Plasmodium histones are inducibly lactylated

We first used western blotting to measure the level of lactyl-lysine in Plasmodium histones after treating cultured parasites with increasing levels of lactate. Trophozoite-stage P. falciparum parasites were exposed to 5-25mM lactate, added to the culture media for 12h. Fig 1B shows that their histones were inducibly and titratably lactylated, with histone H4 showing the strongest lactyl-lysine signal. An antibody identifying a single lactylated residue on histone H4 (K12) confirmed that the general lactyl-lysine antibody recognised this histone, but lactylation of the K12 residue itself was not inducible (Fig 1C). We conducted the same experiments on a second species, P. knowlesi, and showed that histone lactylation was similarly inducible (Fig 1D). It was, however, much less abundant than in P. falciparum, since greater amounts of parasite extract were required to detect comparable signals on P. knowlesi histones, and the specific residue H4K12La was barely detectable (Fig 1E).

We then investigated whether this phenomenon occurs naturally in vivo during malaria. The rodent model species P. yoelii, strain 17XL, can generate severe hyperlactataemia [14], and blood lactate levels in mice infected with P. yoelii 17XL did indeed correlate with histone lactylation in the parasite (Fig 1F). Notably, not all rodent malaria models lead consistently to hyperlactataemia as infection progresses, and in the P. berghei ANKA strain, which does not [14], we did not observe a strong correlation between blood lactate and histone lactylation (S1 Fig). Overall, these results revealed that there is a conserved pathway for histone lactylation in Plasmodium, induced by hyperlactataemia and apparently most prominent in species that do characteristically cause hyperlactataemia.

Histone lactylation varies across the P. falciparum cell cycle

We next investigated the cell-cycle dynamics of the lactyl modification, again using western blotting. Fig 1B–E focuses only within the trophozoite stage, showing that lactylation is induced within 12h and 8h in P. falciparum and P. knowlesi trophozoites respectively. Fig 2A shows that there is a natural cell-cycle-related pattern in histone lactylation, probably because parasite metabolism, which peaks in trophozoites, produces its own endogenous lactate. Histone lactylation peaked in the trophozoite stage in two successive parasite cycles and fell in ring and schizont stages.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Histone lactylation varies across the P. falciparum cell cycle.

A: P. falciparum parasites were grown from 1% parasitaemia without splitting or changing the media for 72h, starting at the early-trophozoite stage and sampling every 12h. Stage of the parasite culture is shown on the x-axis, and example parasite morphology is also shown. By 60h, morphology showed that the culture was starting to crash, with dead (D) pyknotic parasites by 72h. The graph shows the intensity of each KLa signal relative to total H4, quantified from n = 2 biological experiments; one representative western blot is shown. B-E: Lactate levels and parasitaemias were measured every 24h in cultures of P. falciparum (B, C) and P. knowlesi (D, E), starting at 0.5% parasitaemia, either split back to 0.5% every 48h (B, D) or allowed to grow continuously without splitting (C, E). Data shown are the mean of n = 2 biological experiments.

https://doi.org/10.1371/journal.pgen.1011991.g002

To measure the extent of endogenous lactate production in our parasite culture system, we measured lactate levels in the media using a biochemical assay, both when parasites were cultured under standard conditions (reducing the parasitaemia and changing the media every 48h), and also when they were allowed to grow continuously to the point of ‘crashing’ (the term for parasite death caused by high parasitaemia and exposure to spent media [15]). As expected, media lactate levels correlated with parasitaemia, reaching only ~ 5mM within 2 days as a culture of P. falciparum grew from 0.5% to 4.6% parasitaemia (Fig 2B). When the culture was allowed to ‘crash’, lactate levels peaked at ~10mM (Fig 2C). The profile of lactate production was similar in P. knowlesi cultures (Fig 2D), although this parasite is more resistant than P. falciparum to ‘crashing’, so these parasites could grow exponentially for at least 3 days, producing lactate continuously and reaching 11.5mM by day-4 (Fig 2E).

Overall, these data show that routine culture of both parasite species can expose them to significant but modest lactate fluctuations. Routine cultures can reach the clinical threshold for hyperlactataemia (5mM), which is sufficient to induce some histone lactylation (Fig 1B and 1D), but is much lower than the levels attained in human patients with severe malaria [1,2].

Histone acetylation is not strongly affected by exposure to high lactate

In human cells, both acetylation and lactylation have been detected on the same histone lysine residues, and it has been suggested that histone acetyltransferase enzymes may moonlight as lactylases [10]. This raised the possibility that in Plasmodium, histone lactylation could reciprocally affect histone acetylation – a well-established factor in gene expression regulation.

We tested this by western-blotting for acetyl H4 and acetyl H3 under conditions of high lactate exposure, as in Fig 1. Fig 3A–D shows no statistically significant change in acetylation of H3 or H4, despite a trend towards lactate-inducible acetylation. Thus, acetylation was not strongly lactate-inducible, although it is possible that the two histone marks may be somewhat mutually regulated by lactate.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Histone acetylation is not strongly affected by exposure to high lactate.

A,B: Histone 3 lysine acetylation (‘H3Kac’, (A)) and histone 4 lysine acetylation (‘H4Kac’, (B)) were measured in P. falciparum parasites grown in different concentrations of L-lactate for 12h, as in Fig 1. Graphs show quantification of blots from n = 6 independent experiments, with each acetyl-histone signal quantified relative to histone H4 (bars show means, dots show individual values, error bars show SEM). Multiple comparison of means was performed using one-way ANOVA and Tukey HSD post hoc test: none were significantly different at p < 0.05 (lowest p-values 0.35 for H3Kac, 0.52 for H4Kac). A representative western blot from each group of experiments is shown. C,D: Histone 3 lysine acetylation (‘H3Kac’, (C)) and histone 4 lysine acetylation (‘H4Kac’, (D)) were measured in P. knowlesi parasites grown in different concentrations of L-lactate for 8h, as in Fig 1. Data represented and analysed as in (A, B); no trends were significantly different at p < 0.05 (lowest p-values 0.15 for H3Kac, 0.40 for H4Kac). A representative western blot from each group of experiments is shown.

https://doi.org/10.1371/journal.pgen.1011991.g003

Histone lactylation is distributed throughout the nucleus

In P. falciparum, certain histone modications are specifically associated with heterochromatin, which tends to be located at the nuclear periphery, marked by heterochromatin protein 1 (HP1) [16], whereas other modifications associated with euchromatin appear throughout the nucleoplasm. We investigated the location of histone lactylation using immunofluorescence microscopy (Fig 4). We found no evidence for compartmentalisation of lactylated histones, with both the general lactyl-lysine signal (Fig 4A) and the specific signal for H4K12 (Fig 4B) appearing throughout the nuclei of all parasite stages. Neither signal colocalised with HP1, and both appeared similar to the pan-nuclear distribution of histone acetylation (S2 Fig). The great majority of lactyl-lysine signal was nuclear (Fig 4A), suggesting that there is little lactylated protein in the cytoplasm of Plasmodium.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Histone lactylation is distributed throughout the nucleus.

Confocal microscopy images showing the location of lactyl PTMs and of HP1 in P. falciparum 3D7 HP1_HA (this line expresses heterochromatin protein 1 with a haemagglutinin (HA) tag). Lactyl-lysine, red; HP1_HA, green; DAPI, blue. KLa (A) and H4K12La (B) distributions are mostly nuclear, and do not colocalise specifically with HA-tagged HP1. T, trophozoite stage; S, schizont stage. Scale bar (2 µm) applies to all images.

https://doi.org/10.1371/journal.pgen.1011991.g004

Histone lactylation occurs on multiple histone lysine residues

For a comprehensive view of the histone lysine residues that can be lactylated in both P. falciparum and P. knowlesi, we used liquid chromatography-tandem mass spectrometry (LC-MS/MS). Trophozoite-stage parasites were either exposed or not exposed to 25mM lactate and their histones were acid-extracted from chromatin in biological triplicate. In a single pilot experiment, histones were also extracted from mixed stage cultures (+/- 25mM lactate), enabling detection of KLa sites in all stages of the asexual cycle. Histone integrity and the induction of lactylation were confirmed via Coomassie-stained SDS-PAGE and by immunoblotting for KLa (S3 Fig). (25mM lactate, for 12h in P. falciparum or 8h in P knowlesi which has a shorter cell cycle, was chosen for these and all subsequent experiments. From Fig 1, this induced the strongest histone lactylation, without markedly affecting parasite growth (S4 Fig), whereas even higher levels of ≥ 30mM over 48h were previously reported to affect parasite growth in P. falciparum [17]).

Mass spectrometry on the histone-containing region of the gel identified all eight histones (H2A, H2A.Z, H2B, H2B.Z, H3, H3.3, CenH3 and H4), with sequence coverage >67.5% except in the case of CenH3 (S1–S4 Tables). We then mapped KLa sites, as previously reported [10] (S5A Fig), revealing 16 and 21 KLa sites in P. falciparum and P. knowlesi respectively (Fig 5A). These were in P. falciparum: H4 K5, 8, 12, 67; H3 K14, 27, 37, 38; H3.3 K27; H2B K49; H2A.Z K18, 29; and H2B.Z K8, 14, 18, 53. In P. knowlesi, only a subset of the sites were the same: H4 K8, 12, 31, 67; H3 K18, 23; H2B K35, 38, 49, 64, 100, 108; H2A.Z K14, 29; and H2B.Z K8, 14, 29. They were unevenly distributed across the histone tails (for example, none on H2A versus four on H4), and almost all KLa sites could also be acetylated (S5B Fig and S1–S4 Tables). Interestingly, despite the histone sequences being highly conserved between Plasmodium species, the majority of KLa sites were species-specific, suggesting that this modification may have species-specific function(s). As expected from Fig 1, lactylation of some residues was induced after exposure to 25mM lactate, but inducibility varied between residues and also between species (Fig 5B). It also varied in trophozoites versus mixed-stage cultures, implying that some residues may be lactylated stage-specifically.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. Histone lactylation occurs on multiple histone lysine residues in two Plasmodium species.

A: Sequence alignments of histone tails in P. falciparum and P. knowlesi, showing in red all KLa sites detected by mass spectrometry (either in trophozoite or mixed-stage cultures, + /- 25mM lactate). Numbers in pink flanking the peptide sequences are residue numbers of the N- and C-termini of the peptides. The tail of histone H2A is not shown because no KLa was detected on this histone. B: Pie charts representing the distribution of KLa sites from (A), and heatmaps representing normalized intensities of KLa-containing peptides (Z-score transformed). Hierarchical clustering was used to partition the KLa sites into 3 clusters with euclidean distance and ward.D clustering algorithm. The time windows for trophozoite cultures of P. falciparum and P. knowlesi were 24-26 and 12-14 hpi, respectively, and these were treated with lactate for the subsequent 12 or 8h.

https://doi.org/10.1371/journal.pgen.1011991.g005

Cross-talk may occur between histone lactylation, acetylation and methylation

Acetylation and lactylation were often detected on the same lysine residues, both in our data and previously in human cells [10]. (Since the same lysine linkage is used (Fig 1A), the same site on the same histone molecule cannot be simultaneously modified with both, but histones within a population of molecules can be modified with either.) Therefore we sought to investigate whether lactate exposure in the trophozoite stage could induce changes in the distribution of histone acetylation (Kac), as well as di- and tri- methylation (K(me2), K(me3)).

There was no dramatic change in the distribution of Kac sites after exposure to 25mM lactate, and Kac remained the most abundant histone modification, as previously reported [18] (S1–S4 Tables). Almost all KLa sites were also detected as Kac but the converse was not true: ~ 65% of Kac sites were unique (S5B Fig). After lactate exposure, 44% of histone KLa sites in P. falciparum and 17% of those in P. knowlesi were significantly induced (P-value ≤ 0.05) (Fig 6A, top panel and S1–S2 Tables), corroborating the observation from Fig 1 that lactylation was more strongly inducible in P. falciparum than in P. knowlesi. By contrast, only ~12% and ~7% of histone Kac sites, respectively, were significantly up- or down-regulated. Hence, relatively few Kac sites (amongst very many that can be acetylated) must be responsible for the modest rise in total histone acetylation after lactate exposure, which was detected by western blot in Fig 3. Overall, KLa was clearly the most inducible post-translational modification (PTM) after lactate exposure (S6A Fig), and sites that were strongly induced as lactylated were not simultaneously induced as acetylated (S6B Fig).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. Changes in PTMs on both histones and non-histone proteins are induced by exposure to high lactate.

Volcano plots representing (A) histones with differential lysine lactylation (KLa) and acetylation (Kac), in the top panels, and differential tri- or di-methylation (Kme3, Kme2) in the lower panels; (B) non-histone proteins with differential lysine lactylation (KLa). In all plots, data are from trophozoite cultures of P. falciparum or P. knowlesi, + /- 25mM lactate treatment. X-axes represent log₂ fold changes (FC) calculated from normalized intensities, and y-axes represent P-values. Colored dots represent PTMs that changed significantly after lactate exposure, based on their significance threshold indicated by P-value. (Grey dots represent those that did not change significantly.) Symbols for histones and non-histones are hatched and filled circles, respectively. Horizontal and vertical lines represent P-value and log₂FC significance thresholds. In (B), gene name abbreviations in P. falciparum, on the left: RNA MT, RNA methyltransferase; Histone AT, Histone acetyltransferase; LSD1, lysine specific histone demethylase; TsaD, TSaD-domain-containing protein; RBP25, RNA binding protein 25; RP L46, Ribosomal protein L46; G3P, glyceradehyde-3-phosphate dehydrogenase. In P. knowlesi on the right: snRNP PRP4, small nuclear ribonucleoprotein; HDAC SAP18, histone deacetylase subunit SAP18; Ser/Thr PK, serine/threonine protein kinase, CCR4-NOT, CCR4-NOT complex; PRP5 helicase, Pre-mRNA-processing ATP-dependent RNA helicase; SET2, SET2 histone methyltransferase.

https://doi.org/10.1371/journal.pgen.1011991.g006

In contrast to Kac, major changes did occur in K(me2) and K(me3) after lactate exposure: ~ 90% of Kme sites were significantly down-regulated in P. falciparum (Fig 6A, lower panel and S1–S2 Tables). This was not the case in P. knowlesi, where no reciprocal changes in K(me2) and K(me3) were seen.

Non-histone chromatin-associated proteins can be lactylated

Chromatin-associated proteins besides histones appeared in the mass spectrometry data, and many of these were also lactylated, with a subset of the modifications being induced after exogenous lactate exposure (Fig 6B and S3–S4 Tables). In P. falciparum, lactylated sites were found in chromatin-bound proteins such as Alba3, several RNA helicases, proteins implicated in epigenetics (e.g., histone deacetylase 2), in glycolytic metabolism (e.g., glycerol-3-phosphate dehydrogenase) and in transcriptional control (e.g., two ApiAP2 transcription factors). A third AP2 factor, AP2G, which induces gametocytogenesis, was also detected as lactylated, but its coverage fell below the threshold for inclusion, probably due to its very low expression in these asexual-stage cultures. An acetyl-CoA transporter was also lactylated, suggesting again the possibility of cross-talk between acetyl and lactyl epigenetic pathways (S7A Fig and S3 Table). In P. knowlesi, most of the same proteins were not detected as lactylated, although other proteins, including the acetyltransferase GCN5, were lactylated (S7B Fig and S4 Table).

Although almost all the lactylated lysines within histones could also be acetylated (S5B Fig), this was not the case in non-histone proteins. Unique, lactate-induced KLa sites appeared on several chromatin-associated proteins. These included epigenetic regulators (e.g., SAP18, associated with HDAC1) and RNA binding proteins such as Alba4 in the case of P. knowlesi, and RNA helicases (e.g., NAM7, responsible for RNA turnover), RNA binding proteins, AP2 transcription factors and gametocytogenesis factors (e.g., Male Development 1) in the case of P. falciparum (S1–S2 Tables).

Collectively, it appeared that the KLa modification was widespread, both on histones and on other proteins. It was modulated on some proteins by exogenous lactate, but also probably derived from lactate produced by the parasite. The wide range of lactylated proteins suggests that the modification could regulate multiple aspects of Plasmodium biology, from epigenetics to RNA biology to protein function.

Lactyl post-translational modification does not correlate with levels of protein expression

There was no consistent correlation between proteins that were inducibly modified and proteins that changed in abundance after exposure to 25mM lactate (S8 Fig). In P. knowlesi, no histones and very few other proteins showed significant changes (with a sca.P-value ≤ 0.01) after lactate exposure; in P. falciparum, a few proteins did change significantly in abundance, including histone H2A.Z and HP1, and interestingly the changes were mostly downregulation (S8A Fig). Overall, however, there was no correlation between proteins whose PTMs were significantly changed, and proteins that changed in abundance (S8B Fig). Therefore, protein lactylation does not appear to be a consistently ‘stabilizing’ nor ‘destabilizing’ PTM.

Proteins in different biological pathways are lactylated in P. falciparum versus P. knowlesi

It was apparent that the PTMs affected by lactate exposure, both on histones and on non-histone proteins, differed in P. falciparum versus P. knowlesi. Gene ontology (GO) enrichment analysis was therefore performed to gain insight into the pathways most affected in each species.

Amongst the proteins whose PTMs changed after lactate exposure, many different biological functions were found to be enriched (P-value ≤ 0.05), depending on both the species and the epigenetic modification (KLa, Kac or Kme). In P. falciparum, lactate-modulated KLa sites were significantly enriched in proteins involved with metabolism of tRNAs and indole compounds, for example; in P. knowlesi, the top GO terms associated with such KLa sites included ‘cell growth’, ‘cell division’ and ‘autophagy’ (Figs 7A and S9 and S5 Table).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 7. Gene ontology (GO) enrichment analysis of biological processes and cellular components enriched in PTMs after lactate exposure.

Charts represent the top 5 GO terms (biological processes (A), cellular components (B)) with the highest P-value for each PTM. Triangles mark the GO terms associated with proteins whose PTMs were upregulated; circles mark those down-regulated. Shapes are color-coded based on the magnitude of the fold change. The size of the shape, represented by Z-score, indicates the magnitude of the likelihood that the term is up- or down-regulated. P.f and P.k represent P. falciparum (all data on the left of the vertical dashed line) and P. knowlesi (right of dashed line). For simplicity, K(me3) and K(me2) are both presented as K(me). ‘KLa’ refers to all KLa sites detected; KLa.Sig refers to lactylated proteins that changed significantly upon lactate exposure in trophozoites.

https://doi.org/10.1371/journal.pgen.1011991.g007

Amongst GO terms for cellular components, KLa sites, as well as Kac and Kme sites, that responded to lactate exposure were enriched in proteins involved with the nucleosome and protein-DNA complexes. This was not surprising given the known predominance of such PTMs on histones, as well as the location of KLa observed in Fig 4, which was almost exclusively nuclear. In P. knowlesi, there was also an interesting enrichment of lactate-modulated KLa sites in the carbon CCR4/Not complex, which regulates aspects of protein function, RNA metabolism and transcriptional activity during late gametocyte differentiation and post-fertilization development [19] (Figs 7B and S10 and S5 Table). Amongst molecular function GO terms, both species showed enrichment of ‘protein dimerization’ and ‘catalytic activity’, amongst others (S11 Fig and S5 Table).

Blood lactate levels in malaria-infected mice correlate with distinct transcriptomic profiles in the causative parasites

The foregoing data were all derived from parasites in in vitro culture: an environment that may not full mimic the hyperlactataemic host. However, in Fig 1F, we investigated a rodent malaria model that becomes severely hyperlactaemic, P. yoelii 17XL, and showed that histone lactylation in the P. yoelii parasites was strongly induced in hyperlactataemic mice. Rodent malarias are commonly used as in vivo experimental systems, since human malaria patients are less experimentally accessible. Here, the P. yoelii model could reveal the biological pathways that histone lactylation influences in vivo.

RNAseq analysis was previously conducted in groups of mice with early-stage versus late-stage P. yoelii infections (i.e., low versus high blood lactate) [14], and we re-analysed those data for the parasite (rather than the host) transcriptome, seeking changes that correlated with histone lactylation in late-stage infections. For comparison, we analysed the same data from P. berghei ANKA, where late-stage infections did not consistently generate hyperlactataemia [14].

All datasets were first corrected for different mixes of parasite developmental stages [20], and indeed marked differences in the proportions of schizonts were detected in early versus late-stage P. yoelii infections, while gametocytes were found in low proportions in early infections and were undetectable in late infections (S6 Table).

Differential expression (DE) analysis [21] performed on the corrected data revealed distinct expression profiles in late versus early stages of infection in both species (S12A Fig). Large cohorts of genes were differentially expressed (P-value ≤0.05): 502 genes in P. yoelii (238 upregulated and 264 downregulated) and 1521 genes in P. berghei (803 upregulated and 718 downregulated) (S12B Fig and S6 Table). (Using more stringent criteria, P-value ≤0.05 and |log2FC| ≥ 1, expression changes were similarly distributed across fewer total genes (S12B Fig). Late-stage P. yoelii parasites showed a strong downregulation of the lactate dehydrogenase (LDH) gene, as well as the glycolytic enzyme phosphofructokinase (PFK) and several genes putatively involved in gametocytogenesis. Several tRNA synthetases were also downregulated, with the interesting exception of isoleucine tRNA synthetase (isoleucine is the only amino acid that the parasite must scavenge from its host). Indeed, genes with the strongest expression changes were generally downregulated, while fewer – including several histones – were upregulated (Fig 8A and 8B). The opposite was true in late-stage P. berghei, where gene upregulation was most apparent, including the upregulation of phosphofructokinase and glyceraldehyde 3-phosphate dehydrogenase, and of factors involved in DNA replication and repair such as MCMs and RecQ helicase (Fig 8A and 8B).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 8. In rodent malaria parasites in vivo, transcriptomic profiles and lactylation of representative genes both vary with levels of host lactataemia.

A: Volcano plots representing differentially expressed genes detected in late versus early infections of P. yoelii 17XL (top) and P. berghei ANKA (bottom). Data reanalysed from [14]. X-axes represent log₂ fold changes (FC) calculated from normalized CPM values; y-axes represent P-values. Coloured dots represent DE genes, based on their significance threshold indicated by adjusted P-value (FDR) (grey dots represent those below any significance threshold). Horizontal and vertical dashed lines represent P-value and log₂FC significance thresholds. Gene name abbrevations in P. yoelli: SP15, transcription elongation factor SPT5; TIFII, eukaryotic translation initiation factor 3 subunit L; HSP20, HSP20-like chaperone; nPrx, peroxiredoxin; PFK11, ATP-dependent 6-phosphofructokinase; H2B, histone H2B; H4, histone H4; SOD2, superoxide dismutase [Fe]; GEST, gamete egress and sporozoite traversal protein; L-LDH, L-lactate dehydrogenase; LDH, lactate dehydrogenase; FD3, female development protein FD3; H2B.Z, Histone H2B variant; MDV1, male development protein 1; FD1, female development protein 1. Gene name abbrevations in P. berghei: ApiAP2, AP2 domain transcription factor; GSK3beta, glycogen synthase kinase-3 beta; GNAT, N-acetyltransferase in GNAT family; MCM2/5/9, DNA replication licensing factor MCM2/5/9; RECQ1, ATP-dependent DNA helicase RecQ1; SET7, histone-lysine N-methyltransferase SET7; G3PDH, glycerol-3-phosphate dehydrogenase; RIP, PIR protein; Sir2A, histone deacetylase Sir2A; CAF1B, chromatin assembly factor 1 subunit B. B: Heatmaps representing late and early transcriptional profiles in log_2CPM (Z-score transformed) for DE genes (502 in P. yoelii, 1521 in P. berghei). Hierarchical clustering was performed to cluster genes’ expression profiles with euclidean distance and ward.D clustering algorithm. C: Boxplot showing KLa enrichment detected via ChIP in the CDS of five P. yoelii genes whose expression significantly changed in response to hyperlactataemia (sig. regulated), and nine genes whose expression did not (non sig. regulated). Y-axis represents the percentage of chromatin input recovery (% IPed) for each ChIPed region. One biological replicate for each of three mice (i.e., three lactate conditions). Dots represent individual genes, horizontal line represents median, box represents interquartile range, whiskers represent variability outside the interquartile range (quartile -/ + 1.5 * IQR). Multiple comparison of means was performed using ANOVA and Tukey HSD post hoc tests at p-value < 0.05 for between-group comparison. Left boxplot (ANOVA, F = 0.788; d.fsum = 2; p-value = 0.466), right boxplot (ANOVA, F = 23.73; d.fsum = 2; p-value = 6.76e-05). *** p-value ≤ 0.001; ** p-value ≤ 0.01; *p-value ≤ 0.05; ns, not significant. D: Biological process gene ontology (GO) enrichment analysis of DE genes in P. berghei and P. yoelii. Plots show the top 15 GO terms with the highest P-values. The outer circle displays scatterplots of the expression levels (log₂FC) for the genes (dots) in each GO term, whereas the inner circle is a bar plot where the height of the bar indicates the significance of the term (−log10 P-value). The colour corresponds to the Z-score, which indicates whether the term is likely to be decreased (negative value) or increased (positive value).

https://doi.org/10.1371/journal.pgen.1011991.g008

We then sought to test whether a selection of genes whose expression changed in hyperlactataemic hosts were differentially marked with lactyl lysine. For this, we turned from the reanalysis of published transcriptomic data to the gene-specific analysis of newly-generated chromatin, harvested from P. yoelii parasites in mouse hosts with low or high blood lactate. We used chromatin immunoprecipitation (ChIP) to check whether a selection of P. yoelii genes that were differentially expressed in hyperlactataemic hosts were also differentially marked with lactyl lysine (Figs 8C and S12C). Indeed, lactylation in the coding sequences of all five chosen genes correlated strongly with blood lactate. These genes encoded histone H4 (PY17X_0944400), Ap2 transcription factor G (AP2G, PY17X_1440000) and three factors involved in metabolism (Lactate dehydrogenase, PY17X_1345100; ATP-dependent 6-phosphofructokinase, PY17X_0922000; and isoleucine-tRNA ligase (Ile-tRNA), PY17X_1442500). By contrast, there was no correlation between blood lactate and lactylation of chromatin in nine control genes that did not show differential expression. Lactylation within 1kb upstream of the same five genes was more variable and less strongly correlated with blood lactate than lactylation within the gene body (S12D Fig). Thus, it is likely that lactylation of histones and/or other chromatin proteins within gene bodies directly regulates gene transcription, at least in the P. yoelii model.

When GO enrichment analysis was conducted on the up- and down-regulated genes, it was clear that late-stage P. yoelii parasites activated biological (Fig 8D) and molecular (S13 Fig) functions including proteolysis and mono-carboxylic acid metabolism, while late-stage P. berghei parasites activated contrasting functions including DNA replication and repair, chromosome segregation, cell division and actin/cytoskeleton dynamics. Amongst functional pathways, late-stage P. yoelii parasites upregulated fatty acid metabolism whereas P. berghei upregulated sugar metabolism and glycolysis (S13 Fig). Thus, it appeared that P. yoelii parasites in hyperlactaemic hosts showed features of ‘dormancy’. This phenomenon can be induced in cultured P. falciparum by isoleucine starvation and is characterised by slowing of the replicative cycle and proteolysis [22]. By contrast, P. berghei – in hosts without hyperlactataemia – conversely upregulated pathways involved in replication.

Ethics statement

All experimental protocols and procedures were performed in accordance of the UK Animals (Scientific Procedures) Act (PP8697814), and were approved by the Imperial College Animal Welfare and Ethical Review Board (for P. yoelii) or the University of Cambridge AWERB (for P. berghei). The Office of Laboratory Animal Welfare Assurance for the University of Cambridge covers all Public-Health-Service-supported activities involving live vertebrates (no. A5634-01). This study was carried out in compliance with the ARRIVE guidelines (https://arriveguidelines.org/). All studies were in accordance with the Laboratory Animal Science Association’s guidelines for good practice. Prior to any experimental interventions, all mice were acclimatized to animal facilities for one week.

Culture and synchronisation of P. falciparum and P. knowlesi

Plasmodium falciparum (NF54 strain) and Plasmodium knowlesi (A1-H.1 strain) were grown in RPMI 1640 media (Sigma, R4130) with 2.3 g/L sodium bicarbonate, 50 mg/L hypoxanthine (Sigma, H9377) and 25 µg/L gentamicin (Melford Laboratories, G38000-1), supplemented with 5 g/L Albumax II (Invitrogen, 11021037), 10% horse serum (Gibco, 10368902) and 4 g/L glucose for P. knowlesi, or with 2.5 g/L Albumax II and 5% human serum for P. falciparum. P. falciparum and P. knowlesi were cultured in human red blood cells (NHS Blood and Transplant) at 4% or 2% haematocrit respectively in 3% oxygen, 5% CO₂ and 92% nitrogen gas mixture at 37^oC.

For P. falciparum synchronization, parasites at ~ 6% parasitaemia were initially synchronised with 5% sorbitol (w/v in water) and allowed to grow to late schizont stage. ML-10 (LifeArc) [28,29] was then added for 2h at a final concentration of 75 nM. Schizonts were purified via a 65% Percoll (GE Healthcare) gradient (v/v in PBS), washed twice with complete media, and allowed to reinvade in 25% haematocrit in 5 ml of complete media in the abovementioned gas mixture at 37^oC for 3h with agitation at 220 rpm. 5% sorbitol was then used to remove residual schizonts and produce a tightly synchronised culture (0–3 hours post invasion (hpi)).

For P. knowlesi synchronization, parasites at ~ 6% parasitaemia were centrifuged through a 55% Nycodenz gradient (v/v in RPMI) from 100% Nycodenz stock pH 7.5 (27.6% w/v Nycodenz, 5 mM Tris HCl, 3 mM KCl, 0.3 mM CaNa₂·EDTA). Ring stages isolated in the bottom layer were allowed to grow to late schizont stage. ML-10 was then added to a final concentration of 150 nM for 2h. Parasites were centrifuged through a second Nycodenz gradient to collect schizonts from the top layer, which were washed twice with complete media and allowed to reinvade as above. A second Nycodenz cleanup was then performed to remove residual schizonts and produce a tightly synchronised culture (0–3 hpi).

Growth of P. yoelii and P. berghei

Eight-week-old wild-type female C57BL/6J or CD1 mice were obtained from Charles River Laboratories. All mice were specified pathogen-free, housed in groups of five in individually ventilated cages, and provided ad libitum access to food and water.

P. yoelii (17XL strain) was serially passaged through C57BL/6J mice, following previously described protocols [14]. Blood was collected via aseptic cardiac puncture from infected donor mice under non-recovery isoflurane anesthesia and diluted in sterile phosphate-buffered saline to achieve the desired parasite concentration. Experimental mice were subsequently infected with 10⁶ live parasites via intraperitoneal (i.p.) injection.

P. berghei (ANKA 2.34 strain) was maintained by serial passage through CD1 mice as in previously described protocols [30]. Experimental mice were pre-treated with 150µl of 6mg/ml phenylhydrazine i.p. three days prior to infection, to induce reticulocytosis. Blood from donor mice at 1–5% parasitaemia was collected by cardiac puncture under non-recovery Ketaset/Rompun anaesthesia and used to infected experimental mice by passage of 150µl blood, i.p..

Tail capillary blood samples were collected to prepare blood smears for parasitemia assessment and lactate measurement. Parasitemia was quantified through microscopy of thin blood smears stained with 10% Giemsa, following established protocols [14]. Heparinized blood was collected via cardiac puncture under non-recovery anesthesia, into syringes pre-loaded with heparin, and red blood cells were separated from plasma by centrifugation at 4,000g for 10 minutes. Levels of lactate in the cardiac blood of each mouse were immediately measured using the Lactate Pro 2 meter (HAB Direct) for P. yoelii experiments, or the L-lactic acid (L-Lactate) assay kit (Megazyme K-LATE) for P. berghei experiments. Red blood cells were harvested for parasite histone extraction from ten mice per rodent malaria species, of which seven yielded a viable amount of parasite material when infected with P. yoelii 17XL and five yielded an appropriate amount when infected with P. berghei ANKA 2.34.

Sodium L-lactate treatment of P. falciparum and P. knowlesi cultures

For titration experiments, P. knowlesi at 12–14 hpi and P. falciparum at 24–26 hpi, both at 1% parasitaemia, were washed into fresh media and then treated with 0mM, 5mM, 10mM or 25mM sodium L-lactate (Sigma), for 8h (P. knowlesi) and 12h (P. falciparum). Subsequently, samples were harvested and crude histone extractions were performed (as described below).

For mass spectrometry on histones from mixed-stage cultures of both species, parasites at a parasitaemia of ~ 6% were washed into fresh media to remove any endogenously-produced lactate and sodium L-lactate (Sigma), dissolved in incomplete media, was added to a final concentration of 25 mM. For mass spectrometry on histones from trophozoites, P. falciparum at 24–26 hpi and P. knowlesi at 12–14 hpi were treated with sodium L-lactate. Treatments were for 12 and 9h on the two species respectively.

Extraction of Plasmodium histones

P. falciparum and P. knowlesi cultures were collected by centrifugation at 800 x g for 5 minutes, followed by resuspending the pellet in 1.5 volumes 1X PBS. To release parasites from erythrocytes, saponin was added to a final concentration of 0.1% and incubated on ice for 10 minutes, followed by centrifugation at 12,000 x g at 4°C for 10 minutes. Parasite pellets were washed three times by adding ice cold 1X PBS (2 volumes of the saponin lysate) followed by centrifugation at 6000 x g for 2 minutes.

For P. yoelii and P. berghei, infected blood obtained by mouse cardiac puncture was placed in 50 mL RPMI HEPES modification media (‘schizont media’ - no FCS) and spun at 800 xg for 5 minutes, followed by an RPMI wash and centrifugation at the same speed. To remove host leukocytes, blood was passed through Plasmodipur filters (EuroProxima, 8011) using a syringe, followed by washing the filter with RPMI media to elute any remaining blood. The filtration procedure was repeated a second time before collecting the blood pellet by centrifugation at 800 xg for 5 minutes. Filtered blood was then mixed with 0.1% saponin on ice for 15 minutes to release parasites, which were then collected as above.

Parasites from all species were processed either for crude or acid-based histone extraction. For crude histone extraction, parasite cell pellets were lysed by adding 1X RIPA buffer (250 mM Tris-Hcl pH 8.0, 750 mM NaCl, 5% NP-40, 5% sodium deoxycholate, 0.5% SDS, 1mM phenylmethylsulfonyl fluoride, 1X cComplete EDTA-free protease inhibitor cocktail (Roche)) followed by three freeze/thaw cycles on dry ice, 2 minutes each. Laemmli buffer was then added to a final concentration of 1X, boiled at 95°C for 10 minutes and stored at -80°C for western blot analysis.

For mass spectrometry analysis, nuclei were liberated from parasite cell pellets by gently resuspending the parasite pellet twice in two volumes of hypotonic buffer A (10 mM Tris-HCl (pH 8.0), 3 mM MgCl₂, 0.2% v/v Nonidet P-40, 0.25 M sucrose, and a cocktail of EDTA-free protease inhibitors (Roche, 0469313200) followed by centrifugation at 4000 x g at 4°C for 10 minutes. The resulting chromatin pellet was then resuspended in two volumes of hypotonic buffer B (10 mM Tris-HCl, pH 8.0, 0.8 M NaCl, 1 mM EDTA [including protease inhibitor cocktail]) and incubated on ice for 10 minutes followed by centrifugation at 4000 x g at 4°C for 10 minutes. The chromatin pellet was then resuspended with eight volumes of 0.25 M ice-cold HCl followed by vigorous vortexing and rotation for 2h at 4°C. Acid-soluble recovered proteins in the supernatant were collected by centrifugation at 12,000 x g for 30 minutes at 4°C. Histone-containing supernatant was mixed with an equal volume of 30% TCA and incubated overnight rotating at 4°C, followed by centrifugation at 12,000 x g for 15 minutes at 4°C to collect the histone-containing pellet. This was then washed with 500 µl of ice-cold acetone and incubated at -20°C for 3h, followed by centrifugation at 12,000 x g for 15 minutes. Histone pellet was then air-dried and reconstituted in 50 µl ddH₂0. Protein concentration was measured using Qubit protein quantification kit (Invitrogen, Q33211). Purity of extracted histones was then assessed by SDS-PAGE on a 15% acrylamide gel stained with Coomassie blue G-250.

Western blotting

Protein extracts were loaded on 15% SDS polyacrylamide gels. For lactate-titration blots, approximately 10 µg and 100 µg were loaded for P. falciparum and P. knowlesi protein extracts respectively. For validation of histone lactylation prior to mass spectrometry, ~ 10 µg of acid-extracted histones were used. All gels were run at 120 V for 3h and transferred to 0.45 µM nitrocellulose membrane (GE healthcare, 10600020). For both chemiluminescent (Figs 1–3) and fluorescent (S4 Fig) blots, membranes were blocked in PBS-0.1% Tween 20 containing 5% bovine serum albumin, then probed with designated primary antibodies. For chemiluminescent detection, horseradish peroxidase (HRP)-coupled secondary antibodies were used and developed with ECL Plus Western Chemiluminescent HRP Substrate (ThermoScientific, 11527271). For fluorescent detection, fluorescent secondary antibodies were used. Blots were imaged with an Azure biosystems imager 500 and densitometry was performed with ImageJ for all western blots from titration and timecourse experiments. All lactylated histone signals were normalised using a parallel blot for total histone H4 (except for the fluorescence blot in S3 Fig, where lactylated histones were normalized using the H4 signal on the same blot).

Lactate assay

Lactate assays were performed using the L-lactic acid (L-Lactate) assay kit (Megazyme K-LATE). Media samples were taken from 5 ml cultures of P. falciparum NF54 and P. knowlesi A1-H.1, synchronised with Percoll at the ring stage, starting with 0.5% parasitaemia, every 24 h for 5 days. For two biological repeats the cultures were split every 48 h and for another two biological repeats the cultures were not split. The lactate assay was performed according to the manufacturer’s instructions (Megazyme K-LATE) on triplicate media samples of 2µl, plated in the 96-well format. The assays were read on a FLUOstar Microplate Reader (BMG LABTECH). Parasitaemia was verified by microscopy every time samples were collected. Results were analysed using GraphPad Prism.

Immunofluorescence assay

Smears were made from P. falciparum 3D7 HP1_HA + BSD (gift from Prof Till Voss [16]) and NF54, using mixed stage cultures at 4% parasitaemia. Smears were air dried, fixed in 4% paraformaldehyde/PBS for 10 minutes, washed in PBS, air dried and stored at 4°C. Parasites were permeabilised in 0.2% Triton X-100 for 15 minutes, washed in PBS, and blocked in 1% BSA/PBS with 0.1% Tween-20 for 1 h. Primary antibody labelling was done for 1 h, diluted 1:500 in 1% BSA/PBS, followed by three 5-minute washes with block (1% BSA with 0.1% Tween-20). Secondary antibody labelling was done for 1 h, diluted 1:1000 in 1% BSA/PBS, followed by one 5-minute wash with block (see Table 1 for antibodies). The slides were then incubated with DAPI (4’6-diamidino-2-phenylindole) (Thermo Fisher Scientific) at 2 μg/mL in PBS for 5 minutes, followed by one 5-minute wash with block. All incubation and labelling steps were done at room temperature. Slides were cured using ProLong Diamond Antifade Mountant (Invitrogen) at room temperature and stored at 4°C prior to visualisation. Slides were imaged using a confocal fluorescence microscope Zeiss LSM700. Images were pseudo-coloured and merged using Fiji-win64 (Image J).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Antibodies used in western blotting and immunofluorescence microscopy.

https://doi.org/10.1371/journal.pgen.1011991.t001

Mass spectrometry

RNA-seq data analysis

Raw reads from the RNA-Seq experiment generated by [14], belonging to Py17XL and PbANKA early and late infection stages, were downloaded, trimmed and quality checked by Trimmomatic v.0.35 [37]. Low quality reads with Phred quality score values lower than 30 and lengths less than 20 bp were removed. (https://www.bioinformatics.babraham.ac.uk/projects/fastqc/). Filtered reads were further quality controlled by FastQC v.0.11.7 (https://www.bioinformatics.babraham.ac.uk/projects/fastqc/) and used for transcript quantification using SALMON v.0.82 [38]. Py17XL and PbANKA reference transcript datasets were indexed using the quasi-mapping mode (--type quasi) of Salmon to create k-mers of lengths 31 (-K 31). For transcript quantification, input data sequence bias correction and bootstrapped abundance estimate were performed using the “seqBias” and “nom-Bootstraps 30” parameters from Salmon, respectively. Read counts and transcript per million reads (TPMs) were generated using tximport R package version 1.10.0 and lengthScaledTPM method [39] with inputs of transcript quantifications from tool SALMON. Low-expressed transcripts and genes were filtered based on analysing the data mean-variance trend. The expected decreasing trend between data mean and variance was observed when expressed transcripts were determined as those that had 3 of the 6 samples with count per million reads (CPM) 1, which provided an optimal filter for low expression. A gene was considered to be expressed if any of its transcripts with the above criteria was expressed. The TMM method was used to normalise the gene and transcript read counts to –CPM [40].

Prior to DE analysis and to control for heterogeneity within parasite developmental stages, we performed parasite stage deconvolution as follows. A signature matrix was taken from [20], which contains CPM-normalised gene expression measurements calculated across different parasite stages (male and female gametocyte, trophozoite, ring, merozoite, and schizont stages). P. berghei and P. yoelii gene names were converted to one-to-one orthologous P. falciparum gene names using an orthologous gene file obtained from the Malaria Cell Atlas. CPM-normalised counts alongside the signature matrix were then uploaded to CIBERSORT for deconvolution analysis using 100 permutations and disabling quantile normalisation (as recommended by CIBERSORT) to obtain the parasites cell proportions. This was performed separately for reads mapping to P. berghei and P. yoelii.

For DE analysis, limma suite was used [34,41] by fitting a linear model accounting for the cell proportions obtained in CIBERSORT. To compare the expression changes between conditions of experimental design, the contrast groups were set as Late-Early. For DE genes/transcript, the fold change of gene/transcript abundance were calculated based on contrast groups and significance of expression changes were determined using t-test. P-values of multiple testing were adjusted with BH to correct false discovery rate (FDR) [42]. A gene/transcript was considered significantly DE in a contrast group if it had adjusted p-value < 0.01 and |log2FC| ≥ 1.

Gene ontology (GO) pathway enrichment analysis

For RNA-Seq datasets, functional GO and pathway enrichment analysis were performed on significantly DE genes only, using PlasmoDB database [43] for annotation with default parameters. The GO terms (biological process, molecular function and cellular component) and pathways were identified to provide biological insights into the significance of DE using a P-value ≤ 0.05. For visualization of GO and pathways terms, combined with their associated gene expression profiles/significance, GOPlot R package v.1.02 was used [44].

For mass-spectrometry, functional GO enrichment analysis was performed using PlasmoDB database for annotation, with default parameters, on significantly differential proteins of the acid-extracted proteome and PTMs (for the trophozoite-stage experiment only). Exceptionally, for the KLa PTM, GO enrichment analysis was performed on both significantly and non-significantly changing lactylated proteins. Bubble plots were generated using GO terms combined with their associated genes expression profiles/significance across all PTMs and the acid-extracted proteome via in-house R scripts. For both datasets,

Where up and down are the number of genes up-regulated (logFC > 0) in the data or down- regulated (logFC < 0), respectively.

Chromatin immunoprecipitation (ChIP) and quantitative PCR (qPCR)

Blood was harvested from three mice infected with P. yoelii 17XL at different levels of parasitemia and lactatataemia. Purification of erythrocytes from host leukocytes and release of parasites were performed as described above. Parasites were crosslinked immediately with 1% paraformaldehyde (Sigma) by rotating for 10 mins at 37°C, then quenching with 0.125 M glycine for 5 mins at room temperature. Nuclei were liberated by gently resuspending the parasite pellet twice in two volumes of hypotonic buffer A (10 mM Tris-HCl (pH 8.0), 3 mM MgCl₂, 0.2% v/v Nonidet P-40, 0.25 M sucrose, EDTA-free protease inhibitors (Roche, 0469313200)) and holding on ice for 15 mins. This was followed by 15 strokes using a dounce homogenizer and centrifugation at 4000 x g at 4°C for 10 mins. The chromatin pellet was then resuspended in two volumes of hypotonic buffer B (10 mM Tris-HCl, pH 8.0, 0.8 M NaCl, 1 mM EDTA, protease inhibitor cocktail) and incubated on ice for 10 mins followed by centrifugation at 4000 x g at 4°C for 10 mins. 2–3 μg of chromatin were resuspended in 50 μl of shearing buffer (1% SDS, 50 mM Tris pH 8.0, 10 mM EDTA, 1X protease inhibitor cocktail) and sheared for 47 mins using an M220 sonicator (Covaris) at 5% duty cycle, 200 cycles per burst, and 75 W of peak incident power to generate 150–400 bp fragments.

Sheared chromatin was diluted 2x in ChIP dilution buffer (0.01% SDS, 1.1% Triton X-100, 1.2 mM EDTA pH 8.0, 25 mM Tris HCL pH 8.1, 150 mM NaCl, 1X protease inhibitor cocktail) and pre-cleared by incubating with protein A agarose beads (Thermoscientific, 20365) for 1 h at 4°C. 10% of the cleared chromatin was reserved as the input control and 90% was incubated for 4 h at 4°C using 0.75 μg of anti-KLa antibody coupled with protein A agarose beads. After one wash with low salt buffer (0.1%SDS, 1% Triton X-100, 2 mM EDTA, 20 mM Tris-HCL pH 8.1, 150 mM NaCl), one with high salt buffer (0.1% SDS, 1% Triton X-100, 2 mM EDTA pH 8.0, 20 mM Tris-HCl pH 8.0, 500 mM NaCl), and one with lithium chloride (0.25 M LiCl, 1% NP-40, 1% sodium deoxycholate, 1 mM EDTA, 10 mM Tris-HCl pH 8.1), the immunoprecipitate was eluted twice in 85 μl elution buffer (1% SDS, 0.1M NaHCO₃) by rotation at 65°C. ChIPed and input DNA were incubated with 2 μg RNase A (Roche, 10109169001) at 37°C for 30 min with gentle shaking. Samples were then resuspended in decrosslinking buffer (150 mM NaCl, 0.2% SDS, 600 μg Proteinase K (Ambion, AM2546)) by incubation at 65°C for 3 h with interval shaking. DNA was purified overnight at -20°C by adding 0.1 volume of 3M sodium acetate, 4 μl of linear polyacrylamide (Ambion, AM9520) and 2.5 volumes of ice cold ethanol, and collected by centrifugation at 13,000 xg for 30 mins at 4°C. DNA pellets were washed once with 70% ice cold ethanol and resuspended in 20 μl of nuclease free water.

qPCR was conducted with 3–4 ng of DNA in technical duplicate, using a QuantStudio 6 Pro Real-Time PCR System: 2 mins denaturation at 95°C, 40 cycles of 5s 95°C, 30s 60°C. Data were presented as the percentage of DNA immunoprecipitated relative to input (100*2^(Ct_input-Ct_eluates)). Primers were designed in the CDS and within 1kb upstream of the TSS using Primer3 and are listed in Table 2.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Primers used in ChIP-qPCR.

https://doi.org/10.1371/journal.pgen.1011991.t002