Characterization of Dense Granule Metalloproteinase INS-16 in Cryptosporidium parvum

The protozoan pathogen Cryptosporidium parvum infects intestinal epithelial cells and causes diarrhea in humans and young animals. Among the more than 20 genes encoding insulinase-like metalloproteinases (INS), two are paralogs with high sequence identity. In this study, one of them, INS-16 encoded by the cgd3_4270 gene, was expressed and characterized in a comparative study of its sibling, INS-15 encoded by the cgd3_4260 gene. A full-length INS-16 protein and its active domain I were expressed in Escherichia coli, and antibodies against the domain I and an INS-16-specific peptide were produced in rabbits. In the analysis of the crude extract of oocysts, a ~60 kDa fragment of INS-16 rather than the full protein was recognized by polyclonal antibodies against the specific peptide, indicating that INS-16 undergoes proteolytic cleavage before maturation. The expression of the ins-16 gene peaked at the invasion phase of in vitro C. parvum culture, with the documented expression of the protein in both sporozoites and merozoites. Localization studies with antibodies showed significant differences in the distribution of the native INS-15 and INS-16 proteins in sporozoites and merozoites. INS-16 was identified as a dense granule protein in sporozoites and macrogamonts but was mostly expressed at the apical end of merozoites. We screened 48 candidate INS-16 inhibitors from the molecular docking of INS-16. Among them, two inhibited the growth of C. parvum in vitro (EC50 = 1.058 µM and 2.089 µM). The results of this study suggest that INS-16 may have important roles in the development of C. parvum and could be a valid target for the development of effective treatments.


Introduction
Cryptosporidium spp. are gastrointestinal pathogens that can cause severe diarrhea in humans and various animals [1]. Young children in developing countries infected with Cryptosporidium spp. can develop malnutrition and cognitive impairments in addition to clinical illness [2]. In industrialized countries, waterborne outbreaks of cryptosporidiosis are common [3]. Among the more than 40 named Cryptosporidium species, Cryptosporidium parvum is the main species for cryptosporidiosis in farm animals and one of the two dominant species in humans [4]. Currently, there is a lack of effective drugs against Cryptosporidium spp. Although significant progress has been made in the development of novel drugs against cryptosporidiosis in recent years, we still have poor understanding of the biology of the pathogens [5].
The invasion of host cells by apicomplexan parasites is a complex process mediated by receptors and ligands that involve many proteins on both sides [6]. Secreted proteases and protein kinases in secretory organelles of apicomplexans can modify invasion-related 2 of 13 proteins or host cell activities, thus playing important roles in invasion [7]. Comparative genomics analysis of multiple Cryptosporidium species has revealed the presence of numerous genes encoding secreted proteases in the compact genome. Among them, insulinaselike metalloproteinases (INS) are one of the largest protease families with 22 members in C. parvum [8,9]. As the Cryptosporidium genome is only 9 Mb in size and most of thẽ 4000 genes are single copied, INS probably play important roles in the invasion and development of Cryptosporidium spp.
INS are zinc metalloproteinases of the M16 family, which are widely distributed in nature and divided into the M16A, M16B, and M16C subfamilies. The M16 metalloproteinases are characterized by the presence of a functional domain containing an inversion of the thermolysin zinc-binding motif, HXXEH [10]. They have a wide range of substrates and cleave many proteins and small peptides including insulin, β-amyloid, and glucagon [11,12]. In apicomplexans, falcilysin is an M16C insulinase involved in hemoglobin catabolism and may function as two different proteases in two subcellular organelles of Plasmodium falciparum after proteolytic processing [13,14]. An M16A protease, toxolysin-1 (TLN1) of Toxoplasma gondii, is a rhoptry protein that is released during the invasion of host cells. Both the C-terminus and N-terminus of TLN1 undergo cleavage before the maturation of the protease [15]. Another toxolysin, TLN4, is located in the micronemes and involved in parasite fitness [16].
Transcriptome data from C. parvum indicate that some INS genes are highly expressed in the invasion stages of the pathogen [17]. In a recent study, one INS protein of C. parvum, INS20-19, was shown to be potentially involved in invasion or early developmental processes [18]. Another C. parvum INS protein, INS-15, appeared to be post-translationally processed as several fragments and have biological activities similar to toxolysins [19]. Another INS protein, INS-1, is expressed in secretory vesicles within the pathogen and contributes to the formation of macrogamonts [20]. Nevertheless, the role, processing, and trafficking of other INS remain unclear.
In this study, we characterized INS-16 encoded by the cgd3_4270 gene and examined its expression patterns in developmental stages of C. parvum. Two inhibitors of the metalloproteinase were identified through virtual screening, which reduced the growth of the pathogen in vitro. In addition, we compared the expression of INS-16 and INS-15, two metalloproteinases with high sequence identity and encoded by neighboring genes. The results suggest that the two INS are located in different organelles and have different biological functions.

Characteristics of INS-16 and INS-16 Domain I
INS-16 is an M16A secretory metalloproteinase of 1176 amino acids and consists of four classic domains of INS with the key functional motif of HXXEH in the M16 active domain ( Figure 1A). It differs from INS-15 in the number of domains, and it has one more M16 peptidase-like domain at the C-terminus than INS-15. Most of the sequence differences between the two INS are in the N-terminus and C-terminus, especially amino acids 1 to 60 ( Figure 1B). Using DNA extracted from C. parvum oocysts as the template, we successfully cloned the full-length cgd3_4270 gene and its domain I fragment ( Figure S1A,D), producing recombinant proteins in E. coli ( Figure S1B,E). The recombinant proteins were purified using Ni-NTA affinity chromatography, with the purity being confirmed using SDS-PAGE analysis ( Figure S1C,F).

Proteolytical Processing of Native INS-16
To assess the expression of the native INS-16 protein in C. parvum, we used antibodies generated against the INS-16-specific peptide in the Western blot analysis of sporozoite lysates. The antibodies reacted with a ~60 kDa fragment of the protein instead of the expected 134 kDa full length protein ( Figure 2C), indicating that native INS-16 is likely proteolytically processed after its translation.

Differential Expression of INS-15 and INS-16 in Life Cycle Stages of C. parvum
The antibodies against the INS-16 domain I and specific peptide were used to characterize the expression of INS-16 in oocysts, excysted sporozoites, and intracellular stages using immunofluorescence microscopy. The results showed that both antibodies reacted with sporozoites within oocysts, with no significant difference in the staining pattern. However, in excysted sporozoites, anti-INS-16 domain I antibodies reacted with the entire parasites, while antibodies against the INS-16-specific peptide showed a dotty pattern mostly in the middle of the sporozoites in immunofluorescence microscopy. In the immunofluorescence analysis of intracellular developmental stages, both antibodies reacted with meronts with similar staining patterns. The fluorescence signal of antibodies against INS-16 domain I covered almost the entire merozoites, while the fluorescence signal of antibodies against the INS-16-specific peptide was a small dot at the apical end of merozoites. In sexual stage, the antibodies against INS-16 domain I reacted with both the macrogamont and microgamont, while the antibodies against the INS-16-specific peptide reacted mostly with one side of the macrogamont with no reactivity to microgamonts ( Figure 3A,B). To

Proteolytical Processing of Native INS-16
To assess the expression of the native INS-16 protein in C. parvum, we used antibodies generated against the INS-16-specific peptide in the Western blot analysis of sporozoite lysates. The antibodies reacted with a~60 kDa fragment of the protein instead of the expected 134 kDa full length protein ( Figure 2C), indicating that native INS-16 is likely proteolytically processed after its translation.

Differential Expression of INS-15 and INS-16 in Life Cycle Stages of C. parvum
The antibodies against the INS-16 domain I and specific peptide were used to characterize the expression of INS-16 in oocysts, excysted sporozoites, and intracellular stages using immunofluorescence microscopy. The results showed that both antibodies reacted with sporozoites within oocysts, with no significant difference in the staining pattern. However, in excysted sporozoites, anti-INS-16 domain I antibodies reacted with the entire parasites, while antibodies against the INS-16-specific peptide showed a dotty pattern mostly in the middle of the sporozoites in immunofluorescence microscopy. In the immunofluorescence analysis of intracellular developmental stages, both antibodies reacted with meronts with similar staining patterns. The fluorescence signal of antibodies against INS-16 domain I covered almost the entire merozoites, while the fluorescence signal of antibodies against the INS-16-specific peptide was a small dot at the apical end of merozoites. In sexual stage, the antibodies against INS-16 domain I reacted with both the macrogamont and microgamont, while the antibodies against the INS-16-specific peptide reacted mostly with one side of the macrogamont with no reactivity to microgamonts ( Figure 3A,B).
oocysts, INS-16 expression was mainly in the anterior end and middle of the sporozoites, mostly in dense granules, with some gold particles in the oocyst matrix also ( Figure 4A). In contrast, the expression of INS-15 was mainly concentrated around the nucleus of sporozoites, with some gold particles in dense granules ( Figure 4B). In meronts of infected HCT-8 cells, INS-16 expression was mainly in the apical end of merozoites, while the expression of INS-15 covered the entire merozoites ( Figure 4A,B). This result is consistent with immunofluorescence analysis of INS-15 and INS-16, as indicated above.

Anti-Cryptosporidial Effects of Candidate Inhibitors of INS-16
The molecular docking of the simulated structure of INS-16 identified 100 potenti inhibitors of the metalloproteinase. Among them, 48 compounds were evaluated for vitro effects on the invasion and growth of C. parvum in HCT-8 cells using a qRT-PC assay ( Table 1). Ten of the compounds showed >50% growth inhibition at th concentration of 10 µM used in the initial evaluation ( Figure 5A). The efficacy of these wa further assessed in dose-response experiments. Among them, 3805-1518 and F107-194 had inhibition rates of over 80% at the concentration of 5 µM, with EC50 values of 1.05 µM and 2.089 µM, respectively ( Figure 5B,C). The two compounds displayed lo cytotoxicity on HCT-8 cells, with TC50 values > 100 µM ( Figure 5D,E).

Anti-Cryptosporidial Effects of Candidate Inhibitors of INS-16
The molecular docking of the simulated structure of INS-16 identified 100 potential inhibitors of the metalloproteinase. Among them, 48 compounds were evaluated for in vitro effects on the invasion and growth of C. parvum in HCT-8 cells using a qRT-PCR assay ( Table 1). Ten of the compounds showed >50% growth inhibition at the concentration of 10 µM used in the initial evaluation ( Figure 5A). The efficacy of these was further assessed in dose-response experiments. Among them, 3805-1518 and F107-1944 had inhibition rates of over 80% at the concentration of 5 µM, with EC 50 values of 1.058 µM and 2.089 µM, respectively ( Figure 5B,C). The two compounds displayed low cytotoxicity on HCT-8 cells, with TC 50 values > 100 µM ( Figure 5D,E).     [18,[20][21][22][23]. In agreement with data generated from the present study, they seemingly have diverse expression patterns and biological functions.
INS-16, like insulin-degrading enzyme (IDE) in humans, appears to be a classical zinc metalloproteinase that is proteolytically processed before maturation. Domain analysis shows that INS-16 contains one active domain and three inactive or middle domains. This special structure allows the N-terminal zinc-binding active domain of classical zinc metalloproteinase to be connected to the C-terminal domain, forming a closed proteolytic chamber to exert activity [24,25]. In the Western blot analysis, it was shown that INS-16specific antibodies recognize a~60 kDa product in C. parvum sporozoites, which is much smaller than the recombinant protein expressed in E. coli. This proteolytic processing appears to be common in INS of C. parvum. In previous studies, several products of different sizes were observed in the detection of INS4, INS6, INS-15, and INS20-19 in crude extracts of sporozoites despite the fact that some Cryptosporidium INS do not have four domains. This was attributed to the presence of several putative SΦX(E/D) cleavage sites in the sequences [18,19,21]. The proteolytic processing of INS-16 is similar to two toxolysins in Toxoplasma gondii. The Western blot analysis of native TLN4 showed that TLN4 antibodies principally recognized a~55 kDa product in tachyzoite lysate, which contained the active domain and the first inactive domain of TLN4 [16]. Another toxolysin, TLN1, was also shown to go through cleavage at the C-terminal, generating a product smaller than the predicted size. It was believed that the cleavage of the C-terminal domain provided additional flexibility for substrate binding, allowing the enzyme to cleave larger substrates [15].  [21,23]. In immunoelectron microscopy analysis, the expression of INS-16 was mainly detected in dense granules located in the anterior end and middle of sporozoites and merozoites. In contrast, INS-15 expression in sporozoites was mainly confined to areas around the nucleus, as shown in the present ( Figure 4A) and previous studies [19]. INS-15, nevertheless, appears to be a dense granule protein present over the entire merozoites of C. parvum ( Figure 4B). The only other known dense granule protein of C. parvum, CpClec, is expressed in both sporozoites and merozoites and mediates the infection of C. parvum through Ca 2+ -dependent binding with sulfated proteoglycans on host intestinal epithelial cells [26]. In other apicomplexan parasites, dense granule proteins are involved in the formation of parasitophorous vacuoles and the modification of the host cell activities [27]. Further studies are needed to elucidate the precise functions of diverse members of the INS family of C. parvum.
INS-16 is highly expressed in macrogamonts and may play a role in the sexual life stage of C. parvum. Indirect immunofluorescence microscopy showed high INS-16 expression in macrogamonts, and this expression pattern was similar but not identical to the previously reported INS1 expression [20]. INS1 expression is more likely located in small vesicles within macrogamonts, while INS-16 is more likely located in the dense granules within macrogamonts. In addition, the transcriptional activity of the ins-16 gene was high at 72 h of in vitro C. parvum infection ( Figure S2). Therefore, INS-16 may interact with INS1 or other wall-forming proteins to participate in the development of the sexual stages and oocysts of C. parvum.
Inhibitors of INS-16 can effectively inhibit the growth of C. parvum in vitro, indicating that INS-16 may be an important drug target. Although we failed to obtain the full INS-16 protein with enzymatic activity, based on the structure of human IDE together with known inhibitors, we were able to simulate the active structure of INS-16, leading to the identification of 48 potential inhibitors from the ChemDiv database through molecular docking. Among them, only 3805-1518 and F107-1944 effectively inhibited the growth of C. parvum in vitro without significant cytotoxicity to the host cells. Prior to this, there have been no studies on the inhibitors of C. parvum INS. In Plasmodium falciparum, it was known that piperazine-based hydroxamic acids kill parasites by blocking falcilysin (FLN). These inhibitors can competitively bind to active and substrate recognition sites in the protease therefore inhibit FLN activity [28,29]. At present, the mechanism of INS-16 in the invasion and development of C. parvum is not clear. Further structural and genetic manipulation studies are needed to identify the action mechanism of INS-16 and its candidate inhibitors.

Parasite, Host Cells, and In Vitro Infection
Oocysts of the C. parvum IOWA isolates were purchased from Waterborne, Inc. (New Orleans, LA, USA), stored at 4 • C, and used within three months. For infection experiments, oocysts were treated with 0.5% sodium hypochlorite on ice for 10 min and washed three times with PBS via centrifugation at 13,200× g for 3 min. The treated oocysts were resuspended with RPMI 1640 culture medium for in vitro infection. HCT-8 cells (ATCC, CCL-244, Chinese Academy of Sciences Shanghai Branch) were seeded into 12-well plates, cultured to~80% confluence, and inoculated with RPMI 1640 culture medium containing C. parvum oocysts. Free sporozoites were obtained from the sodium-hypochlorite-treated oocysts via incubation in PBS (pH 7.4) containing 0.25% trypsin and 0.5% taurodeoxycholic