum genome using hidden Markov model comparisons. We identified several candidate P. falciparum genes whose predicted protein sequences shared significant homology to those of the plant ALA pathway enzymes (Table 2).

Table 2.

Results of Sequence Ortholog Searches for α-Linolenic Acid (ALA) Pathway Enzymes in Plasmodium falciparum and Toxoplasma gondii Based on ALA Pathway Enzyme Sequences From 8 Plant Genomes

ALA Pathway Enzymes Arabidopsis thaliana Arabidopsis lyrata Populus trichocarpa Ricinus communis Vitis vinifera Oryza sativa japonica Sorghum bicolor Zea mays P. falciparum T. gondii 
3.1.1.4 X X X X X X X X PFI1180w …
X X X X X X X X … TGGT1_071870
X … … X … … … X … TGGT1_093110
… …. … … … … X … … TGME49_089800
3.1.1.32 X X X X X … X … PF14_0250 …
X X X X X … X … … TGME49_077950
DOX1 … X … X … … … … … TGGT1_092680
… X … … X … … … … TGGT1_103890
1.13.11.12 X … X … … X … X PF14_0067 …
X X X X X … … X … TGME49_111250
X X … X … X X X … TGGT1_069270
X X X X X X X X … TGGT1_071620
… … … … … X … … … TGME49_071990
… … … … … X … … … TGGT1_020910
… … X … … … … … … TGGT1_043530
… … X … … X … … … TGVEG_065070
… … … … … X X X … TGVEG_085480
… … … … … … X … … TGME49_088000
HPL1 X X X X X X X X … TGME49_115770
4.2.1.92 X X X X X X X X … TGME49_115770
5.3.99.6 … X … … … … … … … TGVEG_074120
X … … … … … … … … TGME49_067710
… … … X … … … … … TGGT1_014600
… … … … X … … … … TGME49_037480
… … … … … … … X … TGGT1_078240
1.3.1.42 X X … X … … X … PF14_0086 …
X … X … X X X X PFI0920c …
X X X … … … X X … TGGT1_072520
X … … X X X … … … TGGT1_076080
OPCL1 X X X X X X X X PFF0945c …
X X X X X X X X … TGGT1_076330
ACX … X X … X … … … PFE1345c …
X … X … X … … … PF13_0291 …
… … … … X … … … PFF1355w …
X X X X X X X X … TGGT1_025410
X X X X X X X X … TGGT1_115860
MFP2 X X X X X X X X PF14_0232 …
X X X X X X X X PFL1940w …
X X X X X X … X … TGME49_024090
X X X X X X X X … TGGT1_115670
X X X X X X X X … TGGT1_105590
2.3.1.16 X X X X X X X X PF14_0484 …
X X X X X X X X … TGGT1_112820
2.1.1.141 X X X … … … … … MAL13P1.214 …
… … … X X … … … PFE1115 …
X X X X X … … … … TGGT1_076110
Open in a new tab

The first column lists the names and Enzyme Commission numbers of enzymes involved in the ALA pathway, according to their Kyoto Encyclopedia of Genes and Genomes (KEGG) description. Sometimes >1 gene was associated with a specific KEGG functional definition. Columns 2–9 list the 8 plant genomes from which the ALA pathway enzyme sequences were obtained. Columns 10 and 11 list the gene hits in P. falciparum and T. gondii genomes, respectively. In columns 2–9, X indicates plant genomes that yielded the P. falciparum and T. gondii hits listed in columns 10 and 11.

We found orthologs (sometimes >1) in P. falciparum for 9 ALA pathway enzymes in the 8 different plant genomes. Meanwhile, for 4 enzymes from the ALA pathway there were no significant hits in P. falciparum (Table 2). We also scanned the genome of T. gondii for orthologs of plant ALA pathway enzymes and found significant sequence hits for all 13 enzymes. We checked for expression of the 14 genes identified in P. falciparum that were orthologs of the enzymes in the plant ALA pathway (Table 3). RT-PCR analysis showed that all of these P. falciparum genes were transcribed in late asexual stages, strongly suggesting that this pathway might be active in this parasite (Figure 4).

Table 3.

Primer Sequences Used to Verify the Transcription of Genes Identified in Plasmodium falciparum

Forward Primer
 Reverse Primer
 Product Size
P. falciparum gene No. Sequence No. Sequence Without Introns With Introns
PFI1180w p1 TTGTATCACAAGTTAATCCACACG p2 TTGTTGTATATACCATTCAACGTCG 338 549
PF14_0250 p13 AACAAATCTAATATATGAATTAACACC p14 AGTCGGCTTCATTAACCCATCC 692 x
PF14_0067 p15 AAACTATTGATGGACGTGATTGTG p16 AAATGATTCTACATTATTTTGTGTGG 680 x
PFF0945c p21 GATCACAATAATGATCATAATAGTGG p22 TAAGGCAGGATAATGATCAAAGTC 669 x
PF14_0232 p27 TTTATGTAATACAATACAACAATTACC p28 TTAATATATCTAATCGTGGTATACTG 623 x
PFL1940w p31 GTGGATTTGATTAATTTTAATGTGTG p32 TGTAAACTCTTCCTTTATTGACAAC 575 x
PFE1115c p33 GTCAAATGACAGAAAGAGCTATGG p34 CTGCCTGTGTATTTACTATCTCG 706 1952
MAL13P1.214 p35 TAAAGGTTTACGAGTTCATTTTTGG p36 CCATCTTTGCATTTTCCTCTTGC 704 978
PF14_0484 p39 GTTGTGGTCCGGTACCATTGC p40 TGACCTAAAGATAATGCTCCTCC 865 983
PF14_0086 p41 TTAGGTTGTCCTCAACAGATAGC p42 TCCTTTCTGTAACGCCTGTACC 680 x
PFI0920c p43 GCAAGGAAAGCTTGGTTAAAAGG p44 TATTCCTTTTGTAAATTGTAATCGTG 635 x
PFE1345c p45 GTTACTAGTGTAGATGTTAATTATGC p46 CGTAGTTGGTTAATATCAAGTTTGG 503 662
PF13_0291 p47 CTATGAGAATGACTGTGCGTCAG p48 TTGGACCTGCATAATTTGGATGG 552 x
PFF1355w p49 TATACCGAATAAAATCTATATCCATG p50 TCCTTAATCGTAAGTATAAAGACTG 432 556
Open in a new tab

Product was detected for all genes. Reverse transcription–polymerase chain reaction results are shown in Figure 4.

Figure 4.

Figure 4.

Open in a new tab

Plasmodium falciparum α-linolenic acid metabolism pathway (adapted from http://www.kegg.com/) showing metabolites detected in our study. A, Metabolites shown in bold font were detected in our initial liquid chromatography–mass spectrometry analysis of plasma and erythrocyte samples, underlining indicates metabolites with specific formulas detected by 13C labeling, single asterisks represent metabolites with redundant formulas detected by 13C labeling, italicized metabolites were confirmed by tandem mass spectrometry, and double asterisks represent enzyme identification numbers for which the P. falciparum enzyme sequences had domains matching those in enzyme sequences from plants. B, Reverse transcription–polymerase chain reaction data showing expression in the late asexual stage of 14 Plasmodium falciparum genes identified in ortholog searches (Table 2). Primer pairs used are shown in Table 3 and indicated at the top of the lanes. bp, base pairs; M, 1 kb Plus DNA ladder (Invitrogen); −, first-strand complementary DNA (cDNA) synthesis reactions without reverse transcriptase; +, first-strand complementary DNA (cDNA) synthesis reactions with reverse transcriptase.

DISCUSSION

We used an untargeted LC-MS–based approach to study the metabolome of P. falciparum–infected and -uninfected patient plasma, as well as Percoll-enriched infected and uninfected in vitro–cultured erythrocytes. An untargeted approach provides the opportunity to identify metabolites that were not previously predicted to exist. Several metabolites with significantly different abundances between infected and uninfected samples and belonging to a variety of biochemical classes were detected, including products of ALA metabolism not previously known to exist in malaria.

We note that although we used an untargeted LC-MS approach to achieve a wider coverage of novel aspects of parasite biochemistry, all metabolite identities of interest will require further confirmation using their pure forms. PLSDA models of plasma and erythrocyte metabolomic data showed clear segregation between infected and uninfected samples.

P. falciparum requires fatty acids for growth and scavenges a broad range of lipids from host plasma [19]. Mi-Ichi et al reported an enrichment of ALA in infected red cells as compared to human serum [19]. Furthermore, human erythrocytes have been shown to take up fatty acids from the plasma [20].

Our labeling experiments with 13C-ALA revealed that both P. falciparum–infected and -uninfected erythrocytes were capable of taking up 13C-ALA, although the difference was not significant. It is unknown whether fatty acids are taken up through energy-dependent mechanism(s) or by passive diffusion [21].

Several 13C-ALA metabolite species with specific chemical formulas were enriched in the infected samples at the 4-hour time point. At the 8-hour time point, the abundance levels of these metabolite species had increased 2-fold on average as compared to their levels at 4 hours, suggesting either an accumulation of the products or an increase in the activity of the pathway. In addition, a greater number of labeled metabolites were detected at the 8-hour time point. Overall, we detected 35 of the 38 total metabolites downstream of ALA, strongly suggesting that plant-like ALA metabolism might be active in P. falciparum (Figure 4).

Our data comparing the retention time of pure traumatin with the metabolite identified as traumatin in plasma and erythrocyte samples, coupled with the MS/MS data confirming the identity of 3 representative metabolites in the pathway, strongly suggested that P. falciparum parasites can generate plant-like metabolites of ALA. A representative set of metabolites was used for MS/MS because of the need for pure standards to generate MS/MS profiles and the unavailability of pure forms of most ALA pathway metabolites.

Metabolites of ALA belong to a family of oxygenated lipids called oxylipins and are known to have important biological roles in plants and fungi [22]. Synthesis of oxylipins in plants is catalyzed by at least 3 enzyme types, namely, peroxygenase, α-dioxygenase, and 9-/13-lipoxygenase [23]. Lipoxygenases convert ALA to hydroperoxides, which are then metabolized to various oxylipins [22, 24]. We thus used the sequences of enzymes in the ALA pathway of 8 known plant genomes to search for orthologs in the P. falciparum genome.

We detected significant hits in the P. falciparum genome for 9 of the 13 plant enzyme sequences. Of the putative pathway genes that we identified in P. falciparum, many of the transcripts and proteins have been detected during asexual and gametocyte stages [25–27]. Furthermore, RT-PCR analysis showed that the 14 genes identified were transcribed in the late asexual stage of P. falciparum. Stage-specific metabolomic analysis would be required to define stage-related metabolite profile and function(s) of this pathway.

Lipoxygenase is one of the enzyme types that is known to initiate oxylipin synthesis [22]. One of the P. falciparum candidates (PF14_0067) is predicted to contain the PLAT/LH2 (polycystin-1, lipoxygenase, alpha-toxin/lipoxygenase homology 2) and LCCL domains. The former, which is found in a variety of membrane or lipid associated proteins [28, 29], is also present in the lipoxygenase sequences in the plant ALA pathway. Given that the conversion of precursor fatty acids by lipoxygenases to hydroperoxides is a critical step in the formation of oxylipins [24], PF14_0067 might play a role in initiating the formation of oxylipins in P. falciparum. Interestingly, malaria parasites possess several plant-like features, including an apicoplast [30, 31], the fatty acid synthesis II pathway [32, 33], and several transcription factors [34, 35]. Our data now suggest that the ability to generate oxylipins might be another plant-like feature possessed by P. falciparum.

Our 13C labeling and MS/MS data and our sequence analyses suggest that the genes identified in P. falciparum might be the functional counterparts of the ones in the plant ALA pathway. We also detected orthologs in the T. gondii genome for all 13 enzymes in the plant ALA pathway. While our inability to find hits for the other sequences in P. falciparum could potentially be due its use of a noncanonical ALA metabolism pathway, further work is required in this area.

Alternatively, at least some of the biochemical reactions in this pathway may be mediated by nonenzymatic mechanism(s). Schwarzer et al reported that the trophozoite stage of P. falciparum–infected erythrocytes generated oxygenated fatty acids from arachidonic acid and linoleic acid through a nonenzymatic hemozoin-catalyzed lipid peroxidation process because of lack of evidence of lipoxygenase activity or the existence of a lipoxygenase sequence in P. falciparum at the time [36].

We note that linoleic acid (18:2 [n − 6]; ω − 6) and ALA (18:3 [n − 3]; ω − 3) are structurally distinct essential fatty acids that are metabolized into entirely different kinds of long chain polyunsaturated fatty acids [37]. Also, γ-linolenic acid generated from linoleic acid is not a precursor of the ALA pathway. Nevertheless, a small contribution of nonenzymatic mechanisms (from host or parasite) toward the total pool of ALA metabolites could not be completely ruled out.

Oxylipins perform diverse functions in plants, including activation of pathogen-specific defenses, response to environmental stress, development, and reproduction [23, 38]. Similar processes might be performed by oxylipins in malaria, such as quorum sensing, induction of gametocytes, or modulating host responses to infection. However, further investigations will be required to demonstrate the role of oxylipins in P. falciparum infection.

Only a small overlap was evident between the significantly different metabolites in the plasma and erythrocyte samples (Supplementary Table 1). The metabolites detected in both of our sample types could reflect parasite-specific biology. However, the plasma also contains a vast array of host metabolites. Acetone, for example, was the most differentially abundant metabolite in plasma. Acetone is normally present in human serum [39] and is elevated in patients infected with hepatitis E virus [40]. Further investigation will be needed to determine the significance of changes in acetone levels during malaria infection.

Formaldehyde was found in both types of infected samples. This is a highly reactive aldehyde formed during oxidation of hydrocarbons, and it is metabolically converted into formate. Elevated formate levels can have several systemic effects, including metabolic acidosis, circulatory shock, respiratory insufficiency, and acute renal failure [41, 42]. In vitro culture conditions, which differ from the natural host environment, could alter parasite metabolism and also contribute to this minimal overlap between the metabolite identities in the 2 data sets [43]. Although the in vivo and in vitro systems are dissimilar in many respects, the in vitro system allows the identification in the in vivo data set of metabolites that are parasite derived and that may be relevant to human disease.

We note that some of the metabolites, even though exhibiting highly differential abundances between infected and uninfected samples, were nominated either multiple identities (Supplementary Table 1) or none at all because of limitations in our software. Thus, multiple extraction and detection methods would be necessary to achieve a more complete coverage of metabolite identities [44].

To our knowledge, this is the first report on the existence of a plant-like metabolic pathway in P. falciparum. Our approach has enabled us to confirm some of the known metabolites, as well as unravel a hitherto undiscovered aspect of P. falciparum biology. Further studies will be required to more completely dissect the relevance of the ALA pathway in P. falciparum infection.

Our results add to the growing body of literature that describes the application of novel methods to better understand parasite biology and address important biological and clinically relevant questions. Data generated from these undertakings will foster our efforts toward generating accurate disease models, discovering effective antimalarial therapeutics, and developing an effective vaccine.

Supplementary Data

Supplementary materials are available at The Journal of Infectious Diseases online (http://jid.oxfordjournals.org/). Supplementary materials consist of data provided by the author that are published to benefit the reader. The posted materials are not copyedited. The contents of all supplementary data are the sole responsibility of the authors. Questions or messages regarding errors should be addressed to the author.

Supplementary Data
supp_206_2_238__index.html (1.1KB, html)

Notes

Acknowledgments. We thank Dr Morad Hassani, Dr Krishanthi Subramaniam, Anirudh Kumar, Vasiliki Pappa, Catherine Manix, and Adam Goldman-Yassen for their helpful comments.

V. L. was involved in experimental design, metabolite extractions, LC-MS experiments, data collection and analysis, and manuscript preparation. K. Y. R. was involved in experimental design and data analysis. LC-MS experiments were run in K. Y. R.'s laboratory. W. W. synthesized pure traumatin and verified its structure by nuclear magnetic resonance analysis. Y. Y. performed PLSDA and cross-validation analysis. K. K. and A. F. performed bioinformatic analysis to identify orthologous gene sequences in P. falciparum and T. gondii. Traumatin synthesis was performed in P. W.'s laboratory. O. N., S. M., and D. N. collected and provided patient samples for this study. J. P. D. was involved in sample collection, experimental design, data analyses, and manuscript preparation.

Financial support. This work was supported by the National Institute of Allergy and Infectious Diseases (Grant No. R01AI077623).

Potential conflicts of interest. All authors: No reported conflicts.

All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.

References

  • 1.Chuang I, Richie TL. World Malaria Report 2010: documenting progress towards malaria eradication. Expert Rev Vaccines. 2012;11:39–41. [Google Scholar]
  • 2.Fidock DA, Eastman RT, Ward SA, Meshnick SR. Recent highlights in antimalarial drug resistance and chemotherapy research. Trends Parasitol. 2008;24:537–44. doi: 10.1016/j.pt.2008.09.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Richards JS, Beeson JG. The future for blood-stage vaccines against malaria. Immunol Cell Biol. 2009;87:377–90. doi: 10.1038/icb.2009.27. [DOI] [PubMed] [Google Scholar]
  • 4.Speake C, Duffy PE. Antigens for pre-erythrocytic malaria vaccines: building on success. Parasite Immunol. 2009;31:539–46. doi: 10.1111/j.1365-3024.2009.01139.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Daily JP, Scanfeld D, Pochet N, et al. Distinct physiological states of Plasmodium falciparum in malaria-infected patients. Nature. 2007;450:1091–5. doi: 10.1038/nature06311. [DOI] [PubMed] [Google Scholar]
  • 6.Mackinnon MJ, Marsh K. The selection landscape of malaria parasites. Science. 2010;328:866–71. doi: 10.1126/science.1185410. [DOI] [PubMed] [Google Scholar]
  • 7.Besteiro S, Vo Duy S, Perigaud C, Lefebvre-Tournier I, Vial HJ. Exploring metabolomic approaches to analyse phospholipid biosynthetic pathways in Plasmodium. Parasitology. 2010;137:1343–56. doi: 10.1017/S0031182009991934. [DOI] [PubMed] [Google Scholar]
  • 8.Olszewski KL, Llinas M. Central carbon metabolism of Plasmodium parasites. Mol Biochem Parasitol. 2011;175:95–103. doi: 10.1016/j.molbiopara.2010.09.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Lakshmanan V, Rhee KY, Daily JP. Metabolomics and malaria biology. Mol Biochem Parasitol. 2011;175:104–11. doi: 10.1016/j.molbiopara.2010.09.008. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Teng R, Junankar PR, Bubb WA, Rae C, Mercier P, Kirk K. Metabolite profiling of the intraerythrocytic malaria parasite Plasmodium falciparum by (1)H NMR spectroscopy. NMR Biomed. 2009;22:292–302. doi: 10.1002/nbm.1323. [DOI] [PubMed] [Google Scholar]
  • 11.Basant A, Rege M, Sharma S, Sonawat HM. Alterations in urine, serum and brain metabolomic profiles exhibit sexual dimorphism during malaria disease progression. Malar J. 2010;9:110. doi: 10.1186/1475-2875-9-110. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Olszewski KL, Morrisey JM, Wilinski D, et al. Host-parasite interactions revealed by Plasmodium falciparum metabolomics. Cell Host Microbe. 2009;5:191–9. doi: 10.1016/j.chom.2009.01.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Olszewski KL, Mather MW, Morrisey JM, et al. Branched tricarboxylic acid metabolism in Plasmodium falciparum. Nature. 2010;466:774–8. doi: 10.1038/nature09301. [DOI] [PMC free article] [PubMed] [Google Scholar] [Retracted]
  • 14.Couto AS, Kimura EA, Peres VJ, Uhrig ML, Katzin AM. Active isoprenoid pathway in the intra-erythrocytic stages of Plasmodium falciparum: presence of dolichols of 11 and 12 isoprene units. Biochem J. 1999;341(Pt 3):629–37. [PMC free article] [PubMed] [Google Scholar]
  • 15.Tonhosolo R, D'Alexandri FL, de Rosso VV, et al. Carotenoid biosynthesis in intraerythrocytic stages of Plasmodium falciparum. J Biol Chem. 2009;284:9974–85. doi: 10.1074/jbc.M807464200. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Lian LY, Al-Helal M, Roslaini AM, et al. Glycerol: an unexpected major metabolite of energy metabolism by the human malaria parasite. Malar J. 2009;8:38. doi: 10.1186/1475-2875-8-38. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Sana TR, Waddell K, Fischer SM. A sample extraction and chromatographic strategy for increasing LC/MS detection coverage of the erythrocyte metabolome. J Chromatogr B Analyt Technol Biomed Life Sci. 2008;871:314–21. doi: 10.1016/j.jchromb.2008.04.030. [DOI] [PubMed] [Google Scholar]
  • 18.Westerhuis JA, van Velzen EJ, Hoefsloot HC, Smilde AK. Multivariate paired data analysis: multilevel PLSDA versus OPLSDA. Metabolomics. 2010;6:119–28. doi: 10.1007/s11306-009-0185-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Mi-Ichi F, Kita K, Mitamura T. Intraerythrocytic Plasmodium falciparum utilize a broad range of serum-derived fatty acids with limited modification for their growth. Parasitology. 2006;133:399–410. doi: 10.1017/S0031182006000540. [DOI] [PubMed] [Google Scholar]
  • 20.Skeaff CM, Hodson L, McKenzie JE. Dietary-induced changes in fatty acid composition of human plasma, platelet, and erythrocyte lipids follow a similar time course. J Nutr. 2006;136:565–9. doi: 10.1093/jn/136.3.565. [DOI] [PubMed] [Google Scholar]
  • 21.Dechamps S, Shastri S, Wengelnik K, Vial HJ. Glycerophospholipid acquisition in Plasmodium—a puzzling assembly of biosynthetic pathways. Int J Parasitol. 2010;40:1347–65. doi: 10.1016/j.ijpara.2010.05.008. [DOI] [PubMed] [Google Scholar]
  • 22.Tsitsigiannis DI, Keller NP. Oxylipins as developmental and host-fungal communication signals. Trends Microbiol. 2007;15:109–18. doi: 10.1016/j.tim.2007.01.005. [DOI] [PubMed] [Google Scholar]
  • 23.Mosblech A, Feussner I, Heilmann I. Oxylipins: structurally diverse metabolites from fatty acid oxidation. Plant Physiol Biochem. 2009;47:511–7. doi: 10.1016/j.plaphy.2008.12.011. [DOI] [PubMed] [Google Scholar]
  • 24.Vellosillo T, Martinez M, Lopez MA, et al. Oxylipins produced by the 9-lipoxygenase pathway in Arabidopsis regulate lateral root development and defense responses through a specific signaling cascade. Plant Cell. 2007;19:831–46. doi: 10.1105/tpc.106.046052. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Bowyer PW, Simon GM, Cravatt BF, Bogyo M. Global profiling of proteolysis during rupture of Plasmodium falciparum from the host erythrocyte. Mol Cell Proteomics. 2011;10:M110 001636. doi: 10.1074/mcp.M110.001636. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Le Roch KG, Zhou Y, Blair PL, et al. Discovery of gene function by expression profiling of the malaria parasite life cycle. Science. 2003;301:1503–8. doi: 10.1126/science.1087025. [DOI] [PubMed] [Google Scholar]
  • 27.Silvestrini F, Lasonder E, Olivieri A, et al. Protein export marks the early phase of gametocytogenesis of the human malaria parasite Plasmodium falciparum. Mol Cell Proteomics. 2010;9:1437–48. doi: 10.1074/mcp.M900479-MCP200. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Bateman A, Sandford R. The PLAT domain: a new piece in the PKD1 puzzle. Curr Biol. 1999;9:R588–90. doi: 10.1016/s0960-9822(99)80380-7. [DOI] [PubMed] [Google Scholar]
  • 29.Ponting CP, Hofmann K, Bork P. A latrophilin/CL-1-like GPS domain in polycystin-1. Curr Biol. 1999;9:R585–8. doi: 10.1016/s0960-9822(99)80379-0. [DOI] [PubMed] [Google Scholar]
  • 30.Fichera ME, Roos DS. A plastid organelle as a drug target in apicomplexan parasites. Nature. 1997;390:407–9. doi: 10.1038/37132. [DOI] [PubMed] [Google Scholar]
  • 31.Lim L, McFadden GI. The evolution, metabolism and functions of the apicoplast. Philos Trans R Soc Lond B Biol Sci. 2010;365:749–63. doi: 10.1098/rstb.2009.0273. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Surolia N, Surolia A. Triclosan offers protection against blood stages of malaria by inhibiting enoyl-ACP reductase of Plasmodium falciparum. Nat Med. 2001;7:167–73. doi: 10.1038/84612. [DOI] [PubMed] [Google Scholar]
  • 33.Yu M, Kumar TR, Nkrumah LJ, et al. The fatty acid biosynthesis enzyme FabI plays a key role in the development of liver-stage malarial parasites. Cell Host Microbe. 2008;4:567–78. doi: 10.1016/j.chom.2008.11.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Balaji S, Babu MM, Iyer LM, Aravind L. Discovery of the principal specific transcription factors of apicomplexa and their implication for the evolution of the AP2-integrase DNA binding domains. Nucleic Acids Res. 2005;33:3994–4006. doi: 10.1093/nar/gki709. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.De Silva EK, Gehrke AR, Olszewski K, et al. Specific DNA-binding by apicomplexan AP2 transcription factors. Proc Natl Acad Sci U S A. 2008;105:8393–8. doi: 10.1073/pnas.0801993105. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Schwarzer E, Kuhn H, Valente E, Arese P. Malaria-parasitized erythrocytes and hemozoin nonenzymatically generate large amounts of hydroxy fatty acids that inhibit monocyte functions. Blood. 2003;101:722–8. doi: 10.1182/blood-2002-03-0979. [DOI] [PubMed] [Google Scholar]
  • 37.Brenna JT, Salem N, Jr, Sinclair AJ, Cunnane SC. alpha-Linolenic acid supplementation and conversion to n-3 long-chain polyunsaturated fatty acids in humans. Prostaglandins Leukot Essent Fatty Acids. 2009;80:85–91. doi: 10.1016/j.plefa.2009.01.004. [DOI] [PubMed] [Google Scholar]
  • 38.Gobel C, Feussner I. Methods for the analysis of oxylipins in plants. Phytochemistry. 2009;70:1485–503. doi: 10.1016/j.phytochem.2009.07.040. [DOI] [PubMed] [Google Scholar]
  • 39.Psychogios N, Hau DD, Peng J, et al. The human serum metabolome. PLoS One. 2011;6:e16957. doi: 10.1371/journal.pone.0016957. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Munshi SU, Taneja S, Bhavesh NS, Shastri J, Aggarwal R, Jameel S. Metabonomic analysis of hepatitis E patients shows deregulated metabolic cycles and abnormalities in amino acid metabolism. J Viral Hepat. 2011;18:e591–602. doi: 10.1111/j.1365-2893.2011.01488.x. [DOI] [PubMed] [Google Scholar]
  • 41.Pandey CK, Agarwal A, Baronia A, Singh N. Toxicity of ingested formalin and its management. Hum Exp Toxicol. 2000;19:360–6. doi: 10.1191/096032700678815954. [DOI] [PubMed] [Google Scholar]
  • 42.Barceloux DG, Bond GR, Krenzelok EP, Cooper H, Vale JA. American Academy of Clinical Toxicology practice guidelines on the treatment of methanol poisoning. J Toxicol Clin Toxicol. 2002;40:415–46. doi: 10.1081/clt-120006745. [DOI] [PubMed] [Google Scholar]
  • 43.LeRoux M, Lakshmanan V, Daily JP. Plasmodium falciparum biology: analysis of in vitro versus in vivo growth conditions. Trends Parasitol. 2009;25:474–81. doi: 10.1016/j.pt.2009.07.005. [DOI] [PubMed] [Google Scholar]
  • 44.Wishart DS, Lewis MJ, Morrissey JA, et al. The human cerebrospinal fluid metabolome. J Chromatogr B Analyt Technol Biomed Life Sci. 2008;871:164–73. doi: 10.1016/j.jchromb.2008.05.001. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Data
supp_206_2_238__index.html (1.1KB, html)
supp_jis339_jis339suppl.doc (608KB, doc)
supp_jis339_jis339suppl_table1.xls (95KB, xls)

Articles from The Journal of Infectious Diseases are provided here courtesy of Oxford University Press

RESOURCES