An essential amino acid synchronises malaria parasite development with daily host rhythms

The replication of blood-stage malaria parasites is synchronised to the host’s daily feeding rhythm. We demonstrate that a metabolite provided to the parasite from the host’s food can set the schedule for Plasmodium chabaudi’s intraerythrocytic development cycle (IDC). First, a large-scale screen reveals multiple rhythmic metabolites in the blood that match the timing of the IDC, but only one - the amino acid isoleucine - that malaria parasites must scavenge from host food. Second, perturbing the timing of isoleucine provision and withdrawal demonstrate that parasites use isoleucine to schedule and synchronise their replication. Thus, periodicity in the concentration of isoleucine in the blood, driven by host-feeding rhythms, explains why timing is beneficial to the parasite and how it coordinates with host rhythms. Blood-stage replication of malaria parasites is responsible for the severity of disease symptoms and fuels transmission; disrupting metabolite-sensing by parasites offers a novel intervention to reduce parasite fitness.

Introduction long isoleucine is absent, and that isoleucine removal does not elevate parasite mortality in these conditions. Our 80 findings are consistent with parasites using isoleucine as a time-of-day cue, and with the concept that parasite 81 control of the IDC schedule protects them from starvation and facilitates maximal exploitation of host resources. 82

83
Metabolites that associate with host feeding/fasting rhythms and the IDC 84 To identify metabolites whose rhythms correspond to the timing of host feeding and the IDC schedule, we 85 compared four groups of malaria infections in mice that were either wild type (WT) C57BL/6J strain or Per1/2-null 86 circadian clock-disrupted mice (previously backcrossed onto a C57BL/6J background for 10 generations). We generated 3 different groups of hosts whose feeding 90 rhythms differed with respect to the light:dark cycle and whether they had an intact TTFL clock, and a 4 th group of 91 hosts which lacked both feeding rhythms and an intact TTFL clock (Fig 1A). Parasites exhibit high amplitude IDC 92 rhythms in the 3 groups of mice with feeding rhythms (the phase of the IDC coinciding with the phase of host 93 feeding), and parasites in the 4th group show severely dampened rhythms (Fig 2; O'Donnell et al 2019). To explain 94 how the parasite IDC is linked to host rhythms, a time-cue/time-dependent resource must vary or have rhythmicity 95 across the day, with a peak concentration that associates with the timing of host feeding as well as the same parasite 96 IDC stage, across all 3 treatment groups with rhythmic feeding and a rhythmic IDC, yet be arrhythmic in the 4 th group 97 ( Fig 1B). Thus, we identified candidate metabolites by intersecting rhythmic metabolites in each of our treatment 98 groups. Specifically, having verified that the IDC schedules followed the expected patterns across the treatment 99 groups (Fig 2, Supp Table 1) we intersected metabolites rhythmic in dark (i.e. night) feeding (DF, n=18) and light (i.e. 100 day) feeding (LF, n=17) WT mice: in DF and LF mice parasites have inverted IDC timing but host TTFL clocks are intact, 101 and so any rhythmic metabolites in the blood could arise directly from via TTFL clock-driven processes or via TTFL 102 clock-independent food-processing rhythms ( Fig 1B). Next, we intersected those metabolites rhythmic in DF and LF 103 WT mice with rhythmic metabolites in Per1/2-null mice (whose metabolites appear in the blood from TTFL clock 104 independent processing of food (O'Donnell et al 2019)) that feed during a time-restricted 12 h window in constant 105 darkness (TRF mice, n=17; Fig 1B). Finally, we removed metabolites from the intersected list that remained rhythmic 106 in Per1/2-null mice with no feeding rhythm or TTFL clock (ALF mice, n=16) because parasites in these hosts exhibit 107 dampened rhythms (O'Donnell et al 2019). We then determined the phase of each candidate metabolite (using peak 108 time-of-day as a phase marker) in the intersection between the 3 groups with feeding rhythms and tested for an 109 association with the timing (phase) of the parasite IDC.  Fig 1B). This resulted in a list of 42 metabolites, consisting of 3 acylcarnitines, 11 amino acids, 9 biogenic 159 amines and 19 glycerophospholipids ( Fig 3A). We narrowed this list further by comparing whether the peak timing of 160 the patterns exhibited by each of these 42 metabolites corresponded to the timing of the host's feeding rhythm and

221
We found that lysophosphatidylcholine a C18:1 (lysoPC a 18:1), which contains an oleic acid side chain, is associated 222 with IDC rhythms. However, parasites are able to synthesise these fatty acids de novo via type II FA synthase (

236
Thus, both hosts and parasites are reliant on the host's food to acquire isoleucine.

237
Coinciding with the rise in isoleucine concentration during the feeding window, parasites make their transition 238 from trophozoites to schizonts before bursting and beginning development as ring stages at the end of the feeding 239 window (Fig 4). This suggests that the IDC schedule directly follows isoleucine rhythms in the host's blood -put Mean blood-glucose concentration differed between the groups, being higher in DF and TRF mice 278 (DF=8.55±0.14 mmol/L, TRF=8.59±0.13 mmol/L) than in LF and ALF mice (LF=7.90±0.14 mmol/L, ALF=7.68±0.11 279 mmol/L). We also found that glucose concentration varies throughout the day and that patterns differed between 280 our treatment groups. We used a change in the Akaike Information Criterion for small-sample sizes (

309
Specifically, we carried out two experiments in parallel to quantify how P. chabaudi's IDC progression is affected 310 when isoleucine is removed from media, and whether the IDC is then completed (defined as the proportion of 311 parasites that reach the schizont stage, Supp Fig 2) when isoleucine is returned. First, parasites cultured in the 312 absence of isoleucine (n=32 cultures from the blood of 8 mice, which were split equally across both treatments) 313 develop extremely slowly with approx. 3-fold fewer completing the IDC compared to parasites with isoleucine (50 314 mg/L, which is the same concentration as RPMI 1640) in their media (n=32 cultures) (Fig 6A). The best fitting model 315 contained only "treatment" (parasites cultured with or without isoleucine) as a main effect (AICc=0, Supp Table 7-316  A). The reduction in schizonts in isoleucine-free conditions was not due to a higher death rate because the density of 317 parasites remains constant during culture and did not differ between the treatments (Fig 6B)  revealed several metabolites that associate with the timing of host feeding, with isoleucine emerging as the best 351 candidate for a role in coordinating the IDC schedule (Fig 3, 4). Further, parasites are sensitive to the presence and 352 absence of isoleucine in their environment, slowing/pausing development when it is absent and then continuing 353 development as normal when isoleucine deprivation ends (Fig 6). Whilst isoleucine is not the only factor essential for 354 IDC completion, and we do not examine vitamins, cofactors, purines and folates which may also be crucial for 355 successful growth (Sherman 1979), our data revealed that isoleucine alone is sufficient to schedule the IDC.

356
Additionally, whilst the lack of rhythmicity in blood glucose concentration in infections with IDC rhythms does not 357 support glucose as a time-cue or scheduling force for the IDC. However, glucose may be indirectly involved.

358
Parasites that are glucose deprived fail to concentrate isoleucine (Martin and Kirk 2007), likely due to a lack of 359 glycolysis and ATP production needed to operate isoleucine transporters (Fig 7). High concentrations of isoleucine in