Comparative proteomics reveals Cryptosporidium parvum manipulation of the host cell molecular expression and immune response

Cryptosporidium is a life-threating protozoan parasite belonging to the phylum Apicomplexa, which mainly causes gastroenteritis in a variety of vertebrate hosts. Currently, there is a re-emergence of Cryptosporidium infection; however, no fully effective drug or vaccine is available to treat Cryptosporidiosis. In the present study, to better understand the detailed interaction between the host and Cryptosporidium parvum, a large-scale label-free proteomics study was conducted to characterize the changes to the proteome induced by C. parvum infection. Among 4406 proteins identified, 121 proteins were identified as differentially abundant (> 1.5-fold cutoff, P < 0.05) in C. parvum infected HCT-8 cells compared with uninfected cells. Among them, 67 proteins were upregulated, and 54 proteins were downregulated at 36 h post infection. Analysis of the differentially abundant proteins revealed an interferon-centered immune response of the host cells against C. parvum infection and extensive inhibition of metabolism-related enzymes in the host cells caused by infection. Several proteins were further verified using quantitative real-time reverse transcription polymerase chain reaction and western blotting. This systematic analysis of the proteomics of C. parvum-infected HCT-8 cells identified a wide range of functional proteins that participate in host anti-parasite immunity or act as potential targets during infection, providing new insights into the molecular mechanism of C. parvum infection.


Sample preparation and collection
C. parvum oocysts of the Iowa strain were purchased from a commercial source (Waterborne, New Orleans, LA, USA). The HCT-8 human colon adenocarcinoma cell line (ATCC, Manassas, VA, USA) was cultured in Dulbecco's modified Eagle's medium (DMEM) nutrient mix F-12 (Gibco, Grand Island, NY, USA) containing 10% fetal bovine serum (FBS, Gibco). For infection, C. parvum oocysts were suspended in phosphate-buffered saline (PBS) with 20% sodium hypochlorite, centrifuged at 3, 667 × g, and washed twice with PBS. Infection was performed in culture medium containing viable C. parvum oocysts (at an oocyst to host cell ratio of 2:1) at 37˚C in a humidified 5% CO 2 incubator for 36 h. An equal volume of culture medium was added to the control group. After 36 h of incubation, the cells were washed three times with cold PBS and collected using a cell scraper. The cells were centrifuged at 200 × g for 10 min to obtain the cell pellet, which was stored at -80˚C until analysis.

Quantitative real-time reverse transcription PCR (qRT-PCR)
Total RNA from the HCT-8 cells was extracted using the TRIzol reagent according to the manufacturer's protocol (Takara Biotechnology, Inc., Shiga, Japan). The primers were designed by EnzyArtisan Biotech Co., Ltd. (Shanghai, China, Table 1). The cDNA was prepared using a PrimeScript RT-PCR Kit (Takara Biotechnology, Inc.) and then amplified using a SYBR qPCR Mix (EnzyArtisan Biotech). The ACTB (β-actin) gene was used as an endogenous control for normalization.

Protein extraction and trypsin digestion
More than 2 × 10 7 cells of each group were collected and sonicated three times on ice using a high intensity ultrasonic processor (Scientz, Zhejiang, China) in lysis buffer (8 M urea, 1% Protease Inhibitor Cocktail). The lysate was centrifuged at 12, 000 × g at 4˚C for 10 min to separate the insoluble debris. Finally, the soluble supernatant was collected and the protein concentration was determined using a bicinchoninic acid (BCA) kit according to the manufacturer's instructions (Beyotime, Jiangsu, China). Then, the protein supernatant was thoroughly digested using trypsin for further analysis.

LC-MS analysis and database searching
The tryptic peptides were loaded onto a laboratory-prepared reversed-phase analytical column (C18), subjected to a nanospray ionization (NSI) source followed by tandem mass spectrometry (MS/MS) in a Q Exactive Plus apparatus (ThermoScientific, Waltham, MA, USA) coupled online to an ultra-performance liquid chromatography (UPLC) column, and then selected for MS/MS using a normalized collision energy (NCE) setting of 28. The fragments were detected in the Orbitrap at a resolution of 17, 500. Automatic gain control (AGC) was set at 5e4 and the fixed first mass was set as 100 m/z. Then, the Maxquant search engine (v.1.5.2.8, Max Planck Institute of Biochemistry, Munich, German) was used for secondary mass spectral data process. All tandem mass (MS/MS) spectra were searched against human and C. parvum data in the Uniprot database, which was concatenated with a reverse decoy database to calculate the false discovery rate (FDR) caused by random matching. Trypsin/P was specified as the cleavage enzyme allowing up to two missing cleavages and the minimum length of the peptide segment was set as seven amino acid residues. 20 ppm and 5 ppm were set as the mass error tolerance of primary parent ions for the first and main search, respectively, and the mass error tolerance was set as 0.02 Da for secondary fragment ions.

Bioinformatic analysis
The UniProt-GOA database (http://www.ebi.ac.uk/GOA/), InterPro domain database (http:// www.ebi.ac.uk/interpro/), and the Kyoto Encyclopedia of Genes and Genomes (KEGG) database (http://www.genome.jp/kaas-bin/kaas_main; http://www.kegg.jp/kegg/mapper.html) were used separately for Gene Ontology (GO) annotation, characterization of the proteins basic functions, domain functional descriptions, and annotation of biological processes or pathways. Then, a two-tailed Fisher's exact test was employed to test the enrichment of the differentially expressed proteins (DEPs) against all identified proteins. A P-value < 0.05 was considered significant. All the categories obtained after enrichment were collected and then filtered for those categories that were at least enriched in one of the clusters with a Pvalue < 0.05. This filtered P-value matrix was transformed and then clustered using one-way hierarchical clustering in Genesis. Clusters were visualized using the R Package, pheatmap (https://cran.r-project.org/web/packages/cluster/). We used Wolfpsort (http://www.genscript.com/psort/wolf_psort.html) to predict the subcellular locations of the proteins. COG (Clusters of Orthologous Groups of proteins)/KOG (euKaryotic Ortholog Groups) functional classification of the identified proteins was conducted through database comparison and analysis.

Protein-protein interaction network
All DEPs' database accession numbers or sequences were searched against STRING database version 11[26] for protein-protein interaction (PPI) analysis. Only interactions between the proteins belonging to the searched data set were selected, thereby excluding external candidates. The interaction network from STRING was visualized in Cytoscape [27].

Statistical analysis
Data are presented as means ± SD. Statistical tests were performed using the GraphPad Prism 8 Statistics software package (GraphPad Inc., La Jolla, CA, USA). A value of P < 0.05 was considered statistically significant.

Proteomic analysis of DEPs in non-infected and infected samples
A total of 4406 human proteins were identified, of which 3307 contained quantitative information, ranging in molecular weight from 1.9 kDa to 964.8 kDa (S1 Fig). Among them, 33, 812 acquired unique peptides belonged to 3307 proteins, which served as reliable evidence for these predicted proteins (S1 Fig). Most of the identified proteins contained no less than two unique peptides as multiple peptide-level evidence for protein identification (S1

Subcellular localization of DEPs.
A mass spectrometry comparison between infected and non-infected HCT-8 cells identified 121 DEPs (1.5-fold difference in abundance), of which 64 were upregulated and 57 were downregulated after infection (Fig 1A and 1B). Interestingly, among the DEPs, those located on the plasma membrane were all upregulated after infection (S1 Table). Besides, the upregulated and downregulated proteins had various locations in the cellular compartments, including the cytoplasm, nucleus, mitochondria, and extracellular region (Fig 1B). In-depth analysis further demonstrated the different locations of DEPs caused by infection. Functional enrichment analysis indicated that the upregulated proteins were concentrated in the transporters associated with antigen processing (TAP) complex, extracellular space, intrinsic component of plasma membrane, and MHC class I peptide loading complex. The downregulated proteins were found to be mostly located in the organelle lumen, membrane-enclosed lumen, intracellular organelle lumen, microbody lumen, and peroxisomal matrix (S3 Fig). The obvious local distinction between upregulated and downregulated proteins suggested the special effect of C. parvum infection on various components of the cell.
In addition, we conducted bioinformatic analysis on the intracellular parasite components, and 231 C. parvum-derived proteins were identified at 36 h post infection (Figs 1C and S4). These parasitic proteins were enriched significantly in the cytoplasm, extracellular space, nucleus, plasma membrane, nucleus, mitochondria, and endoplasmic reticulum (Fig 1C), which constitute the core components of C. parvum.

GO enrichment and KEGG pathway enrichment of DEPs.
We further annotated and investigated the functions of the identified DEPs, with the aim of discovering important molecules or signaling pathways that play key roles during C. parvum parasitism. According to biological process analysis from the GO classification, the DEPs were closely involved in cellular process, single-organism process, metabolic process, response to stimulus, and biological regulation (Fig 2A). Among the 121 DEPs, 102 were identified as having binding functions and 47 had catalytic activity (Fig 2A). COG/KOG categories provided detailed functional classification of DEPs, suggesting that the C. parvum infection has extensive effects on various physiological activities of host cells, such as RNA processing and modification, signal transduction mechanisms, posttranslational modification, protein turnover, chaperones, energy production and conversion, amino acid transport, and metabolism ( Fig 2B).
Based on further GO enrichment analysis, C. parvum infection seemed to evoke a series of cell defense reactions, in that the upregulated proteins were related to cellular response to type I interferon, and the defense response to virus and other organisms. By contrast, the downregulated proteins were functionally enriched for antibiotic metabolic processes (including aminoglycoside antibiotic and doxorubicin metabolic process), positive regulation of inclusion body assembly, protein oligomerization, and hypoxia response due to infection. Moreover, several signaling pathways were affected by C. parvum infection. KEGG pathway enrichment analysis of upregulated and downregulated proteins was conducted separately, which indicated an obvious distinction in functional clusters between the DEPs (Fig 2C and  2D). Remarkably, the upregulated proteins exhibited significant enrichment for inflammatory signaling pathways, and were involved in many pathogenic pathways of virus and bacterial infections, Malaria, Leishmaniasis, and Toxoplasmosis ( Fig 2C). These findings suggested that in addition to some specific pathways, different parasites or viruses might target common pathways and machineries, which forms the basis for the discovery of broad range vaccines or drugs. By contrast, the downregulated proteins caused by C. parvum infection were found to be markedly enriched in multiple metabolic pathways, including metabolism of many amino acids and propanoate, biosynthesis of folate, unsaturated fatty acids and steroid hormone, and the peroxisome pathway ( Fig 2D). Besides, C. parvum had a significant effect on the host oxidative phosphorylation process, with up to 12 related proteins being differentially abundant after infection (Fig 3). The extensive influence of C. parvum on host metabolic process revealed C. parvum's need for a special cellular microenvironment, and also reflected its dependence on certain nutrients, which provided ideas for identifying effective drugs or inhibitors.

Bioinformatic analysis of C. parvum proteomic data
To gain a deeper understanding of the biological characteristics of C. parvum, we also conducted a bioinformatic analysis on the identified C. parvum-derived proteins. The COG/KOG annotation revealed the predominant involvement of identified C. parvum-derived proteins in however, the detailed mechanisms require further study. Besides, 109 C. parvum-derived proteins were cross referenced with the STRING and Cytoscape databases for protein interaction analysis (S5 Fig). Thirty-two proteins were determined and clustered with a high degree of network interaction, and the hub proteins were identified, which included cgd7_2280, elongation factor 2, and cgd3_2090, identifying a series of parasitic proteins of vital importance, which might represent the core functional proteins of C. parvum.

Validation of the proteomic results
To assess the proteins that were identified as being closely related to the host anti-parasite mechanism, or associated with parasite development and pathogenicity during C. parvum infection, we verified their expression using qRT-PCR and western blotting. Firstly, the mRNA expression level of nine genes, including ISG15, IFITM3 (interferon-induced transmembrane protein 3), PLSCR1 (phospholipid scramblase 1), NMI (N-Myc and STAT interactor), IFI35 (interferon-induced protein 35), HELZ2 (helicase with zinc finger 2), SFPQ (splicing factor, proline-and glutamine-rich), RBMX (RNA-binding motif protein), and ATP5PO (ATP synthase peripheral stalk subunit OSCP) were compared in normal and infected HCT-8 cells, and the results were in accordance with the proteomic analysis ( Fig 5A). Then, several proteins were chosen and further confirmed using western blotting, which showed that the levels of ISG15, IFITM3, PLSCR1, NMI, STAT1, signal transducer and activator of transcription 2 (STAT2), SFPQ, RBMX, and protein S100-A10 (S100A10) accorded with those in the proteomic analysis (Fig 5B and 5C).

Discussion
As an enteric pathogen, C. parvum primarily infects the apical end of small intestinal enterocytes, where they are enveloped by host membranes but remain extracytoplasmic [28,29]. Intestinal epithelial cells (IECs) can sense and participate in the coordination of appropriate immune responses, ranging from tolerance to anti-pathogen immunity [30,31]. Hence, IECs are considered as the frontline sensors for pathogenic microorganism encounters [32,33]. The HCT-8 cell line possesses an excellent ability to support C. parvum growth after infection [34][35][36]. In this study, we established a C. parvum infection model using HCT-8 cells to mimic the pathogenesis and immune reaction during C. parvum invasion and development, aiming to provide key insights into how the host reacts effectively to infection. Our data revealed that the levels of 121 host proteins were altered during C. parvum infection, which possessed diverse functions, and profiled various biological process changes in infected host cells globally.
Based on bioinformatic analysis and recent research evidence, we focused on several significant signaling pathways and molecules that might serve as important targets during the anti-C. parvum immune response. GO analysis revealed that the interferon (IFN) signaling pathway was the most enriched biological process, in that seven (10%) of the 67 significantly upregulated proteins were identified as targets of IFN signaling; specifically, ISG15, 2'-5'-oligoadenylate synthetase 3 (OAS3), NMI, IFITM3, STAT1, STAT2, and IFI35. Interferon-induced protein ISG15 has been implicated as a core player during the host antiviral response by directly disrupting viral replication, budding, and release, as well as by limiting tissue damage and modulating human type I interferon signaling [37][38][39][40][41][42][43]. The biological relevance of ISG15 in parasitic infection has been demonstrated. Novel findings suggest that ISG15 expression was induced during Theileria annulata and Leishmania brazilensis infections, but not in L. amazonensis infection [44,45]. Recently, ISG15 was found to participate in autophagy-mediated ubiquitinoylation of the vacuole during T. gondii infection, as a bridge that links the ATG pathway with IFN-γ-dependent restriction of T. gondii in host cells [46]. In the present study, significantly increased ISG15 expression was confirmed, indicating a potential anti-parasitic function of ISG15 during C. parvum infection. The PPI network further exhibited strong interaction among ISG15 and other upregulated proteins including OAS3, IFITM3, IFI35, intercellular adhesion molecule-1 (ICAM-1), STAT1, STAT2, PLSCR1, eukaryotic translation initiation factor 2 alpha kinase 2 (EIF2AK2), and NMI, most of which are interferon-induced proteins. The upregulation of IFITM3, STAT1, STAT2, PLSCR1, and NMI after infection was validated using western blotting and qRT-PCR. Based on these results, we identified that the prevailing immune response of HCT-8 cells to C. parvum is centered on IFN signaling, revealing the involvement of unique proteins and a distinct immune process in C. parvum infection, which might be key regulators in the anti-C. parvum immune reaction.
ICAM-1 is a cell surface glycoprotein and an adhesion receptor that is expressed widely by endothelial, epithelial, and immune cells, and is responsible for leukocyte recruitment and signal transduction [47]. Under inflammatory conditions or stimulation by injury, the expression of ICAM-1 markedly increased in immune cells and epithelial cells [48][49][50]. Recently, ICAM-1 has also been proven to be an important cytoadherence receptor for Plasmodium chabaudi infection, and binds to Plasmodium falciparum erythrocyte membrane protein 1, suggesting its important role in parasitic protozoa infection [51,52]. Furthermore, ICAM-1 expression was induced in the rat jejunum during acute T. gondii infection [53]. Moreover, Chen et al. demonstrated that the upregulation of ICAM-1 in epithelial cells was modulated by micro-RNA-221 following C. parvum infection. ICAM-1 also enhanced the attachment of infected host cells [54]. In the present study, the robustly enhanced expression of ICAM-1 in C. parvum infected cells suggested that ICAM-1 might serve as a biosensor in the response to C. parvum infection and participate in the recruitment of immune cells. However, the detailed mechanism requires in-depth investigation.
In the present study, we observed that IFITM3 was upregulated at both the mRNA and protein levels after C. parvum infection. IFITM3 is an innate immune response protein generally known to inhibit the entry and replication of many viruses [55][56][57]. Individuals lacking IFITM3 are highly susceptible to infection, even when challenged with influenza virus of low pathogenicity [58]. Therefore, IFITM3 is considered as a first line of cell defense in response to viruses. Besides, the protective effect of IFITM3 is also reflected in the long survival of lung resident memory T cells, in which sustained IFITM3 expression facilitated their survival and protection from viral infection during subsequent exposure [59]. In addition, IFITM3 could mediate the autophagic degradation of interferon regulatory factor 3 (IRF3) and negatively regulates tissue-damaging inflammatory production of type I IFNs induced by the virus [60]. Accordingly, we hypothesized that the enhancement of IFITM3 in response to C. parvum infection is the result of host immune defense and a self-regulating mechanism to prevent fatal inflammation after C. parvum infection; however, the mechanism awaits future exploration.
STAT1 and STAT2 proteins are key mediators of IFN signaling, and are essential components in the cellular antiviral response and adaptive immunity [61]. In various parasite infections, such as those by T. gondii and Schistosoma japonicum, STAT1 and STAT2 are evoked and mediate the immune response [62,63]. NMI and IFI35 are both IFN-induced proteins: NMI enhances STAT-mediated transcription of downstream genes in the Janus kinase (JAK)-STAT pathway [64]. IFI35 is known to interact with N-Myc and NMI [65]. However, IFI35 has been proven to negatively regulate retinoic acid-inducible gene I protein (RIG-I) antiviral signaling and promote the replication of vesicular stomatitis virus. Therefore, IFI35 might serve as a flexible immunological regulator during pathogen defense [66,67]. In the present study, the upregulation of IFN-induced proteins STAT1, STAT2, NMI, and IFI35 was observed and validated, and the strong interactions in the PPI network further demonstrated the IFN-centered immune reaction against C. parvum. The inflammatory molecules mentioned above might be key targets in the anti-C. parvum response.
Apart from IFN-centered signaling pathways, the remarkable downregulation of proteins involved in multiple metabolic pathways also reflects C. parvum's dependency on certain nutrients, which also provides candidates for effective drugs or inhibitors. Possessing highly streamlined metabolic pathways, but an inability to de novo synthesize nucleosides, fatty acids, and any amino acids, C. parvum hijacks and remodels existing host metabolic pathways for its own benefit [10,68]. In the present study, we found that the levels of multiple host metabolic enzymes were downregulated in response to C. parvum infection, and then an in-depth exploration was conducted to characterize the proteins' functions. The data revealed widely decreased levels of host metabolic enzymes such as argininosuccinate synthase (ASS1), glutamine synthetase (GLUL), phosphoglycerate mutase 1 (PGAM1), glyoxylate and hydroxypyruvate reductase (GRHPR), NADH: ubiquinone oxidoreductase subunit B6 (NDUFB6), NFS1 cysteine desulfurase (NFS1), pyruvate dehydrogenase E1 component subunit alpha (PDHA1), SDHB, and serine hydroxymethyltransferase 2 (SHMT2) after infection. Folate and thymidylate biosynthesis are of vital importance during DNA replication, and dihydrofolate reductase (DHFR) is a key enzyme in folate metabolism, catalyzing the oxidation of NADPH and reduction of dihydrofolate to NADP and tetrahydrofolate [69,70]. The important roles of folate metabolism and nucleic acid synthesis in cryptosporidiosis, mean that currently, inosine monophosphate dehydrogenase (IMPDH) and dihydrofolate reductase-thymidylate synthase (DHFR-TS) are common targets for the treatment of cryptosporidiosis [71,72]. However, the parasite has been found to tolerate the loss of these classical targets, possibly via as-yet-undiscovered purine transporters and salvage enzymes. In the present study, the data revealed downregulation of SHMT2 in host cells, which is a pyridoxal phosphate (PLP) binding protein catalyzing the cleavage of serine to glycine, accompanied with the production of 5, 10-methylenetetrahydrofolate (5, 10-CH 2 -THF) [73,74]. The downregulation of SHMT2 by C. parvum revealed not only obstruction of the host metabolism, but also a potential parasitism strategy of C. parvum to sustain an essential nutrient supply. An early study reported the promotion of exogenous purine nucleosides during C. parvum infection [75]. However, genomic analysis indicated the loss of pyrimidine de novo synthesis in C. parvum; therefore, this parasite is almost entirely dependent on import from the host for its purine and pyrimidine requirements [76,77]. SHMT2 is also an essential intermediate for purine biosynthesis; therefore, we concluded that inhibition of SHMT2 caused by C. parvum serves as one of the strategies to secure a purine supply from the host. Similarly, we found that ASS1, a urea cycle enzyme that converts nitrogen from ammonia and aspartate to urea, was downregulated after C. parvum infection. A recent publication on hepatocellular carcinoma showed that downregulation of ASS1 is associated with a more malignant cancerous phenotype and poor prognosis [78]. In-depth investigation further revealed that ASS1 facilitates pyrimidine synthesis during cancerous proliferation by activating CAD (carbamoyl-phosphate synthase 2, aspartate transcarbamylase, and dihydroorotase complex), through the regulation of aspartate levels [79]. Based on these processes, C. parvum would obtain sufficient purines and pyrimidines, which would ensure successful parasitism and replication, possibly by interfering with the host's purine and pyrimidine metabolism.
Besides, we observed significant reductions in the abundances of enzymes involved in the cell oxidative respiratory chain, such as NFS1, PDHA1, and SDHB. Mounting evidence demonstrates the C. parvum is an energy-requiring parasite. It is generally accepted that Cryptosporidium spp. has lost mitochondria-like organelle-derived energy metabolic capabilities via reductive evolution. Lacking constitutive oxidative phosphorylation, C. parvum is unable to complete the tricarboxylic acid (TCA) cycle and cytochrome-based electron transport processes [10,68,80,81]. Thus, gluconeogenesis and glycolysis are the most important energy producing processes in parasites [15,80]. Genomic and biochemical evidence further confirmed the dependence of Cryptosporidium on glycolysis as the main energy source and an overall reliance on the host for basic nutrients [10,[82][83][84][85]. PGAM1 was among the proteins identified in our study, which is an important glycolytic enzyme coordinating glycolysis and biosynthesis, including the pentose phosphate pathway and serine synthesis pathway [86,87]. Evidence showed that some host-parasite homologous genes of the host glycolysis/gluconeogenesis pathways were downregulated, while host-exclusive genes were upregulated during invasion and intracellular development, suggesting parasite-derived competition for metabolic substrates, which might explain the reduction of host enzymes in our study [17]. Another study suggested that sugar transportation occurs between Cryptosporidium and host intestine epithelial cells [88]. Parasites further deprive the host cells of the substrates of glycolysis by transportation, and might correspondingly affect the expression of the host genes in the glycolysis pathway. Recent research showed that C. parvum could interfere with host glucose transporters (GLUT) 1/2 and Na + -coupled glucose transporter (SGLT) 1 expression, and a significantly higher intracellular glucose level was observed in infected cells, which would point to an adaptation of the host cells' glucose uptake after infection [89]. Based on this, PGAM1 might be a new target for C. parvum to inhibit glycolysis and biosynthesis in host cells.
Research has also proven that glutaminolysis and lactate are necessary for parasite replication [15]. In this study, the downregulation of GLUL after C. parvum infection would hinder host glutamine synthesis from glutamate and ammonia, and might even promote parasite replication, which is dependent on glutamine catabolism. All these identified enzymes indicate modification of host energy production and conversion by C. parvum to sustain its own growth, which might also provide new insights into anti-Cryptosporidium targets.
In summary, we observed significant enrichment of interferon-centered signaling pathways and extensive inhibition of metabolism-related enzymes in host cells caused by C. parvum infection, providing a deeper understanding of the molecules and their functions involved in the host-C. parvum interaction. The identified DEPs and signaling pathways narrowed the range of important functional molecules during anti-parasite immune responses and also confirmed C. parvum's deprivation of certain host nutrients, which might be potential targets for Cryptosporidium treatment. Moreover, the mechanism by which the identified factors participate in C. parvum parasitism and the complex interplay between the identified factors require further investigation.