Permalink
Cannot retrieve contributors at this time
Fetching contributors…
Cannot retrieve contributors at this time
| <?xml version="1.0" encoding="UTF-8"?><!DOCTYPE article PUBLIC "-//NLM//DTD JATS (Z39.96) Journal Archiving and Interchange DTD v1.1d1 20130915//EN" "JATS-archivearticle1.dtd"><article xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink" article-type="research-article" dtd-version="1.1d1"><front><journal-meta><journal-id journal-id-type="nlm-ta">elife</journal-id><journal-id journal-id-type="hwp">eLife</journal-id><journal-id journal-id-type="publisher-id">eLife</journal-id><journal-title-group><journal-title>eLife</journal-title></journal-title-group><issn publication-format="electronic">2050-084X</issn><publisher><publisher-name>eLife Sciences Publications, Ltd</publisher-name></publisher></journal-meta><article-meta><article-id pub-id-type="publisher-id">04106</article-id><article-id pub-id-type="doi">10.7554/eLife.04106</article-id><article-categories><subj-group subj-group-type="display-channel"><subject>Research article</subject></subj-group><subj-group subj-group-type="heading"><subject>Genomics and evolutionary biology</subject></subj-group><subj-group subj-group-type="heading"><subject>Microbiology and infectious disease</subject></subj-group></article-categories><title-group><article-title>Genome-wide regulatory dynamics of translation in the <italic>Plasmodium falciparum</italic> asexual blood stages</article-title></title-group><contrib-group><contrib contrib-type="author" id="author-16213" equal-contrib="yes"><name><surname>Caro</surname><given-names>Florence</given-names></name><xref ref-type="aff" rid="aff1">1</xref><xref ref-type="fn" rid="equal-contrib">†</xref><xref ref-type="other" rid="par-1"/><xref ref-type="other" rid="par-2"/><xref ref-type="fn" rid="con1"/><xref ref-type="fn" rid="conf1"/><xref ref-type="other" rid="dataro1"/><xref ref-type="other" rid="dataro2"/><xref ref-type="other" rid="dataro3"/></contrib><contrib contrib-type="author" id="author-16612" equal-contrib="yes"><name><surname>Ahyong</surname><given-names>Vida</given-names></name><xref ref-type="aff" rid="aff1">1</xref><xref ref-type="fn" rid="equal-contrib">†</xref><xref ref-type="other" rid="par-1"/><xref ref-type="fn" rid="con2"/><xref ref-type="fn" rid="conf1"/><xref ref-type="other" rid="dataro1"/><xref ref-type="other" rid="dataro2"/><xref ref-type="other" rid="dataro3"/></contrib><contrib contrib-type="author" id="author-16883"><name><surname>Betegon</surname><given-names>Miguel</given-names></name><xref ref-type="aff" rid="aff1">1</xref><xref ref-type="aff" rid="aff2">2</xref><xref ref-type="other" rid="par-1"/><xref ref-type="fn" rid="con3"/><xref ref-type="fn" rid="conf1"/><xref ref-type="other" rid="dataro1"/><xref ref-type="other" rid="dataro2"/><xref ref-type="other" rid="dataro3"/></contrib><contrib contrib-type="author" corresp="yes" id="author-16616"><name><surname>DeRisi</surname><given-names>Joseph L</given-names></name><xref ref-type="aff" rid="aff1">1</xref><xref ref-type="aff" rid="aff2">2</xref><xref ref-type="corresp" rid="cor1">*</xref><xref ref-type="other" rid="par-1"/><xref ref-type="fn" rid="con4"/><xref ref-type="fn" rid="conf1"/><xref ref-type="other" rid="dataro1"/><xref ref-type="other" rid="dataro2"/><xref ref-type="other" rid="dataro3"/></contrib><aff id="aff1"><label>1</label><institution content-type="dept">Department of Biochemistry and Biophysics</institution>, <institution>University of California, San Francisco</institution>, <addr-line><named-content content-type="city">San Francisco</named-content></addr-line>, <country>United States</country></aff><aff id="aff2"><label>2</label><institution>Howard Hughes Medical Institute, University of California, San Francisco</institution>, <addr-line><named-content content-type="city">San Francisco</named-content></addr-line>, <country>United States</country></aff></contrib-group><contrib-group content-type="section"><contrib contrib-type="editor"><name><surname>Gingeras</surname><given-names>Thomas R</given-names></name><role>Reviewing editor</role><aff><institution>Cold Spring Harbor Laboratory</institution>, <country>United States</country></aff></contrib></contrib-group><author-notes><corresp id="cor1"><label>*</label>For correspondence: <email>joe@derisilab.ucsf.edu</email></corresp><fn fn-type="con" id="equal-contrib"><label>†</label><p>These authors contributed equally to this work</p></fn></author-notes><pub-date publication-format="electronic" date-type="pub"><day>10</day><month>12</month><year>2014</year></pub-date><pub-date pub-type="collection"><year>2014</year></pub-date><volume>3</volume><elocation-id>e04106</elocation-id><history><date date-type="received"><day>19</day><month>07</month><year>2014</year></date><date date-type="accepted"><day>06</day><month>11</month><year>2014</year></date></history><permissions><copyright-statement>Copyright © 2014, Caro et al</copyright-statement><copyright-year>2014</copyright-year><copyright-holder>Caro et al</copyright-holder><license xlink:href="http://creativecommons.org/licenses/by/4.0/"><license-p>This article is distributed under the terms of the <ext-link ext-link-type="uri" xlink:href="http://creativecommons.org/licenses/by/4.0/">Creative Commons Attribution License</ext-link>, which permits unrestricted use and redistribution provided that the original author and source are credited.</license-p></license></permissions><self-uri content-type="pdf" xlink:href="elife04106.pdf"/><abstract><object-id pub-id-type="doi">10.7554/eLife.04106.001</object-id><p>The characterization of the transcriptome and proteome of <italic>Plasmodium falciparum</italic> has been a tremendous resource for the understanding of the molecular physiology of this parasite. However, the translational dynamics that link steady-state mRNA with protein levels are not well understood. In this study, we bridge this disconnect by measuring genome-wide translation using ribosome profiling, through five stages of the <italic>P. falciparum</italic> blood phase developmental cycle. Our findings show that transcription and translation are tightly coupled, with overt translational control occurring for less than 10% of the transcriptome. Translationally regulated genes are predominantly associated with merozoite egress functions. We systematically define mRNA 5′ leader sequences, and 3′ UTRs, as well as antisense transcripts, along with ribosome occupancy for each, and establish that accumulation of ribosomes on 5′ leaders is a common transcript feature. This work represents the highest resolution and broadest portrait of gene expression and translation to date for this medically important parasite.</p><p><bold>DOI:</bold> <ext-link ext-link-type="doi" xlink:href="10.7554/eLife.04106.001">http://dx.doi.org/10.7554/eLife.04106.001</ext-link></p></abstract><abstract abstract-type="executive-summary"><object-id pub-id-type="doi">10.7554/eLife.04106.002</object-id><title>eLife digest</title><p>The genome of an organism includes all of the genes or information necessary to build, maintain, and replicate that organism. However, cells with the same genome—such as a skin cell and a liver cell from the same person—can look and behave very differently depending on which of the genes in their genomes they express, and to what extent.</p><p>For a gene to be expressed, its DNA is ‘transcribed’ to make an RNA molecule, which is then ‘translated’ to make a protein. Efforts to measure the transcription and translation processes in diseased cells, or in the microorganisms that cause infections, may lead to new treatments and preventative medicines. Such work is currently ongoing in the global effort to treat and prevent malaria.</p><p>Malaria is both preventable and curable, yet over 600,000 people are estimated to die from this disease each year. The disease is caused by a single-celled parasite called <italic>Plasmodium</italic>. Mosquitoes carry the parasites in their salivary glands, and when a mosquito bites a human, these parasites are injected into the bloodstream with the mosquito's saliva. <italic>Plasmodium</italic> parasites then travel to and infect the liver, before bursting out of this tissue into the bloodstream. Here, the parasites infect red blood cells and undergo rounds of replication during which the symptoms of the disease are manifested. It is also during this bloodstream phase that parasites can develop into forms capable of infecting another mosquito and continuing the transmission cycle.</p><p>The genes, RNA molecules, and proteins of the <italic>Plasmodium falciparum</italic> parasite—which causes the most serious cases of malaria in humans—have been cataloged to better understand the biology of this parasite. However, the processes that control how, and when, an RNA transcript is translated into a protein are not well understood.</p><p>Now Caro et al. have uncovered which RNA molecules are being translated, and by how much, during <italic>Plasmodium</italic> development within the blood. The transcription and translation of genes in this parasite were found to be tightly linked processes; the expression of only a few genes was controlled more by the translation process than by transcription. These translationally regulated genes were found mainly to be those that encode proteins involved in the parasite's exit from the red blood cells and spread throughout the bloodstream.</p><p>Caro et al. discovered that genetic regulation of the malaria parasite resembles a preset genetic program, rather than a system that responds to changes and external signals. As such, these findings suggest that targeting such a genetic program within <italic>Plasmodium</italic> and preventing its implementation could prove an effective strategy to curb the spread of malaria.</p><p><bold>DOI:</bold> <ext-link ext-link-type="doi" xlink:href="10.7554/eLife.04106.002">http://dx.doi.org/10.7554/eLife.04106.002</ext-link></p></abstract><kwd-group kwd-group-type="author-keywords"><title>Author keywords</title><kwd><italic>Plasmodium falciparum</italic></kwd><kwd>translation</kwd><kwd>ribosome profiling</kwd><kwd>malaria</kwd></kwd-group><kwd-group kwd-group-type="research-organism"><title>Research organism</title><kwd>other</kwd></kwd-group><funding-group><award-group id="par-1"><funding-source><institution-wrap><institution-id institution-id-type="FundRef">http://dx.doi.org/10.13039/100000011</institution-id><institution content-type="university">Howard Hughes Medical Institute</institution></institution-wrap></funding-source><principal-award-recipient><name><surname>Caro</surname><given-names>Florence</given-names></name><name><surname>Ahyong</surname><given-names>Vida</given-names></name><name><surname>Betegon</surname><given-names>Miguel</given-names></name><name><surname>DeRisi</surname><given-names>Joseph L</given-names></name></principal-award-recipient></award-group><award-group id="par-2"><funding-source><institution-wrap><institution-id institution-id-type="FundRef">http://dx.doi.org/10.13039/100000002</institution-id><institution content-type="university">National Institutes of Health</institution></institution-wrap></funding-source><award-id>UO1 AI075517</award-id><principal-award-recipient><name><surname>Caro</surname><given-names>Florence</given-names></name></principal-award-recipient></award-group><funding-statement>The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.</funding-statement></funding-group><custom-meta-group><custom-meta><meta-name>elife-xml-version</meta-name><meta-value>2.0</meta-value></custom-meta><custom-meta specific-use="meta-only"><meta-name>Author impact statement</meta-name><meta-value>Ribosome profiling has bridged the knowledge gap between transcription and translation during malaria blood stage development and provided a comprehensive gene expression resource for this parasite.</meta-value></custom-meta></custom-meta-group></article-meta></front><body><sec sec-type="intro" id="s1"><title>Introduction</title><p>The transcriptome of the intraerythrocytic developmental cycle (IDC) of <italic>P. falciparum</italic> is characterized by a continuous cascade wherein the expression of the majority of genes is maximally induced once per cycle and their timing correlates well with the timing for the respective protein's biological function (<xref ref-type="bibr" rid="bib5">Bozdech et al., 2003</xref>). The apparent lack of dynamic transcriptional regulation suggested that complementary post-transcriptional mechanisms could play an important role in the regulation of parasite gene expression (<xref ref-type="bibr" rid="bib23">Hughes et al., 2010</xref>). This is a reasonable assumption, given that global or gene-specific translational regulation of gene expression is a mechanism that allows fast adaptations during drastic changes in environmental conditions as well as during rapid transitions in developmental programs. Indeed a few examples of translational control in <italic>Plasmodium</italic> have been reported. In sporozoites present in the mosquito salivary gland, phosphorylation of the eukaryotic translation initiation factor eIF2α by the kinase IK2, inhibits translation and causes accumulation of mRNAs into granules. Translational repression is alleviated by eIF2α phosphatase during the transition into the mammalian host, allowing parasites to transform into the liver stages (<xref ref-type="bibr" rid="bib47">Zhang et al., 2010</xref>). Similarly, PK4 kinase activity leads to the reduction of global protein synthesis through phosphorylation of eIF2α in schizonts and gametocytes and is essential for the completion of the parasite's erythrocytic cycle (<xref ref-type="bibr" rid="bib48">Zhang et al., 2012</xref>). Gene-specific translational regulation has also been observed in <italic>P. falciparum</italic> and is mediated by cis-acting sequences in combination with RNA-binding proteins. For example, dihydrofolate reductase-thymidylate synthase (DHFR-TS) binds within the coding region of its own cognate mRNA to repress translation (<xref ref-type="bibr" rid="bib49">Zhang and Rathod, 2002</xref>) and antifolate treatment has been shown to relieve this repressive effect without alteration of mRNA levels (<xref ref-type="bibr" rid="bib37">Nirmalan et al., 2004a</xref>). In <italic>Plasmodium berghei,</italic> storage of translationally repressed mRNAs prior to fertilization is mediated by mRNA binding via the RNA helicase DOZI and the Sm-like factor CITH (<xref ref-type="bibr" rid="bib31">Mair et al., 2006</xref>, <xref ref-type="bibr" rid="bib32">2010</xref>). Upstream open reading frames (uORFs) found on 5′ UTRs of transcripts have been reported to regulate the translation of specific genes (<xref ref-type="bibr" rid="bib34">Morris and Geballe, 2000</xref>). In <italic>P. falciparum</italic>, the only uORF described and functionally characterized to date is a 120 codon region upstream of the <italic>var2csa</italic> (PFL0030c) coding region, a unique variant of the surface antigen PfEMP1 that mediates adhesion to placenta in pregnant women (<xref ref-type="bibr" rid="bib2">Amulic et al., 2009</xref>). In this case, translation of the uORF modulates repression of var2csa translation. Aside from these examples, the extent to which global and gene-specific translational control operates in <italic>P. falciparum</italic> during the IDC remains sparse.</p><p>Since the <italic>P. falciparum</italic> genome was fully sequenced (<xref ref-type="bibr" rid="bib16">Gardner et al., 2002</xref>), several large-scale studies have provided detailed insights into the expression of genes and proteins across the parasite's life cycle. Parallel mass spectrometry-based proteomics and genome-wide expression profiling revealed differences between mRNA abundance and the accumulation of the corresponding protein, supporting the notion that post-transcriptional regulation of gene expression is at play in this parasite (<xref ref-type="bibr" rid="bib28">Le Roch et al., 2004</xref>; <xref ref-type="bibr" rid="bib36">Nirmalan et al., 2004b</xref>; <xref ref-type="bibr" rid="bib14">Foth et al., 2011</xref>). These methods, however, are limited in their ability to measure low abundance proteins and do not capture the underlying relationship between transcriptional activity and translational efficiency. More recently, polysome profiling was used to monitor discrepancies between polysome-associated and steady-state mRNAs in 30% of the <italic>P. falciparum</italic> blood stage transcriptome (<xref ref-type="bibr" rid="bib7">Bunnik et al., 2013</xref>); however, this approach does not reveal the precise localization of the ribosomes, and thus can not be used to accurately assess the translational efficiency of a given mRNA (<xref ref-type="bibr" rid="bib24">Ingolia, 2014</xref>).</p><p>Here, we adapted the ribosome profiling technique (<xref ref-type="bibr" rid="bib25">Ingolia et al., 2009</xref>) to describe the translational dynamics of the <italic>P. falciparum</italic> asexual blood stage transcriptome. We simultaneously evaluate mRNA abundance, gene structure, ribosome positioning, and translational efficiency for genes expressed through five stages of the IDC. We demonstrate that the data are highly reproducible, and we find that the translational efficiency of the majority of mRNAs expressed follows a narrow distribution, exhibiting a tight coupling between transcription and translation. Only 10% of the genes expressed deviate from this trend and are translationally up- or down-regulated. We found a surprising amount of ribosome density associated with 5′ leaders of transcripts particularly in genes with functions associated with merozoite egress and invasion. Overall, the precision and depth of the dataset presented herein add significantly to our understanding of <italic>P. falciparum</italic> gene expression by linking transcriptional and translational dynamics throughout the blood stages.</p></sec><sec sec-type="results" id="s2"><title>Results</title><sec id="s2-1"><title>Overview of ribosome profiling in <italic>P. falciparum</italic> asexual blood stages</title><p>To create whole-genome, high-resolution profiles of mRNA abundance and translation during in vivo blood stage development of <italic>P. falciparum,</italic> we adapted the ribosome profiling technique described by <xref ref-type="bibr" rid="bib25">Ingolia et al. (2009)</xref>. Ribosome profiling is based on the deep sequencing of ribosome protected mRNA fragments obtained by nuclease digestion of polysomes, cycloheximide-arrested ribosomes bound to mRNA. These fragments represent the exact location of the ribosome at the moment the sample was harvested. Five stages representative of the 48-hr IDC of <italic>P. falciparum</italic> were harvested for both mRNA and polysome isolation; ring, early trophozoite, late trophozoite, schizont stages, and purified merozoites. To assess the reproducibility of the data, we harvested independent biological replicates of each stage. Polysomes were isolated in the presence of the translation elongation inhibitor cycloheximide, then nuclease digested to produce monosomes, and sedimented by centrifugation on a sucrose gradient (<xref ref-type="fig" rid="fig1">Figure 1</xref> and <xref ref-type="fig" rid="fig1s1">Figure 1—figure supplement 1</xref>). To minimize isolation of RNA fragments bound by proteins other than 80S ribosomes, RNA was extracted only from the fractions of the sucrose gradient containing the monosome peak. The resulting ∼30 nt fragments of RNA, corresponding to ribosome footprints, were processed into strand-specific deep sequencing libraries in parallel with the mRNA samples, fragmented to ∼30 nt for consistency. Despite the unusually high AT content of the <italic>P. falciparum</italic> genome, over 92% of all 30 nt sequenced reads, derived from coding sequences (CDSs), mapped uniquely to the genome (<xref ref-type="supplementary-material" rid="SD1-data">Figure 1—source data 1</xref> and <xref ref-type="fig" rid="fig1s2">Figure 1—figure supplement 2</xref>).<fig-group><fig id="fig1" position="float"><object-id pub-id-type="doi">10.7554/eLife.04106.003</object-id><label>Figure 1.</label><caption><title>Ribosome profiling of the <italic>P. falciparum</italic> asexual blood stages, experimental outline.</title><p>(<bold>A</bold>) Synchronized parasite cultures were maintained in hyperflasks at 5% hematocrit and maximum 15% parasitemia. Cycloheximide-treated cultures containing ∼10<sup>10</sup> parasites were harvested at ring, early trophozoite, late trophozoite and schizont stages (11, 21, 31, and 45 hpi, respectively) for total RNA or polysome isolation. Merozoites were purified through magnetic isolation of late stage schizonts (see ‘Materials and methods’). Nuclease treated polysomes were fractionated on a sucrose gradient. Ribosome footprints (∼30 nt) derived from the monosome peak (dashed red line) or chemically fragmented polyA purified mRNA (∼30 nt) were used to build sequencing libraries. mRNA and ribosome footprint samples were processed in parallel to create deep sequencing libraries compatible with the Illumina platform. (<bold>B</bold>) Sucrose gradient A260 absorbance profile of polysome extracts derived from late trophozoites treated with micrococcal nuclease (green, +MNase) or untreated controls (gray, No treatment). Red arrow indicates the 80S monosome peak collected for ribosome footprint library preparation.</p><p><bold>DOI:</bold> <ext-link ext-link-type="doi" xlink:href="10.7554/eLife.04106.003">http://dx.doi.org/10.7554/eLife.04106.003</ext-link></p><p><supplementary-material id="SD1-data"><object-id pub-id-type="doi">10.7554/eLife.04106.004</object-id><label>Figure 1—source data 1.</label><caption><title>Illumina sequencing mapping statistics against <italic>P. facliparum</italic> W2 SNP-corrected genome.</title><p><bold>DOI:</bold> <ext-link ext-link-type="doi" xlink:href="10.7554/eLife.04106.004">http://dx.doi.org/10.7554/eLife.04106.004</ext-link></p></caption><media xlink:href="elife04106s001.xlsx" mimetype="application" mime-subtype="xlsx"/></supplementary-material></p></caption><graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="elife04106f001"/></fig><fig id="fig1s1" position="float" specific-use="child-fig"><object-id pub-id-type="doi">10.7554/eLife.04106.005</object-id><label>Figure 1—figure supplement 1.</label><caption><title>Polysome profiles of the <italic>P. falciparum</italic> asexual blood stages.</title><p>Sucrose gradient A260 absorbance profiles of polysome extracts treated with micrococcal nuclease (green, +MNase) and untreated controls (gray, No treatment). Red dotted line indicates monosome peak harvested for ribosome footprint library generation.</p><p><bold>DOI:</bold> <ext-link ext-link-type="doi" xlink:href="10.7554/eLife.04106.005">http://dx.doi.org/10.7554/eLife.04106.005</ext-link></p></caption><graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="elife04106fs001"/></fig><fig id="fig1s2" position="float" specific-use="child-fig"><object-id pub-id-type="doi">10.7554/eLife.04106.006</object-id><label>Figure 1—figure supplement 2.</label><caption><title>Read size influence on mappability.</title><p>Single nucleotide sliding windows ranging from 10 to 50 nt were used to generate in silico libraries of the <italic>P. falciparum</italic> W2 SNP corrected genome. These were uniquely aligned, allowing no mismatches, to either the whole genome (gray, WG) or the coding sequences (blue, CDS) using Bowtie (<xref ref-type="bibr" rid="bib35">Mortazavi et al., 2008</xref>) and the percentage of aligned reads were calculated for each window size. The analysis was repeated using sliding windows generated from the coding genome only (red, CDS to WG) for a more representative mappability estimate of an RNA-seq data set. Read sizes of ≥20 nt asymptotically approach maximum mappability percentages.</p><p><bold>DOI:</bold> <ext-link ext-link-type="doi" xlink:href="10.7554/eLife.04106.006">http://dx.doi.org/10.7554/eLife.04106.006</ext-link></p></caption><graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="elife04106fs002"/></fig><fig id="fig1s3" position="float" specific-use="child-fig"><object-id pub-id-type="doi">10.7554/eLife.04106.007</object-id><label>Figure 1—figure supplement 3.</label><caption><title>Reproducibility and coverage threshold determination using two fully independent biological replicates.</title><p>mRNA abundance measurements (<bold>A</bold>) and ribosome footprint densities (<bold>B</bold>) in terms of rpkM in two fully independent biological replicates of the late trophozoite timepoint. Genes with at least ≥32 total mRNA reads counted (rM) are highly reproducible (r ≥ 0.9) across replicas (A and B red dots, and Figure D) whereas low read counts have a negative effect on rpkM reproducibility (A and B blue dots, and C). (<bold>C</bold>) Genes were binned based on their rM in replica 1. In each bin Pearson correlations of rpkM values of replica 1 and replica 2 were calculated. At 32 rM, <italic>r</italic> values were consistently above 0.9 indicating that rpkMs calculated for genes with ≥32 rM are highly reproducible across replicates, and this is independent of the number of genes in the bin. <italic>r</italic> = Pearson correlation coefficient.</p><p><bold>DOI:</bold> <ext-link ext-link-type="doi" xlink:href="10.7554/eLife.04106.007">http://dx.doi.org/10.7554/eLife.04106.007</ext-link></p></caption><graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="elife04106fs003"/></fig></fig-group></p><p>To quantitatively obtain mRNA abundance and ribosome footprint density measures, we calculated rpkMs (reads per kilobase of exon model per million reads mapped, as in <xref ref-type="bibr" rid="bib35">Mortazavi et al. (2008)</xref> for each gene. We established the minimum number of mRNA reads sequenced per coding region (rM; reads per million reads mapped) required to confidently include genes in downstream analyses, to be ≥ 32 rM (‘Materials and methods’, <xref ref-type="fig" rid="fig1s3">Figure 1—figure supplement 3</xref>). Using this conservative threshold, 3605 genes qualified for further analysis. Between biological replicates, Pearson correlation values were consistently high, ranging from <italic>r</italic> = 0.94 to <italic>r</italic> = 0.99 (<xref ref-type="fig" rid="fig2">Figure 2A</xref>), highlighting the quality and reproducibility of our data. In addition, we compared the RNA-seq transcriptome of the five stages sampled to our previously published transcriptome data set, originally generated using long oligo microarrays (<xref ref-type="bibr" rid="bib5">Bozdech et al., 2003</xref>). The RNA-Seq transcription profiles of the set of genes shared by the two data sets (n = 1829) were highly correlated (average <italic>r</italic> = 0.7) to the corresponding 11, 21, 31, 45, and 2 hr post merozoite invasion time points of the microarray data set, despite the use of different methodologies (microarray vs RNA-seq) and the use of different <italic>P. falciparum</italic> strains (HB3 vs W2, respectively). Because of the higher sensitivity of RNA-seq, we were able to accommodate an additional 743 genes into the cascade-like transcriptome extending it to a total of 3110 genes (<xref ref-type="fig" rid="fig2">Figure 2B</xref>, <xref ref-type="supplementary-material" rid="SD2-data">Figure 2—source data 1</xref>). The remaining 495 genes in our RNA-seq data set lacked sufficient variation over the five time points for inclusion within the phaseogram. These genes, referred to as non-phasic genes, are nevertheless included in all analyses.<fig id="fig2" position="float"><object-id pub-id-type="doi">10.7554/eLife.04106.008</object-id><label>Figure 2.</label><caption><title>Ribosome profiling through the <italic>P. falciparum</italic> IDC.</title><p>(<bold>A</bold>) Reproducibility among biological replicates. Two fully independent biological replicas of each stage were sampled for RNA-seq (left panels, blue) and ribosome profiling (right panels, green). Each dot represents the log<sub>2</sub> rpkM measured for each gene in each stage. <italic>r</italic> = Pearson correlation coefficient. (<bold>B</bold>) Gene expression and translation are tightly coupled during the <italic>P. falciparum</italic> IDC. Phaseograms of mRNA (left heatmap) and ribosome footprint density (right heatmap) as a function of development for 3110 phasic and 495 non-phasic genes organized in the same order in the left and right heatmap. Data represent mean centered log<sub>2</sub> mRNA and ribosome footprint rpkM values for each gene (rows) in each sampled stage (columns). R = rings, ET = early trophozoites, LT = late trophozoites, S = schizonts, M = merozoites. (<bold>C</bold>) log<sub>2</sub> rpkM of mRNA abundance vs ribosome footprint density for all genes expressed (rM ≥ 32) across the IDC. Pearson correlation coefficients <italic>r</italic> ≥ 85. n = total number of genes.</p><p><bold>DOI:</bold> <ext-link ext-link-type="doi" xlink:href="10.7554/eLife.04106.008">http://dx.doi.org/10.7554/eLife.04106.008</ext-link></p><p><supplementary-material id="SD2-data"><object-id pub-id-type="doi">10.7554/eLife.04106.009</object-id><label>Figure 2—source data 1.</label><caption><title><italic>P. falciparum</italic> ribosome profiling data set.</title><p><bold>DOI:</bold> <ext-link ext-link-type="doi" xlink:href="10.7554/eLife.04106.009">http://dx.doi.org/10.7554/eLife.04106.009</ext-link></p></caption><media xlink:href="elife04106s002.xlsx" mimetype="application" mime-subtype="xlsx"/></supplementary-material></p></caption><graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="elife04106f002"/></fig></p></sec><sec id="s2-2"><title>Gene expression and translation are tightly coupled during the <italic>P. falciparum</italic> IDC</title><p>While RNA-Seq reveals the abundance and architecture of individual mRNAs, ribosome profiling provides a complementary and quantitative measure of mRNA translation. Ribosome occupancy along the CDS results in a profile that indicates the timing and magnitude of translation of a given mRNA, thus quantitatively delineating regions of each mRNA molecule that are actually bound by 80S ribosomes (<xref ref-type="bibr" rid="bib25">Ingolia et al., 2009</xref>). To inspect translation on a genome-wide scale, ribosome density values of each gene expressed in the data set were organized in the same order as the transcriptome. The translational profile of each gene displayed a cascade-like quality strikingly similar to the transcriptome (<xref ref-type="fig" rid="fig2">Figure 2B</xref>). Much like mRNA abundance, translation of phasic genes reaches a single maximum and a single minimum during the IDC. To determine the exact level of correlation between transcription and translation, we directly compared mRNA and ribosome footprint density measurements (<xref ref-type="fig" rid="fig2">Figure 2C</xref>). In general, translation is tightly correlated with transcription for all phasic and non-phasic genes in rings (<italic>r</italic> = 0.85), early trophozoites (<italic>r</italic> = 0.93), late trophozoites (<italic>r</italic> = 0.91), schizonts (<italic>r</italic> = 0.89), and purified merozoites (<italic>r</italic> = 0.86). This indicates that when an mRNA is detected in one stage it is associated proportionally with ribosomes within the same stage. An example pair of genes is shown in <xref ref-type="fig" rid="fig3">Figure 3A</xref>. Here, mRNA abundance profiles of eukaryotic translation initiation factor eIF2 gamma subunit (PF14_0104) and the conserved protein PF14_0105, show that peak mRNA abundance for these two genes occurs at two different stages, early and late, respectively. Examination of ribosome occupancy of both genes reveals a ribosome density accumulation profile within the coding sequence that mirrors their respective mRNA profiles. As for the majority of genes, ribosome footprint density and mRNA abundance for these two genes are highly correlated (<italic>r</italic> = 0.98 and 0.93 for PF14_0104 and PF14_0105, respectively), indicating that mRNA translation occurs proportionally during the same stages at which these genes are transcribed (<xref ref-type="fig" rid="fig3">Figure 3B</xref>; <xref ref-type="supplementary-material" rid="SD6-data">Supplementary file 1</xref>). Globally, 77% of genes expressed in at least three stages of the IDC display high Pearson correlation (<italic>r</italic> ≥ 0.7) between mRNA abundance and translation (<xref ref-type="fig" rid="fig3s1">Figure 3—figure supplement 1</xref>). Thus, our genome-wide analysis of translation establishes that for the majority of genes expressed during the IDC, transcription and translation occur proportionally.<fig-group><fig id="fig3" position="float"><object-id pub-id-type="doi">10.7554/eLife.04106.010</object-id><label>Figure 3.</label><caption><title>Transcription and translation are highly correlated.</title><p>(<bold>A</bold>) Ribosome footprint (green) and mRNA (blue) coverage profiles of two neighbor genes, the eIF2 gamma subunit (PF14_0104) and the conserved protein PF14_0105 (CDS, white boxes; HMM-defined UTRs, black lines) in rings (R), early trophozoites (ET), late trophozoites (LT), schizonts (S), and merozoites (M). Mappability = mappability score at that position; range 0 (white) to 30 (black). rM = coverage (reads per million reads mapped). (<bold>B</bold>) mRNA and ribosome footprint density of the genes in (<bold>A</bold>) correlate during development. <italic>r</italic> = Pearson correlation coefficient between ribosome footprint density and mRNA abundance of each gene.</p><p><bold>DOI:</bold> <ext-link ext-link-type="doi" xlink:href="10.7554/eLife.04106.010">http://dx.doi.org/10.7554/eLife.04106.010</ext-link></p></caption><graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="elife04106f003"/></fig><fig id="fig3s1" position="float" specific-use="child-fig"><object-id pub-id-type="doi">10.7554/eLife.04106.011</object-id><label>Figure 3—figure supplement 1.</label><caption><title>mRNA abundance and ribosome footprint density are highly correlated for the majority of genes expressed during the IDC.</title><p>Pearson correlation of mRNA abundance and ribosome footprint density of every gene expressed in at least three stages (2412 genes).</p><p><bold>DOI:</bold> <ext-link ext-link-type="doi" xlink:href="10.7554/eLife.04106.011">http://dx.doi.org/10.7554/eLife.04106.011</ext-link></p></caption><graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="elife04106fs004"/></fig></fig-group></p></sec><sec id="s2-3"><title>Ribosome profiling reveals instances of translational control of gene expression</title><p>Ribosome profiling allows the monitoring of translation rates through the simultaneous quantitative measure of mRNA abundance and ribosome density on mRNAs. The ratio of the footprint rpkM to the mRNA rpkM for any given gene represents its relative translational efficiency (TE) (<xref ref-type="bibr" rid="bib25">Ingolia et al., 2009</xref>). To assess the dynamics of translational control and detect variations in control within and between developmental stages, we calculated the relative TE of all expressed genes in our data set (<xref ref-type="fig" rid="fig4">Figure 4A</xref>). The shape and the range of TE distributions obtained for each stage sampled is comparable to those seen in other eukaryotes (<xref ref-type="bibr" rid="bib25">Ingolia et al., 2009</xref>; <xref ref-type="bibr" rid="bib10">Dunn et al., 2013</xref>). Absolute mean translational efficiencies in all stages (log<sub>2</sub>TE µ<sub>Rings</sub> = −0.43, log<sub>2</sub>TE µ<sub>E.trophs.</sub> = −0.56, log<sub>2</sub>TE µ<sub>L.trophs</sub>. = −0.31, log<sub>2</sub>TE µ<sub>Schizonts</sub> = −0.16 and log<sub>2</sub>TE µ<sub>Merozoites</sub> = −0.68) had a maximum difference of 1.47-fold observed between early trophozoites and schizonts. Translational efficiencies display a roughly 100-fold range in absolute values in each of the stages with the exception of the ring and merozoite stages, which exhibit more extreme values. In these stages, the distribution of absolute TE values displays an approximately fourfold larger spread than in early trophozoites, late trophozoites, or schizonts (<xref ref-type="fig" rid="fig4">Figure 4A</xref>, <xref ref-type="supplementary-material" rid="SD2-data">Figure 2—source data 1</xref>). In rings the gene with the largest TE is the merozoite surface protein 9 (PFL1385c, log<sub>2</sub>TE = 4.1) and the gene with the lowest TE is the FIKK family serine/threonine protein kinase (PF14_0734, log<sub>2</sub>TE = −5.1). In merozoites the largest and lowest TE values correspond to the serine repeat antigen 5 (SERA5, PFB0340c, log<sub>2</sub>TE = 4.0) and the alpha adenylyl cyclase (PF14_0788a-c, log<sub>2</sub>TE = −4.7), respectively.<fig id="fig4" position="float"><object-id pub-id-type="doi">10.7554/eLife.04106.012</object-id><label>Figure 4.</label><caption><title>Genome-wide measurements of translation.</title><p>(<bold>A</bold>) Translational efficiency distributions in each stage. Rings and merozoites have most extreme TE values; ± 2 SD above (yellow bars) and below (blue bars) the mean. TE values of translationally up-regulated merozoite surface protein (MSP6) and the eukaryotic initiation factor 2 alpha kinase 1 (IK1) (blue arrowhead) across the time course remain high and low, respectively. μ = mean log<sub>2</sub>TE, n = total number of genes. (<bold>B</bold>) mRNA abundance and translational efficiency heatmap of translationally up- and down-regulated genes (upper panel and lower panel, respectively). Note TE is independent of changes in mRNA abundance for all genes including MSP6 and IK1 (<bold>C</bold>). R = rings, ET = early trophozoites, LT = late trophozoites, S = schizonts, M = merozoites. n = number of genes, μ = mean, SD = standard deviation.</p><p><bold>DOI:</bold> <ext-link ext-link-type="doi" xlink:href="10.7554/eLife.04106.012">http://dx.doi.org/10.7554/eLife.04106.012</ext-link></p></caption><graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="elife04106f004"/></fig></p><p>To determine the contribution of translational efficiency to the dynamic range of gene expression, we examined the genes lying at the extremes of the TE distribution. For the purpose of this analysis, genes with a translational efficiency of two standard deviations above or below the mean in any of the stages were considered translationally up- or down-regulated, respectively. A total of 301 genes, 8.3% of the transcriptome, are translationally regulated by this metric, with 124 genes translationally down-regulated and 177 genes translationally up-regulated (<xref ref-type="fig" rid="fig4">Figure 4B</xref>, <xref ref-type="supplementary-material" rid="SD2-data">Figure 2—source data 1</xref>). The timing of maximum mRNA expression does not influence TE for either of these two groups. Translational efficiencies remain high for the translationally up-regulated and low for the translationally down-regulated genes in all the stages at which they are expressed, regardless of the stage of peak mRNA abundance, suggesting that translational efficiency is largely, but not completely, programmed by the mRNA sequence itself, rather than global factors. For example, translational efficiency of the merozoite surface protein 6 (MSP6, PF10_0346) remains high (log<sub>2</sub>TE ≥ 2) across all stages irrespective of variations in its mRNA abundance. In contrast TE values for the eukaryotic initiation factor 2alpha kinase 1 (IK1, PF14_0423) are among the lowest measured despite high mRNA abundance across all stages (<xref ref-type="fig" rid="fig4">Figure 4C</xref>).</p><p>An examination of the 124 translationally down-regulated genes yielded some expected, and in some cases, unexpected findings. As would be expected, two pseudogenes, the ring-infected erythrocyte surface antigen 2 (RESA-2, PF11_0512) and reticulocyte binding protein homologue 3 (PfRh3, PFL2520w), represent a clear example of low translational efficiency. The PfRh3 pseudogene ribosome profile shows that translation of the 5′ end of this transcript occurs up until the encounter of several in-frame stop codons, causing the reduction in ribosome density from this point on (<xref ref-type="fig" rid="fig5">Figure 5A</xref>, <xref ref-type="fig" rid="fig5s1">Figure 5—figure supplement 1</xref>). This suggests that a truncated version of the PfRh3 protein is being produced in the W2 strain studied here. Evidence for peptides corresponding to the 5′ end of PfRh3 has been found in gametocytes and sporozoites (however not during the asexual stages) using mass spectrometry (<xref ref-type="bibr" rid="bib13">Florens et al., 2002</xref>; <xref ref-type="bibr" rid="bib27">Lasonder et al., 2002</xref>). We note that low levels of ribosomes can still be detected along the full length of this transcript in schizonts and merozoites. Whether these footprints derive from a low level of stop-codon read-through or accumulate via another unknown mechanism remains to be determined.<fig-group><fig id="fig5" position="float"><object-id pub-id-type="doi">10.7554/eLife.04106.013</object-id><label>Figure 5.</label><caption><title>Translationally down-regulated genes have decreased CDS ribosome density.</title><p>(<bold>A</bold>) Ribosome footprint (green) and mRNA (blue) profiles of the PfRh3 pseudogene (PFL2520w) in merozoites (M). In the detail the bars above the gene model indicate AUG, stop, and any other codon, in green, red, and gray, respectively. Boxes indicate the mapping location of peptides identified by mass spectrometry in gametocytes and sporozoites (<xref ref-type="bibr" rid="bib13">Florens et al., 2002</xref>; <xref ref-type="bibr" rid="bib27">Lasonder et al., 2002</xref>). Reduction of ribosome footprint coverage occurs upon encounter of consecutive stop codons (extended red lines). (<bold>B</bold>) eIF2α kinase (PF14_0423) gene in rings (R) showing ribosome footprint accumulation on the 5′ leader, 3′ UTR, and low translational efficiency of the CDS. (CDS, white boxes; HMM-defined UTRs, black lines. Mappability = mappability score at that position; range 0 (white) to 30 (black). rM = coverage (reads per million reads mapped).</p><p><bold>DOI:</bold> <ext-link ext-link-type="doi" xlink:href="10.7554/eLife.04106.013">http://dx.doi.org/10.7554/eLife.04106.013</ext-link></p></caption><graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="elife04106f005"/></fig><fig id="fig5s1" position="float" specific-use="child-fig"><object-id pub-id-type="doi">10.7554/eLife.04106.014</object-id><label>Figure 5—figure supplement 1.</label><caption><title>Translation of a truncated form of PfRh3 during the IDC.</title><p>Ribosome footprint (green) and mRNA (blue) profiles of the PfRh3 pseudogene (PFL2520w) in rings (R), early trophozoites (ET), late trophozoites (LT), schizonts (S), and merozoites (M). (<bold>A</bold>) Translation of PfRh3 occurs until ribosomes dissociate upon the encounter of several consecutive in-frame stop codons (visible in <bold>B</bold>). (<bold>B</bold>) The vertical bars above the gene model indicate AUG, stop, and any other codon, in green, red, and gray, respectively. Boxes indicate the mapping location of peptides identified by mass spectrometry in gametocytes and sporozoites (<xref ref-type="bibr" rid="bib13">Florens et al., 2002</xref>; <xref ref-type="bibr" rid="bib27">Lasonder et al., 2002</xref>).</p><p><bold>DOI:</bold> <ext-link ext-link-type="doi" xlink:href="10.7554/eLife.04106.014">http://dx.doi.org/10.7554/eLife.04106.014</ext-link></p></caption><graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="elife04106fs005"/></fig><fig id="fig5s2" position="float" specific-use="child-fig"><object-id pub-id-type="doi">10.7554/eLife.04106.015</object-id><label>Figure 5—figure supplement 2.</label><caption><title>Translationally down-regulated genes have decreased CDS ribosome density.</title><p>(<bold>A</bold>) Ribosome footprint (green) and mRNA (blue) profiles of the ring-infected erythrocyte surface antigen 2, RESA2 pseudogene (PF11_0512) in rings (R) and merozoites (M). Both, the annotated isoform from PlasmoDB version 7.1 and the gene model for the alternate isoform inferred using ribosome profiling and W2 genomic DNA sequencing data from this study is depicted (CDS, white boxes; HMM-defined UTRs, black lines). The red star indicates a homopolymeric tract in which a single base deletion causes a premature stop codon (red triangle), which coincides with the site of ribosome drop off. (<bold>B</bold>) Ribosome footprint and mRNA profiles of erythrocyte vesicle protein 1, EVP1 (PFD0495c). This gene is transcribed in all stages yet translational efficiencies are relatively low, as evidenced by a depletion of ribosomes on the CDS of the gene particularly in early trophozoites (log2TE = −2.6, −2.9, −1.0, −2.1, −1.4 in rings, early trophozoites, late trophozoites, schizonts, and merozoites, respectively). (CDS, white boxes; HMM-defined UTRs, black lines. Mappability = mappability score at that position; range 0 (white) to 30 (black). rM = coverage (reads per million reads mapped).</p><p><bold>DOI:</bold> <ext-link ext-link-type="doi" xlink:href="10.7554/eLife.04106.015">http://dx.doi.org/10.7554/eLife.04106.015</ext-link></p></caption><graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="elife04106fs006"/></fig></fig-group></p><p>Ring-infected erythrocyte surface antigen 2, RESA2 (PF11_0512) was first described as a pseudogene based on the presence of an internal stop codon (<xref ref-type="bibr" rid="bib8">Cappai et al., 1992</xref>). Since then, transcription of this gene has been demonstrated both in vivo (<xref ref-type="bibr" rid="bib45">Vazeux et al., 1993</xref>) and in vitro (<xref ref-type="bibr" rid="bib5">Bozdech et al., 2003</xref>). RESA-2 is transcribed but poorly translated in rings, early trophozoites and merozoites (log<sub>2</sub>TE −3.2, −2.7, −2.9, respectively). Accordingly, the ribosome profile of this gene in merozoites shows a general depletion of ribosomes along the CDS (<xref ref-type="fig" rid="fig5s2">Figure 5—figure supplement 2</xref>). In rings, ribosome density diminishes at the second exon. To validate the RESA2 gene model, we used genomic DNA sequencing data derived from the <italic>P. falciparum</italic> W2 strain used in this study. We found that 69% (n = 151) of reads mapping to this locus support a single base deletion that creates a premature stop codon exactly at the site of ribosome footprint drop-off (<xref ref-type="supplementary-material" rid="SD7-data">Supplementary file 2</xref>). These data suggest that RESA2 is transcribed and actually translated into a shorter protein product of 461 amino acids. Whether or not the protein product is functional or undergoes post-translational degradation remains to be determined.</p><p>In addition to expected instances of translational regulation, our data permit the discovery of previously uncharacterized translational regulation, especially at the extremes of the TE distributions. One of the most notable examples of translational silencing is the eIF2α kinase IK1 (PF14_0423) for which ribosome footprints accumulate at the 5′ leader and 3′ UTR but not on the CDS, resulting in an extremely low translational efficiency (log<sub>2</sub>TE = −3.6) despite relatively high transcript abundance across all stages (<xref ref-type="fig" rid="fig5">Figure 5B</xref>). The mechanism by which this gene is maintained in a translationally down-regulated state is unknown. Another example is the erythrocyte vesicle protein 1 (EVP1, PFD0495c) for which abundant transcript levels can be detected across all stages, with peak mRNA abundance occurring in rings and schizonts (<xref ref-type="fig" rid="fig5s2">Figure 5—figure supplement 2</xref>). Protein levels, however, have been shown to be undetectable (<xref ref-type="bibr" rid="bib44">Tamez et al., 2008</xref>). Here, we find that translational efficiencies of this gene were low across all stages and lowest in rings and early trophozoites (log<sub>2</sub>TE = −2.6 and −2.9, respectively) demonstrating that post-transcriptional regulation at the level of translation is, at least in part, responsible for its scarcity as a protein.</p><p>Thus, our ribosome profiling data set highlights instances of translational control of genes that may not be detected by proteomic methods. Indeed a search for mass spectrometric data showed no evidence for ∼70% of genes in this category (<xref ref-type="bibr" rid="bib3">Aurrecoechea et al., 2009</xref>). Including the aforementioned examples, our data set describes a total 124 translationally down-regulated genes (listed in <xref ref-type="supplementary-material" rid="SD2-data">Figure 2—source data 1</xref>) for which translational efficiency values lie at the lower extremes of the distribution.</p><p>Protein products of translationally up-regulated genes are likely to be abundant and readily detected using mass spectrometry. Previous proteomic studies show protein evidence in the blood stages for almost all (171 of 177) well-translated genes identified here (<xref ref-type="bibr" rid="bib3">Aurrecoechea et al., 2009</xref>; <xref ref-type="bibr" rid="bib40">Pease et al., 2013</xref>). Mass spectrometric evidence for the remaining six genes is either absent (PFL0245w, PFL2510w, PF11_0204) or has only been found in sporozoites (PFE1615c, MAL7P1.300, PF13_0069a). Despite the lack of proteomic data, our data indicate that these genes are both transcribed and translated during the blood stages of the parasite. Whether post-translational control points exist for these proteins is unknown.</p><p>Among the top ten most highly translated genes are proteins involved in merozoite egress and invasion MSP3, 6, 7, and 9 (merozoite surface proteins PF10_0345, PF10_0346, PF13_0197, and PFL1385c), serine repeat antigen 5 (SERA5, PFB0340c), and RAP1, 2, and 3 (PF14_0102, PFE0080c, and PFE0075c, respectively) (<xref ref-type="supplementary-material" rid="SD2-data">Figure 2—source data 1</xref>). Interestingly, 73 (41%) of all translationally up-regulated genes can be assigned to the repertoire of canonical functions for merozoite egress and invasion described to date (<xref ref-type="bibr" rid="bib46">Yeoh et al., 2007</xref>; <xref ref-type="bibr" rid="bib4">Blackman, 2008</xref>; <xref ref-type="bibr" rid="bib22">Hu et al., 2010</xref>; <xref ref-type="bibr" rid="bib12">Farrow et al., 2011</xref>). Strikingly, for all genes in this set, maximum mRNA abundance is found during the late stages of the IDC (69 schizont and 4 merozoite stage mRNAs) yet for the majority (50, 70%) peak translational efficiency occurs in rings. Consistent with this, peptides for most of these merozoite function proteins (58 of 73) have been detected in rings (<xref ref-type="bibr" rid="bib38">Oehring et al., 2012</xref>; <xref ref-type="bibr" rid="bib40">Pease et al., 2013</xref>). This mode of translational regulation whereby late stage transcripts are highly translated in rings was not exclusively limited to genes related to merozoite egress and invasion. We found evidence for an additional 14 genes with this profile, including, aquaglyceroporin (PF11_0338, log<sub>2</sub>TE = 3.8), tubulin beta chain (PF10_0084, log<sub>2</sub>TE = 1.8), and early transcribed membrane protein 2 (PFB0120w, log<sub>2</sub>TE = 2.5).</p><p>Taken together these data demonstrate that transcription and translation are tightly correlated for the majority of genes expressed during the asexual life cycle of <italic>P. falciparum</italic> with few exceptions. These apply to a small subset of translationally down- and up-regulated genes for which translational efficiencies appear to be inherent properties of the mRNA, independent of changes in mRNA abundance. Genes in this category, especially those that exhibit high translational efficiencies, are enriched with functions associated with merozoite egress and invasion during the transition from late stages into rings.</p></sec><sec id="s2-4"><title>Ribosome occupancy of 5′ leaders is commonly found on genes expressed during the IDC</title><p>Ribosome profiling provides position specific information along each transcript allowing the detection of changes in ribosome distribution on the mRNA and their relationship to translational efficiency. To look for ribosome occupancy features beyond the CDS of transcripts, we first took advantage of the deep coverage and strand specificity of our RNA-seq data to identify 5′ leaders and 3′ UTRs of the <italic>P. falciparum</italic> transcriptome. We constructed a hidden Markov model (HMM) to automatically delineate the boundaries of both 5′ leaders and 3′ UTRs for known gene models (see ‘Materials and methods’). Within the limits imposed by our data, we were able to describe 5′ mRNA leaders and/or 3′ UTRs for 3569 genes in at least one of the stages (<xref ref-type="fig" rid="fig6s1">Figure 6—figure supplement 1</xref>, <xref ref-type="supplementary-material" rid="SD2-data">Figure 2—source data 1</xref>). 5′ leaders are on average longer than 3′ UTRs in each of the stages and median lengths across stages vary to a larger degree for 5′ leaders (from 607 to 1040 nt) than for 3′ UTRs (518–622 nt). The longest 5′ mRNA leader was measured in late trophozoites (8229 nt) for the Ap2 transcription factor, PF11_0404, and the longest 3′ UTR stretched 4773 nt for 60S ribosomal protein L7-3, PF14_0231, in rings. An example pair of genes with mapped 5′ leaders and 3′ UTRs is shown in <xref ref-type="fig" rid="fig6">Figure 6</xref>. Here, our HMM predicts a 636 nt and a 781 nt 5′ leader and a 468 nt and 423 nt 3′ UTR for the Myb2 transcription factor (PF10_0327) and the bromodomain protein (PF10_0328), respectively. These genes, encoded on opposite strands, share a 1536-nt intergenic sequence; however, the span between the region delimited by their 5′ leader sequence is only 120 nt and presumably harbors their respective promoters.<fig-group><fig id="fig6" position="float"><object-id pub-id-type="doi">10.7554/eLife.04106.016</object-id><label>Figure 6.</label><caption><title>Example of extended transcript annotations using the HMM.</title><p>5′ leaders and 3′ UTRs of the gene pair Myb2 (PF10_0327) and bromodomain protein (PF10_0328) were defined using the HMM designed (see ‘Materials and methods’). The sizes of 5′ leaders and 3′ UTRs of these genes in the schizont stage are indicated. The intergenic region is 1536 nt and the spanning distance separating the 5′ leaders is 120 nt. Mappability = mappability score at that position; range 0 (white) to 30 (black). rM = coverage (reads per million reads mapped).</p><p><bold>DOI:</bold> <ext-link ext-link-type="doi" xlink:href="10.7554/eLife.04106.016">http://dx.doi.org/10.7554/eLife.04106.016</ext-link></p></caption><graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="elife04106f006"/></fig><fig id="fig6s1" position="float" specific-use="child-fig"><object-id pub-id-type="doi">10.7554/eLife.04106.017</object-id><label>Figure 6—figure supplement 1.</label><caption><title>HMM-defined 5′ leader and 3′ UTR characteristics.</title><p>5′ leader (<bold>A</bold>) and 3′ UTR (<bold>B</bold>) length distribution and their statistics (<bold>C</bold>) per stage.</p><p><bold>DOI:</bold> <ext-link ext-link-type="doi" xlink:href="10.7554/eLife.04106.017">http://dx.doi.org/10.7554/eLife.04106.017</ext-link></p></caption><graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="elife04106fs007"/></fig></fig-group></p><p>Next, using mRNA boundaries derived from our data, we analyzed ribosome distribution along each transcript during life cycle progression. More than 80% of the ribosome footprints in rings, early trophozoites, late trophozoites, and schizonts, were mapped to CDS regions of the genome, except in merozoites, where only 68% were mapped to the CDS (<xref ref-type="fig" rid="fig7">Figure 7A</xref>). On average less than 1% of all reads obtained were mapped to 3′ UTRs in each stage, and most transcripts had no observed footprints past the stop codon. In contrast, footprints were far more common in 5′ leaders (9.1%, 4.8%, 7.5%, and 4.8% in rings, early trophozoites, late trophozoites, and schizonts, respectively) particularly in merozoites (23%). Footprint enrichment is specific to 5′ leaders and not due to non-specific background since this would result in an increase of footprints mapping evenly along the length of the transcript, including the 3′ UTR, and not just the 5′ leader. Furthermore, these footprints most likely represent ribosomes because they derive from the 80S monosome fraction of the sucrose gradient, and their footprint read length distributions are equal to those of CDS mapping footprints, whereas they are significantly divergent from rRNA or tRNA read length distributions (<xref ref-type="fig" rid="fig7s1">Figure 7—figure supplement 1</xref>).<fig-group><fig id="fig7" position="float"><object-id pub-id-type="doi">10.7554/eLife.04106.018</object-id><label>Figure 7.</label><caption><title>Transcripts accumulate ribosome density within the 5′ leader.</title><p>(<bold>A</bold>) Proportion of mRNA or ribosome footprint reads mapping to CDS, to HMM-defined 5′ leaders and 3′ UTRs, antisense to annotated coding genes or to other regions of the genome such as mitochondria, plastid, tRNA, rRNA, ncRNA, and 5′ leader and 3′ UTR regions not defined by the HMM. (<bold>B</bold>) Proportion of ribosome footprints mapping inside or outside predicted uORFs in the HMM-defined 5′ leaders. (<bold>C</bold>) Ribosome footprint (green) and mRNA (blue) profiles of the EBA-175 (MAL7P1.176) gene in rings (R) showing ribosome footprint accumulation on the 5′ leader. In the detail, the bars above the gene model indicate AUG, stop, and any other codon, in green, red, and gray, respectively and in all three possible frames. Gray bars indicate the 9 uORFs present in the 5′ leader, starting with an AUG (green line) and ending with a stop codon (red line). Black bar inside CDS indicates a deletion specific to the W2 strain used in this study. CDS, white boxes; HMM-defined UTRs, black lines. Mappability = mappability score at that position; range 0 (white) to 30 (black). rM = coverage (reads per million reads mapped).</p><p><bold>DOI:</bold> <ext-link ext-link-type="doi" xlink:href="10.7554/eLife.04106.018">http://dx.doi.org/10.7554/eLife.04106.018</ext-link></p><p><supplementary-material id="SD3-data"><object-id pub-id-type="doi">10.7554/eLife.04106.019</object-id><label>Figure 7—source data 1.</label><caption><title>Predicted uORFs.</title><p><bold>DOI:</bold> <ext-link ext-link-type="doi" xlink:href="10.7554/eLife.04106.019">http://dx.doi.org/10.7554/eLife.04106.019</ext-link></p></caption><media xlink:href="elife04106s003.xlsx" mimetype="application" mime-subtype="xlsx"/></supplementary-material></p></caption><graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="elife04106f007"/></fig><fig id="fig7s1" position="float" specific-use="child-fig"><object-id pub-id-type="doi">10.7554/eLife.04106.020</object-id><label>Figure 7—figure supplement 1.</label><caption><title>5′ leader footprints are derived from ribosomes.</title><p>Ribosome footprint read length distributions for reads mapping either to CDSs, 5′ leaders, 3′ UTRs, antisense, rRNAs or tRNAs are plotted. Read lengths of rRNA and tRNA mapping footprints are significantly different than those mapping the 5′ leader, the CDS, or the 3′ UTR of transcripts in all stages. KS = Kolmogorov–Smirnov test. D = KS test statistic. R = rings, ET = early trophozoites, LT = late trophozoites, S = schizonts, M = merozoites.</p><p><bold>DOI:</bold> <ext-link ext-link-type="doi" xlink:href="10.7554/eLife.04106.020">http://dx.doi.org/10.7554/eLife.04106.020</ext-link></p></caption><graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="elife04106fs008"/></fig><fig id="fig7s2" position="float" specific-use="child-fig"><object-id pub-id-type="doi">10.7554/eLife.04106.021</object-id><label>Figure 7—figure supplement 2.</label><caption><title>Distribution of uORF coverage on 5′ leaders of genes expressed during the IDC.</title><p>The proportion of ribosome footprints mapping inside predicted uORFs was calculated for each gene expressed in each stage. The median of each of these distributions is ∼0.5.</p><p><bold>DOI:</bold> <ext-link ext-link-type="doi" xlink:href="10.7554/eLife.04106.021">http://dx.doi.org/10.7554/eLife.04106.021</ext-link></p></caption><graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="elife04106fs009"/></fig><fig id="fig7s3" position="float" specific-use="child-fig"><object-id pub-id-type="doi">10.7554/eLife.04106.022</object-id><label>Figure 7—figure supplement 3.</label><caption><title>uORFs present on 5′ leaders have no effect on TE.</title><p>Translational efficiency (log2TE) for all genes expressed in each stage is plotted against the proportion of reads mapping within uORFs, the number of predicted uORFs, or the length of predicted uORFs in the 5′ leader. No direct relationship between these parameters can be observed.</p><p><bold>DOI:</bold> <ext-link ext-link-type="doi" xlink:href="10.7554/eLife.04106.022">http://dx.doi.org/10.7554/eLife.04106.022</ext-link></p></caption><graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="elife04106fs010"/></fig><fig id="fig7s4" position="float" specific-use="child-fig"><object-id pub-id-type="doi">10.7554/eLife.04106.023</object-id><label>Figure 7—figure supplement 4.</label><caption><title>Detection of ribosome density on uORFs.</title><p>(<bold>A</bold>) Ring stage mRNA (blue) and ribosome footprint (green) profiles of VAR2CSA (PFL0030c) are shown. There is virtually no ribosome density on transcript CDS (log2TERings = −4.2). Ribosomes do accumulate on the previously described (<xref ref-type="bibr" rid="bib2">Amulic et al., 2009</xref>) 360 nt uORF (white box). This region is depicted in more detail in the panel below where the amino acids, AUGs and stop codons of each of the three reading frames are denoted with gray, green, and red bars, respectively. Note that ribosomes start accumulating upstream of the previously described uORF. Mappability at the 3′ end of this antigenic variation gene is poor and therefore no mRNA read coverage can be detected here. (<bold>B</bold>) Ring stage mRNA (blue) and ribosome footprint (green) profiles of PFE1550w (unknown function) are shown. Translational efficiency of the CDS is log2TERings = −3.6 in rings. 90% of ribosome footprints that map to the 5′ leader of this gene accumulate on one of the six predicted uORFs (detailed figure below). The predicted uORF is 168 nt (56 aa). Mappability = mappability score at that position; range 0 (white) to 30 (black). rM = coverage (reads per million reads mapped).</p><p><bold>DOI:</bold> <ext-link ext-link-type="doi" xlink:href="10.7554/eLife.04106.023">http://dx.doi.org/10.7554/eLife.04106.023</ext-link></p></caption><graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="elife04106fs011"/></fig><fig id="fig7s5" position="float" specific-use="child-fig"><object-id pub-id-type="doi">10.7554/eLife.04106.024</object-id><label>Figure 7—figure supplement 5.</label><caption><title>uORFs present on 5′ leaders have no effect on TE.</title><p>Ribosome density on the 5′ leader (log25′RD) for all genes expressed in each stage is plotted against the proportion of reads mapping within uORFs, the number of predicted uORFs, or the length of predicted uORFs in the 5′ leader. No direct relationship between these parameters can be observed.</p><p><bold>DOI:</bold> <ext-link ext-link-type="doi" xlink:href="10.7554/eLife.04106.024">http://dx.doi.org/10.7554/eLife.04106.024</ext-link></p></caption><graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="elife04106fs012"/></fig><fig id="fig7s6" position="float" specific-use="child-fig"><object-id pub-id-type="doi">10.7554/eLife.04106.025</object-id><label>Figure 7—figure supplement 6.</label><caption><title>5′ ribosome density can be found on 5′ leaders devoid of AUGs.</title><p>Ring stage mRNA (blue) and ribosome footprint (green) profiles of (<bold>A</bold>) aquaglyceroporin (PF11_0338) and (<bold>B</bold>) PFC0486c (unknown function) are shown. Both genes display high ribosome density on their 5’ leaders and these are devoid of AUGs. Mappability = mappability score at that position; range 0 (white) to 30 (black). rM = coverage (reads per million reads mapped).</p><p><bold>DOI:</bold> <ext-link ext-link-type="doi" xlink:href="10.7554/eLife.04106.025">http://dx.doi.org/10.7554/eLife.04106.025</ext-link></p></caption><graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="elife04106fs013"/></fig></fig-group></p><p>Upstream open reading frames are a major source for 5′ leader ribosome density found from yeast to humans (<xref ref-type="bibr" rid="bib25">Ingolia et al., 2009</xref>; <xref ref-type="bibr" rid="bib6">Brar et al., 2012</xref>), and these have been shown to play a role in translational regulation of the downstream ORF in a few well-studied examples (<xref ref-type="bibr" rid="bib34">Morris and Geballe, 2000</xref>). In <italic>P. falciparum</italic>, ribosomes have been suggested to accumulate on 5′ leaders of genes displaying a delay in translation presumably due to long uORFs (<xref ref-type="bibr" rid="bib7">Bunnik et al., 2013</xref>).</p><p>We defined 36,086 possible uORF regions in the 5′ leaders of genes expressed during the <italic>P. falciparum</italic> IDC using a liberal definition that includes any stretch of at least two amino acids, starting with an AUG codon (<xref ref-type="supplementary-material" rid="SD3-data">Figure 7—source data 1</xref>). Regardless of stage, half of the total ribosome footprint coverage in 5′ leaders, in aggregate, or on a gene-by-gene basis did not overlap with these predicted uORFs (<xref ref-type="fig" rid="fig7">Figure 7B</xref>, <xref ref-type="fig" rid="fig7s2">Figure 7—figure supplement 2</xref>). We could find no significant correlation between the number of uORFs per gene, the uORF lengths, or the degree to which ribosome density was enriched in uORFs with translational efficiency (<xref ref-type="fig" rid="fig7s3">Figure 7—figure supplement 3</xref>). For example, erythrocyte binding antigen-175 (EBA-175, MAL7P1.176) is well translated in rings (log<sub>2</sub>TE = 1.4) and displays a large amount of 5′ leader ribosome occupancy. Half (49%) of the reads map within the nine predicted uORFs on the 5′ leader of this gene, the other half maps outside these uORFs (<xref ref-type="fig" rid="fig7">Figure 7C</xref>). Using this liberal definition of an uORF, the data do not support an association between ribosome occupancy in these regions, nor does it support an association between the presence of these regions and translational efficiency.</p><p>Nonetheless, there exist at least two clear exceptions. First, we were able to validate translation of the reported uORF present in the 5′ leader sequence of the var2csa mRNA which is expressed only in rings (<xref ref-type="bibr" rid="bib2">Amulic et al., 2009</xref>). The majority of ribosome footprint density localizes to this uORF, and to a second one just upstream, while the var2csa coding sequence is largely devoid of footprints (log<sub>2</sub>TE<sub>Rings</sub> = −4.2, <xref ref-type="fig" rid="fig7s4">Figure 7—figure supplement 4</xref>), consistent with its translational repression during growth in the absence of plancental tissue. Second, another striking example of uORF translation was found on PFE1550w (unknown function) for which the ratio of uORF to total 5′ leader mapping reads is 0.9 (<xref ref-type="fig" rid="fig7s4">Figure 7—figure supplement 4</xref>). Indeed, ribosome footprint density is concentrated on one of the 6 uORFs predicted in the 5′ leader of this gene, 56 amino acids long. This gene is also translationally down-regulated in all stages (log<sub>2</sub>TE = −2.7 on average). These two genes represent exceptional cases for which uORF translation negatively correlates with translation of the downstream ORF.</p><p>Aside from these two exceptions, for the vast majority of genes, ribosome occupancy appears spread along 5′ leaders and not preferentially concentrated within possible uORFs. For this reason, we calculated 5′ leader ribosome density (5′RD) for each gene expressed during the IDC, defined as upstream ribosome occupancy normalized for mRNA expression level and size of the leader sequence (5′ leader ribosome footprint rpkM/5′ leader mRNA rpkM) (<xref ref-type="supplementary-material" rid="SD2-data">Figure 2—source data 1</xref>). No positive correlation exists between the number of uORFs per gene, the uORF lengths, or the degree to which ribosome density is enriched in uORFs and 5′RD, reinforcing the notion that uORFs are not a requisite for ribosome association to 5′ leaders (<xref ref-type="fig" rid="fig7s5">Figure 7—figure supplement 5</xref>). In fact 5′ ribosome density can be found on transcripts with 5′ leaders completely devoid of AUGs, and thus, without uORFs by definition, such as the highly translated aquaglyceroporin (log<sub>2</sub>TE = 3.8 and log<sub>2</sub>5’RD = 2.9 in rings), and PFC0486c (unknown function, log<sub>2</sub>TE = 1.6 and log<sub>2</sub>5′RD = 1.1 in rings) (<xref ref-type="fig" rid="fig7s6">Figure 7—figure supplement 6</xref>).</p><p>Overall, rings and merozoite stage parasites were found to express transcripts with the highest 5′RD (mean log<sub>2</sub>5′RD −0.03, and 0.11, respectively) relative to early trophozoites, late trophozoites, and schizonts (mean log<sub>2</sub>5′RD −1.11, −0.26, −0.83), where the range of 5′RD values is also narrower (<xref ref-type="fig" rid="fig8">Figure 8A</xref>). Interestingly, among genes at the extremes of the 5′RD distributions (mean ± 1 SD), we also found many of our identified translationally up- and down-regulated transcripts (66% and 40%, respectively). On average, 5′RD was enriched on translationally up-regulated transcripts (mean log<sub>2</sub>5′RD = 0.83) and depleted for translationally down-regulated transcripts in all stages (mean log<sub>2</sub>5′RD = −1.11), suggesting the possibility that 5′RD is a byproduct of translational efficiency itself (<xref ref-type="fig" rid="fig8">Figure 8B</xref>).<fig id="fig8" position="float"><object-id pub-id-type="doi">10.7554/eLife.04106.026</object-id><label>Figure 8.</label><caption><title>5′ ribosome density is commonly found on genes expressed during the IDC.</title><p>(<bold>A</bold>) 5′ RD distributions in each stage. Transcripts in rings and merozoites have on average higher 5′ RD values; ± 2 SD values lie outside gray shade. μ = mean log<sub>2</sub>5′RD, n = total number of genes. (<bold>B</bold>) 5′RD values of the translationally up-regulated set of genes (yellow boxes) are relatively higher (average log<sub>2</sub>5′RD R = 1.73, ET = −0.26, LT = 0.78, S = 0.30, M = 1.16.) than the rest (white boxes) or the set of down-regulated (blue buxes) genes. (<bold>C</bold>) 5′RD weakly correlates with translational efficiency. The translationally up-regulated gene set (yellow circles) is associated with high 5′RD, particularly in rings. The translationally up-regulated genes merozoite surface protein (MSP6), aquaglyceroporin (AQP), serine repeat antigen (SERA5), and the reticulocyte binding protein homologue 3 (PfRh3) are pointed out. <italic>r</italic> = Pearson correlation coefficient.</p><p><bold>DOI:</bold> <ext-link ext-link-type="doi" xlink:href="10.7554/eLife.04106.026">http://dx.doi.org/10.7554/eLife.04106.026</ext-link></p></caption><graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="elife04106f008"/></fig></p><p>In order to determine whether a direct relationship between 5′RD and translational efficiency of the downstream ORF exists, we compared these values for each gene. 5′RD positively correlates, albeit moderately, with translational efficiency in all stages, particularly in rings and merozoites (<italic>r</italic> = 0.51 and 0.49, respectively). We focused on the subset of genes with highest and lowest 5′RD values (mean ± 2 SD) and found that only a fraction of the translationally up- and down-regulated genes overlap with this category of extreme 5′RDs in each stage (<xref ref-type="fig" rid="fig8">Figure 8C</xref>). The largest overlap occurred in rings where the highest 5′RD values were found in 43% (31 genes) of the translationally up-regulated genes, including MSP6, AQP, and SERA5. These results indicate that while in general a correspondence between 5′RD and translational efficiency exists, one is not necessarily predictive of the other and exceptions apply. This is the case, for example, of the translationally down-regulated transcript of the pseudogene PfRh3, which in rings has the second highest 5′RD value (log<sub>2</sub>5′RD = 5.1).</p><p>In summary, our data establish ribosome accumulation on 5′ leaders as a common feature of transcripts expressed during the IDC. Ribosome density is not restricted to predicted uORFs present within these regions and, with few exceptions, the uORF number, length, or coverage level, is not a requirement for 5′ ribosome density and has no measurable effects on the translation of the downstream ORF. Even though 5′RD is more commonly found on 5′ leaders of highly translated transcripts, this is not a universal trend since only a moderate correlation exists between 5′RD and the translational efficiency of the downstream ORF.</p></sec><sec id="s2-5"><title>3′ UTR ribosome occupancy is rare</title><p>While our data showed 3′ UTRs to be relatively depleted of ribosomes, we searched for rare cases of high 3′ UTR ribosome density, possibly arising from stop codon read-through, alternative stop codon usage, or re-initiation of downstream ORFs (<xref ref-type="bibr" rid="bib10">Dunn et al., 2013</xref>; <xref ref-type="bibr" rid="bib19">Guydosh and Green, 2014</xref>). We systematically searched for transcripts for which coverage, in a sliding window of 30 nt, was greater in the 3′ UTR than the CDS, and found 19 genes meeting this criterion. These genes could be qualitatively divided into two categories: 14 with putative stop codon read-through and/or alternate stop codon usage and 5 genes for which the origin of the 3′ UTR density is unclear (listed in <xref ref-type="supplementary-material" rid="SD4-data">Figure 9—source data 1</xref>). An example of stop codon read-through is the conserved plasmodium protein (PF13_0160), shown in <xref ref-type="fig" rid="fig9">Figure 9A</xref>. Ribosomes not only extend beyond the annotated stop codon of this transcript but also skip subsequent in-frame stop codons present on the predicted 644 nt 3′ UTR. Interestingly, ribosome footprints accumulate in a single large peak approximately 130 nt downstream of the annotated stop codon. On the 1290 nt 3′ UTR of the sodium-dependent phosphate transporter (MAL13P1.206), two large peaks of ribosome footprint density, one approximately 560 nt and the other 860 nt from the stop codon, can be observed (<xref ref-type="fig" rid="fig9">Figure 9B</xref>). The origin of these footprints is unclear, and it is possible that these are the product of nuclease protection by RNA-binding proteins that co-sediment with the 80S monosome. To confirm that 3′ UTR mapping reads are derived from ribosome footprints, we compared their cumulative read length distributions against a typical CDS footprint read length distribution (<xref ref-type="fig" rid="fig9s1">Figure 9—figure supplement 1</xref>). For the 16 of the 19 genes we observed no significant difference in footprint size distributions localized to the CDS compared with the 3′ UTR. For the remaining three genes, the sodium-dependent phosphate transporter (MAL13P1.206), the acyl-Coa synthetase (PFD0085c), and the conserved plasmodium protein (PF13_0160), 3′ UTR footprint size distributions were divergent from those on the CDS, implying that footprints found on these genes' 3′ UTRs may be produced by nuclease protection of these regions by factors other than ribosomes that co-sediment with 80S ribosomes.<fig-group><fig id="fig9" position="float"><object-id pub-id-type="doi">10.7554/eLife.04106.027</object-id><label>Figure 9.</label><caption><title>3′ UTR ribosome density.</title><p>(<bold>A</bold>) Late trophozoite stage mRNA (blue) and ribosome footprint (green) profiles of the conserved plasmodium protein, PF13_0160. Ribosomes can be detected up to ∼130 nt beyond the stop codon on the 3′ UTR and accumulate in a single large peak. Red lines indicate in-frame stop codons on the 3′ UTR. (<bold>B</bold>) Two large peaks of ribosome footprint density can be detected 560 nt and 860 nt downstream from the stop codon in the 3′ UTR of the sodium-dependent phosphate transporter, MAL13P1.206.</p><p><bold>DOI:</bold> <ext-link ext-link-type="doi" xlink:href="10.7554/eLife.04106.027">http://dx.doi.org/10.7554/eLife.04106.027</ext-link></p><p><supplementary-material id="SD4-data"><object-id pub-id-type="doi">10.7554/eLife.04106.028</object-id><label>Figure 9—source data 1.</label><caption><title>Genes with 3′ UTR ribosome occupancy.</title><p><bold>DOI:</bold> <ext-link ext-link-type="doi" xlink:href="10.7554/eLife.04106.028">http://dx.doi.org/10.7554/eLife.04106.028</ext-link></p></caption><media xlink:href="elife04106s004.xlsx" mimetype="application" mime-subtype="xlsx"/></supplementary-material></p></caption><graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="elife04106f009"/></fig><fig id="fig9s1" position="float" specific-use="child-fig"><object-id pub-id-type="doi">10.7554/eLife.04106.029</object-id><label>Figure 9—figure supplement 1.</label><caption><title>3′ UTR ribosome footprint size distribution.</title><p>Cummulative read length distributions of all reads mapping to the 3′ UTR of the 19 genes with 3′ ribosome density identified compared to the read length distributions of reads mapping to all CDSs in the late trophozoite stage (black line). Footprint length distributions for MAL13P1.206, PF13_0160, and PFD0085c are least similar to the ribosome footprints that map to the CDS.</p><p><bold>DOI:</bold> <ext-link ext-link-type="doi" xlink:href="10.7554/eLife.04106.029">http://dx.doi.org/10.7554/eLife.04106.029</ext-link></p></caption><graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="elife04106fs014"/></fig></fig-group></p></sec><sec id="s2-6"><title>Antisense detection</title><p>Antisense transcription plays an important role in gene regulation from bacteria to humans, and while its role is increasingly studied in these organisms (<xref ref-type="bibr" rid="bib11">Faghihi and Wahlestedt, 2009</xref>), less is known about its relevance in <italic>P. falciparum</italic>. Previous serial analysis of gene expression (SAGE) (<xref ref-type="bibr" rid="bib39">Patankar et al., 2001</xref>), nuclear run-on experiments (<xref ref-type="bibr" rid="bib33">Militello et al., 2005</xref>), and more recently antisense splicing events detected by RNA-seq (<xref ref-type="bibr" rid="bib30">López-Barragán et al., 2011</xref>; <xref ref-type="bibr" rid="bib43">Sorber et al., 2011</xref>), suggest that antisense RNAs are synthesized by RNA pol II and may constitute up to ∼12% of the erythrocytic-stage steady-state RNA (<xref ref-type="bibr" rid="bib18">Gunasekera et al., 2004</xref>), yet their presence and biological role, if any, remains unclear. A more recent study found no correlation between natural antisense transcript levels and protein abundance (<xref ref-type="bibr" rid="bib42">Siegel et al., 2014</xref>).</p><p>The 30 nt fragmentation and RNA-ligase-based library preparation method employed here affords exquisite strand specificity by minimizing artifacts associated with random priming during reverse transcription. As evidence of this specificity, the highest expressed gene in our data set, histone h2a (PFF0860c), yielded a total of 765,510 reads on the sense strand, and only two reads on the antisense strand, corresponding to a sense:antisense ratio greater than 10<sup>5</sup>. Furthermore, our HMM mapping of 5′ leaders and 3′ UTRs facilitates the differentiation between independently transcribed antisense RNA and transcripts that occur by virtue of being part of an adjacent gene. We took advantage of the nature of our data set to identify antisense transcripts and looked for effects on sense mRNA translation.</p><p>For this analysis only, we relaxed our stringent coverage threshold from ≥32 rM to ≥16 rM for inclusion of antisense transcripts. We based our threshold on the presence of an antisense transcript to the sexual stage specific gene pfs16 (PFD0310w) confirmed by strand-specific RT-PCR (<xref ref-type="fig" rid="fig10s1">Figure 10—figure supplement 1</xref>). This antisense is predicted by the HMM to be ∼4 kb, extending over the complete coding sequence and beyond and with a coverage level of 23 rM over the sense CDS. Using the 16 rM threshold, we detected 84 antisense transcripts to several known ORFs (listed in <xref ref-type="supplementary-material" rid="SD5-data">Figure 10—source data 1</xref>), including the nucleoside transporter pfNT4 (PFA0160c) depicted in <xref ref-type="fig" rid="fig10">Figure 10A</xref>. The merozoite stage contained the highest number of antisense transcripts (46), and the fewest (13) were found in early trophozoites. Manual inspection revealed that in 63% of these instances, the putative antisense transcript actually emanates from the 5′ leader or 3′ UTR of a neighboring gene (not defined by the HMM). Antisense reads for the para-hydroxybenzoate polyprenyltransferase (PFF0370w), for example, are actually derived from the 3′ UTR of the neighboring conserved protein PFF0375c (<xref ref-type="fig" rid="fig10">Figure 10B</xref>).<fig-group><fig id="fig10" position="float"><object-id pub-id-type="doi">10.7554/eLife.04106.030</object-id><label>Figure 10.</label><caption><title>Strand-specific libraries can distinguish antisense from sense gene transcription.</title><p>(<bold>A</bold>) Schizont stage mRNA (blue) and ribosome footprint (green) profiles of the nucleoside transporter pfNT4 (PFA0160c). The antisense transcript covers the full extent of the sense transcript and displays ribosome density. (<bold>B</bold>) An example of antisense reads originating from a neighboring UTR in the schizont stage. The antisense reads in the para-hydroxybenzoate polyprenyltransferase (PFF0370w) stem from the 3′ UTR of the neighboring conserved plasmodium protein (PFF0375c).</p><p><bold>DOI:</bold> <ext-link ext-link-type="doi" xlink:href="10.7554/eLife.04106.030">http://dx.doi.org/10.7554/eLife.04106.030</ext-link></p><p><supplementary-material id="SD5-data"><object-id pub-id-type="doi">10.7554/eLife.04106.031</object-id><label>Figure 10—source data 1.</label><caption><title>Antisense transcripts.</title><p><bold>DOI:</bold> <ext-link ext-link-type="doi" xlink:href="10.7554/eLife.04106.031">http://dx.doi.org/10.7554/eLife.04106.031</ext-link></p></caption><media xlink:href="elife04106s005.xlsx" mimetype="application" mime-subtype="xlsx"/></supplementary-material></p></caption><graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="elife04106f010"/></fig><fig id="fig10s1" position="float" specific-use="child-fig"><object-id pub-id-type="doi">10.7554/eLife.04106.032</object-id><label>Figure 10—figure supplement 1.</label><caption><title>Strand-specific RT-PCR detection of the antisense to Pfs16.</title><p>Read coverage on the plus and minus strand of the stage-specific protein precursor Pfs16 (PFD0310w) locus. The gene is encoded on the plus strand and the antisense transcript covers and extends beyond the sense transcript ∼3.7 kb. The strand-specific primer was used for both reversetranscription and as forward primer for the PCR (blue arrowhead). The 5 PCR primers (black arrowhead) and the expected amplicon sizes are shown next to the strand-specific RT-PCR results. 18S rRNA primers were used in the control reactions. a.u. = arbitrary units.</p><p><bold>DOI:</bold> <ext-link ext-link-type="doi" xlink:href="10.7554/eLife.04106.032">http://dx.doi.org/10.7554/eLife.04106.032</ext-link></p></caption><graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="elife04106fs015"/></fig></fig-group></p><p>We next interrogated the impact of this set of antisense transcripts. Overall, antisense transcripts showed no effect on mRNA abundance and translational efficiencies of the cognate sense transcript. These observations parallel those described for antisense transcripts in yeast (<xref ref-type="bibr" rid="bib6">Brar et al., 2012</xref>). Thus, at first approximation antisense transcripts do not appear to play a role in translational regulation. However, these observations could be confounded due to the small number of genes in this set, and we cannot exclude the possibility of sense/anti-sense heterogeneity at the single cell level, obscured here at the population level.</p></sec></sec><sec sec-type="discussion" id="s3"><title>Discussion</title><p>Herein we present, for the first time, a comprehensive view of the coupled transcriptional and translational dynamics of the <italic>P. falciparum</italic> IDC by determination of transcript abundance and architecture together with ribosomal density and positioning. The quality of our data relies on several critical features: (1) high temporal specificity and reproducibility of fully independent biological replicas of five strictly staged cultures; (2) purified merozoites to allow discrete measurements in this stage without confounding contributions from schizonts or rings; (3) monosome isolation from sucrose gradients to specifically enrich for ribosome-derived footprints and avoid potential complications that can arise with methods like sucrose cushions which are prone to mRNA contamination; (4) sufficient sequencing depth of biological replicates to set a statistical threshold for minimum read coverage and to demonstrate reproducibility; (5) stringent strand specificity to facilitate an HMM for the description of transcript boundaries and the detection of antisense transcription.</p><p>Previous studies of the transcript abundance in the malaria blood stages revealed a periodic cascade of gene expression, whereby the majority of expressed genes exhibit one peak of expression per cell cycle (<xref ref-type="bibr" rid="bib5">Bozdech et al., 2003</xref>). The global profile of transcriptional expression was subsequently found to be highly stereotypical across strains and appeared to lack dynamic responses to perturbation (<xref ref-type="bibr" rid="bib29">Llinás et al., 2006</xref>; <xref ref-type="bibr" rid="bib15">Ganesan et al., 2008</xref>). It has been suggested that translational control of protein expression could compensate for the lack of transcriptional dynamics. Proteomic studies described delays in peak mRNA and corresponding protein abundance implicating translational or post-translational mechanisms in the modulation of gene expression (<xref ref-type="bibr" rid="bib28">Le Roch et al., 2004</xref>; <xref ref-type="bibr" rid="bib14">Foth et al., 2011</xref>).</p><p>Our ribosome profiling results reveal a tight coupling of transcription and translation for the majority of expressed genes, indicating that most protein products are translated with highly similar timing and in proportion to their corresponding mRNA transcripts. Synthesized proteins are likely to exert their functions immediately upon translation but post-translational regulation, not captured by our data, could still be at play. Direct correlations of translational efficiencies measured in this study along with proteomic data sets are hampered by the reduced sensitivity of the latter and differences in temporal resolution and staging of the parasites between data sets. However, the available proteomic evidence is largely consistent with the results presented here, particularly for highly translated proteins. The simultaneous capture of mRNA abundance and translation is expected to be a more accurate proxy for protein levels than measurements of mRNA abundance alone (<xref ref-type="bibr" rid="bib25">Ingolia et al., 2009</xref>) and provides a critical resource for the identification of instances of post-translational regulation of gene expression. However, we note that this data set only provides a direct measure of relative changes in translational efficiencies rather than changes in bulk transcription and translation.</p><p>While no up- or down-regulation of global translation efficiencies were observed in any particular stage, more extreme translational efficiencies were measured in subsets of genes expressed in rings and merozoites. We find 177 translationally up-regulated genes with functions predominantly related to merozoite egress and invasion, with peak mRNA in schizonts and peak translational efficiency in rings. It is likely that the genes with unknown functions, regulated in an analogous way during the merozoite to ring transition, are also associated with this process. Our data support a model whereby the transcripts of proteins necessary for merozoite structure and function are made in the previous stage in large abundance, are translationally up-regulated during the invasion process, and remain highly translated well into the ring stage despite rapid mRNA decay during this stage (<xref ref-type="bibr" rid="bib41">Shock et al., 2007</xref>). Whether the accumulation of 5′ leader ribosome density is a mechanism that assists in this process or is it merely a byproduct of more efficient ribosomal initiation on these templates remains to be tested. With the emergence of genome editing tools such as CRISPR/Cas9 (<xref ref-type="bibr" rid="bib17">Ghorbal et al., 2014</xref>), it may be possible to create versions of genes with altered cis-acting sequences to test for modulation of 5′ ribosome density and its effect on translational efficiency.</p><p>The global nature of ribosome accumulation within the 5′ leader sequences of many transcripts during the IDC and the lack of an association between 5′RD and the number or length of uORFs suggests that ribosomes accumulate on 5′ leaders through means other than a uORF model. For comparison, in yeast under starvation conditions the fraction of ribosome footprints derived from 5′ leaders is increased by sixfold and in some cases no single uORF can account for the entire distribution of ribosomes on the 5′ leader of a gene (<xref ref-type="bibr" rid="bib25">Ingolia et al., 2009</xref>), much like <italic>P. falciparum</italic>. What mechanism could account for global ribosome accumulation in the 5′ leader? The presence of apparent 80S ribosomes within the 5′ leader sequence, regardless of whether they cover uORFs or not, suggests an engagement mode in which the fidelity of start codon recognition is altered or suspended. Current models propose that the 43S pre-initiation complex loads onto the mRNA with the assistance of other initiation factors near the 5′ cap and proceeds to scan down the length of the mRNA until it encounters an AUG codon. This is followed by the assembly of the 48S preinitiation complex and then finally the 80S complex (for review, see <xref ref-type="bibr" rid="bib20">Hinnebusch, 2011</xref>). The AUG that is ultimately chosen is not always the first one encountered, and its sequence context is important for recognition. The factors eIF1, eIF1A, and eIF5 have been implicated in recognition of the ‘correct’ AUG (<xref ref-type="bibr" rid="bib1">Aitken and Lorsch, 2012</xref>). In the case of <italic>P. falciparum</italic>, differential regulation or modification of these factors could plausibly result in altered start codon selection and 80S assembly. Whether prematurely initiated complexes are able to scan without synthesizing a peptide or are required to assemble and reassemble until encountering the right start codon remains an open question. Large 5′ ribosome accumulation on translationally up-regulated genes in the ring stage suggests that premature initiation on these transcripts is not detrimental. The development of an in vitro translation system that recapitulates upstream 80S assembly on <italic>P. falciparum</italic> 5′ leaders will allow direct testing of premature initiation and its effect on translational efficiency in this parasite.</p><p>Our ribosome profiling data add an important component to the rich compendium of genome-wide data, including transcript abundance (<xref ref-type="bibr" rid="bib5">Bozdech et al., 2003</xref>), mRNA decay (<xref ref-type="bibr" rid="bib41">Shock et al., 2007</xref>), splicing (<xref ref-type="bibr" rid="bib43">Sorber et al., 2011</xref>), and proteomics for this parasite (<xref ref-type="bibr" rid="bib28">Le Roch et al., 2004</xref>; <xref ref-type="bibr" rid="bib14">Foth et al., 2011</xref>). Features such as 5′ leaders, 3′ UTRs, introns, and antisense transcripts are clearly visible and often well delineated. While experimental validation of transcriptional start sites, terminators, and promoters is required, spanning regions between transcripts, such as the one shown in <xref ref-type="fig" rid="fig6">Figure 6</xref>, can be used for the search and identification of such functional sites in a reduced sequence space. The data are available at NCBI GEO (accession #GSE58402) to facilitate future queries and normalized read coverage plots for all 5 timepoints are available packed as a single Mochiview file (<xref ref-type="bibr" rid="bib9">Caro et al., 2014</xref>). Together our results describe a simplified regulatory architecture of gene translation, albeit one that includes peculiar and potentially unique mechanisms specialized for its highly structured and coordinated lifecycle within erythrocytes. Further biochemical dissection of translational initiation mechanisms and determinants of translational efficiency unique to <italic>Plasmodium</italic> may reveal weaknesses that could be exploited for possible therapeutic intervention.</p></sec><sec sec-type="materials|methods" id="s4"><title>Materials and methods</title><sec id="s4-1"><title>Cell culture</title><p>W2 strain cultures were maintained in Hyperflasks (Corning, Corning, NY) in 500 ml RPMIc (RPMI 1640 media supplemented with 0.25% Albumax II (GIBCO, Grand Island, NY), 2 g/L sodium bicarbonate, 0.1 mM hypoxanthine, 25 mM HEPES (pH 7.4), and 50 μg/L gentamycin), at 37°C, 5% O<sub>2</sub>, and 5% CO<sub>2</sub>, to maximum 10–15% parasitemia at 5% hematocrit (HC) and frequent media changes (at least every 6–8 hr). Cells were synchronized by two consecutive sorbitol treatments for three generations, for a total of six treatments. Maximum invasion, point at which half of the culture is either rings or schizonts, was defined as hour zero and independent time points containing ∼10<sup>10</sup> parasites were harvested 11, 21, 31, and 45 hr later.</p></sec><sec id="s4-2"><title>Polysome isolation and library generation</title><p>Cultures were incubated for 5 min in 500 ml 37°C RPMIc, 100 µg/ml cycloheximide (Acros Organics, Bridgewater, NJ) and harvested by centrifugation for 8 min at 3.65×<italic>g</italic> at room temperature. An aliquot was removed and flash frozen in liquid nitrogen for total RNA purification, followed by poly(A)-purification and chemical fragmentation with Zn<sup>2+</sup> to ∼30 nt for consistency in mRNA-Seq library preparation. The remaining culture was treated with ice-cold 0.1% saponin in 1X PBS, 100 µg/ml cycloheximide, for RBC lysis. Parasites were resuspended in ice-cold parasite lysis buffer (15 mM KOAc, 15 mM MgOAc, 10 mM Tris HCl pH 7.4, 0.5 mM DTT, 0.5% Triton X-100, 100 μg/ml cycloheximide) and dripped into a conical tube filled with, and immersed in, liquid nitrogen. Frozen cells transferred placed in liquid nitrogen pre-chilled chambers and pulverized for 3 min at 15 Hz, on a Retsch MM301 mixer mill. Pulverized cells were thawed on ice, and cell debris was removed by centrifugation at 4°C, 16,000×<italic>g</italic> for 10 min. The supernatant was treated with 2.88 U/µg Micrococcal nuclease for 30 min at room temperature and immediately loaded onto sucrose gradients for ultracentrifugation at 35,000 rpm for 3 hr at 4°C in a L8-60 M Beckman centrifuge. Monosome fractions only, were collected to generate ribosome footprint libraries for deep sequencing using the HiSeq 2000 (Illumina, San Diego, CA), as described (<xref ref-type="bibr" rid="bib25">Ingolia et al., 2009</xref>).</p></sec><sec id="s4-3"><title>Merozoite purification</title><p>Late stage schizonts (40–44 hpi) were magnetically purified using MACS LD columns (Miltenyi Biotec, San Diego, CA) and resuspended in RPMIc without blood addition. After reaching maximum invasion (1:1 schizont to ring ratio), cultures were harvested by centrifugation at 1500 rpm at room temperature for 5 min. Pelleted cultures were resuspended into fresh RPMIc and placed at 37°C. Merozoites in the supernatant were treated with 100 µg/ml cycloheximide for 5 min at room temperature, harvested at 4000 rpm at 4°C for 5 min and resuspended in RPMIc and passaged again through a MACS LD column. Parasite lysis buffer was added to the merozoite-enriched flow-through and flash frozen in liquid nitrogen. This procedure was repeated three times every 45 min using the original culture. The same procedure described above was used for RNA extraction and library preparation.</p></sec><sec id="s4-4"><title>SNP-corrected W2 genome</title><p>W2 strain genomic DNA was isolated from >90% synchronized ring stage cultures. Paired end libraries were constructed using the Nextera DNA Sample Prep Kit (Epicenter Biotechnologies, Madison, WI) according to the manufacturer's instructions reducing PCR cycles from nine to six and using 80% A/T dNTPs. Libraries were sequenced using the HiSeq 2000 (Illumina). Reads were aligned to the <italic>P. falciparum</italic> PlasmoDB 3D7 version 7.1 genome using Bowtie 0.12.1 (<xref ref-type="bibr" rid="bib26">Langmead et al., 2009</xref>) with parameters –v1 –m 1 (one mismatch allowed, unique mapping). A SNP was called when five or more W2 reads supported, with over 90% agreement, a different base than the one found in the <italic>P. falciparum</italic> 3D7 7.1 genome. 19401 SNPs (0.08% of total bases) were detected and used to produce the SNP-corrected W2 genome based on the 3D7 genome. Fastq files are available for download at NCBI SRA, accession #SRP042946.</p></sec><sec id="s4-5"><title>Software pipeline, mappability, and rpkM calculation</title><p>Quality-filtered ribosome footprints and mRNA sequencing reads were trimmed to remove library adapter sequence, filtered for <italic>P. falciparum</italic> rRNA using blast, and aligned uniquely to the W2 SNP-corrected genome using Bowtie 0.12.1 (<xref ref-type="bibr" rid="bib26">Langmead et al., 2009</xref>) allowing no mismatches. The percentage mappability was calculated using an <italic>in silico</italic> library of the <italic>P. falciparum</italic> W2 SNP-corrected genome created using a single nucleotide sliding window of 30 nt. The <italic>in silico</italic> library was uniquely aligned to the genome allowing no mismatches. The mappability score is given by the number of 30 nt sequences covering each nt position in the genome, such that any position has a score that ranges from 0 to 30. Both mRNA and ribosome footprint rpkMs were calculated as in <xref ref-type="bibr" rid="bib35">Mortazavi et al. (2008)</xref>, excluding the first 50 bases of each gene to eliminate bias introduced by the observed ribosome accumulation peak near the start codon. Genes with fewer than 80% mappable bases (248 genes) or any overlapping non-CDS feature on the same strand (77 genes) were excluded from this calculation. Data are available for download at NCBI GEO, accession #GSE58402. MochiView genome browser data tracks are available in <xref ref-type="supplementary-material" rid="SD6-data">Supplementary file 1</xref> (<xref ref-type="bibr" rid="bib21">Homann et al., 2010</xref>).</p></sec><sec id="s4-6"><title>Extended phaseogram</title><p>The genes of the RNA-seq transcriptome obtained in this work were listed in the same phaseogram order as the previously published microarray transcriptome (<xref ref-type="bibr" rid="bib5">Bozdech et al., 2003</xref>). The criteria for inclusion of a gene into the phaseogram was mRNA ≥ 32 rM, >2 peak to trough ratios, and Pearson correlation coefficient >0.8 with the expression profiles of the two neighboring genes.</p></sec><sec id="s4-7"><title>Hidden Markov model to describe transcript boundaries</title><p>The HMM was built using RNA-Seq data obtained in this study and two states: transcript (<italic>t</italic>) or intergenic (<italic>i</italic>) with three possible emissions: (1) read present, (2) read not present but position is unmappable, (3) read not present but the position is mappable. Both state and emission probabilites were calculated using ∼30 kb training set of manually identified transcript and non-transcript regions. The initial probabilities were set to 0.5. Transition probabilities were estimated from the median length of intergenic regions of (1252 nt) and median lengths of CDS regions (2545 nt), where the P<sub><italic>t->i</italic></sub> = (1/2545), P <sub><italic>t->t</italic></sub> = (2544/2545), P<sub><italic>i->t</italic></sub> = (1251/1252), and P<sub><italic>i->i</italic></sub> = (1/1252). We applied the Viterbi algorithm to predict the optimal path of transcript tracks per time point with a 10 nt window resolution. HMM-defined 5′ leader and 3′ UTR coordinates are available for download at NCBI GEO, accession #GSE58402.</p></sec><sec id="s4-8"><title>Strand-specific RT-PCR</title><p>Total RNA from late stage parasites was isolated and reverse transcribed using SuperScript III (Invitrogen, Carlsbad, CA) according to manufacturer's instructions, using either an antisense-specific primer to Pfs16 (PFD0310w) or a random nonamer. cDNA was amplified using the Pfs16 antisense-specific primer as a forward primer in combination with one of five reverse primers (<xref ref-type="supplementary-material" rid="SD8-data">Supplementary file 3</xref>). 18S rRNA primers were used in the control reactions with the random nonamer-derived cDNA.</p></sec></sec></body><back><ack id="ack"><title>Acknowledgements</title><p>The authors would like to thank Dr Graham Ruby for technical expertise in devising the HMM for UTR description and Dr. Giselle Knudsen for technical support. We would also like to thank members of the Weissman Lab: Dr Eugene Oh, and Dr Gloria Brar for helpful experimental advice and suggestions for the analysis of data and Dr Jonathan Weissman for the use of a density gradient fractionator and for helpful discussions during article preparation.</p></ack><sec sec-type="additional-information"><title>Additional information</title><fn-group content-type="competing-interest"><title>Competing interests</title><fn fn-type="conflict" id="conf1"><p>The authors declare that no competing interests exist.</p></fn></fn-group><fn-group content-type="author-contribution"><title>Author contributions</title><fn fn-type="con" id="con1"><p>FC, Conception and design, Acquisition of data, Analysis and interpretation of data, Drafting or revising the article</p></fn><fn fn-type="con" id="con2"><p>VA, Conception and design, Acquisition of data, Analysis and interpretation of data, Drafting or revising the article</p></fn><fn fn-type="con" id="con3"><p>MB, Conception and design, Analysis and interpretation of data</p></fn><fn fn-type="con" id="con4"><p>JLDR, Conception and design, Analysis and interpretation of data, Drafting or revising the article</p></fn></fn-group></sec><sec sec-type="supplementary-material"><title>Additional files</title><supplementary-material id="SD6-data"><object-id pub-id-type="doi">10.7554/eLife.04106.033</object-id><label>Supplementary file 1.</label><caption><p>MochiView genome browser data tracks.</p><p><bold>DOI:</bold> <ext-link ext-link-type="doi" xlink:href="10.7554/eLife.04106.033">http://dx.doi.org/10.7554/eLife.04106.033</ext-link></p></caption><media xlink:href="elife04106s006.zip" mimetype="application" mime-subtype="zip"/></supplementary-material><supplementary-material id="SD7-data"><object-id pub-id-type="doi">10.7554/eLife.04106.034</object-id><label>Supplementary file 2.</label><caption><p>RESA2 mapping reads.</p><p><bold>DOI:</bold> <ext-link ext-link-type="doi" xlink:href="10.7554/eLife.04106.034">http://dx.doi.org/10.7554/eLife.04106.034</ext-link></p></caption><media xlink:href="elife04106s007.xlsx" mimetype="application" mime-subtype="xlsx"/></supplementary-material><supplementary-material id="SD8-data"><object-id pub-id-type="doi">10.7554/eLife.04106.035</object-id><label>Supplementary file 3.</label><caption><p>Strand-specific RT-PCR primers.</p><p><bold>DOI:</bold> <ext-link ext-link-type="doi" xlink:href="10.7554/eLife.04106.035">http://dx.doi.org/10.7554/eLife.04106.035</ext-link></p></caption><media xlink:href="elife04106s008.xlsx" mimetype="application" mime-subtype="xlsx"/></supplementary-material><sec sec-type="datasets"><title>Major datasets</title><p>The following datasets were generated:</p><p><related-object content-type="generated-dataset" source-id="http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE58402" source-id-type="uri" id="dataro1"><collab collab-type="author">Caro F</collab>, <collab collab-type="author">Ahyong V</collab>, <collab collab-type="author">Betegon M</collab>, <collab collab-type="author">DeRisi JL</collab>, <year>2014</year><x>, </x><source>Ribosome Profiling in P. falciparum asexual blood stages</source><x>, </x><ext-link ext-link-type="uri" xlink:href="http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE58402">http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE58402</ext-link><x>, </x><comment>Publicly available at NCBI Gene Expression Omnibus.</comment></related-object></p><p><related-object content-type="generated-dataset" id="dataro2"><collab collab-type="author">Caro F</collab>, <collab collab-type="author">Ahyong V</collab>, <collab collab-type="author">Betegon M</collab>, <collab collab-type="author">DeRisi JL</collab>, <year>2014</year><x>, </x><source>W2 Plasmodium falciparum genome sequences</source><x>, </x><object-id pub-id-type="art-access-id">SRP042946</object-id><x>, </x><comment>Publicly available at NCBI SRA.</comment></related-object></p><p><related-object content-type="generated-dataset" source-id="http://dx.doi.org/10.5061/dryad.vb855" source-id-type="uri" id="dataro3"><collab collab-type="author">Caro F</collab>, <collab collab-type="author">Ahyong V</collab>, <collab collab-type="author">Betegon M</collab>, <collab collab-type="author">DeRisi JL</collab>, <year>2014</year><x>, </x><source>Data from: Genome-wide Regulatory Dynamics of Translation in the Plasmodium falciparum Asexual Blood Stages</source><x>, </x><ext-link ext-link-type="uri" xlink:href="http://dx.doi.org/10.5061/dryad.vb855">http://dx.doi.org/10.5061/dryad.vb855</ext-link><x>, </x><comment>Available at Dryad Digital Repository under a CC0 Public Domain Dedication.</comment></related-object></p></sec></sec><ref-list><title>References</title><ref id="bib1"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Aitken</surname><given-names>CE</given-names></name><name><surname>Lorsch</surname><given-names>JR</given-names></name></person-group><year>2012</year><article-title>A mechanistic overview of translation initiation in eukaryotes</article-title><source>Nature Structural & Molecular Biology</source><volume>19</volume><fpage>568</fpage><lpage>576</lpage><pub-id pub-id-type="doi">10.1038/nsmb.2303</pub-id></element-citation></ref><ref id="bib2"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Amulic</surname><given-names>B</given-names></name><name><surname>Salanti</surname><given-names>A</given-names></name><name><surname>Lavstsen</surname><given-names>T</given-names></name><name><surname>Nielsen</surname><given-names>MA</given-names></name><name><surname>Deitsch</surname><given-names>KW</given-names></name></person-group><year>2009</year><article-title>An upstream open reading frame controls translation of var2csa, a gene implicated in placental malaria</article-title><source>PLOS Pathogens</source><volume>5</volume><fpage>e1000256</fpage><pub-id pub-id-type="doi">10.1371/journal.ppat.1000256</pub-id></element-citation></ref><ref id="bib3"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Aurrecoechea</surname><given-names>C</given-names></name><name><surname>Brestelli</surname><given-names>J</given-names></name><name><surname>Brunk</surname><given-names>BP</given-names></name><name><surname>Dommer</surname><given-names>J</given-names></name><name><surname>Fischer</surname><given-names>S</given-names></name><name><surname>Gajria</surname><given-names>B</given-names></name><name><surname>Gao</surname><given-names>X</given-names></name><name><surname>Gingle</surname><given-names>A</given-names></name><name><surname>Grant</surname><given-names>G</given-names></name><name><surname>Harb</surname><given-names>OS</given-names></name><name><surname>Heiges</surname><given-names>M</given-names></name><name><surname>Innamorato</surname><given-names>F</given-names></name><name><surname>Iodice</surname><given-names>J</given-names></name><name><surname>Kissinger</surname><given-names>JC</given-names></name><name><surname>Kraemer</surname><given-names>E</given-names></name><name><surname>Li</surname><given-names>W</given-names></name><name><surname>Miller</surname><given-names>JA</given-names></name><name><surname>Nayak</surname><given-names>V</given-names></name><name><surname>Pennington</surname><given-names>C</given-names></name><name><surname>Pinney</surname><given-names>DF</given-names></name><name><surname>Roos</surname><given-names>DS</given-names></name><name><surname>Ross</surname><given-names>C</given-names></name><name><surname>Stoeckert</surname><given-names>CJ</given-names><suffix>Jr</suffix></name><name><surname>Treatman</surname><given-names>C</given-names></name><name><surname>Wang</surname><given-names>H</given-names></name></person-group><year>2009</year><article-title>PlasmoDB: a functional genomic database for malaria parasites</article-title><source>Nucleic Acids Research</source><volume>37</volume><fpage>D539</fpage><lpage>D543</lpage><pub-id pub-id-type="doi">10.1093/nar/gkn814</pub-id></element-citation></ref><ref id="bib4"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Blackman</surname><given-names>MJ</given-names></name></person-group><year>2008</year><article-title>Malarial proteases and host cell egress: an “emerging” cascade</article-title><source>Cellular Microbiology</source><volume>10</volume><fpage>1925</fpage><lpage>1934</lpage><pub-id pub-id-type="doi">10.1111/j.1462-5822.2008.01176.x</pub-id></element-citation></ref><ref id="bib5"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Bozdech</surname><given-names>Z</given-names></name><name><surname>Llinás</surname><given-names>M</given-names></name><name><surname>Pulliam</surname><given-names>BL</given-names></name><name><surname>Wong</surname><given-names>ED</given-names></name><name><surname>Zhu</surname><given-names>J</given-names></name><name><surname>DeRisi</surname><given-names>JL</given-names></name></person-group><year>2003</year><article-title>The transcriptome of the intraerythrocytic developmental cycle of <italic>Plasmodium falciparum</italic></article-title><source>PLOS Biology</source><volume>1</volume><fpage>E5</fpage><pub-id pub-id-type="doi">10.1371/journal.pbio.0000005</pub-id></element-citation></ref><ref id="bib6"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Brar</surname><given-names>GA</given-names></name><name><surname>Yassour</surname><given-names>M</given-names></name><name><surname>Friedman</surname><given-names>N</given-names></name><name><surname>Regev</surname><given-names>A</given-names></name><name><surname>Ingolia</surname><given-names>NT</given-names></name><name><surname>Weissman</surname><given-names>JS</given-names></name></person-group><year>2012</year><article-title>High-resolution view of the yeast meiotic program revealed by ribosome profiling</article-title><source>Science</source><volume>335</volume><fpage>552</fpage><lpage>557</lpage><pub-id pub-id-type="doi">10.1126/science.1215110</pub-id></element-citation></ref><ref id="bib7"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Bunnik</surname><given-names>EM</given-names></name><name><surname>Chung</surname><given-names>DW</given-names></name><name><surname>Hamilton</surname><given-names>M</given-names></name><name><surname>Ponts</surname><given-names>N</given-names></name><name><surname>Saraf</surname><given-names>A</given-names></name><name><surname>Prudhomme</surname><given-names>J</given-names></name><name><surname>Florens</surname><given-names>L</given-names></name><name><surname>Le Roch</surname><given-names>KG</given-names></name></person-group><year>2013</year><article-title>Polysome profiling reveals translational control of gene expression in the human malaria parasite <italic>Plasmodium falciparum</italic></article-title><source>Genome Biology</source><volume>14</volume><fpage>R128</fpage><pub-id pub-id-type="doi">10.1186/gb-2013-14-11-r128</pub-id></element-citation></ref><ref id="bib8"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Cappai</surname><given-names>R</given-names></name><name><surname>Kaslow</surname><given-names>DC</given-names></name><name><surname>Peterson</surname><given-names>MG</given-names></name><name><surname>Cowman</surname><given-names>AF</given-names></name><name><surname>Anders</surname><given-names>RF</given-names></name><name><surname>Kemp</surname><given-names>DJ</given-names></name></person-group><year>1992</year><article-title>Cloning and analysis of the RESA-2 gene: a DNA homologue of the ring-infected erythrocyte surface antigen gene of <italic>Plasmodium falciparum</italic></article-title><source>Molecular and Biochemical Parasitology</source><volume>54</volume><fpage>213</fpage><lpage>221</lpage><pub-id pub-id-type="doi">10.1016/0166-6851(92)90113-X</pub-id></element-citation></ref><ref id="bib9"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Caro</surname><given-names>F</given-names></name><name><surname>Ahyong</surname><given-names>V</given-names></name><name><surname>Betegon</surname><given-names>M</given-names></name><name><surname>DeRisi</surname><given-names>JL</given-names></name></person-group><year>2014</year><article-title>Data from: genome-wide regulatory dynamics of translation in the plasmodium falciparum asexual blood stages</article-title><source>Dryad Digital Repository</source><pub-id pub-id-type="doi">10.5061/dryad.vb855</pub-id></element-citation></ref><ref id="bib10"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Dunn</surname><given-names>JG</given-names></name><name><surname>Foo</surname><given-names>CK</given-names></name><name><surname>Belletier</surname><given-names>NG</given-names></name><name><surname>Gavis</surname><given-names>ER</given-names></name><name><surname>Weissman</surname><given-names>JS</given-names></name></person-group><year>2013</year><article-title>Ribosome profiling reveals pervasive and regulated stop codon readthrough in <italic>Drosophila melanogaster</italic></article-title><source>eLife</source><volume>2</volume><fpage>e01179</fpage><pub-id pub-id-type="doi">10.7554/eLife.01179</pub-id></element-citation></ref><ref id="bib11"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Faghihi</surname><given-names>MA</given-names></name><name><surname>Wahlestedt</surname><given-names>C</given-names></name></person-group><year>2009</year><article-title>Regulatory roles of natural antisense transcripts</article-title><source>Nature Reviews Molecular Cell Biology</source><volume>10</volume><fpage>637</fpage><lpage>643</lpage><pub-id pub-id-type="doi">10.1038/nrm2738</pub-id></element-citation></ref><ref id="bib12"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Farrow</surname><given-names>RE</given-names></name><name><surname>Green</surname><given-names>J</given-names></name><name><surname>Katsimitsoulia</surname><given-names>Z</given-names></name><name><surname>Taylor</surname><given-names>WR</given-names></name><name><surname>Holder</surname><given-names>AA</given-names></name><name><surname>Molloy</surname><given-names>JE</given-names></name></person-group><year>2011</year><article-title>The mechanism of erythrocyte invasion by the malarial parasite, <italic>Plasmodium falciparum</italic></article-title><source>Seminars in Cell & Developmental Biology</source><volume>22</volume><fpage>953</fpage><lpage>960</lpage><pub-id pub-id-type="doi">10.1016/j.semcdb.2011.09.022</pub-id></element-citation></ref><ref id="bib13"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Florens</surname><given-names>L</given-names></name><name><surname>Washburn</surname><given-names>MP</given-names></name><name><surname>Raine</surname><given-names>JD</given-names></name><name><surname>Anthony</surname><given-names>RM</given-names></name><name><surname>Grainger</surname><given-names>M</given-names></name><name><surname>Haynes</surname><given-names>JD</given-names></name><name><surname>Moch</surname><given-names>JK</given-names></name><name><surname>Muster</surname><given-names>N</given-names></name><name><surname>Sacci</surname><given-names>JB</given-names></name><name><surname>Tabb</surname><given-names>DL</given-names></name><name><surname>Witney</surname><given-names>AA</given-names></name><name><surname>Wolters</surname><given-names>D</given-names></name><name><surname>Wu</surname><given-names>Y</given-names></name><name><surname>Gardner</surname><given-names>MJ</given-names></name><name><surname>Holder</surname><given-names>AA</given-names></name><name><surname>Sinden</surname><given-names>RE</given-names></name><name><surname>Yates</surname><given-names>JR</given-names></name><name><surname>Carucci</surname><given-names>DJ</given-names></name></person-group><year>2002</year><article-title>A proteomic view of the <italic>Plasmodium falciparum</italic> life cycle</article-title><source>Nature</source><volume>419</volume><fpage>520</fpage><lpage>526</lpage><pub-id pub-id-type="doi">10.1038/nature01107</pub-id></element-citation></ref><ref id="bib14"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Foth</surname><given-names>BJ</given-names></name><name><surname>Zhang</surname><given-names>N</given-names></name><name><surname>Chaal</surname><given-names>BK</given-names></name><name><surname>Sze</surname><given-names>SK</given-names></name><name><surname>Preiser</surname><given-names>PR</given-names></name><name><surname>Bozdech</surname><given-names>Z</given-names></name></person-group><year>2011</year><article-title>Quantitative time-course profiling of parasite and host cell proteins in the human malaria parasite <italic>Plasmodium falciparum</italic></article-title><source>Molecular & Cellular Proteomics</source><volume>10</volume><fpage>M110.006411</fpage><pub-id pub-id-type="doi">10.1074/mcp.M110.006411</pub-id></element-citation></ref><ref id="bib15"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Ganesan</surname><given-names>K</given-names></name><name><surname>Ponmee</surname><given-names>N</given-names></name><name><surname>Jiang</surname><given-names>L</given-names></name><name><surname>Fowble</surname><given-names>JW</given-names></name><name><surname>White</surname><given-names>J</given-names></name><name><surname>Kamchonwongpaisan</surname><given-names>S</given-names></name><name><surname>Yuthavong</surname><given-names>Y</given-names></name><name><surname>Wilairat</surname><given-names>P</given-names></name><name><surname>Rathod</surname><given-names>PK</given-names></name></person-group><year>2008</year><article-title>A genetically hard-wired metabolic transcriptome in <italic>Plasmodium falciparum</italic> fails to mount protective responses to lethal antifolates</article-title><source>PLOS Pathogens</source><volume>4</volume><fpage>e1000214</fpage><pub-id pub-id-type="doi">10.1371/journal.ppat.1000214</pub-id></element-citation></ref><ref id="bib16"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Gardner</surname><given-names>MJ</given-names></name><name><surname>Hall</surname><given-names>N</given-names></name><name><surname>Fung</surname><given-names>E</given-names></name><name><surname>White</surname><given-names>O</given-names></name><name><surname>Berriman</surname><given-names>M</given-names></name><name><surname>Hyman</surname><given-names>RW</given-names></name><name><surname>Carlton</surname><given-names>JM</given-names></name><name><surname>Pain</surname><given-names>A</given-names></name><name><surname>Nelson</surname><given-names>KE</given-names></name><name><surname>Bowman</surname><given-names>S</given-names></name><name><surname>Paulsen</surname><given-names>IT</given-names></name><name><surname>James</surname><given-names>K</given-names></name><name><surname>Eisen</surname><given-names>JA</given-names></name><name><surname>Rutherford</surname><given-names>K</given-names></name><name><surname>Salzberg</surname><given-names>SL</given-names></name><name><surname>Craig</surname><given-names>A</given-names></name><name><surname>Kyes</surname><given-names>S</given-names></name><name><surname>Chan</surname><given-names>MS</given-names></name><name><surname>Nene</surname><given-names>V</given-names></name><name><surname>Shallom</surname><given-names>SJ</given-names></name><name><surname>Suh</surname><given-names>B</given-names></name><name><surname>Peterson</surname><given-names>J</given-names></name><name><surname>Angiuoli</surname><given-names>S</given-names></name><name><surname>Pertea</surname><given-names>M</given-names></name><name><surname>Allen</surname><given-names>J</given-names></name><name><surname>Selengut</surname><given-names>J</given-names></name><name><surname>Haft</surname><given-names>D</given-names></name><name><surname>Mather</surname><given-names>MW</given-names></name><name><surname>Vaidya</surname><given-names>AB</given-names></name><name><surname>Martin</surname><given-names>DM</given-names></name><name><surname>Fairlamb</surname><given-names>AH</given-names></name><name><surname>Fraunholz</surname><given-names>MJ</given-names></name><name><surname>Roos</surname><given-names>DS</given-names></name><name><surname>Ralph</surname><given-names>SA</given-names></name><name><surname>McFadden</surname><given-names>GI</given-names></name><name><surname>Cummings</surname><given-names>LM</given-names></name><name><surname>Subramanian</surname><given-names>GM</given-names></name><name><surname>Mungall</surname><given-names>C</given-names></name><name><surname>Venter</surname><given-names>JC</given-names></name><name><surname>Carucci</surname><given-names>DJ</given-names></name><name><surname>Hoffman</surname><given-names>SL</given-names></name><name><surname>Newbold</surname><given-names>C</given-names></name><name><surname>Davis</surname><given-names>RW</given-names></name><name><surname>Fraser</surname><given-names>CM</given-names></name><name><surname>Barrell</surname><given-names>B</given-names></name></person-group><year>2002</year><article-title>Genome sequence of the human malaria parasite <italic>Plasmodium falciparum</italic></article-title><source>Nature</source><volume>419</volume><fpage>498</fpage><lpage>511</lpage><pub-id pub-id-type="doi">10.1038/nature01097</pub-id></element-citation></ref><ref id="bib17"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Ghorbal</surname><given-names>M</given-names></name><name><surname>Gorman</surname><given-names>M</given-names></name><name><surname>Macpherson</surname><given-names>CR</given-names></name><name><surname>Martins</surname><given-names>RM</given-names></name><name><surname>Scherf</surname><given-names>A</given-names></name><name><surname>Lopez-Rubio</surname><given-names>JJ</given-names></name></person-group><year>2014</year><article-title>Genome editing in the human malaria parasite <italic>Plasmodium falciparum</italic> using the CRISPR-Cas9 system</article-title><source>Nature Biotechnology</source><volume>32</volume><fpage>819</fpage><lpage>821</lpage><pub-id pub-id-type="doi">10.1038/nbt.2925</pub-id></element-citation></ref><ref id="bib18"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Gunasekera</surname><given-names>AM</given-names></name><name><surname>Patankar</surname><given-names>S</given-names></name><name><surname>Schug</surname><given-names>J</given-names></name><name><surname>Eisen</surname><given-names>G</given-names></name><name><surname>Kissinger</surname><given-names>J</given-names></name><name><surname>Roos</surname><given-names>D</given-names></name><name><surname>Wirth</surname><given-names>DF</given-names></name></person-group><year>2004</year><article-title>Widespread distribution of antisense transcripts in the <italic>Plasmodium falciparum</italic> genome</article-title><source>Molecular and Biochemical Parasitology</source><volume>136</volume><fpage>35</fpage><lpage>42</lpage><pub-id pub-id-type="doi">10.1016/j.molbiopara.2004.02.007</pub-id></element-citation></ref><ref id="bib19"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Guydosh</surname><given-names>NR</given-names></name><name><surname>Green</surname><given-names>R</given-names></name></person-group><year>2014</year><article-title>Dom34 rescues ribosomes in 3’ untranslated regions</article-title><source>Cell</source><volume>156</volume><fpage>950</fpage><lpage>962</lpage><pub-id pub-id-type="doi">10.1016/j.cell.2014.02.006</pub-id></element-citation></ref><ref id="bib20"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Hinnebusch</surname><given-names>AG</given-names></name></person-group><year>2011</year><article-title>Molecular mechanism of scanning and start codon selection in eukaryotes</article-title><source>Microbiology and Molecular Biology Reviews</source><volume>75</volume><fpage>434</fpage><lpage>467</lpage><pub-id pub-id-type="doi">10.1128/MMBR.00008-11</pub-id></element-citation></ref><ref id="bib21"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Homann</surname><given-names>OR</given-names></name><name><surname>Johnson</surname><given-names>AD</given-names></name></person-group><year>2010</year><article-title>MochiView: versatile software for genome browsing and DNA motif analysis</article-title><source>BMC Biology</source><volume>8</volume><fpage>49</fpage><pub-id pub-id-type="doi">10.1186/1741-7007-8-49</pub-id></element-citation></ref><ref id="bib22"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Hu</surname><given-names>G</given-names></name><name><surname>Cabrera</surname><given-names>A</given-names></name><name><surname>Kono</surname><given-names>M</given-names></name><name><surname>Mok</surname><given-names>S</given-names></name><name><surname>Chaal</surname><given-names>BK</given-names></name><name><surname>Haase</surname><given-names>S</given-names></name><name><surname>Engelberg</surname><given-names>K</given-names></name><name><surname>Cheemadan</surname><given-names>S</given-names></name><name><surname>Spielmann</surname><given-names>T</given-names></name><name><surname>Preiser</surname><given-names>PR</given-names></name><name><surname>Gilberger</surname><given-names>TW</given-names></name><name><surname>Bozdech</surname><given-names>Z</given-names></name></person-group><year>2010</year><article-title>Transcriptional profiling of growth perturbations of the human malaria parasite <italic>Plasmodium falciparum</italic></article-title><source>Nature Biotechnology</source><volume>28</volume><fpage>91</fpage><lpage>98</lpage><pub-id pub-id-type="doi">10.1038/nbt.1597</pub-id></element-citation></ref><ref id="bib23"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Hughes</surname><given-names>KR</given-names></name><name><surname>Philip</surname><given-names>N</given-names></name><name><surname>Starnes</surname><given-names>GL</given-names></name><name><surname>Taylor</surname><given-names>S</given-names></name><name><surname>Waters</surname><given-names>AP</given-names></name></person-group><year>2010</year><article-title>From cradle to grave: RNA biology in malaria parasites</article-title><source>Wiley Interdisciplinary Reviews RNA</source><volume>1</volume><fpage>287</fpage><lpage>303</lpage><pub-id pub-id-type="doi">10.1002/wrna.30</pub-id></element-citation></ref><ref id="bib24"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Ingolia</surname><given-names>NT</given-names></name></person-group><year>2014</year><article-title>Ribosome profiling: new views of translation, from single codons to genome scale</article-title><source>Nature Reviews Genetics</source><volume>15</volume><fpage>205</fpage><lpage>213</lpage><pub-id pub-id-type="doi">10.1038/nrg3645</pub-id></element-citation></ref><ref id="bib25"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Ingolia</surname><given-names>NT</given-names></name><name><surname>Ghaemmaghami</surname><given-names>S</given-names></name><name><surname>Newman</surname><given-names>JR</given-names></name><name><surname>Weissman</surname><given-names>JS</given-names></name></person-group><year>2009</year><article-title>Genome-wide analysis in vivo of translation with nucleotide resolution using ribosome profiling</article-title><source>Science</source><volume>324</volume><fpage>218</fpage><lpage>223</lpage><pub-id pub-id-type="doi">10.1126/science.1168978</pub-id></element-citation></ref><ref id="bib26"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Langmead</surname><given-names>B</given-names></name><name><surname>Trapnell</surname><given-names>C</given-names></name><name><surname>Pop</surname><given-names>M</given-names></name><name><surname>Salzberg</surname><given-names>SL</given-names></name></person-group><year>2009</year><article-title>Ultrafast and memory-efficient alignment of short DNA sequences to the human genome</article-title><source>Genome Biology</source><volume>10</volume><fpage>R25</fpage><pub-id pub-id-type="doi">10.1186/gb-2009-10-3-r25</pub-id></element-citation></ref><ref id="bib27"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Lasonder</surname><given-names>E</given-names></name><name><surname>Ishihama</surname><given-names>Y</given-names></name><name><surname>Andersen</surname><given-names>JS</given-names></name><name><surname>Vermunt</surname><given-names>AM</given-names></name><name><surname>Pain</surname><given-names>A</given-names></name><name><surname>Sauerwein</surname><given-names>RW</given-names></name><name><surname>Eling</surname><given-names>WM</given-names></name><name><surname>Hall</surname><given-names>N</given-names></name><name><surname>Waters</surname><given-names>AP</given-names></name><name><surname>Stunnenberg</surname><given-names>HG</given-names></name><name><surname>Mann</surname><given-names>M</given-names></name></person-group><year>2002</year><article-title>Analysis of the <italic>Plasmodium falciparum</italic> proteome by high-accuracy mass spectrometry</article-title><source>Nature</source><volume>419</volume><fpage>537</fpage><lpage>542</lpage><pub-id pub-id-type="doi">10.1038/nature01111</pub-id></element-citation></ref><ref id="bib28"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Le Roch</surname><given-names>KG</given-names></name><name><surname>Johnson</surname><given-names>JR</given-names></name><name><surname>Florens</surname><given-names>L</given-names></name><name><surname>Zhou</surname><given-names>Y</given-names></name><name><surname>Santrosyan</surname><given-names>A</given-names></name><name><surname>Grainger</surname><given-names>M</given-names></name><name><surname>Yan</surname><given-names>SF</given-names></name><name><surname>Williamson</surname><given-names>KC</given-names></name><name><surname>Holder</surname><given-names>AA</given-names></name><name><surname>Carucci</surname><given-names>DJ</given-names></name><name><surname>Yates</surname><given-names>JR</given-names><suffix>III</suffix></name><name><surname>Winzeler</surname><given-names>EA</given-names></name></person-group><year>2004</year><article-title>Global analysis of transcript and protein levels across the <italic>Plasmodium falciparum</italic> life cycle</article-title><source>Genome Research</source><volume>14</volume><fpage>2308</fpage><lpage>2318</lpage><pub-id pub-id-type="doi">10.1101/gr.2523904</pub-id></element-citation></ref><ref id="bib29"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Llinás</surname><given-names>M</given-names></name><name><surname>Bozdech</surname><given-names>Z</given-names></name><name><surname>Wong</surname><given-names>ED</given-names></name><name><surname>Adai</surname><given-names>AT</given-names></name><name><surname>DeRisi</surname><given-names>JL</given-names></name></person-group><year>2006</year><article-title>Comparative whole genome transcriptome analysis of three <italic>Plasmodium falciparum</italic> strains</article-title><source>Nucleic Acids Research</source><volume>34</volume><fpage>1166</fpage><lpage>1173</lpage><pub-id pub-id-type="doi">10.1093/nar/gkj517</pub-id></element-citation></ref><ref id="bib30"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>López-Barragán</surname><given-names>MJ</given-names></name><name><surname>Lemieux</surname><given-names>J</given-names></name><name><surname>Quiñones</surname><given-names>M</given-names></name><name><surname>Williamson</surname><given-names>KC</given-names></name><name><surname>Molina-Cruz</surname><given-names>A</given-names></name><name><surname>Cui</surname><given-names>K</given-names></name><name><surname>Barillas-Mury</surname><given-names>C</given-names></name><name><surname>Zhao</surname><given-names>K</given-names></name><name><surname>Su</surname><given-names>XZ</given-names></name></person-group><year>2011</year><article-title>Directional gene expression and antisense transcripts in sexual and asexual stages of <italic>Plasmodium falciparum</italic></article-title><source>BMC Genomics</source><volume>12</volume><fpage>587</fpage><pub-id pub-id-type="doi">10.1186/1471-2164-12-587</pub-id></element-citation></ref><ref id="bib31"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Mair</surname><given-names>GR</given-names></name><name><surname>Braks</surname><given-names>JA</given-names></name><name><surname>Garver</surname><given-names>LS</given-names></name><name><surname>Wiegant</surname><given-names>JC</given-names></name><name><surname>Hall</surname><given-names>N</given-names></name><name><surname>Dirks</surname><given-names>RW</given-names></name><name><surname>Khan</surname><given-names>SM</given-names></name><name><surname>Dimopoulos</surname><given-names>G</given-names></name><name><surname>Janse</surname><given-names>CJ</given-names></name><name><surname>Waters</surname><given-names>AP</given-names></name></person-group><year>2006</year><article-title>Regulation of sexual development of Plasmodium by translational repression</article-title><source>Science</source><volume>313</volume><fpage>667</fpage><lpage>669</lpage><pub-id pub-id-type="doi">10.1126/science.1125129</pub-id></element-citation></ref><ref id="bib32"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Mair</surname><given-names>GR</given-names></name><name><surname>Lasonder</surname><given-names>E</given-names></name><name><surname>Garver</surname><given-names>LS</given-names></name><name><surname>Franke-Fayard</surname><given-names>BM</given-names></name><name><surname>Carret</surname><given-names>CK</given-names></name><name><surname>Wiegant</surname><given-names>JC</given-names></name><name><surname>Dirks</surname><given-names>RW</given-names></name><name><surname>Dimopoulos</surname><given-names>G</given-names></name><name><surname>Janse</surname><given-names>CJ</given-names></name><name><surname>Waters</surname><given-names>AP</given-names></name></person-group><year>2010</year><article-title>Universal features of post-transcriptional gene regulation are critical for Plasmodium zygote development</article-title><source>PLOS Pathogens</source><volume>6</volume><fpage>e1000767</fpage><pub-id pub-id-type="doi">10.1126/science.1125129</pub-id></element-citation></ref><ref id="bib33"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Militello</surname><given-names>KT</given-names></name><name><surname>Patel</surname><given-names>V</given-names></name><name><surname>Chessler</surname><given-names>AD</given-names></name><name><surname>Fisher</surname><given-names>JK</given-names></name><name><surname>Kasper</surname><given-names>JM</given-names></name><name><surname>Gunasekera</surname><given-names>A</given-names></name><name><surname>Wirth</surname><given-names>DF</given-names></name></person-group><year>2005</year><article-title>RNA polymerase II synthesizes antisense RNA in <italic>Plasmodium falciparum</italic></article-title><source>RNA</source><volume>11</volume><fpage>365</fpage><lpage>370</lpage><pub-id pub-id-type="doi">10.1261/rna.7940705</pub-id></element-citation></ref><ref id="bib34"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Morris</surname><given-names>DR</given-names></name><name><surname>Geballe</surname><given-names>AP</given-names></name></person-group><year>2000</year><article-title>Upstream open reading frames as regulators of mRNA translation</article-title><source>Molecular and Cellular Biology</source><volume>20</volume><fpage>8635</fpage><lpage>8642</lpage><pub-id pub-id-type="doi">10.1128/MCB.20.23.8635-8642.2000</pub-id></element-citation></ref><ref id="bib35"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Mortazavi</surname><given-names>A</given-names></name><name><surname>Williams</surname><given-names>BA</given-names></name><name><surname>McCue</surname><given-names>K</given-names></name><name><surname>Schaeffer</surname><given-names>L</given-names></name><name><surname>Wold</surname><given-names>B</given-names></name></person-group><year>2008</year><article-title>Mapping and quantifying mammalian transcriptomes by RNA-Seq</article-title><source>Nature Methods</source><volume>5</volume><fpage>621</fpage><lpage>628</lpage><pub-id pub-id-type="doi">10.1038/nmeth.1226</pub-id></element-citation></ref><ref id="bib37"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Nirmalan</surname><given-names>N</given-names></name><name><surname>Sims</surname><given-names>PFG</given-names></name><name><surname>Hyde</surname><given-names>JE</given-names></name></person-group><year>2004a</year><article-title>Translational up-regulation of antifolate drug targets in the human malaria parasite <italic>Plasmodium falciparum</italic> upon challenge with inhibitors</article-title><source>Molecular and Biochemical Parasitology</source><volume>136</volume><fpage>63</fpage><lpage>70</lpage><pub-id pub-id-type="doi">10.1016/j.molbiopara.2004.02.013</pub-id></element-citation></ref><ref id="bib36"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Nirmalan</surname><given-names>N</given-names></name><name><surname>Sims</surname><given-names>PFG</given-names></name><name><surname>Hyde</surname><given-names>JE</given-names></name></person-group><year>2004b</year><article-title>Quantitative proteomics of the human malaria parasite <italic>Plasmodium falciparum</italic> and its application to studies of development and inhibition</article-title><source>Molecular Microbiology</source><volume>52</volume><fpage>1187</fpage><lpage>1199</lpage><pub-id pub-id-type="doi">10.1111/j.1365-2958.2004.04049.x</pub-id></element-citation></ref><ref id="bib38"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Oehring</surname><given-names>SC</given-names></name><name><surname>Woodcroft</surname><given-names>BJ</given-names></name><name><surname>Moes</surname><given-names>S</given-names></name><name><surname>Wetzel</surname><given-names>J</given-names></name><name><surname>Dietz</surname><given-names>O</given-names></name><name><surname>Pulfer</surname><given-names>A</given-names></name><name><surname>Dekiwadia</surname><given-names>C</given-names></name><name><surname>Maeser</surname><given-names>P</given-names></name><name><surname>Flueck</surname><given-names>C</given-names></name><name><surname>Witmer</surname><given-names>K</given-names></name><name><surname>Brancucci</surname><given-names>NM</given-names></name><name><surname>Niederwieser</surname><given-names>I</given-names></name><name><surname>Jenoe</surname><given-names>P</given-names></name><name><surname>Ralph</surname><given-names>SA</given-names></name><name><surname>Voss</surname><given-names>TS</given-names></name></person-group><year>2012</year><article-title>Organellar proteomics reveals hundreds of novel nuclear proteins in the malaria parasite <italic>Plasmodium falciparum</italic></article-title><source>Genome Biology</source><volume>13</volume><fpage>R108</fpage><pub-id pub-id-type="doi">10.1186/gb-2012-13-11-r108</pub-id></element-citation></ref><ref id="bib39"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Patankar</surname><given-names>S</given-names></name><name><surname>Munasinghe</surname><given-names>A</given-names></name><name><surname>Shoaibi</surname><given-names>A</given-names></name><name><surname>Cummings</surname><given-names>LM</given-names></name><name><surname>Wirth</surname><given-names>DF</given-names></name></person-group><year>2001</year><article-title>Serial analysis of gene expression in <italic>Plasmodium falciparum</italic> reveals the global expression profile of erythrocytic stages and the presence of anti-sense transcripts in the malarial parasite</article-title><source>Molecular Biology of the Cell</source><volume>12</volume><fpage>3114</fpage><lpage>3125</lpage><pub-id pub-id-type="doi">10.1091/mbc.12.10.3114</pub-id></element-citation></ref><ref id="bib40"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Pease</surname><given-names>BN</given-names></name><name><surname>Huttlin</surname><given-names>EL</given-names></name><name><surname>Jedrychowski</surname><given-names>MP</given-names></name><name><surname>Talevich</surname><given-names>E</given-names></name><name><surname>Harmon</surname><given-names>J</given-names></name><name><surname>Dillman</surname><given-names>T</given-names></name><name><surname>Kannan</surname><given-names>N</given-names></name><name><surname>Doerig</surname><given-names>C</given-names></name><name><surname>Chakrabarti</surname><given-names>R</given-names></name><name><surname>Gygi</surname><given-names>SP</given-names></name><name><surname>Chakrabarti</surname><given-names>D</given-names></name></person-group><year>2013</year><article-title>Global analysis of protein expression and phosphorylation of three stages of <italic>Plasmodium falciparum</italic> intraerythrocytic development</article-title><source>Journal of Proteome Research</source><volume>12</volume><fpage>4028</fpage><lpage>4045</lpage><pub-id pub-id-type="doi">10.1021/pr400394g</pub-id></element-citation></ref><ref id="bib41"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Shock</surname><given-names>JL</given-names></name><name><surname>Fischer</surname><given-names>KF</given-names></name><name><surname>DeRisi</surname><given-names>JL</given-names></name></person-group><year>2007</year><article-title>Whole-genome analysis of mRNA decay in <italic>Plasmodium falciparum</italic> reveals a global lengthening of mRNA half-life during the intra-erythrocytic development cycle</article-title><source>Genome Biology</source><volume>8</volume><fpage>R134</fpage><pub-id pub-id-type="doi">10.1186/gb-2007-8-7-r134</pub-id></element-citation></ref><ref id="bib42"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Siegel</surname><given-names>TN</given-names></name><name><surname>Hon</surname><given-names>CC</given-names></name><name><surname>Zhang</surname><given-names>Q</given-names></name><name><surname>Lopez-Rubio</surname><given-names>JJ</given-names></name><name><surname>Scheidig-Benatar</surname><given-names>C</given-names></name><name><surname>Martins</surname><given-names>RM</given-names></name><name><surname>Sismeiro</surname><given-names>O</given-names></name><name><surname>Coppée</surname><given-names>JY</given-names></name><name><surname>Scherf</surname><given-names>A</given-names></name></person-group><year>2014</year><article-title>Strand-specific RNA-Seq reveals widespread and developmentally regulated transcription of natural antisense transcripts in <italic>Plasmodium falciparum</italic></article-title><source>BMC Genomics</source><volume>15</volume><fpage>150</fpage><pub-id pub-id-type="doi">10.1186/1471-2164-15-150</pub-id></element-citation></ref><ref id="bib43"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Sorber</surname><given-names>K</given-names></name><name><surname>Dimon</surname><given-names>MT</given-names></name><name><surname>DeRisi</surname><given-names>JL</given-names></name></person-group><year>2011</year><article-title>RNA-seq analysis of splicing in <italic>Plasmodium falciparum</italic> uncovers new splice junctions, alternative splicing and splicing of antisense transcripts</article-title><source>Nucleic Acids Research</source><volume>39</volume><fpage>3820</fpage><lpage>3835</lpage><pub-id pub-id-type="doi">10.1093/nar/gkq1223</pub-id></element-citation></ref><ref id="bib44"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Tamez</surname><given-names>PA</given-names></name><name><surname>Bhattacharjee</surname><given-names>S</given-names></name><name><surname>van Ooij</surname><given-names>C</given-names></name><name><surname>Hiller</surname><given-names>NL</given-names></name><name><surname>Llinás</surname><given-names>M</given-names></name><name><surname>Balu</surname><given-names>B</given-names></name><name><surname>Adams</surname><given-names>JH</given-names></name><name><surname>Haldar</surname><given-names>K</given-names></name></person-group><year>2008</year><article-title>An erythrocyte vesicle protein exported by the malaria parasite promotes tubovesicular lipid import from the host cell surface</article-title><source>PLOS Pathogens</source><volume>4</volume><fpage>e1000118</fpage><pub-id pub-id-type="doi">10.1371/journal.ppat.1000118</pub-id></element-citation></ref><ref id="bib45"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Vazeux</surname><given-names>G</given-names></name><name><surname>Scanf</surname><given-names>CL</given-names></name><name><surname>Fandeur</surname><given-names>T</given-names></name></person-group><year>1993</year><article-title>The RESA-2 gene of <italic>Plasmodium falciparum</italic> is transcribed in several independent isolates</article-title><source>Infection and Immunity</source><volume>61</volume><fpage>4469</fpage><lpage>4472</lpage></element-citation></ref><ref id="bib46"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Yeoh</surname><given-names>S</given-names></name><name><surname>O'Donnell</surname><given-names>RA</given-names></name><name><surname>Koussis</surname><given-names>K</given-names></name><name><surname>Dluzewski</surname><given-names>AR</given-names></name><name><surname>Ansell</surname><given-names>KH</given-names></name><name><surname>Osborne</surname><given-names>SA</given-names></name><name><surname>Hackett</surname><given-names>F</given-names></name><name><surname>Withers-Martinez</surname><given-names>C</given-names></name><name><surname>Mitchell</surname><given-names>GH</given-names></name><name><surname>Bannister</surname><given-names>LH</given-names></name><name><surname>Bryans</surname><given-names>JS</given-names></name><name><surname>Kettleborough</surname><given-names>CA</given-names></name><name><surname>Blackman</surname><given-names>MJ</given-names></name></person-group><year>2007</year><article-title>Subcellular Discharge of a serine Protease mediates Release of Invasive malaria parasites from host erythrocytes</article-title><source>Cell</source><volume>131</volume><fpage>1072</fpage><lpage>1083</lpage><pub-id pub-id-type="doi">10.1016/j.cell.2007.10.049</pub-id></element-citation></ref><ref id="bib47"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Zhang</surname><given-names>M</given-names></name><name><surname>Fennell</surname><given-names>C</given-names></name><name><surname>Ranford-Cartwright</surname><given-names>L</given-names></name><name><surname>Sakthivel</surname><given-names>R</given-names></name><name><surname>Gueirard</surname><given-names>P</given-names></name><name><surname>Meister</surname><given-names>S</given-names></name><name><surname>Caspi</surname><given-names>A</given-names></name><name><surname>Doerig</surname><given-names>C</given-names></name><name><surname>Nussenzweig</surname><given-names>RS</given-names></name><name><surname>Tuteja</surname><given-names>R</given-names></name><name><surname>Sullivan</surname><given-names>WJ</given-names><suffix>Jr</suffix></name><name><surname>Roos</surname><given-names>DS</given-names></name><name><surname>Fontoura</surname><given-names>BM</given-names></name><name><surname>Ménard</surname><given-names>R</given-names></name><name><surname>Winzeler</surname><given-names>EA</given-names></name><name><surname>Nussenzweig</surname><given-names>V</given-names></name></person-group><year>2010</year><article-title>The plasmodium eukaryotic initiation Factor-2α kinase IK2 controls the latency of sporozoites in the mosquito salivary glands</article-title><source>The Journal of Experimental Medicine</source><volume>207</volume><fpage>1465</fpage><lpage>1474</lpage><pub-id pub-id-type="doi">10.1084/jem.20091975</pub-id></element-citation></ref><ref id="bib48"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Zhang</surname><given-names>M</given-names></name><name><surname>Mishra</surname><given-names>S</given-names></name><name><surname>Sakthivel</surname><given-names>R</given-names></name><name><surname>Rojas</surname><given-names>M</given-names></name><name><surname>Ranjan</surname><given-names>R</given-names></name><name><surname>Sullivan</surname><given-names>WJ</given-names><suffix>Jr</suffix></name><name><surname>Fontoura</surname><given-names>BM</given-names></name><name><surname>Ménard</surname><given-names>R</given-names></name><name><surname>Dever</surname><given-names>TE</given-names></name><name><surname>Nussenzweig</surname><given-names>V</given-names></name></person-group><year>2012</year><article-title>PK4, a eukaryotic initiation factor 2α(eIF2α) kinase, is essential for the development of the erythrocytic cycle of plasmodium</article-title><source>Proceedings of the National Academy of Sciences of USA</source><volume>109</volume><fpage>3956</fpage><lpage>3961</lpage><pub-id pub-id-type="doi">10.1073/pnas.1121567109</pub-id></element-citation></ref><ref id="bib49"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Zhang</surname><given-names>K</given-names></name><name><surname>Rathod</surname><given-names>PK</given-names></name></person-group><year>2002</year><article-title>Divergent regulation of dihydrofolate reductase between malaria parasite and human host</article-title><source>Science</source><volume>296</volume><fpage>545</fpage><lpage>547</lpage><pub-id pub-id-type="doi">10.1126/science.1068274</pub-id></element-citation></ref></ref-list></back><sub-article article-type="article-commentary" id="SA1"><front-stub><article-id pub-id-type="doi">10.7554/eLife.04106.036</article-id><title-group><article-title>Decision letter</article-title></title-group><contrib-group content-type="section"><contrib contrib-type="editor"><name><surname>Gingeras</surname><given-names>Thomas R</given-names></name><role>Reviewing editor</role><aff><institution>Cold Spring Harbor Laboratory</institution>, <country>United States</country></aff></contrib></contrib-group></front-stub><body><boxed-text><p>eLife posts the editorial decision letter and author response on a selection of the published articles (subject to the approval of the authors). An edited version of the letter sent to the authors after peer review is shown, indicating the substantive concerns or comments; minor concerns are not usually shown. Reviewers have the opportunity to discuss the decision before the letter is sent (see <ext-link ext-link-type="uri" xlink:href="http://elifesciences.org/review-process">review process</ext-link>). Similarly, the author response typically shows only responses to the major concerns raised by the reviewers.</p></boxed-text><p>Thank you for sending your work entitled “Genome-wide Regulatory Dynamics of Translation in the <italic>Plasmodium falciparum</italic> Asexual Blood Stages” for consideration at <italic>eLife</italic>. Your article has been favorably evaluated by Richard Losick (Senior editor) and 3 reviewers, one of whom is a member of our Board of Reviewing Editors.</p><p>The Reviewing editor and the other reviewers discussed their comments before we reached this decision, and the Reviewing editor has assembled the following comments to help you prepare a revised submission.</p><p>This manuscript presents genome-wide transcriptional maps and association of the detected steady-state RNAs to ribosomes as a means of studying the possible use of translation as a post-transcriptional control mechanism during 5 stages of development in <italic>P. falciparum</italic>. The use of such a regulatory strategy would help to explain why there is a continuous flow of transcription from the majority of genes in <italic>P. falciparum</italic> throughout the development cycle in place of a more regulated and dynamic gene expression strategy. In addition, the authors seek to enhance our understanding of how this regulatory strategy is enacted in this organism. Thus, this is a timely and potentially and important manuscript. However, overall, the manuscript contains several important issues that require attention to help readers better appreciate the messages communicated.</p><p>1) Given the critiques of the riboseq approaches (Chew et al, 2013;119 Guttman et al, 2013) it would be important to firmly establish that the RNAs isolated in the ribosome protocol are part of a productive polysome translation complex and instead of individual monosomes. This concern is given support in <xref ref-type="fig" rid="fig1s1">Figure 1–figure supplement 1</xref> which shows considerable size 80s peaks in the untreated samples for the Late Troph, Schizonts and merozoites stages. In short, the detection of RNA protected by ribosomes in these footprint experiments cannot always be with associated with RNA destined for translation. Wilson & Masel, 2011 have reported that non-productive binding of single ribosomes to mRNA do not undergo productive translation. Thus, RNAseq results obtained from the 80S peak might be compared to RNAseq results obtained from RNAs isolated from the polysome peak. This would act as a control against this problematic possibility.</p><p>2) Since the authors do not generate a proteomic dataset for their samples, it is not convincing that the translation efficiency values that they infer from their ribosomal occupancy profiles are direct indicators of protein synthesis. This is especially true given the delay in protein as compared to transcript detection observed during the <italic>P. falciparum</italic> IDC (Le Roch et al., Genome Biol., 2004 and Foth et al., Mol. Cell. Proteomics, 2011). It is also important to note that select noncoding RNAs show a translational efficiency similar to coding regions, even though they are not translated (Ingolia et al., Cell, 2011, and Guttman et al., Cell, 2013); these noncoding RNAs can be distinguished from coding regions only by their ribosomal release score, which was not calculated in this study. Together, the translation efficiency measured here is not enough to indicate if a protein product is being made. The authors have to provide additional evidence or analyses to conclusively state that there is a lack of “overt” translational control.</p><p>3) Another reason for the detection of a transcript in the ribosomal occupancy profile could be due to its presence in other ribonucleoprotein complexes that associate or co-sediment with the 80S monosome fractions. The authors infer the same when they discuss the 3' UTRs of transcripts (<xref ref-type="fig" rid="fig9">Figure 9</xref>). Could this be the case for other regions of apparent ribosomal occupancy such as the 5' leader sequence? The authors should discuss this in more detail. A straightforward experiment would be to determine the protein composition of the 80S monosome fractions isolated for ribosome profiling in different stages of the IDC. A related question: Is there a lack of polysomes in merozoite stages? This appears to be the case in <xref ref-type="fig" rid="fig1s1">Figure 1–figure supplement 1</xref>.</p><p>4) Another concern is that the authors do not discuss their data in light of the polysome occupancy study published by Bunnik et al., Genome Biology, 2013, and the delayed translation phenomenon that has been observed by Le Roch et al., Genome Biol., 2004, and Foth et al., Mol. Cell. Proteomics, 2011.</p><p>5) The antisense RNAs section is not 'up-to-date' with the literature. Three recent studies have detected pervasive transcription of the parasite genome – López-Barragán et al., BMC Genomics, 2011; Siegel et al., BMC Genomics, 2014; Wei et al., PLOS One, 2014 – none of which are referenced and discussed here.</p><p>6) The authors establish a threshold read count, in terms of reads per million mapped (rM), for downstream analysis. The threshold chosen here seems reasonable for this data set, but I do think there is an important clarification to be made. There are two separate questions that might be addressed by a read count threshold: a minimum expression level for biologically significant effects and a minimum count for statistically reliable analysis. An rM threshold is reasonable for the former, while a raw read count seems more appropriate for the latter. Doubling sequence coverage would permit statistically reliable detection of lower expression levels.</p><p>7) At several points the authors discuss a correlation between mRNA abundance and translation, or between transcription and translation; these seem to be used interchangeably. The authors seem to mean that the amount of protein produced correlates well with the abundance of the mRNA that encodes it – but this phrase could also suggest that more abundant mRNAs are translated more highly. The authors should clarify this point when it's first mentioned.</p><p>8) The authors discuss relative translational efficiency (TE). Measures of overall translation and TE in different samples are all scaled by the per-million-mapped factor to correct for differences in sequencing depth, etc. This correction obscures changes in the bulk mRNA content or translational activity between different stages. These effects are discussed in e.g. Lovén et al., Cell 2012. The authors state that “Absolute mean translational efficiencies were the same in all stages” but this conclusion is almost an arithmetic identity after applying per-million-mapped normalization. The polysome profiles presented in <xref ref-type="fig" rid="fig1s1">Figure 1– figure supplement 1</xref> indicate large changes in overall translational activity at different developmental stages.</p><p>9) In discussing RESA2 I was confused by <xref ref-type="fig" rid="fig5s2">Figure 5–figure supplement 2</xref> until I realized that only the truncated reading frame was shown. It would be great to indicate the full-length reading frame on the graph.</p><p>10) In discussing the eIF2alpha kinase IK1, the authors note that the gene is translationally silenced, perhaps reflecting one of the few uORF-mediated translational silencing events in Plasmodium. As eIF2alpha phosphorylation can induce a bypass of such silencing, this seems like a potentially interesting auto regulatory system – in animals, the eIF2alpha phosphatase GADD34 is regulated by uORF silencing in order to provide feedback; here it would seem that positive feedback may act instead.</p><p>11) The authors discuss qualitatively the observation that many abundant but poorly translated mRNAs are not detectable by proteomics, in contrast to well-translated messages. It would be valuable to quantify detectability, controlling for mRNA abundance, and test less anecdotally whether detectability, as a proxy for protein abundance, follows translation better than mRNA abundance.</p><p>12) It is striking that the 5' UTRs in Plasmodium were longer than the 3' UTRs – this differs from the situation in other eukaryotes, over a wide range of absolute UTR length. The authors should comment on this.</p></body></sub-article><sub-article article-type="reply" id="SA2"><front-stub><article-id pub-id-type="doi">10.7554/eLife.04106.037</article-id><title-group><article-title>Author response</article-title></title-group></front-stub><body><p><italic>1) Given the critiques of the riboseq approaches (Chew et al, 2013;119</italic> <italic>Guttman et al, 2013) it would be important to firmly establish that the RNAs isolated in the ribosome protocol are part of a productive polysome translation complex and instead of individual monosomes. This concern is given support in</italic> <xref ref-type="fig" rid="fig1s1"><italic>Figure 1–figure supplement 1</italic></xref> <italic>which shows considerable size 80s peaks in the untreated samples for the Late Troph, Schizonts and merozoites stages. In short, the detection of RNA protected by ribosomes in these footprint experiments cannot always be with associated with RNA destined for translation. Wilson & Masel, 2011 have reported that non-productive binding of single ribosomes to mRNA do not undergo productive translation. Thus, RNAseq results obtained from the 80S peak might be compared to RNAseq results obtained from RNAs isolated from the polysome peak. This would act as a control against this problematic possibility</italic>.</p><p>We thank the reviewers for pointing out this important issue. Our response to this issue and #2 are linked; therefore please refer to the section below for our response.</p><p><italic>2) Since the authors do not generate a proteomic dataset for their samples, it is not convincing that the translation efficiency values that they infer from their ribosomal occupancy profiles are direct indicators of protein synthesis. This is especially true given the delay in protein as compared to transcript detection observed during the P. falciparum IDC (Le Roch et al., Genome Biol., 2004 and Foth et al., Mol. Cell. Proteomics, 2011). It is also important to note that select noncoding RNAs show a translational efficiency similar to coding regions, even though they are not translated (Ingolia et al., Cell, 2011, and Guttman et al., Cell, 2013); these noncoding RNAs can be distinguished from coding regions only by their ribosomal release score, which was not calculated in this study. Together, the translation efficiency measured here is not enough to indicate if a protein product is being made. The authors have to provide additional evidence or analyses to conclusively state that there is a lack of “overt” translational control</italic>.</p><p>To summarize the questions raised by reviewer points #1 and #2:</p><p>A) Do monosome fragments actually represent productive translating ribosomes?</p><p>B) Given that there are footprints that apparently map to ncRNAs, which, by definition, are not expected to be translated, what is evidence that the TE values faithfully reflect translational efficiency?</p><p>C) What about the Wilson & Masel findings with regard to unproductive ribosome binding?</p><p>D) Given that we do not have a proteomic dataset to accompany this study, how do we reconcile our findings with the apparent delay in protein translation that is observed during the IDC as published by Le Roch and Foth?</p><p>Issues A & B are inherently linked. First, a number of methods have recently arisen to assist in the discrimination of RNA fragments that derive from ribosome protection as opposed to anything else.</p><p>In a recently published study by Ingolia et al. (<italic>Cell Rep</italic>., September 2014), the authors propose a simple metric to distinguish true ribosome footprints from either spurious RNA fragments or non-ribosomal RNA binding proteins by comparing the cumulative read size distribution between protein coding genes and classical non-coding sequences (such as telomerase RNA or RNAse P). According to this study, comparison of read size distribution alone is sufficient to identify true ribosome footprints.</p><p>Our manuscript actually uses this exact metric of read size distribution comparisons in Figure S8. Our data clearly shows a difference between footprints mapping to the 5’ leader and CDS, as opposed to noncoding RNA derived fragments (rRNAs, tRNAs). Thus, for each region with sufficient numbers of mapped fragments, we assess whether the cumulative distribution of fragment lengths is consistent with ribosome binding or not.</p><p>In another published paper by Guttman et al., the authors use a metric called the ribosomal release score (RRS) which is similar to the disengagement score (DS) from Chew et al, to distinguish true coding regions from noncoding RNAs by determining whether there is evidence for a substantial decrease in the number of ribosome footprints following the stop codon, a feature that would be expected for a legitimately translated region.</p><p>We have further analyzed our dataset using the ribosomal release score (RSS) metric for a number of non-coding RNAs for which we have derived TE values and compared these ncRNAs with the rest of our protein coding genes (below). We find that TE values for protein coding genes are well correlated with the ribosome release scores (black dots), yet nearly all ncRNAs yield poor release scores, regardless of TE value (red dots).<fig id="fig11" position="float"><label>Author response image 1.</label><graphic xlink:href="elife04106f011"/></fig></p><p>Thus, both the method using cumulative distributions of fragment lengths and the RRS metric are consistent with ongoing translation in protein coding regions, as opposed to regions within classic ncRNAs (like tRNAs). However, it should be noted that not all regions fit neatly into one category or another. Regions such as the 5’ leader sequences reveal ribosome density with fragment length distributions that are indistinguishable from those within coding regions, yet in such regions, an RRS metric cannot be calculated. Ingolia et al. (Cell Rep., September 2014), argue that translation on transcripts outside of protein-coding regions can in fact produce polypeptides and therefore a TE value is valid for many non-coding sequences such as 5’UTRs or lncRNAs.</p><p>Within coding regions, the legitimacy of ribosome protected fragments and their relationship to translational efficiency is now widely supported in a variety of organisms [Stern-Ginossar et al., Science 2012, Oh et al., Cell 2011, Dunn et al., <italic>eLife</italic> 2013, Ingolia et al. Cell 2011, Smith et al., Cell Rep. 2014, reviewed in Ingolia, Nat. Rev Genet. 2014], yet clearly there is still some debate in the field with regard to unexpected ribosome density outside of coding regions. Realizing the complexity of this issue, we have chosen a conservative path for our manuscript, whereby we use TE to only refer to protein coding regions, and RD (ribosomal density) to refer to those outside of the coding regions (like 5’ RD).</p><p>C) What about the Wilson & Masel findings with regard to unproductive ribosome binding?</p><p>With specific regard to the Wilson & Masel 2011 paper cited by the reviewer, these authors report ribosomal association to stable unannotated transcripts, yet did not actually assess whether these complexes were engaged in actual translation and synthesis of a peptide. In fact they state:</p><p><bold>“</bold>While we do not know the extent to which this ribosomal association leads to translation, these SUTs, apart from RDT1, do not appear to encode functional, protein-coding genes.<bold>”</bold></p><p>The authors speculate that perhaps ribosome association to unannotated or short ORFs may indeed lead to translation “by accident at least at low levels” and “provide raw material for <italic>de novo</italic> birth of protein-coding genes.”</p><p>D) How do we reconcile our findings with the apparent delay in protein translation that is observed during the IDC as published by Le Roch and Foth?</p><p>With regards to a comparison between the data presented in this manuscript and previous mass spectrometry based studies, the reviewer cites two articles which report a delay in mRNA abundance with respect to protein detection, Le Roch et al., Genome Biol., 2004 and Foth et al., Mol. Cell. Proteomics, 2011.</p><p>In Le Roch et al., the authors analyzed a total of 459 mRNA-protein pairs and found that for 171 genes transcript abundance in one stage correlated better with protein levels of the next stage (>0.5 increase in Spearman rank correlation). However, it should be noted that 45 of these 171 transcripts were also better correlated to the proteins in the <italic>previous</italic> stage. By excluding these genes, a delay between transcript abundance and the detection of its corresponding protein may be attributed to only 129 transcripts.</p><p>In Foth et al. the authors analyze 125 mRNA-protein pairs and detect multiple isoforms for some of the proteins. A delay between mRNA abundance and protein detection is observed for 63.7% of 237 protein isoforms (it is unclear how many of the 125 genes analyzed this represents).</p><p>Here, we report 177 and 124 translationally up- and down-regulated genes, respectively. This represents 295 genes, which is more than twice the number detected in Le Roch or Foth. However, these 295 genes represent only 8% of the 3,605 genes present in our analysis, and for remaining 92%, we find that transcript abundance is proportional to ribosome occupancy throughout the lifecycle.</p><p>Thus, while we identify more instances of translational control than all prior studies, these instances appear to be the exception, and not the rule, as the vast majority of genes appear to lack overt translational control as measured by this assay.</p><p><italic>3) Another reason for the detection of a transcript in the ribosomal occupancy profile could be due to its presence in other ribonucleoprotein complexes that associate or co-sediment with the 80S monosome fractions. The authors infer the same when they discuss the 3' UTRs of transcripts (</italic><xref ref-type="fig" rid="fig9"><italic>Figure 9</italic></xref><italic>). Could this be the case for other regions of apparent ribosomal occupancy such as the 5' leader sequence? The authors should discuss this in more detail. A straightforward experiment would be to determine the protein composition of the 80S monosome fractions isolated for ribosome profiling in different stages of the IDC. A related question: Is there a lack of polysomes in merozoite stages? This appears to be the case in</italic> <xref ref-type="fig" rid="fig1s1"><italic>Figure 1–figure supplement 1</italic></xref>.</p><p>The issue of distinguishing between fragments that derive from actual ribosome binding as opposed to other RNA binding proteins is discussed above. In short, we use the cumulative distribution of read lengths in a manner identical to Ingolia et al. (Cell Rep., September 2014), as a discrimination method. Here, we find that the fragments deriving from the 5’ leaders show a perfectly overlapping cumulative distribution to coding regions. This is not the case for known ncRNAs, like tRNAs.</p><p>Regardless, whether our 80S fractions contain binding activities other than ribosomes can be assessed by mass spectrometry, as the reviewer points out. During the course of this study, we in fact prepared schizont and merozoite monosome fractions and a pooled schizont polysome fraction for peptide sequencing by tandem mass spectrometry (LC-MS/MS). The results of this analysis are shown below:<table-wrap id="tblu1" position="anchor"><table frame="hsides" rules="groups"><thead><tr><th/><th>total spectral counts</th><th>% human</th><th>% plasmodium</th><th>% of plasmodium ribosomal</th><th>% of plasmodium non-ribosomal</th></tr></thead><tbody><tr><td>merozoite monosomes</td><td>241</td><td>1.7</td><td>98.3</td><td>100.0</td><td>0.0</td></tr><tr><td>schizont monosomes</td><td>2937</td><td>35.5</td><td>64.5</td><td>89.6</td><td>10.4</td></tr><tr><td>schizont polysomes</td><td>3128</td><td>8.7</td><td>91.3</td><td>95.7</td><td>4.3</td></tr></tbody></table></table-wrap></p><p>These results indicate that the monosomes and polysome fractions that we collected contain mostly ribosomal components (89–100%). The remainder 4.3-10.4% of Plasmodium proteins that are non-ribosomal constitute highly expressed proteins such as heat shock proteins, exported proteins, membrane proteins, GAPDH, and other housekeeping genes.</p><p>Only 0.26% of spectral counts in the schizont monosome fraction mapped to putative non-ribosomal RNA-binding proteins (such as the DNA/RNA-binding protein Alba). While we cannot exhaustively account for every protein that co-sediments with the 80S peak, this proteomic evidence confirms that the vast majority of the 80S proteins are ribosomal.</p><p>With regard to the polysome profiles from merozoites, we agree that the relative peak height of the polysome region compared to the 80S appears reduced. This could be the result of a global reduction in overall translation at this stage. In this study, it should be noted that we do not directly assess bulk or absolute levels of overall translation activity. Rather, this study examines the relative translational efficiency of transcripts on a per-timepoint, per message basis. In merozoites, we do observe a slightly broader distribution of efficiencies, but not dramatically so.</p><p><italic>4) Another concern is that the authors do not discuss their data in light of the polysome occupancy study published by Bunnik et al., Genome Biology, 2013, and the delayed translation phenomenon that has been observed by Le Roch et al., Genome Biol., 2004, and Foth et al., Mol. Cell. Proteomics, 2011</italic>.</p><p>For a comparison with Le Roch et al., Genome Biol., 2004, and Foth et al., Mol. Cell. Proteomics, 2011 please see our response to points #1 & #2, part D.</p><p>The reviewers ask for a comparison with the work published by Bunnik et al. where polysome-associated mRNA was compared to total steady state mRNA. A major conclusion of this paper is that there exists strong translational control for 471 genes, where the peak of transcriptional abundance is in trophozoite or schizont stage, but they are most highly associated with polysomes in the ring stage.</p><p>There are a number of confounding factors that obstruct a fair comparison on the work presented here and that of Bunnik et al. These factors include the strain, the time points sampled, the methodology for isolation of polysomes, and most importantly, the overall sequencing coverage. We discuss each of these factors below.</p><p>In Bunnik et al., the authors used the 3D7 strain of <italic>P. falciparum</italic> to assess hours 0, 18, and 36 hours post invasion whereas we used the W2 strain to sample hours 11, 21, 31, 45 in addition to purified merozoites.</p><p>A critical step in ribosome profiling is the isolation of intact polysomes. In our hands, directly using sucrose gradients yields polysome profiles enriched in higher order polysomes, with more than 3 ribosomes. Polysome profiles in Bunnik et al. are obtained by pelleting parasite lysates on sucrose cushions, and re-loading the ribosomal pellets on sucrose gradients. This additional step may cause the dissociation of ribosomes from mRNAs, and could explain the reduced amount of higher order polysomes in their profiles, which in turn may skew their assessment of translation towards lower values. (Compare Bunnik <xref ref-type="fig" rid="fig1">Figure 1 A-D</xref>)</p><p>Finally, sequencing coverage and biological replicates are the most important determinant of obtaining reliable data in our experiments. In Bunnik et al, there is only a single biological replicate (just the schizont stage). Their deepest exome-wide coverage for polysomes is 12.7x (Table 1, Bunnik et al), and their average coverage for polysomes is 5.3x. The average exome-wide coverage of our ribosome footprint data is 159.9x, which is, on average, 30 times greater than Bunnik et al. (table below).<table-wrap id="tblu2" position="anchor"><table frame="hsides" rules="groups"><tbody><tr><td/><td>X-fold exome-wide coverage</td></tr><tr><td>timepoint</td><td>ribosome footprints</td></tr><tr><td>rings</td><td>315.3</td></tr><tr><td>early troph.</td><td>205.8</td></tr><tr><td>late troph.</td><td>55.3</td></tr><tr><td>schizonts</td><td>53.6</td></tr><tr><td>merozoites</td><td>169.7</td></tr><tr><td><bold>average</bold></td><td><bold>159.9</bold></td></tr><tr><td><bold>average all</bold></td><td><bold>137.4</bold></td></tr></tbody></table></table-wrap></p><p>More importantly, the major difference between the Bunnik dataset and our own is the set of 471 genes that the authors identify as being translationally up-regulated in rings. Approximately 76% of these genes fall below our threshold for minimum read coverage (32rM) for analysis, despite the fact that our ring stage coverage is 456-fold greater than the Bunnik zero hour polysome sample (315x coverage compared to Bunnik’s 0.69x).</p><p>Of the remaining genes (excluding apicoplast genes, var, rifin, and stevors), we identified only five genes that we would consider to be translationally up-regulated in rings.</p><p>Through our own biological replicates of each time point and our assessment of reliability with respect to sequencing coverage (<xref ref-type="fig" rid="fig1s3">Figure 1–figure supplement 3</xref>), we believe it would be inappropriate to draw conclusions for genes or whole datasets that lack sufficient coverage, and thus it is not possible to cleanly compare the two datasets.</p><p><italic>5) The antisense RNAs section is not 'up-to-date' with the literature. Three recent studies have detected pervasive transcription of the parasite genome – López-Barragán et al., BMC Genomics, 2011; Siegel et al., BMC Genomics, 2014; Wei et al., PLOS One, 2014 – none of which are referenced and discussed here</italic>.</p><p>We thank the reviewers for pointing this out. We updated our antisense section to reference López-Barragán et al., BMC Genomics, 2011 and Siegel et al., BMC Genomics, 2014. The text now reads “Previous serial analysis of gene expression (SAGE) (<xref ref-type="bibr" rid="bib39">Patankar et al. 2001</xref>), nuclear run-on experiments (<xref ref-type="bibr" rid="bib33">Militello et al. 2005</xref>) and more recently antisense splicing events detected by RNA-seq (<xref ref-type="bibr" rid="bib43">Sorber et al. 2011</xref>; <xref ref-type="bibr" rid="bib30">López-Barragán et al. 2011</xref>), suggest that antisense RNAs are synthesized by RNA pol II and may constitute up to ∼12% of the erythrocytic-stage steady-state RNA (<xref ref-type="bibr" rid="bib18">Gunasekera et al. 2004</xref>), yet their presence and biological role, if any, remains unclear. A more recent study found no correlation between natural antisense transcript levels and protein abundance (<xref ref-type="bibr" rid="bib42">Siegel et al. 2014</xref>).”</p><p>The work by Wei et al. analyzes uncapped/non-adenylated intermediate size ncRNAs, which we do not focus on.</p><p><italic>6) The authors establish a threshold read count, in terms of reads per million mapped (rM), for downstream analysis. The threshold chosen here seems reasonable for this data set, but I do think there is an important clarification to be made. There are two separate questions that might be addressed by a read count threshold: a minimum expression level for biologically significant effects and a minimum count for statistically reliable analysis. An rM threshold is reasonable for the former, while a raw read count seems more appropriate for the latter. Doubling sequence coverage would permit statistically reliable detection of lower expression levels</italic>.</p><p>We took a conservative approach by choosing a minimum of rM of 32 per gene. We agree with the reviewer that a reliable statistical analysis could still be obtained by lowering this threshold and we did so for the antisense analysis where we chose a ≥16rM bar. Future studies into these low count reads will be necessary to establish their biological significance.</p><p><italic>7) At several points the authors discuss a correlation between mRNA abundance and translation, or between transcription and translation; these seem to be used interchangeably. The authors seem to mean that the amount of protein produced correlates well with the abundance of the mRNA that encodes it – but this phrase could also suggest that more abundant mRNAs are translated more highly. The authors should clarify this point when it's first mentioned</italic>.</p><p>We fully agree with the reviewers and clarified the ambiguity by changing the text from “To determine the exact level of correlation between mRNA abundance and translation we directly compared mRNA and ribosome footprint density measurements (<xref ref-type="fig" rid="fig2">Figure 2C</xref>). In general, translation is tightly correlated with transcription for all phasic and non-phasic genes in rings (r = 0.85), early trophozoites (r = 0.93), late trophozoites (r = 0.91), schizonts (r = 0.89) and purified merozoites (r = 0.86)”, to “To determine the exact level of correlation between transcription and translation we directly compared mRNA and ribosome footprint density measurements (<xref ref-type="fig" rid="fig2">Figure 2C</xref>). In general, translation is tightly correlated with transcription for all phasic and non-phasic genes in rings (r = 0.85), early trophozoites (r = 0.93), late trophozoites (r = 0.91), schizonts (r = 0.89) and purified merozoites (r = 0.86). This indicates that when an mRNA is detected in one stage it is associated proportionally with ribosomes within the same stage.”</p><p>We changed: “The simultaneous capture of transcription and translation is expected to be…” to “The simultaneous capture of mRNA abundance and translation is expected to be…”</p><p>We changed the phrase “… peak transcription occurring in rings and schizonts” to “…peak mRNA abundance occurring in rings and schizonts”.</p><p><italic>8) The authors discuss relative translational efficiency (TE). Measures of overall translation and TE in different samples are all scaled by the per-million-mapped factor to correct for differences in sequencing depth, etc. This correction obscures changes in the bulk mRNA content or translational activity between different stages. These effects are discussed in e.g. Lovén et al., Cell 2012. The authors state that “Absolute mean translational efficiencies were the same in all stages” but this conclusion is almost an arithmetic identity after applying per-million-mapped normalization. The polysome profiles presented in</italic> <xref ref-type="fig" rid="fig1s1"><italic>Figure 1–figure supplement 1</italic></xref> <italic>indicate large changes in overall translational activity at different developmental stages</italic>.</p><p>This study measures the relative change in translational efficiency on a per-gene, per-time point basis, and does not measure absolute changes in bulk translation or transcription, as the reviewer correctly points out. To clarify this for the reader, we have added a statement in the Discussion highlighting this important point “However, we note that this dataset only provides a direct measure of relative changes in translational efficiencies rather than changes in bulk transcription and translation.”</p><p>With respect to the statement that the reviewer highlights regarding mean translational efficiencies, we have corrected and amended this sentence, since the mean TEs are in fact not identical in all stages (for example, the mean TE in schizonts compared to early trophs are 1.47-fold different). As shown in <xref ref-type="fig" rid="fig4">Figure 4</xref>, the shape of the TE distributions are not identical, and thus we would indeed expect some variation in the mean.</p><p>The text now reads:</p><p>“Absolute mean translational efficiencies in all stages (log<sub>2</sub>TE <italic>µ</italic><sub>Rings</sub> = -0.43, log<sub>2</sub>TE <italic>µ</italic><sub>E.trophs.</sub> = -0.56, log<sub>2</sub>TE <italic>µ</italic><sub>L.trophs.</sub> = -0.31, log<sub>2</sub>TE <italic>µ</italic><sub>Schizonts</sub> = -0.16 and log<sub>2</sub>TE <italic>µ</italic><sub>Merozoites</sub> = -0.68) had a maximum fold difference of 1.47-fold observed between early trophozoites and schizonts.”</p><p>9) In discussing RESA2 I was confused by <xref ref-type="fig" rid="fig5s2"><italic>Figure 5–figure supplement 2</italic></xref> until I realized that only the truncated reading frame was shown. It would be great to indicate the full-length reading frame on the graph.</p><p>We thank the reviewers for indicating to us how to make the figure clearer. As advised, we added the annotated gene model described in PlasmoDB version 7.1, which is the version used for reference in this work.<fig id="fig12" position="float"><label>Author response image 2.</label><graphic xlink:href="elife04106f012"/></fig></p><p><italic>10) In discussing the eIF2alpha kinase IK1, the authors note that the gene is translationally silenced, perhaps reflecting one of the few uORF-mediated translational silencing events in Plasmodium. As eIF2alpha phosphorylation can induce a bypass of such silencing, this seems like a potentially interesting auto regulatory system – in animals, the eIF2alpha phosphatase GADD34 is regulated by uORF silencing in order to provide feedback; here it would seem that positive feedback may act instead</italic>.</p><p>We fully agree with the reviewers. Ribosome footprints partially overlap with 2 of the 18 uORFs counted on the 5’ leader of the eIF2 alpha kinase transcript. Whether these uORFs or other 5’ leader motifs are responsible for the silencing of the downstream coding region remains an important question. Additionally, this gene presents the lowest ribosomal release score calculated, which strongly suggests that it is not being translated.</p><p><italic>11) The authors discuss qualitatively the observation that many abundant but poorly translated mRNAs are not detectable by proteomics, in contrast to well-translated messages. It would be valuable to quantify detectability, controlling for mRNA abundance, and test less anecdotally whether detectability, as a proxy for protein abundance, follows translation better than mRNA abundance</italic>.</p><p>We did not wish to make the statement that genes that have high mRNA abundance would not be detected by proteomic methods, and we thank reviewers for highlighting this misunderstanding. We have therefore revised our manuscript as following:</p><p>We changed the phrase from: “Thus, our ribosome profiling dataset highlights instances of translational control of genes for which, in the presence of abundant mRNA, the protein counterpart is unlikely to be detected by proteomic methods” to “Thus, our ribosome profiling dataset highlights instances of translational control of genes that may not be detected by proteomic methods.”</p><p><italic>12) It is striking that the 5' UTRs in Plasmodium were longer than the 3' UTRs – this differs from the situation in other eukaryotes, over a wide range of absolute UTR length. The authors should comment on this</italic>.</p><p>A review of the current literature reveals a wide distribution of different 5’UTR sizes in different types of organisms. Previous studies found that 5’UTRs of mRNAs encoding genes which need to be strongly and finely regulated are often longer than average (Kozak 1987; Hurowitz and Brown 2003; David et al. 2006; Nagalakshmi et al. 2008; Bruno et al. 2010). On the other hand, Lynch et al. (2005) proposed that the length of the 5’UTR is driven mainly by random genetic drift and mutational processes that cause stochastic turnover in transcription-initiation sites and premature start codons, combined with a selective constrain against the premature initiation of translation. Additional sequencing and mapping of transcript architecture in other eukaryotes will help establish in the future whether or not <italic>P. falciparum</italic> truly lies at the extreme of the distribution.</p></body></sub-article></article> |