Pleiotropic Roles for the Plasmodium berghei RNA Binding Protein UIS12 in Transmission and Oocyst Maturation

Colonization of the mosquito host by Plasmodium parasites is achieved by sexually differentiated gametocytes. Gametocytogenesis, gamete formation and fertilization are tightly regulated processes, and translational repression is a major regulatory mechanism for stage conversion. Here, we present a characterization of a Plasmodium berghei RNA binding protein, UIS12, that contains two conserved eukaryotic RNA recognition motifs (RRM). Targeted gene deletion resulted in viable parasites that replicate normally during blood infection, but form fewer gametocytes. Upon transmission to Anopheles stephensi mosquitoes, both numbers and size of midgut-associated oocysts were reduced and their development stopped at an early time point. As a consequence, no salivary gland sporozoites were formed indicative of a complete life cycle arrest in the mosquito vector. Comparative transcript profiling in mutant and wild-type infected red blood cells revealed a decrease in transcript abundance of mRNAs coding for signature gamete-, ookinete-, and oocyst-specific proteins in uis12(-) parasites. Together, our findings indicate multiple roles for UIS12 in regulation of gene expression after blood infection in good agreement with the pleiotropic defects that terminate successful sporogony and onward transmission to a new vertebrate host.


INTRODUCTION
During the complex life cycle of Plasmodium, malarial parasites switch between different host cells, extra-and intracellular life cycle forms, and an invertebrate and a vertebrate host. As the parasites progress in their developmental program, and particularly after switching host, adaptation to new environmental conditions, including temperature and immune defense, is central. For instance, Plasmodium gametocytes must remain infectious in the blood of the vertebrate host until they are eventually transmitted to the vector during a blood meal. The gametocytes need to quickly adapt to new environmental factors, leave the digestive environment of the blood meal, and colonize the mosquito midgut (Billker et al., 1998;Billker et al., 2004). To swiftly accomplish a host switch, transcription and translation are tightly coordinated.
Transcriptional reprogramming is controlled by the apicomplexan Apetala-2 (Api-AP2) transcription factors (Balaji et al., 2005;Painter et al., 2011). Many genes needed for gametocytogenesis and early events of mosquito colonization are transcribed by the interplay of the Plasmodium transcription factors AP2-G and AP2-I, while asexual gene expression is repressed by AP2-G2 and AP2-FG controls female-specific gene regulation (Sinha et al., 2014;Yuda et al., 2015;Josling et al., 2020;Yuda et al., 2020). This transcriptional switch leads to the expression of mRNAs which are required for the multiple steps from gametocytogenesis to zygote formation until the motile ookinete stage is reached and the next checkpoint in transcription control is initiated by the AP2-O 1-4 family Modrzynska et al., 2017). Therefore, a tight fine-tuning is essential to synthesize tailored proteins at the right time.
Translational repression occurs during host switch to keep gametocytes and sporozoites in a latent state. Accordingly, translation of pre-synthesized mRNAs, which are required for mosquito colonization, occurs only after transmission and constitutes a mechanism to be prepared for the new host environment. In gametocytes a large number of genes have been described to be translationally repressed (Le Roch et al., 2004;Mair et al., 2006;Lasonder et al., 2008;Mair et al., 2010;Zhang et al., 2010;Lasonder et al., 2016). Translational repression is particularly dominant in female gametocytes and affects a wide range of mRNAs, including those that encode surface proteins, e.g., P25, P28, and CCP2, transcriptional regulators, e.g., AP2-O, and proteases, e.g., plasmepsin IV. A conserved U-rich 47 nucleotide-long cis-acting consensus motif in the 5' and 3' UTRs of a subset of translationally repressed transcripts in P. berghei gametocytes was shown to be important in translational regulation in gametocytes (Hall et al., 2005;Braks et al., 2008). In the female gametocyte, the DDX6-class DEAD box RNA-helicase DOZI (development of zygote impaired) is bound to its target transcripts in a large storage mRNP complex and silences their translation (Mair et al., 2006;Mair et al., 2010;Guerreiro et al., 2014). The Pumilio-family RNA binding protein Puf2 increases proliferation and decreases stage differentiation. As a consequence, deletion of P. berghei and P. falciparum Puf2 resulted in higher gametocyte rates and premature transformation of sporozoites in the mosquito salivary glands (Miao et al., 2010;Gomes-Santos et al., 2011;Muller et al., 2011;Lindner et al., 2013). In pfpuf2(-) parasites, an up-regulation of many gametocyte specific transcripts, mostly independent of those that are regulated in dozi(-)/cith(-), was observed (Miao et al., 2010). Specific binding of pfs25, pfs28, and Plasmepsin VI mRNA by the PfPuf2 protein was shown by RNA immunoprecipitation (Miao et al., 2013).
During host switch to the mosquito vector Puf2 negatively regulates gametocyte formation, and a post-transcriptional activator of gametocyte formation still remains elusive. In a complex with other proteins DOZI control of mRNA translation occurs much later, shortly after fertilization of female gametocytes inside the mosquito. Therefore, it is likely that additional factors participate in the regulation of posttranscriptional control of gene expression during mosquito transmission. In the present study we characterized an RNA binding protein, originally identified as up-regulated in infectious sporozoites gene 12 (UIS12), which is also highly expressed in gametocytes. A suppression subtractive hybridization screen aimed at discovering genes which are upregulated in salivary gland sporozoites first described the P. berghei RNA binding protein UIS12 (PBANKA_0506200) (Matuschewski et al., 2002). As expected, protein prediction programs indicate an intracellular localization. Comparison of transcriptome and proteome data indicate that P. berghei UIS12 is expressed in sporozoites, but the corresponding protein is not detectable, suggesting that UIS12 expression is regulated posttranscriptionally (Lasonder et al., 2008;Le Roch et al., 2004;Lindner et al., 2019). Due to the presence of the two RRM motifs, UIS12 is likely involved in post-transcriptional control of gene expression in gametocytes and pre-erythrocytic stages. Employing experimental genetics, the role of UIS12 for parasite life cycle progression was analyzed with an emphasis on a potential function in post-transcriptional regulation.

RESULTS
The RNA Binding Protein PbUIS12 Is Expressed at Multiple Life Cycle Stages We initiated our analysis by an NCBI conserved domain database (CDD) search (Marchler-Bauer et al., 2015), which revealed two~70 amino acid long RNA-recognition motifs (RRM), belonging to the class of RNA-binding domains (RBD) (Dreyfuss et al., 1988), in the P. berghei UIS12 protein sequence. The UIS12 gene, and particularly the two RRM domains, are highly conserved between different Plasmodium UIS12 orthologs ( Figure 1A), while in a candidate Toxoplasma gondii ortholog (TGME49_268380) the conservation is limited to the two RRM domains only. In general, the amino-and carboxy-terminal regions are less conserved, and all Plasmodium ortholog proteins have similar sizes of~1,300 amino acid residues. The PbUIS12 coding region has a total length of 4,070 base pairs and consists of two exons ( Figure 2A). Of note, the two regions encoding the RRM domains are separated by the single intron, which also exhibits a degree of conservation. We next measured UIS12 transcript abundance throughout the P. berghei life cycle by quantitative RT-PCR ( Figure 1B). In good agreement with the previous notion of up-regulation in salivary gland sporozoites, UIS12 mRNA steady state levels are low in midgut oocysts and highest in salivary gland sporozoites. High UIS12 transcript levels were also detected in early (24h) liver stages. In mixed asexual blood stages UIS12 mRNA levels are low, whereas in schizonts, gametocytes and ookinetes, considerable expression was detected. Our observations mostly corroborate published data (Howick et al., 2019;Otto et al., 2014). However, we did not specifically analyze ring stages or distinguished between male and female gametocytes (Yeoh et al., 2017). Together, UIS12 is expressed at multiple points of the P. berghei life cycle, indicative of multiple functions during life cycle progression and parasite stage conversion.

Targeted Gene Disruption of the RNA-Binding Protein PbUIS12
In order to study the gene function, we generated a UIS12 lossof-function mutant. PbUIS12 deletion was generated by double homologous recombination employing a replacement strategy, as described previously (van Dijk et al., 1995) (Figure 2A).
The UIS12 targeting construct (pDPBANKA_UIS12) comprised 500 bp each of the 5' and 3' non-coding sequences flanking a TgDHFR/TS pyrimethamine resistance cassette for positive selection with the anti-malarial drug pyrimethamine. Upon a double homologous recombination event, the UIS12 gene is expected to be replaced by the selectable marker ( Figure 2A). As recipient parasites we used P. berghei (strain: ANKA; clone: cl507), which expresses the fluorescent protein GFP under the control of the constitutive eIF1a promoter (Janse et al., 2006a). After transfection we recovered parental uis12(-) blood stage populations, from which clonal lines were generated by injection of limiting serial dilutions into mice. The desired recombination event was confirmed by diagnostic PCR, which confirmed the presence of 5' and 3' integration in the parental and the clonal uis12(-) populations, and by the absence of wild-type-specific PCR fragments in the clonal uis12(-) populations. During the course of our UIS12 knockout analysis and to further corroborate our findings we successfully generated an independent second uis12(-) knockout line (uis12(-) clone 2), in a different P. berghei ANKA recipient line, Bergreen, which constitutively expresses GFP under the control of the HSP70 promoter (Kooij et al., 2012) (Figure 2B: clone 2; Figure S1A: clone 1).
At this stage, we could already conclude that the successful generation of clonal uis12(-)-knockout lines demonstrates that the gene is not essential for normal progression of asexual blood stages, although UIS12 is expressed considerably in schizonts ( Figure 1B).

Impaired Gametocyte Development of uis12(-) Parasites
After successful selection of uis12(-) parasites, we first examined asexual and sexual blood stage growth. To this end, C57BL/6 mice (n=5) were infected by intravenous injection of either 10,000 uis12(-) or WT-infected erythrocytes. Starting three days after infection the mice were monitored by daily microscopic examination of thin blood films, and parasitemia and gametocytemia were quantified (Figures 2C,D). Parasitemia increased daily in wild-type-and uis12(-)-infected mice to a similar extent. On day 7, wild-type and uis12 (-)-infected mice started developing symptoms indicative of onset of experimental cerebral malaria, supporting the notion that uis12(-) parasites display virulence typically seen in infections with P. berghei ANKA parasites.
The blood films were also analyzed for the presence of male and female gametocytes ( Figure 2D). This analysis revealed that, in contrast to normal asexual blood infection, gametocytemia was significantly reduced in uis12(-) parasites as compared to the wildtype. Uis12(-) gametocyte numbers were already slightly reduced on day four after infection, when the first fully mature gametocytes became visible. The difference between WT and uis12(-) parasites increased further over time. Examination of size and morphology of male or female gametocytes and ookinetes revealed no difference between the two parasite lines ( Figure 2E). The presence of ookinetes indicates that uis12(-) gametocytes are able to fertilize and subsequently form a zygote that develops into an ookinete.
Both, male and female gametocytes were found in the expected female biased proportions. Together, the data indicate that UIS12 is dispensable for asexual growth, but plays an important role during gametocytogenesis. Analysis of the second uis12(-) clone corroborated these findings (Figures S1B, C).

Exflagellation Is Markedly Reduced in uis12(-) Parasites
The potential of male gametocytes to produce motile microgametes is an active process and termed exflagellation. In both uis12(-) clonal lines, exflagellation events were detected very rarely by microscopic examination of undiluted blood from uis12(-) infected mice at any given time point. Yet, this already showed that uis12(-) male gametocytes are able to produce motile male gametes, albeit at very low frequencies. To quantify the exflagellation centers, C57BL/ 6 mice (n=4) were injected intravenously with a high dose of 10 7 blood stages. Four days later, parasitemia, gametocytemia and the number of exflagellation centers were enumerated ( Figure 2F, uis12 (-) clone 1). Interestingly, in this experiment, the parasitemia was slightly, but significantly higher in uis12(-)-infected mice. In , clonal uis12(-) clone 2, and wild-type parasites served as templates. 5'-and 3'-integration-specific primer combinations amplify the predicted fragment only in the recombinant locus. Wild-typespecific primer pairs that do not produce a PCR fragment in the recombinant locus amplified the wild-type locus (right panel). Absence of residual wild-type confirms a clonal uis12(-) line (central panel). Time courses of parasitemia (C) and gametocytemia (D) of uis12(-) (clone 1, red) and wild-type (WT) (black) parasites, starting three days after intravenous injection of 10,000 mixed blood stages into C57BL/6 mice (n=5). Daily microscopic analysis of Giemsa-stained blood films was used to determine parasitemia and gametocytemia. (C) Parasitemia, percentage of asexual parasites per total erythrocytes (D) gametocytemia, percentage of gametocytes per total erythrocytes. Means of parasitemia and gametocytemia ( ± SD) are shown. *p < 0.05 (multiple t-tests, one per row). (E) Gametocytes and ookinetes of uis12(-) (clone 1) are morphologically indistinguishable from wild-type (WT). Shown are microscopic images of Giemsa-stained female (♀) and male (♂) gametocytes and ookinetes (Ook., from ookinete cultures). Brightfield microscopy, magnification 1,000-fold. Scale bars, 5 µm. (F) Asexual and sexual blood stage development. Infected blood (C57BL/6 mice; n=4) was analyzed four days after injection of 10 7 uis12(-) (clone 1, red) or WT (black)-infected erythrocytes, respectively. Plots show parasitemia (left), gametocytemia (center), and numbers of exflagellation centers per microliter of mouse tail blood (right). Results are expressed as means ( ± SD). *p < 0.05 (Mann-Whitney test). agreement with our first experiment ( Figure 2D), gametocytemia was 8.6-fold reduced in uis12(-) parasite-infected mice. Exflagellation was not visible under these conditions, i.e., 1:25 diluted blood, in uis12(-)-infected mice ( Figure 2F). Even after an extended incubation or when examined the following days, exflagellation was not detectable. In conclusion, male uis12(-) gametocytes retain the ability to progress to functional microgametes, but at a very low rate. The defect in exflagellation is more severe than the reduction in gametocytemia indicating additive defects in absence of UIS12 upon progression from blood infection to transmission to the insect vector.

Reduced Oocyst Formation By uis12(-)
Gametocytes, But Failure to Produce Sporozoites The severe defect in in vitro exflagellation implied a large impact in mosquito colonization. However, after transmission to Anopheles stephensi mosquitos, uis12(-) parasites were able to establish midgut infections and produce oocysts, showing that early events of midgut colonization are not abolished in vivo ( Figure 3). The mean mosquito infection prevalences were 54% ( ± 18%) for wild-type (n=5) and 46% ( ± 23%) for uis12(-) (n=5). This finding prompted us to quantify oocyst maturation. To this end, midguts were isolated daily starting on day 4 after mosquito infection up until day 14. Oocyst numbers and morphology as well as their ability to produce sporozoites were analyzed by fluorescence microscopy (Figure 3, uis12(-) clone 2). In uis12(-)-infected mosquitoes, the number of oocysts per midgut was more than 10-fold reduced as compared to wildtype, consistent with reduced gametocyte numbers and the very low exflagellation rate ( Figure 3B). Strikingly, in the absence of UIS12 oocysts were markedly smaller in size ( Figures 3A, C). While up until day 4 the oocyst size was comparable between both groups, over time uis12(-) oocysts displayed an abnormal growth pattern. Their size remained unchanged, whereas the WT oocyst diameter increased~10-fold over the following ten days. This finding indicates that the early events of oocyst development, including transformation of the motile ookinete to the spherical oocyst, and the initial increase in biomass (Carter et al., 2007) are reduced, but not completely abolished. An additional defect leads to arrested oocyst maturation. The analysis of the second uis12(-) clone corroborated these findings (Figures S1D-F).
Upon close examination at later time points, uis12(-) oocysts displayed little or no nuclear staining, harbored many large vacuoles, and frequently were surrounded by an apparently wrinkled oocyst wall ( Figure 3A, Figure S1F). As a consequence, no sporozoites were seen inside uis12(-) oocysts. This lack of sporozoite formation was confirmed by analysis of extracted midguts; no midgut-associated sporozoites were detectable. In WT-infected midguts sporoblast formation was observed inside the oocysts starting around day 11 after infection, and from day 12 onwards oocysts started to fill with sporozoites (Matuschewski, 2006)  In good agreement, initial microscopic screening of salivary gland extracts of uis12(-)-infected mosquitoes for presence of sporozoites did not return a single sporozoite. To obtain independent support for the complete absence of sporozoites in uis12(-)-infected mosquitoes, we inoculated C57BL/6 mice with salivary gland extracts of 20 uis12(-)-infected mosquitoes. In addition, we exposed C57BL/6 mice to~100 uis12(-)-infected mosquitoes ( Table 1, both clones). In sharp contrast to control infections with wild-type-infected mosquitoes under similar conditions, none of the attempts to transmit uis12(-) parasites resulted in blood stage infections ( Table 1). This defect in transmission confirms that uis12(-) parasites are unable to complete their development inside the mosquito.

Deregulation of Transcripts Important for Mosquito Transmission in uis12(-) Blood Stages
We hypothesized that due to the presence of the two RRM motifs, UIS12 is likely involved in post-transcriptional control of gene expression. Reduced gametocyte production, exflagellation, and oocyst development detected in uis12(-) parasites suggests that UIS12 might be regulating target mRNAs and alter expression of the respective proteins, which in turn play important roles in these stages.
To profile the transcriptome of uis12(-) parasites, we performed a comparative microarray analysis on uis12(-) and wild-type mixed blood stage RNAs (Figure 4, Figure S2). Since the first defect is detected during gametocyte formation, we limited our analysis to mixed blood stages, that include the asexual blood stages, from which gametocytes originate. Attempts to retrieve sufficient uis12(-) gametocytes for a microarray failed. The microarray analysis was performed with two biological replicates (R1 and R2), both with uis12(-) clone 1. For both replicates, parasitemia of the samples were comparable (R1: WT 6.2%; uis12(-) 6.3%; R2: WT 6.4%; uis12(-) 8.7%). As expected, gametocytemia was reduced in uis12(-) parasites (R1: WT 0.2%; uis12(-) 0.06%). Each microarray experiment was performed by dual-color hybridizations, and a color-swap dyereversal was included in order to compensate for dye-specific effects. In this microarray analysis, the -fold change of the expression levels of 2,890 P. berghei genes could be successfully compared ( Figure 4A, Figures S2A, D).
Overall, absence of PbUIS12 correlated with a global perturbation of 30% of all mRNA expression levels (threshold, <-2 or >2). More transcripts were down-regulated (646, 23%) than up-regulated (199, 7%). The mean down-regulation was -6.4-fold, whereas the mean up-regulation was +2.5-fold (GEO Series accession number: GSE152686). As expected, the UIS12knockout was independently confirmed by the microarray analysis; UIS12 transcripts were down-regulated by -22.3-fold ( Table 2). Within the 25 most down-regulated transcripts known proteins required for the switch to the mosquito vector, CCp1 (LAP2), LAP5, P48/45, and P28, as well as numerous previously unrecognized proteins are represented ( Table 2). The list of the 25 most up-regulated genes contains, among others, PIR proteins (Table S1). Remarkably, many transcripts that were downregulated more than 10-fold in uis12(-) blood stages, are reported to be expressed in gametocytes (Hall et al., 2005). We also analyzed the proportions of up-, down-, and nonregulated transcripts and found distinctions between stagespecific RNAs, as defined by RNA sequencing data . For instance, gametocyte transcripts were defined by >2fold up-regulation in gametocytes as compared to asexual blood a C57BL/6 mice were exposed to the bites of~100 infected mosquitoes for 15-20 min or injected intravenously (i.v.) with salivary gland extract isolated from 20 mosquitoes (uis12(-)) and, as a control, 10,000 sporozoites (WT). Data are pooled from two separate experiments. b the pre-patent period represents the days until detection of the first infected erythrocytes by microscopic thin blood film examination after sporozoite inoculation. N. A. not applicable. The colors were used to highlight different conditions. stages ( Figure 4B, Table 3, Figures S2B, E). In good agreement with normal replication rates, very few asexual blood stagespecific transcripts were down-regulated in uis12(-) parasites. Instead, the proportion of up-regulated transcripts was highest for asexual blood stage-specific mRNAs, which might reflect over-representation of these stages due to defects in gametocytogenesis. However, signature transcripts, like AMA1 (-0.3), HSP70 (-1.6), MSP1 (1.7) and MSP8 (1.5), were neither  up-nor down-regulated in the absence of UIS12 ( Figure 4B, Table 3). We also noted that the mRNA levels of many signature mosquito stage-specific proteins were down-regulated in uis12(-) mixed blood stages. In addition, we also compared the proportions of up-, down-and non-regulated transcripts and found distinctions between male-and female-specific transcripts, as defined by RNA sequencing data (Yeoh et al., 2017) ( Figure  4C, Figures S2C, F). While 67% of the female gametocyte transcripts were down-regulated in absence of UIS12, only 43% of male-specific transcripts were down-regulated, indicating differential impact of UIS12 on male and female gametocytespecific transcripts in mixed blood stages. Next, we wanted to independently confirm the microarray results by quantitative RT-PCR of selected transcripts on blood stage and gametocyte mRNA from uis12(-) and WT samples ( Figure 5). This analysis revealed a good agreement with the mean -fold changes determined in the microarray for gametocyteand blood stage-specific transcripts. The gametocyte-expressed transcripts Puf1, DOZI, UIS1/IK2, SET, and Actin II (Mair et al., 2006;Pace et al., 2006;Deligianni et al., 2011;Gomes-Santos et al., 2011;Muller et al., 2011) showed similar -fold changes in both assays. Similarly, the blood stage-expressed transcripts AMA1 and MSP1 remained un-regulated. For some transcripts, e.g., male development factor (MDV1) (Lal et al., 2009), the reduction in uis12(-) parasites was more pronounced by quantitative RT-PCR in comparison to the microarray data. qRT-PCR on gametocyte mRNA samples were consistent with the data from asexual blood stages ( Figure 5, uis12(-) clone 1). Overall, we could confirm the results of the microarray on these samples, indicating that the down-regulation of the gametocyte transcripts is a direct result of a specific UIS12-dependent regulation and not merely a consequence of the absence of gametocytes in uis12(-) mixed blood stages.

Gene Ontology Analysis of the Deregulated Transcripts in uis12(-) Parasites
We next performed a gene ontology (GO) enrichment analysis (Ashburner et al., 2000) (Tables S2 and S3). According to this analysis the 199 up-regulated and the 646 down-regulated transcripts are involved in 21 and 42 different biological processes, as described by distinct GO terms, respectively. Examples for GO terms of the set of down-regulated genes include reproduction, cell communication and phosphorylation (Table S2), in good agreement with the known roles of kinases in gametogenesis (Invergo et al., 2017). Examples for GO terms of the set of up-regulated genes include RNA processing and histone acetylation (Table S3), further indicative of perturbations in gene expression and mRNA translation in absence of UIS12.

A 10-Nucleotide Motif Shared By Up-Regulated Transcripts in uis12(-) Parasites
Since UIS12 contains two RNA recognition motifs and the combination of at least two RRM domains allows the continuous recognition of a 8-10 nucleotide-long sequence Shown are exemplary blood and transmission stage transcripts. The mean -fold change represents steady-state mRNA levels in uis12(-) parasites compared to the corresponding WT parasite stages. *genes with a knock-out phenotype similar to uis12(-). motif with high affinity (Maris et al., 2005), we screened all ORFs, and 1,000 base pairs of the flanking 5' UTR and 3' UTR sequences of genes up-regulated >2.5-fold in replicate 1 for common 8-10 nucleotide-long motifs using MEME (Bailey et al., 2009) using the respective portions of the non-regulated transcripts as reference. While this search did not return a motif in either the 5' or 3' UTR sequences, we found a 10 nucleotidelong motif (TTTYTTTTTC) in the ORF ( Figure 6A). This signature was found 751 times in 174 of the 220 up-regulated sequences (e-value: 4.2x10 -4 ). We restricted our search to upregulated transcripts, since down-regulation can at least be partially attributed to fewer gametocytes in mixed uis12(-) blood stages.

DISCUSSION
In this study, deletion of the candidate RNA binding protein UIS12 led to the discovery of a potential post-transcriptional regulator of gametocytogenesis and mosquito transmission. Our results underscore the key roles of a fine-tuned transcriptional control and post-transcriptional regulation during transmission of the malaria parasite between hosts. When UIS12 is absent, fewer gametocytes are produced and in vitro exflagellation is severely impaired, yet not abolished. The defects of uis12(-) parasites during host switching cannot be compensated for in subsequent steps in the Plasmodium life cycle, and uis12(-) infection leads to fewer oocysts, which fail to undergo oocyst maturation. We can correlate these defects with a set of signature gametocyte and mosquito-stage transcripts that are downregulated in uis12(-) parasites already during blood infection. Parasites lacking UIS12 produce~80% fewer gametocytes. Whether this reduction is due to a defect in sexual conversion or later during the maturation and/or survival of the sexual stages cannot be distinguished in the P. berghei in vivo model, where gametocytes are constantly formed. Both male and female gametocytes account for this reduction, indicative of a role of UIS12 in early events of gametocytogenesis, perhaps in conjunction with transcriptional control by AP2-G, AP2-I, AP2-FG, and AP2-G2. Gene deletion of AP2-G resulted in a complete loss of gametocytes and the deletion of AP2-G2 in a major reduction of gametocytes (Sinha et al., 2014). While AP2-G expression was unaffected in uis12(-) parasites, AP2-G2 was not included in our microarray screen. Plasmodium falciparum possesses at least 189 candidate RNA-binding proteins (RBP), which comprise 3.5% of all annotated genes. In good agreement with the complex regulation network required for stage conversion, a third of all RBPs show elevated expression during the gametocyte stage (Reddy et al., 2015). In contrast to a broad role in gene regulation, such as translational inhibition by phosphorylation of eIF2a, RRM-containing RBPs are mainly known for their tight RNA-specific control of RNA processing, export, and stability. The presence of tandem RRM domains, a feature observed in UIS12, ensures high sequence specificity (Maris et al., 2005).
Whether UIS12 is a positive regulator of gametocytogenesis and how it influences this process awaits further molecular and biochemical analysis, including RNA-pull down studies to capture UIS12 target transcripts. The role of UIS12 in promoting gametocytogenesis shares similarities with another RNA-binding protein, the pumilio-familiy RNA-binding protein Puf1 which was shown to function in P. falciparum gametocyte maturation, and especially of female gametocytes (Shrestha et al., 2016). Both proteins likely inhibit asexual propagation and, thereby, promote gametocyte formation. A related pumiliofamiliy RNA-binding protein, Puf2, exerts the corresponding A B function in sporozoites and prepares them for transmission to the vertebrate host (Miao et al., 2010;Gomes-Santos et al., 2011;Muller et al., 2011). We note that UIS12 likely executes a second function, similar to the one analyzed in the present study, in preerythrocytic stages, the stage where it was originally identified (Matuschewski et al., 2002), but the complete block in sporogony prevented us from studying this by classical reverse genetics. It, thus, still remains unclear if UIS12 also plays a role in the host switch back to the mammalian host. Of note, in sporozoites translational repression of UIS12 was described, and the protein is only detectable in liver stages (Lindner et al., 2019). Accordingly, a stage-specific knockout, which depletes UIS12 during sporogony is required to study its cellular roles for sporozoite transmission and liver infection. The few remaining uis12(-) gametocytes were able to form male and female gametes, fertilize, and form ookinetes. Accordingly, numbers of oocysts are extremely low in uis12 (-)-infected Anopheles mosquitoes. Complete life cycle arrest in uis12(-) infections occurs at an early time-point after oocysts are formed, indicating that the events leading to oocyst formation are not impaired (Carter et al., 2007).
Our microarray analysis revealed a sharp drop in the steady state mRNA levels of numerous well-characterized Plasmodium genes. The P. berghei LCCL protein family members CCp1 and lectin adhesive-like protein 5 (LAP5) are expressed in female gametocytes and ookinetes. PbLAP proteins are important for sporogenesis inside oocysts, yet known to be expressed already in gametocytes (Pradel et al., 2004;Trueman et al., 2004;Raine et al., 2007;Lavazec et al., 2009;Saeed et al., 2010). The surface proteins of the 6-cysteine repeat protein family P48/45 are essential for male gametocyte fertility (van Dijk et al., 2001). P25 and P28 are surface proteins of sexual stages, with distinct roles in ookinete to oocyst transformation (Siden-Kiamos et al., 2000;Tomas et al., 2001). Interestingly, some transcripts that are severely depleted, including P28, AP2-O, CCP2, and CCP4, have been reported to be translationally controlled (Vervenne et al., 1994;Saeed et al., 2013;Lasonder et al., 2016). Strikingly, the function of some of these repressed targets extends far beyond the ookinete stage, and the corresponding proteins are important for oocyst and sporozoite formation, illustrating the complexity of regulation of post-transcriptional control of gene expression in gametocytes (Lasonder et al., 2016). A deregulation of mRNAs already in mixed blood stages in absence of UIS12 might partially explain the observed inability to complete oocyst development. However, the abundant UIS12 mRNA levels in ookinetes suggest UIS12 presence at that stage and a potential additional checkpoint of post-transcriptional control by UIS12 possibly after transcription of mosquito stage-specific target genes is orchestrated by a family of closely related Apetala-2 proteins, AP2-O 1-4, in ookinetes Modrzynska et al., 2017). And this additional checkpoint in ookinetes might be the reason for the pleiotropic effects observed upon absence of UIS12, finally resulting in the aberrant maturation of oocysts. Due to the low UIS12 transcript levels in oocysts a direct role at this point in the life cycle is less likely. We, thus, propose that UIS12 exerts multiple regulatory roles at consecutive life cycle stages, and, hence, is together with Puf2 one of the first known Plasmodium RNA binding proteins with pleiotropic functions for life cycle progression (Gomes-Santos et al., 2011;Lindner et al., 2013;Miao et al., 2013;Miao et al., 2010;Muller et al., 2011).
The GO-term enrichment of down-regulated transcripts in uis12(-) shows perturbation of many kinases and GO-term enrichment of up-regulated transcripts revealed perturbation of gene expression. A MEME search in the ORFs, 5' UTRs and 3' UTRs of the up-regulated transcripts identified in the microarray did reveal a 10-nucleotide U-rich motif in the ORFs unique to the analyzed subsets when compared to sequences that were not perturbated in absence of UIS12. This 10 nucleotide-long U-rich motif might represent a signature of UIS12 interaction with target mRNAs during blood infection, which might be required for proper preparation for host switching. We wish to point out that down-regulation of gametocyte signature transcripts in uis12(-) blood stages can be largely attributed to the at least 3.5-fold reduction in gametocyte numbers, yet the levels of down-regulation ranged from a low (-2, e.g., Puf2) to more than 15-fold (e.g., CCP1 and CCP2). By qPCR we could confirm the down-regulation of some of the gametocyte specific transcripts such as P28 in a gametocyte enriched mRNA preparation of uis12(-) blood stage parasites, indicating that the down-regulation of gametocyte transcripts might not only be a consequence of the absence of gametocytes in uis12(-) mixed blood stages, but a direct result of a specific UIS12dependent regulation.
Further analyses employing RNA-Seq or single cell transcriptomics on synchronized and purified stages, ranging from ring stages to ookinetes, are warranted to gain an in-depth understanding of the mRNA repertoire that is under UIS12 control. Since the shared signature motif was identified in upregulated transcripts, this would argue against a role of UIS12 as an mRNA stabilizing factor. At least three alternative scenarios emerge (Figure 7). One potential, albeit unusual, mechanism of UIS12 could be binding and stabilizing of large set of target mRNAs that encode gametocyte-and mosquito stage-specific proteins. Alternatively, UIS12 acts on one or few mRNAs, which in turn drive expression control, for instance by stabilizing transcripts that encode checkpoint activators. A reverse, and perhaps more likely function would be an inhibitory role towards target mRNAs or a partner protein, which in turn modulate the half-lives of target mRNAs. According to such a scenario, upregulation of transcripts in the absence of UIS12 might be a consequence of increased mRNA stability. We consider the identification of a nucleotide motif a good starting point for future work, and are hesitant to propose a molecular mechanism at this early stage of discovery.
In gametocytes, the conserved DOZI/CITH RNA storage complex (Mair et al., 2006;Mair et al., 2010;Guerreiro et al., 2014) is central to translational repression. For instance, the DOZI/CITH complex is involved in translationally repressing P25, P28, PlasmepsinVI, AP2-O, and GAP45. These transcripts are also contained on our list of down-regulated transcripts in uis12(-) blood stages. Thus, we were curious how many targets overlap when we compared the reported 117 genes downregulated in both dozi(-) and cith(-) as determined by microarray on gametocytes (Mair et al., 2010) with the 646 down-regulated transcripts in uis12(-) blood stages (<-2-fold change), we found a surprisingly high overlap, of 102 of the 117 transcripts ( Figure 6B). The RNA-helicase DOZI was previously shown to be a key regulator of P25 and P28 transcript stability, as well as of 731 other transcripts, and they all share a conserved U-rich nucleotide motif mainly in the 5' and 3' UTR (Hall et al., 2005;Braks et al., 2008;Guerreiro et al., 2014). The abundance of U residues in the DOZI and UIS12 motifs could hint at a related mechanism. The key differences are that the U-rich motif in our study is found in the ORFs of target genes, while the U-rich motif bound by the DOZI complex is located in 5' and 3' regions of its targets (Braks et al., 2008). Also, the U-rich motif identified for DOZI binding has a length of 47 nucleotides, whereas the motif identified herein is only 10 nucleotides long. Despite these clear distinctions a potential functional link of these two proteins remains an attractive possibility, since UIS12 and DOZI have >100 overlapping down-regulated targets ( Figure 6B).
We note that, in marked contrast to UIS12, gene deletion of DOZI leads to a developmental arrest of the fertilized female gamete, a process that is not abolished in the absence of UIS12. In all three knockout lines, uis12(-), dozi(-), and puf2(-) parasites the mRNA levels of P25 and P28 were perturbated. While in dozi(-) and uis12(-) the transcripts were down-regulated, in puf2(-) an up-regulation was described, indicating involvement in distinct shared and opposing regulatory processes. On the other hand, PlasmepsinVI is translationally repressed by PfPuf2 (Miao et al., 2013) and reduced in uis12(-) parasites, similar to the meiotic recombinase DMC1 (Mlambo et al., 2012). Both knockouts as well as the CCP knockouts resemble the defects of uis12(-). Clearly, further work is warranted to decipher the molecular mechanisms that orchestrate the sequential post-transcriptional control events and contributions of the regulatory molecules.
In conclusion, our experimental genetics analysis of the Plasmodium berghei candidate RNA binding protein UIS12 uncovered broad perturbations of mRNA expression during blood infection, including up-regulation of several hundred transcripts, which share a U-rich 10 nucleotide-long motif in their open reading frames. The observed mRNA deregulation correlates with pleiotropic defects after blood stage propagation during sexual stage development, mosquito midgut colonization and oocyst maturation, ultimately leading to a complete block in the Plasmodium life cycle inside the Anopheles vector.
FIGURE 7 | Potential roles of Plasmodium UIS12 in expression control of gametocyte-and mosquito stage-specific mRNAs. UIS12 appears to influence parasite stage conversion at three different points in the complex Plasmodium life cycle (top). During blood infection commitment to onward transmission is initiated and Plasmodium parasites undergo a series of developmental transitions (center); female (blue) and male gametocytes (purple) are formed in the mammalian host; inside the mosquito midgut male gametes (red) fertilize female gametes (green), leading to formation of a motile zygote, the ookinete (yellow); oocyst development leads to formation of sporozoites (blue), which upon another mosquito bite invade hepatocytes (pink) and form liver stages (blue). Note that the last checkpoint (grey) could not be addressed experimentally because of complete life cycle arrest prior to sporozoite formation. The candidate RNA-binding protein UIS12 might modulate expression of gametocyte-and mosquito stage-specific transcripts by one of at least three alternative mechanisms (bottom); i) stabilization and/or translational repression of a large set of target mRNAs (left); ii) binding and stabilization of few mRNAs that in turn encode checkpoint activators (center); iii) inhibition of partner protein(s), which in turn stabilize target mRNAs (right). In the absence of UIS12 post-transcriptional expression of a large set of gametocyte and mosquito stage -specific genes is impaired resulting in reduced differentiation and, ultimately, arrest of life cycle progression.

Ethics Statement
All animal work conducted in this study was carried out in strict accordance with the German "Tierschutzgesetz in der Fassung vom 18. Mai 2006 (BGBl. I S. 1207)", which implements the directive 86/609/EEC from the European Union and the European Convention for the protection of vertebrate animals used for experimental and other scientific purposes. The protocol was approved by the ethics committee of MPI-IB and the Berlin state authorities (LAGeSo Reg# G0469/09, G0294/15).

Parasites and Experimental Animals
P. berghei ANKA cl507 parasites that constitutively express GFP under the PbEF1a promoter and P. berghei ANKA Bergreen parasites that express the green fluorescent protein GFP under the PbHSP70 promoter throughout their life cycle have been used in our experiments Kooij et al., 2012).
All mouse experiments were performed with female NMRI or C57BL/6 mice purchased from Charles River Laboratories (Sulzfeld, Germany) or bred in house. NMRI mice were used for transfection experiments, blood stage infections, and transmission to Anopheles stephensi mosquitos. C57BL/6 mice were used for sporozoite and blood stage infections, exflagellation essays and ookinete cultures.

Gene Deletion of P. berghei UIS12
The P. berghei UIS12 gene deletion was performed by double crossover homologous recombination of a replacement pB3Dknockout plasmid (van Dijk et al., 1995). The gene knockout construct targeting the endogenous UIS12 locus comprised a 647 bp UIS12 5' UTR and a 524 bp UIS12 3' UTR sequence flanking a T. gondii DHFR/TS pyrimethamine-resistance cassette. Table S4 lists the oligonucleotides used for generation of the knockout construct. Prior to transfection, plasmids were linearized by restriction digest. A digest with SacII and KpnI produced a fragment consisting of the pyrimethamine resistance cassette flanked by the UIS12 5' and 3' UTRs. Transfection by electroporation of cultured schizonts was performed as previously described (Janse et al., 2006b). Two independent parasite lines, one in P. berghei Bergreen and one in P. berghei cl507 recipient lines were generated and phenotypically characterized. For genotyping of recombinant parasites, knockoutand wild-type-specific PCR fragments were amplified from genomic DNA of parental and clonal lines with the oligonucleotides listed in Table S4.

Quantitative Real-Time PCR (qRT-PCR)
Quantitative real-time PCR was used to quantify UIS12 gene expression and for comparing gene expression in uis12(-) and wild-type (WT) mixed blood stages and gametocytes. For mixed blood stage analyses, C57BL/6 mice were infected i.v. with 10,000 iRBC, and 7 days after infection whole blood was isolated. The gametocyte samples were generated by infecting phenylhydrazinetreated NMRI mice with 10 6 iRBC. Three to four days later, when gametocytemia reached the highest levels, 12.5 mg sulfadiazine was supplemented to 1 liter of drinking water. Within 48 h the sulfadiazine treatment eliminated the asexual stages, which was confirmed by a Giemsa-stained thin blood film. Next, the infected blood was harvested. To isolate mixed blood stages or gametocytes, whole blood was cleaned from white blood cells and erythrocytes using cellulose-glass bead columns and saponin (0.3%; Sigma-Aldrich) lysis. Parasites were stored in 1 ml TRIzol (Invitrogen) or RLT-Buffer with b-mercaptoethanol (Qiagen). Parasite RNA was isolated with the RNeasy Kit (Qiagen) or TRIzol-Reagent (Sigma-Aldrich). DNase treatment with Quiagen RNase free DNase Set or Ambion Turbo DNA free Kit was performed to remove residual genomic DNA.
Reverse transcription to cDNA was done using the RETROscript Kit (Ambion). The qPCR was performed on parasite cDNA that was generated from total RNA of selected P. berghei life cycle stages. Quantitative real-time PCR was performed as previously described (Silvie et al., 2008) using Power SYBR Green PCR Master Mix and the StepOnePlus Real-Time PCR System (both Applied Biosystems). All procedures were performed according to the manufacturer's instructions. Life cycle expression data were normalized to GFP that is constitutively expressed under the EF1a promoter in P. berghei 507cl. Mixed blood stage and gametocyte expression data for comparison of knockout and WT parasites were normalized to PbHSP70. The relative transcript abundance was determined by the 2 DDCT method. For oligonucleotide sequences refer to Table S4.
Plasmodium Life Cycle: Parasite Asexual and Sexual Growth, Mosquito Infection, and Sporozoite Transmission For asexual blood stage and gametocyte growth curves, C57BL/6 mice were injected intravenously with 10,000 infected red blood cells. From day 3 to 7, Giemsa (VWR)-stained thin blood films were microscopically examined to determine parasitemia and gametocytemia. Parasitemia and gametocytemia are defined as the percentage of red blood cells infected with asexual stages or gametocytes, respectively.
To analyze male exflagellation, C57BL/6 mice were intravenously injected with 10 7 infected red blood cells (iRBC). After four days, parasitemia and gametocytemia were determined. Next, five microliters of tail blood were diluted 1:25 in ookinete medium (RPMI 1640, 10% FCS and 50 mM xanthurenic acid, pH 8.0) and incubated for 12 min at room temperature. For the next 6 min exflagellation centers were counted with help of a Neubauer chamber.
Ookinete cultures were set up as previously described (Weiss et al., 1997). Four days after the intravenous injection of 10 6 iRBC into C57BL/6 mice pre-treated one day earlier with phenylhydrazine 97% (Sigma-Aldrich, 100µl of 6.18 µl/ml) 1 ml of fresh mouse blood was mixed with 10 ml of ookinete medium. After 20 h incubation at 20°C, at 80% humidity the culture was examined for development of ookinetes and ookinetes were purified with anti-P28-coupled magnetic beads.
Anopheles stephensi mosquitoes were kept at 28°C (noninfected) or 20°C (infected) under a 14 h light/10 h dark cycle at 80% humidity. Four days after the transfer of 10 7 iRBC, when exflagellation reached a maximum, mosquitoes were infected as described (Vanderberg, 1975). After 4 to 14 days, midguts were isolated and oocyst development analyzed by fluorescent microscopy (Zeiss AxioObserver Z1). For visualization of nuclei, HOECHST 33342 (Invitrogen) was added at a 1:1,000 dilution. Images were analyzed using the FIJI processing (oocyst diameter) and/or Photoshop (oocyst numbers) software.
Midguts and salivary glands were analyzed at day 14-18 and day 18-30, respectively, for presence of sporozoites.
To test Plasmodium transmission, age-matched female C57BL/6 mice were either subjected to the bites of~100 infected mosquitos or inoculated with 10,000 sporozoites, or for uis12(-) the salivary gland extract of 20 mosquitos, in RPMI intravenously. After three days Giemsa-stained thin blood films were analyzed for presence of infected red blood cells.

Microarray Expression Profiling of P. berghei Blood Stages
The microarray was performed on two biological replicate samples of uis12(-) and wild-type with comparable parasitemia. One replicate was generated by infecting one C57BL/6 mouse each with 10,000 uis12(-) or wild-type blood stages intravenously (R1, biological replicate 1). After 7 days of infection, mixed blood stages were isolated from whole blood. For both samples, the parasitemia was comparable (uis12(-) 6.3%, WT 6.2%). The other replicate (R2, biological replicate 2) was generated by injecting one NMRI mouse with 10 7 uis12(-) or wild-type blood stages and mixed blood stage parasites were harvested after 3 days (uis12(-) 8.7%, WT 6.4%).
Whole blood was isolated and cleaned from white blood cells and erythrocytes using cellulose-glass bead columns and saponin (0.3%; Sigma-Aldrich) lysis. About 10 8 parasites were stored in 1ml TRIzol (Invitrogen). Total RNA was isolated using glycogen as a carrier according to the 'TRIzol Reagent RNA preparation method' (Invitrogen). The amount of RNA was determined by OD 260/280 nm measurement using a 'NanoDrop 1000' (peQlab) spectrophotometer. The RNA size, integrity and the amount of total RNA were measured with a 'Bioanalyzer 2100' using an 'RNA Nano 6000 Microfluidics Kit' (both Agilent genomics). RNA labeling was performed with the two color 'Quick Amp Labeling Kit' (Agilent genomics) using a 1:5 mixture of 'FullSpectrum MultiStart Primer' (Systembio). mRNA was reverse transcribed and amplified using the primer mixture. The RNA was split and labeled with Cyanine 3-CTP and Cyanine 5-CTP, respectively. After precipitation, purification and quantification, labeled samples were hybridized to the Agilent '4x44K custom-commercial microarrays' according to the manufacturers protocol. Microarray experiments were performed as dual-color hybridizations. A color-swap dyereversal was performed in order to compensate specific effects of the dyes and to ensure statistically relevant data analysis (Churchill, 2002;Wernersson et al., 2007). Scanning of the microarrays was performed with 5 µm resolution and the extended mode using a 'High Resolution Microarray Laser Scanner' (G2505, Agilent Technologies). Raw microarray image data were extracted and analyzed with the 'G2567AA Image Analysis/Feature Extraction software' (Version A.10.5.1.1, Agilent Technologies). On reporter and sequence level, the extracted MAGE-ML files were further analyzed with the 'Rosetta Resolver Biosoftware' (7.2.2.0 SP1.31). All subsequent data analysis was performed with the Microsoft Excel, Prism and the PlasmoDB (http://plasmodb.org/plasmo/) database. P. berghei microarrays were designed on the OligoWiz (http:// www.cbs.dtu.dk/services/OligoWiz/) server that included annotated open reading frames (ORF) from the whole genome of P. berghei. ORF specific probes were designed as sense oligonucleotides of the coding strand (Wernersson et al., 2007). The aimed oligonucleotide length was between 50 and 60 bases. To avoid self-hit, the maximal homology was 97%, the cross hybridization maximum length was at 80% and random primed position preference scores within OligoWiz were chosen. Depending on these parameters and the sequence input length, each ORF was covered by different numbers of specific oligonucleotide probes. In total 41,133 specific probe sequences were uploaded to eArray (Agilent Technologies, https://earray. chem.agilent.com/earray/) as expression application with a customer specified feature layout.
The microarray data discussed in this publication have been deposited in NCBI's Gene Expression Omnibus (Edgar et al., 2002) and are accessible through GEO Series accession number GSE152686.

Statistical Analysis and Online Bioinformatic Tools
Statistical significance was determined with the help of Graph Pad Prism software, using Mann-Whitney test for comparison of parasitemia, gametocytemia, and exflagellation dot-plots and multiple t-tests (one per row) for all growth curves.
The identification and positioning of the two RRM domains of the UIS12 orthologous proteins has been performed by NCBI WEB batch-CD Search Tool (Marchler-Bauer et al., 2015). Protein comparison was done by NCBI protein blast (blastp) with default settings, no compositional adjustment and without low complexity filter. Sequence coverage of all Plasmodium orthologues was higher than 90%.
Gene ontology (GO) enrichment analysis (Ashburner et al., 2000) was performed with help of the Gene Ontology enrichment analysis tool from PlasmoDB, with the default settings and the redundant GO terms were removed with help of the REVIGO software.
To screen for common 8-10 nucleotide-long motifs we used the MEME webtool (Bailey et al., 2009), screened all ORFs, 5'UTR sequences and 3'UTR sequences (1,000 nucleotides per gene for ORFs) of transcripts up-regulated >2.5-fold and compared them to the respective gene parts of the nonregulated (-1.5 to -1.5) transcripts from our microarray analysis of biological replicate 1.
Parts of the illustrations used in Figure 7 were obtained from Servier Medical Art, (http://www.servier.fr/servier-medical-art) licensed under a Creative Commons Attribution 3.0 Unported License.