Assessment of genetic mutation frequency induced by oxidative stress in Trypanosoma cruzi

Abstract Trypanosoma cruzi is the etiological agent of Chagas disease, a public health challenge due to its morbidity and mortality rates, which affects around 6-7 million people worldwide. Symptoms, response to chemotherapy, and the course of Chagas disease are greatly influenced by T. cruzi‘s intra-specific variability. Thus, DNA mutations in this parasite possibly play a key role in the wide range of clinical manifestations and in drug sensitivity. Indeed, the environmental conditions of oxidative stress faced by T. cruzi during its life cycle can generate genetic mutations. However, the lack of an established experimental design to assess mutation rates in T. cruzi precludes the study of conditions and mechanisms that potentially produce genomic variability in this parasite. We developed an assay that employs a reporter gene that, once mutated in specific positions, convert G418-sensitive into G418-insenstitive T. cruzi. We were able to determine the frequency of DNA mutations in T. cruzi exposed and non-exposed to oxidative insults assessing the number of colony-forming units in solid selective media after plating a defined number of cells. We verified that T. cruzi‘s spontaneous mutation frequency was comparable to those found in other eukaryotes, and that exposure to hydrogen peroxide promoted a two-fold increase in T. cruzi‘s mutation frequency. We hypothesize that genetic mutations in T. cruzi can arise from oxidative insults faced by this parasite during its life cycle.


Introduction
Trypanosoma cruzi is the etiological agent of Chagas disease, a complex zoonosis that affects more than seventy genera of mammalian hosts (Zingales et al., 2012;Baptista et al., 2014). According to the World Health Organization (WHO), around 6-7 million people are affected by this disease in 21 countries, most of them in Latin America [WHO Chagas disease (American trypanosomiasis) fact sheet, 2017]. Also noteworthy is the fact that nowadays this disease is spreading to non-endemic regions due to human migration (Schmunis, 2007).
The life cycle of T. cruzi is complex and involves two hosts: an invertebrate and a mammalian. Humans are considered accidental hosts, in which the classic vectorial infection generally occurs at night when the blood-sucking triatomines defecate during feeding (Frasch, 2000). Once the feces droplets expelled by the triatomine reach the bloodstream or get in contact with eyes, nose or mouth mu-cosa, the infection is then perpetrated (Prata, 2001). Humans may also be infected with T. cruzi through blood transfusion, organ transplantation, from mother to infant during pregnancy, laboratory accidents, as well as through ingestion of food contaminated with triatomine feces (Shikanai-Yasuda et al., 1991;de Noya and González, 2015).
Following the infection by T. cruzi, a short acute phase characterized by high parasitemia takes place, along with unspecific symptoms . During its chronic phase, Chagas disease presents a large spectrum of symptoms and low parasitemia. Interestingly, 30% of infected humans will develop cardiomyopathy, digestive implications or both (Rassi Jr and Marin-Neto, 2010), and a small percentage of them may still develop neurological symptoms (Prata, 2001). Although the mechanisms and factors influencing this clinical unpredictability have not been fully elucidated, the variability in the course of Chagas disease seems to be related to a number of factors such as parasite strain, host age, reinfection, and genetic factors of both host and parasite (Prata, 2001).
Since 2009, T. cruzi strains have been divided into six discrete taxonomic units, namely T. cruzi I -VI, based on its intra-specific genetic variability (Zingales et al., 2009;Baptista et al., 2014). Unquestionably, diverse tissue tropisms, response against immune system, and responsiveness to chemotherapy have been frequently observed in Chagas disease (Revollo et al., 1998;Andrade et al., 2010). In fact, genetic factors are able to strictly regulate infection capacity of parasites, as there is a correlation between genetic diversity and rate of success in escaping the host immune response (Frasch, 2000;Burgos et al., 2013).
It has long been known that several microorganisms display intrinsic, spontaneous mutability events that lead to intra-specific genetic diversity (Steinberg et al., 1971;Taddei et al., 1997;Rosche and Foster, 2000). The generation of spontaneous mutation is a very complex subject since several intrinsic and extrinsic factors might be involved in the process -like the environment in which the organism is found , location of mutation-prone sites in the genome (Patrushev and Minkevich, 2008), and the behavior of the DNA repair system (Hoeijmakers, 2001). However, a number of studies have already shed light on the mechanisms and importance of spontaneous mutation rate in bacteria (Choi et al., 2011;Ford et al., 2013), yeast (Magni and von Borstel, 1962;Glassner et al., 1998;Bensasson, 2011), and in other non-disease causing eukaryotes (Provan et al., 1999;Shikazono et al., 2003). Also, it has already been shown that certain T. cruzi haplogroups display mutations in microssatelite alleles after being cultured in media supplemented with hydrogen peroxide (H 2 O 2 ) (Augusto-Pinto et al., 2003).
Therefore, the study of the mechanisms related to the generation of genetic mutations and diversity in T. cruzi is imperative since they may play a role in how this parasite deals with genotoxic stress and drug response; in fact, experimental analysis of the antigenic diversity generation remains a challenge since few works tried to investigate T. cruzi's mutation rate. In this work, we developed a model that allows the detection of mutational events through the selection of T. cruzi resistant to the aminoglycoside G418. We found that the mutation frequency in this parasite is similar to other eukaryotic cells, being substantially increased by challenging T. cruzi with exogenous H 2 O 2 . Since T. cruzi has to cope with oxidative stress situations during its complex life cycle (Piacenza et al., 2009;Machado-Silva et al., 2016), we hypothesize that immunologic evasion and chemotherapy resistance in Chagas disease could be associated to the generation of genetic variability in T. cruzi enhanced by oxidative stress conditions.

Plasmid construction and bacterial transformation
Wild-type Neo (Neo WT ) and its mutant variants -Neo D90 , Neo D180 , Neo D270 , Neo stop , and Neo stopT ®G -were amplified by PCR from the pROCK_Neo vector (da Rocha et al., 2004), using the primers indicated in Table 1. All resultant amplicons (Table 1) were digested with XhoI and XbaI and then ligated to pMAL-c2G (New England Biolabs Inc., Massachusetts, USA) previously digested with the same endonucleases. Electrocompetent Escherichia coli DH5a (Gonzales et al., 2013) were transformed with ligation products and plated onto 2xYT medium [1.6% tryptone, 1.0% yeast extract, 0.5% NaCl (pH 7.0)] supplemented with 100 mg/mL ampicillin. Bacterial positive clones were screened using the colony PCR method (Bergkessel and Guthrie, 2013) and further isolated.

Bacterial kanamycin resistance assay
DH5a positive clones for all Neo constructs (Table 1) were grown in 2xYT liquid medium supplemented with 100 mg/mL ampicillin, under orbital agitation (180 rpm) at 37°C for 16 h. Bacterial cells were then subject of a serial dilution (suspensions with final concentrations of 10 -4 , 10 -6 , 10 -8 , and 10 -10 cells/mL), and 2.5 mL of each suspension were added onto plates containing 2xYT solid medium (liquid 2xYT plus 2.0% agar) supplemented with 100 mg/mL ampicillin and 0.1 mM isopropyl b-D-1-thiogalactopyranoside (IPTG) in the presence or absence of either 10 mg/mL kanamycin or 10 mg/mL neomycin. Plates were incubated at 37°C for 18 h at the end of which they were photo-documented.

T. cruzi transfection, selection, and genotyping of transfected clones
Epimastigotes of T. cruzi clone CL Brener were grown in liver infusion tryptose medium [0.9% liver infusion broth, 0.5% tryptose, 0.1% NaCl, 0.8% Na 2 HPO 4 , 0.04% KCl, 0.2% hemin, 10% fetal bovine serum; 200 mg/mL streptomycin; 200 mg/mL penicillin (LIT); pH 7.2], at 28°C. Cells were transfected by electroporation, as described elsewhere (da Rocha et al., 2004), using the pROCK_Hygro-Neo stop construct generated as described in Results, Item 2, and then selected in liquid LIT medium supplemented with 200 mg/mL hygromycin B -cells were transferred to fresh hygromycin B-added LIT weekly, for 4-5 weeks. Then, exponentially-grown transfected cells were plated onto blood-agar medium [48.4% LIT, 48.4% brain-heart infusion and 2.5% defibrinated blood (Gomes et al., 1991)] supplemented with 200 mg/mL hygromycin B for selection of transfected clones. Colony forming units (CFU) were then picked and cultured in hygromycin-added liquid LIT medium, and after 7 days were subjected to genomic DNA extraction as follows: 10 8 from each T. cruzi culture was centrifuged at 5000 x g for 5 min and pelleted cells were resuspended in 100 mL Milli-Q water and incubated at 95°C for 10 min. After another centrifugation, the supernatants were collected and genotyping was conducted by PCR using primers 5 and 8, listed in Table 1.

Determination of T. cruzi growth rate and survival
A defined number (5x10 6 /mL) of transfected T. cruzi cells (T. cruzi Neostop ) were cultured for 2 days in fresh hygromycin B-added LIT, until they reached logarithmic growth phase, with cellular concentration around 2x10 7 /mL. After repeating this procedure three times, T. cruzi Neostop cells had their growth rate monitored for 7 or 42 days. After that, transfected cells were transferred to hygromycin B-added LIT supplemented with either 200 or 400 mg/mL G418 and cultured for 2 days. The number of viable cells was determined using a hemocytometry chamber by the use of erythrosine as a vital stain for differentiation between live and dead cells. All experiments were performed in biological triplicates and results are reported in mean ± standard deviation. Statistical analyses (one-way ANOVA) were performed using GraphPad Prism v6.0 (GraphPad Software, Inc.).

T. cruzi genomic DNA extraction
T. cruzi genomic DNA was extracted through cellular lysis, deproteination and precipitation, as described in Andrade et al. (1999). Briefly, a defined number of exponentially-grown T. cruzi cells (10 8 ) were washed three times with PBS and incubated in 200 mL of lysis solution [(0.5% SDS, 100 mM EDTA, and 10 mM Tris-HCl (pH 8.0)] with 20 mg/mL RNase, for 1 h, at 37°C. Then, 100 mg/mL proteinase K was added to the lysate, which was incubated at 50°C for 3 h. Deproteination was conducted by the addition of 200 mL saturated phenol followed by gentle homogenization and centrifugation; the organic phase was then dispose -the same procedures were repeated for the addition of 200 mL of phenol/chloroform 1:1 (v/v) and 200 mL of chloroform. DNA precipitation was carried out using absolute isopropanol at -80°C overnight. The isopropanolic suspension of DNA was then centrifuged at 16,000 x g, for 10 min, and pelleted DNA was washed twice with ethanol 70% before being dried and resuspended in sterile MilliQ water.

T. cruzi genomic DNA sequencing
Genomic DNA from T. cruzi was sequenced through the Sanger method using a MegaBACE 1000 DNA Sequencing System (GE Healthcare). For each reaction, DYEnamic ET Dye Terminator MegaBACE kit and the specific set of primers were used. Sequences were analyzed by the Phred-Phrap algorithm (Ewing et al., 1998) and examined with MultAlin for multiple sequence alignment (Corpet, 1988).

Mutation frequency assay
A defined number of T. cruzi Neostop epimastigotes (10 7 ) was cultured for 42 days in hygromycin-added LIT in the presence or absence of 50 mM H 2 O 2 . Cells were then washed and resuspended in PBS, and counted as described in Materials and Methods, item 4. A volume of suspension containing 10 8 cells were plated onto hygromycin B-added solid blood-agar, either in the presence or absence of G418. After 8 weeks, CFUs were counted, and mutation frequency was determined by dividing the number of CFUs observed on the plate per the number of cells/mL present in the liquid LIT culture from which epimastigotes were collected.

Development of the Neo stop reporter
We developed a methodology to assess DNA mutation rates in T. cruzi based on a system that carries a Neo gene variant unable to encode an amino 3'-glycosyl phosphotransferase [APH(3')-II] that displays its biological activity, unless a genomic mutational event takes place and restores this ability. APH(3')-II is an enzyme responsible for microbial resistance against aminoglycosides such as neomycin, kanamycin, and G418 (Hächler et al., 1996).
First, we sought to determine which segments from APH(3')-II were essential to its activity. For such, we generated three Neo gene shorter variants lacking their first 90, 180 and 270 nucleotides, using primers 2 -4, indicated in Table 1. Each Neo gene variant were ligated into pMAL c-2G (which harbors the lac promoter; Walker et al., 2010), giving rise to Neo D90 -pMAL, Neo D180 -pMAL and Neo D270 -pMAL constructs ( Figure 1A). We next transformed E. coli DH5a with all aforementioned constructs, as well as with the wild type Neo construct (Neo WT -pMAL) ( Figure 1A), and bacterial transformants were selected from 2xYT plates supplemented with 100 mg/mL ampicillin. DH5a transformants were cultivated overnight in liquid ampicillinadded 2xYT, and then plated onto ampicillin-added solid 2xYT supplemented with 0.1 mM IPTG, in the presence or absence of 10 mg/mL kanamycin. We then verified that, unlike Neo WT , none of the three obtained Neo gene variants (Neo D90 , Neo D180 , and Neo D270 ) were able to confer DH5a resistance against kanamycin ( Figure 1A, B). We therefore concluded that the first 30 amino acids of the N-terminal portion of APH(3')-II are essential to its biological activity.
Once we determined that the Neo gene is required to promote resistance against aminoglycosides, we decided to introduce a premature stop codon right after the Neo WT

468
Mutation frequency in Trypanosoma cruzi gene start codon using primers 5 and 7 listed in Table 1, creating the Neo stop variant, in which a G -its fourth base -is substituted by a T, generating the stop codon TGA ( Figure  2A). This premature stop codon prevents the formation of APH(3')-II, completely abrogating the growth capacity of DH5a in the presence of kanamycin ( Figure 1B). We next manually performed an in silico prediction of possible mutations that would restore the translation of the N-terminal portion of APH(3')-II, and thus provide resistance against aminoglycosides. Interestingly, from all predicted mutational events ( Figure 2B), two of them -G®T at position 5, and T®G at position 15 -are classic mutations generated by cellular exposure to H 2 O 2 (Shibutani et al., 1991).

Long-term cultivation induces mutational events in T. cruzi
After (i) observing that DH5a transformed with the Neo stop -pMAL construct did not exhibit growth in 2xYT supplemented with kanamycin ( Figure 1B), and (ii) that oxidation could lead to mutational events that might restore the translation of APH(3')-II from the Neo stop variant (Figure 2B), we sought to transfect T. cruzi with the Neo stop gene variant. As expected, we were unable to observe, through erythrosine vital stain assay, visible growth of clones #1 and #5 of T. cruzi Neostop cultured in hygromycinadded liquid LIT supplemented with G418 ( Figure 3A). We then investigate if long-term cultivation -i.e., 42 days -of T. cruzi Neostop was capable of generating G418-insensitive clones for such, clones #1 and #5 were subject to the same experimental design described above, being cultured for 42 days, instead. Surprisingly, upon the increase of the cultivation period, we were able to verify the presence of G418resistant T. cruzi cells from Neo Stop clones #1 and #5 in hygromycin-added liquid LIT supplemented with 200 or 400 mM G418 ( Figure 3B).

Oxidative stress increases mutational events in T. cruzi
Given the fact that long-term cultivation allows the observation of mutational events in T. cruzi, we decided to take advantage of the established protocol for isolation of T. cruzi clones using solid blood-agar to determine the number of CFUs of G418-insensitive T. cruzi Neostop generated from a defined number of plated cells -this would allow us to determine the frequency of mutation of T. cruzi. Then, 1x10 8 cells from Neo Stop clones #1 and #5, previously cultured in hygromycin B-added liquid LIT for 42 days, in the presence or absence of 50 mM H 2 O 2 , were plated onto hygromycin B-added solid blood-agar, and the number of CFUs were determined, as described in Material and Methods, item 7. We verified that T. cruzi Neostop cultured in the presence of H 2 O 2 , showed a mutation frequency of 1.56x10 -7 , while parasites cultured in control conditions, i.e., in the absence of H 2 O 2 , exhibited a mutation frequency of 0.71x10 -7 . This observation indicated that there is a two-fold increase in mutation frequency when T. cruzi faces situations of environmental oxidative stress. Besides, the experimental design was sensitive enough to allow us to identify the basal frequency of genomic mutations of T. cruzi Neostop , i.e., the frequency of mutational events observed in parasites that were not exposed to H 2 O 2 during this assay. This basal frequency -lower than the one observed in the presence of H 2 O 2 -may indicate the rate of oxidation-independent mutational events that probably take place spontaneously in T. cruzi.

The Neo stopT ®G reporter confers kanamycin resistance to DH5a
We next designed a forward primer carrying a guanine in its 4 th position (#6, Table 1) to artificially obtain the Neo stop gene variant mimicking the oxidation-induced T®G mutation, which creates a downstream start codon, as found in Neo stop clones #18-2, #34-2 and #40-2 ( Figure  4A). The resultant amplicon (Neo stopT ®G) was ligated into pMAL c-2G plasmid, generating the Neo stopT ®G-pMAL construct, which was used to transform DH5a, whose transformants were selected from ampicillin-supplemented 470 Mutation frequency in Trypanosoma cruzi 2xYT plates. After isolation, the Neo stopT ®G-pMAL construct was used to obtain DH5a transformants from solid ampicillin-added 2xYT plates. Once selected, one clone from these bacterial transformants was cultured overnight in liquid ampicillin-added 2xYT, and then plated onto ampicillin-added solid 2xYT supplemented with 0.1 mM IPTG, in the presence or absence of 10 mg/mL kanamycin. We were then able to verify that DH5a harboring Neo stopT ®G became G418-resistant ( Figure 4B), confirming that the aminoglycoside resistance observed in the T. cruzi Neostop clones #18-2, #34-2 and #40-2 is in fact promoted by the T®G mutation, a nucleotide transversion classically induced by oxidants (Shibutani et al., 1991).

Discussion
Genetic diversity is an important factor that is directly related to adaptation and survival of T. cruzi in its hosts; in fact, DNA metabolism and mutagenesis may allow this parasite to increase the chances to adapt to different environments during its complex life cycle (Machado-Silva et al., 2016). In this sense, the study of mechanisms that govern this phenomenon is crucial for the understanding of how T. cruzi evade the immune system and show resistance against drugs, and for the development of new therapeutic strategies. However, currently, other than a restricted number of studies employing in silico approaches to study mutagenesis and variability in T. cruzi (Azuaje et al., 2007a,b), there is scarce information regarding the exact cellular events that may generate intra-specific genomic variability and few biological assays that allow the determination and detection of mutation rates in this parasite.
The Neo gene encodes APH(3')-II, a phosphotransferase that contains 267 amino acids, and is responsible for conferring microbial resistance against aminoglycosides (Hächler et al., 1996). APH(3')-II displays an ATP biding-site and can transfer the g-phosphoryl group from an ATP molecule to the aminoglycoside, converting the latter to its phosphorylated, inactive form (Eustice and Wilhelm, 1984;Shaw et al., 1993;Thompson et al., 2002). We generated a number of mutations in the Neo gene, which gave rise to shorter APH(3')-II variant forms (Table, 1, Figure  1A) that were ineffective in conferring DH5a resistance against kanamycin ( Figure 1B). Then, once we determined    , Table 1) -which generates a start codon into the Neo stop -to obtain the Neo stopT ®G reporter, that confers kanamycin-and neomycin-resistance to DH5a.
that the N-terminal segment of Neo was required to provide resistance against kanamycin and G418, we introduced a premature stop-codon right after Neo's ATG through a G®T mutation at position 4 (Figure 2A), creating a variant (Neo stop ) that would re-establish resistance against aminoglycosides if the mutated codon underwent a mutational event. This was observed when T. cruzi Neostop was cultivated for 42 days in hygromycin B-added liquid LIT (Figure 3).
During its life cycle, T. cruzi undergoes an obligatory intracellular amastigote stage in which the immune system promotes the release of reactive oxygen and nitrogen species to halt the infection (Piacenza et al., 2009); thus, replication of amastigotes under a scenario of oxidative stress can promote a condition from which mutated cells can ultimately increase the pool of mutated T. cruzi, which could lead to intra-specific genetic diversity. Although epimastigotes and amastigotes are subjected to different extents of oxidative stress, data from the literature (Aguiar et al., 2013), as well as unpublished observations from our group, suggest that both aforementioned T. cruzi life forms are equally affected by oxidative stress and share the same responses against this biological condition. Therefore, the observation that epimastigotes treated with H 2 O 2 display a 2-fold increase in mutational events (Results, item 3) suggests that oxidative stress promoted by the host may play a direct role in genetic variability of T. cruzi amastigotes. In fact, for several other organisms, including E. coli, Helicobacter pylori, Salmonella typhimurium, Bacillus subtilis, Pseudomonas, Clostridium, Saccharomyces cerevisiae, and Candida albicans, increased mutation rates are often correlated with increased survival and infection rates in adverse conditions (Wang et al., 2001;Foster, 2000;Linz et al., 2014). In this manner, the increase in the number of G418-resistant T. cruzi Neostop clones after longterm oxidative insult ( Figure 3) suggests that this type of stress could stimulate intra-specific genetic variability.
It is well-established that oxidative stress promotes a range of modifications in nucleic acids, such as doublestrand breaks and nitrogenous base modification (Friedberg et al., 2006). Interestingly, the generation of 8-oxoG, one of the most frequent lesions derived from oxidative stress, has a high mutagenic potential, since the oxidized guanine, if localized in the genomic DNA, promotes a mismatched pairing with adenine resulting in G®T or C®A transversions. In addition, the generation of 8-oxoG in the nucleotide pool also promotes a T®G transversion, consequently leading to nucleotide mismatches (Dizdaroglu et al., 2002;van Loon et al., 2010). In fact, the severity of effects that can arise from the formation of 8-oxoG became evident when the GO system -a pathway specialized in preventing mutagenicity promoted by 8-oxoG, comprised of three enzymes, namely MYH (MutY homologue), MTH (MutT homologue), and OGG1 (FPG homologue) -was first described David et al., 2007).
In T. cruzi, long-term exposure to H 2 O 2 induced DNA mutations related to the generation of 8-oxoG, as clones #18-2, #34-2 and #40-2 showed mutations that are likely consequence of a guanine oxidation ( Figure 4A). Likewise, clones #36-2 and #43-2 also presented formation of 8-oxoG mutations, since guanine in DNA undergoes a mispairment with adenine during replication ( Figure 4A). The mispairing observed in clones #1-2 and #5-2 -which were not exposed to H 2 O 2 -are possibly products of an impaired replication process induced by a wobble conformation, although the DNA template and protein conformation are not disturbed (Johnson and Beese, 2004). These mismatches allow the formation of a structure closer to Watson-Crick base pair than that one observed in G:A and A:G mismatches. Altogether, these verifications indicate that mutations observed in T. cruzi cells exposed to H 2 O 2 are products of generation or misincorporation of 8-oxoG in the DNA, since those mutations are deleterious and do not easily arise in normal environments, considering the abnormalities they cause to the polymerase structure (Johnson and Beese, 2004). Alterations in DNA metabolism can also increase genetic mutation frequency (Castillo-Acosta et al., 2012). Organisms like yeast seem to preferentially insert cytosine opposing apurinic/apyrimidinic sites, and this mechanism could lead to the increase of AT®GC transversions (Thomas et al., 1997).
As suggested for T. cruzi, the presence of mutations, to some extent, are possibly related to the survival of some other organisms. In fact, Trypanosoma brucei strain relies on variant surface glycoproteins (VSG) switching to escape from the host immune system, a process in which recombination plays a crucial role (Hartley and McCulloch, 2008;Horn and McCulloch, 2010). Deletion of deoxyuridin 5'triphosphate pyrophosphatase (dUTPase) can cause a 9fold increase in spontaneous mutation, and the appearance of double strand breaks in T. brucei, which could lead to a recombination process, increasing VSG switching (Castillo-Acosta et al., 2012).
In this work, through a novel assay to assess mutational events in T. cruzi, we demonstrated that oxidative stress increases the mutation frequency in this parasite. We hypothesize that the 2-fold increase in mutation frequency after exposure to H 2 O 2 -which mimics the reactive oxygen species released by human macrophages -indicates that this mutational mechanism, combined with the GO repair system -could generate T. cruzi's intra-specific genetic diversity that can be important to help this trypanosomatid to evade the immune system and be resistant to drug therapy, ultimately allowing this parasite to survive in stressful environments. 472 Mutation frequency in Trypanosoma cruzi