CPP-Ts: a new intracellular calcium channel modulator and a promising tool for drug delivery in cancer cells

Scorpion sting envenoming impacts millions of people worldwide, with cardiac effects being one of the main causes of death on victims. Here we describe the first Ca2+ channel toxin present in Tityus serrulatus (Ts) venom, a cell penetrating peptide (CPP) named CPP-Ts. We show that CPP-Ts increases intracellular Ca2+ release through the activation of nuclear InsP3R of cardiomyocytes, thereby causing an increase in the contraction frequency of these cells. Besides proposing a novel subfamily of Ca2+ active toxins, we investigated its potential use as a drug delivery system targeting cancer cell nucleus using CPP-Ts’s nuclear-targeting property. To this end, we prepared a synthetic CPP-Ts sub peptide14–39 lacking pharmacological activity which was directed to the nucleus of specific cancer cell lines. This research identifies a novel subfamily of Ca2+ active toxins and provides new insights into biotechnological applications of animal venoms.

disulfide bond formation occurs between the residues C10-C24, C17-C30, and C23-C39 in the mature protein, following similar toxins with known structures (UniProt Q8I6X9; UniProt B8XH22; UniProt P86399). The mature CPP-Ts sequence of 45 amino acids with the referred disulfide bonds was chemically synthesized and used in biological assays.

CPP-Ts increases Ca 2+ transient in neonatal rat cardiomyocytes. Synthetic CPP-Ts and Ts venom
altered the amplitude and kinetics of Ca 2+ transients in cardiomyocytes. They provoked a significant increase in Ca 2+ release frequency, thus increasing the cells contraction frequency. Synthetic CPP-Ts doubled cardiomyocytes' contraction frequencies, whereas Ts venom promoted a 5 times increase ( Fig. 2A-D).
CPP-Ts has nuclear localization. We evaluated the subcellular localization of synthetic CPP-Ts in multiple internalization times. It showed a diffuse localization in cardiomyocytes with an internalization time of 1 min (Fig. 3A), at 10 min it was concentrated in the perinuclear region (Fig. 3B), and it was finally almost completely internalized into the cell nucleus within 20 min (Fig. 3C). The cellular images collected in Z-series and analyzed by the Volocyte program showing CPP-Ts colocalized with nuclear markers (Fig. 3D-E), thus validating the intranuclear localization of CPP-Ts. Therefore, CPP-Ts crosses the cellular membrane and is quickly directed to the intranuclear region of cardiomyocytes (Fig. 3).
Biodistribution of 99m Tc-CPP-Ts after intravenous administration in mice is shown in Fig. 5. High uptake was observed in the kidneys as this is probably the main elimination route, and this is consistent with the hydrophilic nature of 99m Tc-CPP-Ts. Because the thyroid only accumulates free 99m Tc, the low uptake of 99m Tc-CPP-Ts suggests that the complex is highly stable, as predicted by the in vitro stability assay detailed above. Scintigraphic images of these mice corroborated the biodistribution findings, showing a very high uptake by the kidneys (Fig. 5B). Additionally, some tissues, such as heart, liver, and lungs showed significantly higher uptake than other organs, indicating the preference of 99m Tc-CPP-Ts to those tissues. Tissues-to-blood ratios increased over time for heart, lungs, and liver (Fig. 5C). After 60 min, ratios reached values higher than 1.5, indicating that these organs had more than 50% of the 99m Tc-CPP-Ts uptake compared to blood. Thus, 99m Tc-CPP-Ts has high affinity to the heart, liver, and lungs, since the uptake of the complex by these organs is not a consequence of blood circulation. In contrast, the uptake values of a non-specific tissue such as muscle showed that the ratios of all timeframes were much lower than those presented by the heart, lungs, and liver (Fig. 5C).
CPP-Ts is not lethal to mice, and anti-CPP-Ts serum neutralizes toxic effects of the Ts venom. Anti-CPP-Ts serum reactivity was tested against synthetic CPP-Ts and Ts venom using rabbit pre-immune serum as the control. While anti-CPP-Ts serum recognized the CPP-Ts toxin, only small interaction with Ts venom was detected by ELISA ( Supplementary Fig. 2). The low reactivity may be related to the low concentration of CPP-Ts in Ts venom.
Anti-CPP-Ts serum neutralized 50% of 2 LD 50 and 75% of 1.5 LD 50 of Ts venom in mice (Supplementary Table 1). This indicates that, in spite of its low concentration, CPP-Ts plays an important role in the lethality of Ts envenomation. However, in vivo toxicity assays showed that synthetic CPP-Ts toxin was not lethal to mice (100% survival), even when a high dose was used (72.5 µg/20 g that is equivalent to 5.5 LD 50 of Ts venom). Considering that anti-CPP-Ts serum reduced the mortality of animals exposed to the whole venom, this toxin probably acts synergistically with other venom components. Sub peptide 14-39 has nuclear localization but lacks pharmacological activity. The cellular sub localization of CPP-Ts was predicted using the PSORT-II program with a 78.3% nuclear localization probability (Fig. 6A). To identify sub peptides that carry the necessary information for nuclear internalization of the molecule, we designed sub peptides from the mature CPP-Ts sequence for in silico analyses (Fig. 6B). Five were less likely to undergo nuclear internalization when compared to the complete CPP-Ts (Fig. 6B). However, the sub peptide 6, corresponding to CPP-Ts amino acids 14 to 39 (sub peptide  ), presented the same nuclear localization probability (78.3%) as the complete toxin (Fig. 6A,B). This sub peptide was then synthesized.
Cardiomyocyte internalization assays showed that the sub peptide 14-39 is directed to the nuclear region within 20 min, similarly to synthetic CPP-Ts (Fig. 6C). However, unlike the complete toxin, it did not alter the Ca 2+ transient and the contraction frequency ( Fig. 6D-G). Thus, the sub peptide 14-39 carries the nuclear internalization properties of CPP-Ts, but not its pharmacological activity.
In addition, sub peptide  does not have cytotoxic effects in cardiomyocytes, as well as synthetic CPP-Ts and Ts venom (Supplementary Fig. 3).

Discussion
Animal venoms are valuable sources of biologically active molecules with selective and specific biological actions. Therefore, they are important tools for the discovery of new drugs and biotechnological products 21 . Herein we present the first description of a scorpion toxin that acts on the intracellular InsP3 receptors. CPP-Ts integrates the list of venomous animal's toxins that bind specifically to biological targets, which corroborates the potential use of Ts venom in biotechnological applications and as a tool to study ion channels and receptors 22,23 . Scorpion toxins active on Ca 2+ channels are members of the scorpionic calcine family, which consists of basic peptides with 33 amino acid residues, stabilized by three disulfide bonds, and containing an "inhibitor cystine-knot" (ICK). Ca 2+ stores in the cytoplasm and nucleus are released through activation of InsP3R and RyR [24][25][26] . These toxins activate the RyRs, which provide most of the intracellular Ca 2+ for muscular contraction 19,[27][28][29] . Calcines bind to the RyR1 receptor through their basic amino acid residue agglomerate followed by Ser or Thr [54-KKCKRRGT-61] 30,31 (Fig. 1C).
CPP-Ts, on the other hand, has features that are distinct from toxins that act on RyRs. Indeed, it lacks the typical RyR binding basic amino acid region and has instead negatively-charged, neutral, and apolar amino acids   (Fig. 1C). In fact, CPP-Ts is most similar to the calcium channel toxin-like BmCa1 from Mesobuthus martensii (63% of similarity) and, together, they may form a new subfamily of toxins able to affect Ca 2+ channel 32,33 (Fig. 1B).
Studies on Ca 2+ transient in cardiomyocytes revealed that scorpion calcines induce a significant increase in the Ca 2+ signal duration and amplitude 18,19 . Likewise, we observed that synthetic CPP-Ts induce Ca 2+ transient alterations in neonatal rat cardiomyocytes (Fig. 2), thus significantly increasing the frequency of cellular contractions. In addition, CPP-Ts is addressed to the nuclear region (Fig. 3), thereby activating Ca 2+ signaling through InsP3R (Fig. 4), a previously unrecognized mechanism of scorpionic toxins.
As intracellular Ca 2+ concentration triggers heart muscle contraction 34,35 , Ca 2+ channel toxins seem to be active in the cardiac system. Cardiac alterations caused by Ts envenoming are one of the most notable and potentially hazardous symptoms 3 . The effect of Ts envenoming in the cardiovascular system includes cardiac arrhythmias, arterial hypertension or hypotension, and circulatory failure 36,37 .
Silveira et al. showed that the complex effects in the cardiac frequency and contraction force evoked by Ts venom were due to simultaneous release of acetylcholine and catecholamine in the postganglionic nerve fibers of the guinea pig heart 37 . However, Teixeira et al. concluded that the increased contractility was neurotransmitter independent and caused by a direct effect of Ts venom in rat cardiomyocytes 38 . Here we show that Ts venom and CPP-Ts have a direct effect on the increase of contraction frequency of neonatal rat cardiomyocytes, corroborating the hypothesis by Teixeira et al. 38 . The biodistribution assays suggest that 99m Tc-CPP-Ts has a high affinity to heart, liver, and lungs (Fig. 5), which are targets that match with the clinical manifestations frequently observed in Ts scorpion envenomation 3 .
The study of new components from Ts venom is important for the understanding of systemic envenoming and may help with the development of alternative anti-scorpion sera 39 . Our research group is involved on the characterization of the immunoreactivity of isolated toxins to foster alternative serum production against venoms [40][41][42][43][44][45][46] .
In the current work, the antisera produced against synthetic CPP-Ts neutralized 1 LD 50 of Ts venom in mice (Supplementary Table 1). The inhibition achieved by the anti-CPP-Ts sera shows the relevance of CPP-Ts effects in Ts envenoming. Cardiac effects are one of the main causes of death on Ts envenoming victims 3 . Since CPP-Ts has a direct cardiac effect, it also has a high potential to be used for alternative sera production, especially if employed along with other components of Ts venom.
Toxins belonging to the calcine family show cell internalization properties. Toxins such as IpTxA and Mca fold into compact, mostly hydrophobic molecules, with a cluster of positively-charged basic residues polarized on one side of the molecule that possibly interacts with cell membrane phospholipids 13,18,47 . Since clathrin-mediated endocytosis inhibitors do not affect their internalization process, they probably cross the plasma membrane passively 48 .
We found that CPP-Ts also has cell penetration properties. Within 20 min, CPP-Ts crossed both the cellular and nuclear membranes and concentrated in the nucleus (Fig. 3). Mca concentrated in the cytoplasm after 2 h of incubation and, only between 4 and 24 h, it was localized in the perinuclear or nuclear regions 13 . Therefore, although CPP-Ts and Mca have similar internalization properties, CPP-Ts has a higher nuclear specificity and lower internalization time than Mca. Although the mechanisms of membrane translocation and nuclear internalization of CPP-Ts have not been studied, we believe that it occurs by active means, since CPP-Ts has internalization specificity (Fig. 7) and is negatively charged, which greatly hinders its interaction with the membrane lipids. In contrast, although IpTxA quickly penetrates in the cell cytoplasm, this toxin is not detected in the nucleus 18 .
There is a growing interest regarding CPPs as potential tools for the delivery of "cargo" to their action sites. Indeed, there is an ongoing search for CPPs sub peptides carrying cell internalization properties while lacking pharmacological activity. Mca was the first example of an animal venom peptide with efficient cell penetration properties and has been shown to carry the anticancer drug doxorubicin 49 . Mca analogs lack the pharmacological activity while keeping cell penetration properties [50][51][52][53] . Herein we present the CPP-Ts synthetic sub peptide  that lacks pharmacological function (Fig. 6), but maintains the nuclear localization property of the original CPP-Ts.
Peptides with cell line-specific internalization properties, which is a characteristic relevant for therapeutic applications, are very rare. The main limitation of classic CPPs such as TAT, Penetratin, Polyarginine, or natural CPPs like Mca, is internalization promiscuity, although the efficiency of non-specific CPPs has been improved 47,54,55 . Surprisingly, the sub peptide  has cellular internalization specificity since it translocates through the membranes of primary cultures of its natural targets, such as cardiomyocytes and hepatocytes. Also, the sub peptide  shows specificity to cancer cell lines of different human tissues such as melanoma, colon/ breast adenocarcinoma, hepatocellular/lung carcinoma, and prostate carcinoma. Remarkably, CPP-Ts was not internalized by the various normal cell lines tested herein (Fig. 7). The mechanism of specificity will need further investigation.
Crotamine-derived nucleolar-targeting peptides translocate rapidly and efficiently to the nucleus of cells that actively proliferate at a given G1/S cell cycle phase and bind to centrosomes and chromosomes 56 . It is possible that a similar mechanism is taking place for the CPP-Ts synthetic sub peptide  although, given its negative net charge (as opposed to the positively charged crotamine analogs and Mca), it is also possible to envisage an entirely different nuclear entrance system such as a specific receptor. Future research is needed to clarify this matter.
The nuclear internalization of CPP-Ts observed in all cancer cell lines tested positions the peptide as a very promising tool for the delivery of antitumor drugs, given the nuclear sensitivity for drug-induced DNA damage 57 and the lack of specificity of the currently available drugs. As the most widely used anticancer drugs belong to the cationic group of anthracyclines 58 , the negative charge of CPP-Ts provides a great advantage because the noncovalent binding of positively charged molecules is considered safer and more effective for drug delivery in blood circulation 59 . Such a specific nuclear drug delivery tool may increase therapeutic efficacy and minimize side effects in cancer therapy. Further studies on the cellular internalization properties of CPP-Ts and efficiency in the transport of molecules should reveal the potential of this toxin as a new CPP for drug delivery in the cellular nucleus of cancer cells.
Concluding, the CPP-Ts that we described here is the first characterized scorpion toxin active in nuclear InsP3R. Its action involves intracellular Ca 2+ release and consequent alteration in the cardiac frequency, thus explaining the symptomatology of Ts envenomation. This natural peptide presents selective internalization properties, with specific nuclear addressing. In comparison with other known natural CPPs 19,47 , CPP-Ts is the quickest to reach the nucleus, showing high nuclear specificity. The ability of CPP-Ts to be internalized by cancer cells and not by normal cell lines as well as its nuclear addressing property make this peptide a potential intranuclear delivery tool to target cancerous cells.

Methods
Experimental animals. Neonatal Wistar rats (1-3 days old, 5-7 g) and female Swiss CF1 mice (4-5 weeks old, 18-22 g or 6-8 weeks old, 24-28 g), from the animal care facilities (CEBIO) of the Federal University of Minas Gerais (UFMG), were used. Adult female New Zealand white rabbits (12 weeks old, 2.5 Kg) were obtained from the Animal Facilities Center of the School of Veterinary Medicine, UFMG. All animals had free access to water and food and were kept under controlled environmental conditions. Animal experiments were performed according to the Brazilian Council for Animal Care guidelines and approved by the Ethics Committee of UFMG (Comissão de Ética no Uso de Animais -CEUA, Protocol number 145/2014, and 05/2016).

Scorpions, RNA and venom extraction.
Ts scorpions were collected as previously described 40 . Scorpion venom was extracted from 40 female scorpions by electrical stimulation 40 . Protein concentration was measured 60 and the venom samples were stored at −20 °C.
RNA extraction was performed two days later. The telson containing venom glands was removed and triturated in TRI reagent (Sigma-Aldrich, MO, USA) to isolate RNA following a previously described protocol 61  Library construction and RNA sequencing. The RNA library was assembled from total Ts telson RNA, using the TruSeq RNA Sample Preparation Kit v2 (Illumina, CA, USA), according to the manufacturer's instructions. The library was sequenced in a paired-end strategy, using the Reagent kit v3 600 cycles (2 × 300) and run in the Illumina Miseq sequencer (Illumina, CA, USA).
The obtained paired-end reads were trimmed using Prinseq-Lite 0.20.4 62 using 30 as a quality score threshold. Reads shorter than 40 bp were also excluded. The trimmed sequences were assembled de novo using the Trinity assembler 63 . The obtained sequence had its identity confirmed by a BLASTx search against the UniProtKB/Swiss-Prot database. Amino acid sequence was analyzed as previously described 40 , and disulfide bonds were predicted by DISULFIND server (http://disulfind.dsi.unifi.it/). BLASTp search 64 was used to find the sequences that most closely matched CPP-Ts, which were then aligned using the ClusalW on BioEdit Software 65 . A second alignment, using the same methods, was performed using other well-characterized toxins belonging to the scorpionic calcines family.

Sequence computational analysis.
To verify the subcellular location of CPP-Ts, we performed an in silico prediction using the software PSORT II (http://psort.hgc.jp/form2.html). Mature CPP-Ts sequence, in addition to six sub peptides (21-26 residues in length), were analyzed.
Peptide synthesis. The CPP-Ts peptide containing 45 amino acid residues and three properly positioned disulfide bonds and the CPP-Ts sub peptide containing 26 residues 14-39 were chemically synthesized by LifeTein, LLC (New Jersey, USA).
Primary culture, cell lines and cell culture. Neonatal cardiomyocytes were freshly isolated from Wistar rats (n = 12, 1-3 days old, 5-7 g) as previously described 66 , and used for transient Ca 2+ analysis and internalization assays. Rat hepatocytes primary culture was obtained as previously described 67  Intracellular Ca 2+ transient analysis. Ca 2+ transients were monitored in neonatal cardiomyocytes as previously described 68 , using Fluo-4/AM fluorescence (excitation at 488 nm and emission at 515 nm) in the line scan detection mode with intervals of 9 sec immediately after treatment (n = 20 cells per treatment). Treatments were synthetic CPP-Ts (2 µg/ml), sub peptide 14-39 (2 µg/ml), and Ts venom (12.8 µg/ml).
Images were collected in a Zeiss Axiovert (Zeiss, CA, USA) confocal microscope. Three lasers were utilized, as follows: excitation at 488 nm and emission at 505-550 nm for Alexa 488; excitation at 568 nm and emission at 585-615 nm for Alexa 555; and excitation at 633 nm and emission at 650 nm for TO-PRO-3.

InsP3 sponge NLS virus transfection.
Neonatal cardiomyocytes were incubated with 100 MOI of InsP3 sponge NLS virus in 1 ml DMEM medium enriched with 10% FBS, which prevents InsP3 binding to the nuclear InsP3R 20 , as previously described 24 . The transfection was verified by mRFP fluorescence emission (excitation at 584 nm and emission at 630 nm) in a fluorescence microscope (Zeiss, CA, USA).
Scintigraphic images were acquired at 10, 30, and 60 min post-injection of 11 MBq 99m Tc-CPP-Ts in healthy Swiss mice (n = 3, 6-8 weeks old, 24-28 g). Image acquisition was performed as previously described 69 . CPP-Ts biological characterization: anti-CPP-Ts serum production, serum neutralization assay and in vivo toxicity. Anti-CPP-Ts serum was produced as previously described 40 . Female New Zealand rabbits (n = 2, 12 weeks old, 2.5 Kg) were injected subcutaneously using two boosters of 100 µg of synthetic CPP-Ts and two boosters of 150 µg of synthetic CPP-Ts at 15-day intervals. Serum titration was performed by ELISA, as previously described 41 . ELISA plates (BD, NJ, USA) were pre-coated with Ts venom (5 µg/ml) or synthetic CPP-Ts (5 µg/ml), and anti-CPPTs rabbit serum was titrated using dilutions ranging from 1:100 to 1:102,400. Pre-immune serum was used as control. Absorbance was measured at 492 nm.
Ts venom LD 50 was previously established as 13.2 µg per 20 g mouse 40 . For in vivo neutralization assays, Ts venom samples (1.5 LD 50 = 19.80 μg or 2 LD 50 = 26.72 μg) were incubated for one hour at 37 °C with 150 μL of anti-CPP-Ts rabbit serum or pre-immune serum. After incubation, the samples were applied by subcutaneous injection in randomized groups of Swiss mice (n = 8 per group, 4-5 weeks old, 18-22 g). Within 24 h, the surviving mice were counted. All analyses were single-blinded.
To evaluate the toxicity of CPP-Ts in live animals, a single high dose of synthetic CPP-Ts (72.5 µg, equivalent to 5.5 LD 50 of Ts venom) was subcutaneously injected in Swiss mice (n = 8, 4-5 weeks old, 18-22 g). Animals were observed and the surviving ones were counted within 24 h.
Cells were plated in coverslips at a confluence of 70-80% and submitted to the internalization assay. For the assay, cells were treated with sub peptide 14-39 (20 µg/ml) for 1 h. After that, cells were washed and fixed with paraformaldehyde 4% (v/v). The immunofluorescence protocol was followed as described (see section: Immunofluorescence). Images were collected in a confocal microscope (Zeiss, CA, USA) (n = 300 cells).
Statistical analyses. Data were expressed as mean ± S.E.M. Normality and equal variance were evaluated by Shapiro-Wilk and Levene's tests, respectively. Means were compared by repeated-measures analysis of variance followed by Mauchly sphericity test or One-way ANOVA. In case of multiple comparisons, a post hoc Bonferroni correction was used. Significance level was set at 0.05 and tests were performed two-sided when possible. All data were analyzed by GraphPad PRISM version 5.00 software (La Jolla, CA, USA) and exact p values were calculated using R (version 3.3.0). Sample size calculations were performed using G Power version 3.1. (Supplementary Data 1).

Data Availability
All the data supporting the findings of this study are included in the article and its Supplementary Information files or are available from the corresponding author upon request. CPP-Ts cDNA and protein sequences are available in GenBank database under the accession number MH061344.