A redesigned TetR-aptamer system to control gene expression in Plasmodium falciparum

One of the most powerful approaches to understanding gene function involves turning genes on and off at will and measuring the impact at the cellular or organismal level. This particularly applies to the cohort of essential genes where traditional gene knockouts are inviable. In Plasmodium falciparum, conditional control of gene expression has been achieved by using multi-component systems in which individual modules interact with each other to regulate DNA recombination, transcription or posttranscriptional processes. The recently devised TetR-DOZI aptamer system relies on the ligand-regulatable interaction of a protein module with synthetic RNA aptamers to control the translation of a target gene. This technique has been successfully employed to study essential genes in P. falciparum and involves the insertion of several aptamer copies into their 3’ untranslated regions (UTRs) which provide control over mRNA fate. However, aptamer repeats are prone to recombination and one or more copies can be lost from the system, resulting in a loss of control over target gene expression. We rectified this issue by redesigning the aptamer array to minimize recombination while preserving the control elements. As proof of concept, we compared the original and modified arrays for their ability to knock down the levels of a putative essential apicoplast protein (PF3D7_0815700) and demonstrated that the modified array is highly stable and efficient. This redesign will enhance the utility of a tool that is quickly becoming a favored strategy for genetic studies in P. falciparum. Importance Malaria elimination efforts have been repeatedly hindered by the evolution and spread of multidrug-resistant strains of Plasmodium falciparum. The absence of a commercially available vaccine emphasizes the need for a better understanding of Plasmodium biology in order to further translational research. This has been partly facilitated by targeted gene deletion strategies for the functional analysis of parasite genes. However, genes that are essential for parasite replication in erythrocytes are refractory to such methods, and require conditional knockdown or knockout approaches to dissect their function. One such approach is the TetR-DOZI system that employs multiple synthetic aptamers in the untranslated regions of target genes to control their expression in a tetracycline-dependent manner. Maintaining modified parasites with intact aptamer copies has been challenging since these repeats are frequently lost by recombination. By interspacing the aptamers with unique sequences, we created a stable genetic system that remains effective at controlling target gene expression.


Abstract: 24
One of the most powerful approaches to understanding gene function involves turning 25 genes on and off at will and measuring the impact at the cellular or organismal level. This 26 particularly applies to the cohort of essential genes where traditional gene knockouts are 27 inviable. In Plasmodium falciparum, conditional control of gene expression has been achieved 28 by using multi-component systems in which individual modules interact with each other to 29 regulate DNA recombination, transcription or posttranscriptional processes. The recently 30 devised TetR-DOZI aptamer system relies on the ligand-regulatable interaction of a protein 31 module with synthetic RNA aptamers to control the translation of a target gene. This technique 32 has been successfully employed to study essential genes in P. falciparum and involves the 33 insertion of several aptamer copies into their 3' untranslated regions (UTRs) which provide 34 control over mRNA fate. However, aptamer repeats are prone to recombination and one or more 35 copies can be lost from the system, resulting in a loss of control over target gene expression. We 36 interfere with aptamer function, as demonstrated by efficient TC-dependent knockdown of the 129 PfAUBL protein that subsequently results in parasite death, indicating an essential role for 130 PfAUBL. The enhanced stability of the aptamer array will expand the usage of TetR-DOZI for 131 gene function studies in P. falciparum. 132 133

Results: 134
Aptamer arrays undergo truncation in E. coli and in P. falciparum 135 The TetR-DOZI aptamer system has been successfully used for the conditional 136 knockdown of several P. falciparum proteins. We implemented this approach to study a putative 137 apicoplast protein called PfAUBL (PF3D7_0815700). PfAUBL was identified as essential for 138 blood-stage survival in the genome-wide piggyBac mutagenesis screen conducted in P. 139 falciparum (43). The protein contains an N-terminal apicoplast transit peptide and a ubiquitin-140 like domain (44). We utilized a previously published plasmid pMG74 for the insertion of a 10X 141 aptamer array downstream of the PfAUBL coding region (24). Plasmids from individual 142 bacterial colonies were analyzed by PCR to verify the size of the aptamer array. While some 143 plasmids contained a full-length ( 1 kb) array, other plasmids appeared to have lost a few 144 aptamers from their arrays (Fig. 1A). We selected a plasmid with an intact sequence-confirmed 145 array to generate PfAUBL Apt parasites. Correct insertion of the aptamer array at the 3' UTR of 146 PfAUBL Apt was confirmed by PCR (Fig. 1B, Fig. S1 and Fig. S2A). 147 Polyclonal parasite populations from two independent transfections were tested for 148 growth in the presence and absence of anhydrotetracycline (aTC), a tetracycline analog. Both 149 parasite lines exhibited moderate growth inhibition 4-5 days after the removal of aTC, but 150 continued to multiply over the course of the growth assay ( Fig. 1C and Fig. S2B). To determine if the parasites replicating without aTC still contained all 10 aptamers, we performed PCR to 152 determine the size of the array before and after aTC removal. An expected 1 kb band was the 153 predominant product from aTC-treated parasites ( Fig. 1D and Fig. S2C). However, one or more 154 smaller bands were also observed, which indicated that both transfections yielded a mixed 155 population of parasites containing varying sizes of aptamer arrays. Post aTC removal, only the 156 smaller PCR products corresponding to truncated arrays could be amplified from the surviving 157 parasites. This suggests that PfAUBL is an essential protein and that parasites with a fully 158 functional aptamer array died upon withdrawal of aTC. Presumably, parasites with arrays that 159 contained fewer aptamers produced enough PfAUBL to survive in the absence of aTC. 160 The extended time interval between parasite transfection and phenotypic analysis of 161 transgenic lines could enable the gradual accumulation of parasites with recombined aptamer 162 arrays. We therefore attempted to isolate a clonal parasite population with an intact array from 163 one of the transfections. PCR analysis of five putative clones revealed aptamer arrays of varying 164 sizes. Parasites from well E11 were found to be nonclonal, and none of the remaining samples 165 contained full-length arrays ( Fig. 2A). Parasites from clone E10 had the largest array with 8 166 aptamers and were therefore analyzed further. Clone E10 parasites grown in the absence of aTC 167 exhibited severe growth inhibition beginning on Day 5, and this defect was much more 168 pronounced than the previously analyzed mixed populations (Fig. 2B). However, a small 169 proportion of aTC-depleted parasites continued to grow and replicate. When examined by PCR 170 on Day 10, the surviving parasites were found to contain a truncated array with 5 aptamers, 171 which was confirmed by sequencing (Fig. 2C). This indicates that aptamer loss can occur within 172 a short time frame and confers a selective advantage to parasites in non-permissive growth 173 conditions.

Modification of spacers between RNA aptamers 175
The TetR-aptamer array contains 10 copies of an aptamer separated by identical spacers 176 (86 bp each and 860 bp in total). Such repetitive sequences are prone to recombination events 177 that can lead to truncations. To minimize recombination between individual aptamers and 178 spacers, we made the following modifications to the array. The original aptamers are 51 bp in 179 length and are predicted to fold into stem-loop structures (Fig. 3A). Bases 1-10 and 42-51 are 180 palindromic and can pair with each other to form a stem with a predicted melting temperature of 181 33 C. Bases 1-6 and 45-51 also encode BamHI restriction enzyme sites (Fig. 3A). To facilitate 182 future cloning reactions that may require BamHI, we altered these bases while maintaining the 183 same GC content (and therefore a similar melting temperature). Each modified aptamer now 184 contained a unique palindromic 6 bp sequence on either end. The loop regions containing 185 conserved sequence motifs that are known to be important for TetR binding were left unaltered. 186 The aptamers in the original array were separated by identical 35 bp spacers. We reduced the 187 spacer length to 11 bp and altered the sequence such that each spacer was unique and formed no 188 significant secondary structure with the 10 stem-loop aptamers (Fig. 3B). Thus, each of the array 189 elements now contains 23 bp of unique sequence corresponding to the spacer region (11 bp) and 190 part of the double stranded stem (12 bp). The redesigned spacers have higher GC content than 191 the original spacer (52% versus 44%), thus facilitating the design of unique primers for 192 sequencing or for future genetic manipulations. They also do not contain any common restriction 193 endonuclease sites that could interfere with downstream cloning. We replaced the original 860 194 bp array in the pMG74 plasmid with the redesigned 619 bp array (Fig. 3C) to generate the pKD 195

vector. 196
The redesigned aptamer array exhibits enhanced stability in E. coli and P. falciparum pKD plasmids with PfAUBL homology arms were isolated from a few bacterial colonies 198 and their aptamer arrays were examined by PCR. All tested plasmids were found to contain an 199 intact array ( 0.7 kb) (Fig. 4A). Parasites were transfected as before with a sequence-confirmed 200 pKD plasmid to generate PfAUBL NewApt parasites. The insertion of the new aptamer array at the 201 3'UTR of PfAUBL was confirmed by PCR ( Fig. 4B and Fig. S1). To determine if the aptamers 202 in the redesigned array retained responsiveness to aTC, PfAUBL NewApt parasites were subjected 203 to a growth assay in the presence or absence of aTC. Growth was completely abrogated 4 days 204 after the withdrawal of aTC, and parasites were undetectable in cultures beyond day 5 (Fig. 4C). 205 The flow cytometric results were confirmed by manual inspection of blood smears. PCR analysis 206 of parasites 4 days post aTC removal did not indicate any loss of aptamers from the array (Fig.  207   4D). Cultures depleted of aTC were maintained for 21 additional days, but no parasite 208 recrudescence was observed. These results implied that the use of unique spacers between 209 identical aptamers was sufficient to prevent recombination within the array and that the changes 210 that we made to the aptamer array did not affect the functionality of the knockdown system. 211 PfAUBL NewApt parasites maintained in permissive culture conditions for several months did not 212 exhibit any loss of aptamers. 213 The pMG74 plasmid was designed to append 1X HA and 1X FLAG tags onto the C-214 terminus of the target protein. However, we failed to detect PfAUBL in PfAUBL Apt parasites 215 using anti-HA or anti-FLAG antibodies. To enhance detection capabilities, we used a 3X HA tag 216 in the pKD plasmid. Probing for PfAUBL with anti-HA antibody demonstrated robust protein 217 detection in aTC-treated PfAUBL NewApt parasites (Fig. 4E). To determine the efficiency of 218 protein knockdown, we examined the levels of PfAUBL in aTC-depleted parasites over a 4-day period. Protein levels were undetectable 24 hours post aTC removal, indicating that the 220 knockdown was quick and efficient. 221 thus permitting the evaluation of null as well as hypomorphic phenotypes. However, its broad 229 implementation has been limited by two key issues. One is that large dynamic range control of 230 gene expression is achieved only when aptamers are introduced on either side of the target 231 gene's coding region (24). While this is easy to accomplish with transgenes, modifying both the 232 5' and 3' UTRs of native genes is less straightforward. The use of a 10X aptamer array in the 3' 233 UTR alone has been sufficient to knock down the expression of several genes, but it may not be 234 enough to achieve complete conditional control of genes that are expressed at high levels. A 235 recent publication overcomes this issue by utilizing a linear vector that carries a recodonized 236 copy of the target gene and homology arms corresponding to the 5' and 3' UTRs, thereby 237 allowing simultaneous insertion of aptamers into the upstream and downstream sequences (42). 238 This study addresses a second drawback of the TetR-DOZI system, namely the instability 239 of the 3' UTR aptamer array due to its repetitive nature. Our results show that truncated arrays 240 arise frequently and that arrays below a certain threshold of aptamers are unable to exert 241 sufficient control over gene expression. This is supported by the original study describing the 242 TetR-RNA aptamers in which a parallel analysis of arrays containing 5 or 10 aptamers showed 243 that the 5X array was unable to regulate reporter gene expression in yeast (24). By replacing the 244 identical spacers in between aptamers with unique sequences, we were able to avoid array 245 truncations. Importantly, modification of the spacers did not interfere with the functionality of 246 the aptamers. We demonstrated this by successfully knocking down the expression of the 247 ubiquitin-like protein PfAUBL, which led to parasite death. The redesigned aptamer array can 248 also be employed in the linear vector described above to facilitate the installation of aptamers in 249 both the 5' and 3' UTRs. The combination of these techniques will greatly expand the 250 application of the TetR-DOZI system in efforts to assign functions to the hundreds of 251 uncharacterized essential P. falciparum genes. 252 253

Plasmid construction 262
The pMG74 plasmid was digested with AatII and XmaI to remove 1X HA and 1X FLAG 263 tags and the existing aptamer array (24). A small synthetic DNA fragment containing cloning 264 sites for the aptamer array (PspOMI and XmaI) and residues encoding a 3X HA tag was inserted into pMG74 using the AatII and XmaI sites (LifeSCT). The redesigned 619 bp aptamer array 266 ( Fig. 3C) was cut with PspOMI and XmaI from a synthetic plasmid (LifeSCT) and ligated into 267 the same sites in pMG74. The resulting plasmid was called pKD and contains the homology arm 268 cloning sites, 3X HA tag and aptamer array shown in Fig. S3A. 269 To create PfAUBL knockdown constructs, pMG74 and pKD were digested with AscI and 270 AatII. Homology arms HA1 and HA2 of the PfAUBL gene were amplified from P. falciparum 271 genomic DNA using the primer pairs AUBL HA1.kd-F + AUBL HA1.kd-R and AUBL HA2.kd-272 F + AUBL HA2.kd-R, respectively ( Table S1). The homology arms were fused in an additional 273 PCR reaction using the primer pair AUBL HA2.kd-F + AUBL HA1.kd-R, creating a combined 274 HA2-HA1 fragment separated by two EcoRV sites. This fragment was inserted into digested 275 pMG74 and pKD plasmids to generate pMG74 PfAUBL and pKD PfAUBL , respectively. Prior to 276 transfection, these plasmids were linearized with EcoRV. 277 The Cas9 enzyme and the guide RNA were expressed from a modified version of pUF1-278 Cas9 (46). Plasmid pUF1-Cas9 (11,096 bp) was modified in several steps to accommodate the 279 insertion of the guide RNA expression cassette found in pRS-LacZ (47). In the first step, the 280 promoter driving Cas9 expression (Hsp86) was removed using XhoI and AflII and replaced with 281 an adaptamer formed from 5'-phosphorylated primers pCas.XhoBtg.F and pCas.XhoBtg.R 282 (Table S1). This plasmid was then digested with XhoI/BtgZI, blunted with T4 polymerase and 283 ligated. The resulting plasmid is smaller (10,221bp) and brings Cas9 expression under the 284 control of the bidirectional calmodulin promoter. In the second step, a 340 bp region of the 285 bluescript plasmid backbone containing several endonuclease sites was removed through 286 digestion with EcoRI/AatII, blunted with T4 polymerase and ligated. In the third step, digestion 287 with XmaI/SapI was used to remove the 3' UTR of Cas9 (PbDHFR-TS) and a complete lac 288 operon found in the bluescript plasmid backbone. This region was replaced with an adaptamer 289 formed from the 5'-phosphorylated primers pUF.NotI.F and pUF.NotI.R (Table S1) containing a 290 NotI site for insertion of the guide RNA cassette found in pRS-LacZ. Before this insertion could 291 take place, however, two BsaI sites in the plasmid backbone had to be removed (BsaI will be 292 used later to insert guide RNA sequences). The plasmid was digested with BsaI, removing a 293 1,992 bp region containing both recognition sites for this Type IIs endonuclease. PCR primers 294 BsaMut.GG.F and BsaMut.GG.R (Table S1) were designed with flanking BsaI sites for Golden 295 Gate insertion back into the plasmid, but with two single base changes to alter the original two 296 BsaI recognition sites. Thus, after digestion of the PCR product with BsaI, and ligation back into 297 the cut plasmid, the resulting plasmid was identical in size (8,757 bp) and contained two single 298 nucleotide changes that removed both BsaI recognition sites. The guide RNA expression cassette 299 from plasmid pAIO (29) was then excised with NotI and inserted into the same site 300 bidirectionally, generating a forward and a reverse orientation. The orientation with the U6 5' 301 regulatory element adjacent to Cas9 was chosen with the expectation that this region would act 302 as a 3' UTR for Cas9 (it contains the 3' UTR of the ribosomal L18 protein, PF3D7_1341200). 303 Finally, the plasmid was digested with BtgZI to remove the pre-guide RNA and the LacZ 304 expression cassette from pRS-LacZ (47) was excised with BsaI and inserted into this site. The 305 final pCasG-LacZ plasmid (10,886) is smaller than the original pUF1-Cas9 plasmid and contains 306 a guide RNA expression cassette that can be used for blue/white colony screening after insertion 307 of a guide RNA sequence with BsaI (Fig. S3B). 308 To generate a pCasG-LacZ plasmid with a guide RNA sequence that targets PfAUBL, 309 pCasG-LacZ was digested with BsaI. PfAUBL-specific guide RNAs were synthesized as oligos 310 (Table S1), annealed, and inserted into the digested plasmid using In Fusion.

Parasite transfections 312
P. falciparum NF54 attB parasites were transfected using a previously described method with 313 minor modifications (48). Briefly, 350 L of red blood cells were electroporated with 75 g each 314 of the pCasG PfAUBL plasmid and either the linearized pMG74 PfAUBL or pKD PfAUBL plasmids. The 315 electroporated RBCs were mixed with 1.5 mL of a mixed-stage culture at 4% parasitemia. 316 Transfected cultures were maintained in CMA with 0.5 M anhydrotetracycline (aTC) (Cayman 317 Chemical) for the first 48 hours, after which they were cultured in selective media containing 2.5 318 g/mL Blasticidin (Corning) and 1.5 M DSM-1 (Calbiochem) along with aTC for 7 days. 319 Cultures were then switched to CMA with aTC until the reemergence of parasites, upon which 320 they were transferred to CMA containing Blasticidin and aTC. Single clones of PfAUBL Apt 321 parasites were isolated by limiting dilution in a 96-well plate. 322

Genotyping and aptamer PCR 323
A 3 L volume of parasite culture or 10 ng of plasmid was used in 25 L PCR reactions 324 with Phusion High-Fidelity DNA polymerase (NEB). To confirm pMG74 PfAUBL insertion into the 325 3' UTR of PfAUBL, the following primer pairs were used: AUBL 5' F + pMG74 R for the 5' 326 integration locus, and TetR-seq + AUBL 3' R for the 3' integration locus. To confirm pKD PfAUBL 327 insertion into the 3' UTR of PfAUBL, the following primer pairs were used: AUBL 5' F + 328 NewApt-5R for the 5' integration locus, and TetR-seq + AUBL 3' R for the 3' integration locus. 329 To determine if any wild-type parasites were still present in the transfected populations, the 330 primer pair AUBL 5' F + AUBL WT 3'R was used. 331 The original aptamer array in the pMG74 plasmid and PfAUBL Apt parasites was 332 amplified using the primer pair Apt-1F + Apt-10R. The redesigned aptamer array in the pKD plasmid and PfAUBL NewApt parasites was amplified using the primer pair NewApt-1F + 334 NewApt-10R. 335

Growth assays 336
Parasites were washed with 10 mL CMA three times to remove aTC from the culture 337 medium and then cultured in CMA alone or CMA with 0.5 M aTC. They were seeded in a 338 96 well plate (Corning) at a 0.5% starting parasitemia and 2% hematocrit in a total volume of 339 250 L per well, in quadruplicate for each medium condition. The plates were incubated in 340 chambers gassed with 94% N2, 3% O2, 3% CO2 at 37°C. A 10 L volume of each parasite 341 sample was collected every day for blood smear analysis or flow cytometry, and the culture 342 medium was exchanged every other day for eight days. 343 Attune Nxt Flow Cytometer (Thermo Fisher Scientific), with a 50 L acquisition volume, and a 350 running speed of 25 L/minute with 10,000 total events collected. The process was repeated on 351 Day 8 for samples collected on days 5-8. 352

Western blotting 353
Parasites were washed with CMA three times to remove aTC from the culture medium. 354 Samples were collected immediately post wash, and then every 24 hours for the next 4 days. The 355 collected samples were centrifuged at 500g for 5 min at room temperature (RT) and pellets were stored at -20°C until the time of protein extraction. To isolate parasites from RBCs, the pellets 357 were thawed and resuspended in 0.15% saponin for 5 min at RT. Lysed RBCs were removed by 358 washing three times with 1X PBS. Parasite pellets were resuspended in NuPAGE LDS sample 359 buffer (Thermo Fisher) containing 2% ß-mercaptoethanol and boiled for 5 min. Proteins were 360 resolved by SDS-PAGE on 4-12% gradient gels and transferred to nitrocellulose membranes. 361 Membranes were blocked in 5% milk and probed overnight at 4°C with 1:2,500 rat anti-HA 362 mAb 3F10 (Roche). They were then incubated for an hour at RT with 1:5,000 anti-rat 363 horseradish peroxidase conjugated antibody (GE Healthcare). Protein bands were detected on X-364 ray film using SuperSignal West Pico Chemiluminescent Substrate (Thermo Scientific), 365 according to the manufacturer's protocol. For loading controls, membranes were stripped of 366 antibody with 200 mM glycine (pH 2.0) for 5 min and reprobed with 1:25,000 mouse anti-367 aldolase mAB 2E11. After incubation with 1:10,000 anti-mouse horseradish peroxidase 368 conjugated antibody (GE Healthcare), proteins were detected as described above.  PCR amplification of the aptamer array from six bacterial colonies transformed with 537 pMG74 PfAUBL plasmid. PCR products of full-length arrays are 1 kb in size. Smaller products in 538 some lanes signify truncated arrays. B) Integration of pMG74 PfAUBL plasmid into the 3' UTR of 539 the PfAUBL gene was confirmed by PCR amplification of the 5' and 3' loci. PCR for the 540 unmodified locus failed to detect any residual wild-type parasites (red) in the PfAUBL Apt 541 population; wild-type parasites served as a control (blue) C) Growth of PfAUBL Apt parasites in 542 the presence or absence of aTC was monitored by daily blood smears for 8 days. The parasites 543 were cut 1:10 on day 4. Removal of aTC caused moderate growth inhibition of PfAUBL Apt 544 parasites beginning on day 4 (two way ANOVA, followed by Bonferroni's correction; ****, P < 545 0.0001; **, P < 0.01; *, P < 0.05). Error bars represent the standard error of the mean from two 546 independent experiments, each conducted in quadruplicate. D) PCR amplification of the aptamer 547 array from the aTC-treated PfAUBL Apt population shows 1 kb (full-length) and 0.5 kb products, 548 indicating that some parasites have lost a few aptamers from their array (D0). Surviving parasites