Figure 1. Structural comparison of parasporins. (
A) Structural model of higher-molecular-weight PS3Aa1 with its three domains. (
B) Low-molecular-weight PS2Aa1 structural model. (
C–
E) Structural comparison between parasporin-2, the 26-kDa nontoxic protein, and aerolysin-like β-PFT. Membrane-binding-related domain I is colored yellow. The membrane-insertion and pore-formation regions are colored blue (domain II) and red (Domain III). It is suggested that the purple amphipathic β-hairpin is necessary for pore formation (
C–
E). Parasporin 4 (PS4) was modeled using the 26-kDa nontoxic protein as an adapted template from Xu et al.
[9], modified by the authors.
3. Effects of Parasporins on Cancer Cells
The mechanism of action of pore-forming proteins (PFPs) is dynamic, with three main steps: (1) the formation of water solubility, (2) self-assembly, and (3) insertion into the membrane, which leads to a pore suspected to be highly destructive for membrane integrity
[34][13]. The points at which these proteins anchor to the membrane probably occur at specific receptors located in the microdomains rich in cholesterol and sphingolipids (lipid rafts), since these are requirements for GPI-anchored proteins, and the glucan region may be required for the binding and assembly in the membrane (
Figure 2, part 1)
[9]. Similarly, it was reported that the cell membrane receptor Beclin-1 could be important in the binding of three-domain parasporins (Parts 2 and 3,
Figure 2) and that the Beclin-1 receptor is present in the mammary epithelium and epithelial carcinoma cells (
Figure 2)
[9,34,35][9][13][14].
The rearrangement of the domains typical for the classic protein model of three pore-forming domains does not occur for PS1
[36][15]. Therefore, its activity is not oriented to forming pores in the membrane
[21,36][16][15]. PS1 was proposed to function as an activator of the apoptotic signaling pathway
[14,19,37][17][18][19]. Selective cytotoxicity has been reported for the HeLa, Sawano, HepG2, HL-60, and MOLT-4 cell lines after PS1’s proteolytic activation by trypsin
(Table 1) [6,38][6][20]. The activity of PS1 mainly involves modulating the influx of Ca
2+ levels
[6,21,39][6][16][21].
The PS2 mechanism of action likely starts with recognizing and binding to a receptor located in the cancer cells’ membranes
[24][22], identifying lipid rafts, and anchoring the protein monomers in the periphery. The oligomers, resistant to sodium dodecyl sulfate (SDS), are embedded in the membrane, leading to its permeabilization
[18,24][23][22]. Although PS2 is considered a selective pore-forming toxin, its primary receptors have not been fully elucidated
[18][23].
Cells exposed to PS2 show morphological changes, including inflammation, blisters and lysis, microtubule disassembly, actin filament coiling, and fragmentation of the mitochondria and endoplasmic reticulum. PS2 resides in the plasma membrane and has been shown to activate apoptosis through caspases
[14][17], triggering increased permeability
[14,18,31][17][23][24]. These effects are induced by the accumulation of PS2 by large oligomers in the membrane’s lipid rafts
[8,18,24,39][8][23][22][21]. In turn, the association of PS2 with GPI is required for cytolytic action. By contrast, membrane cholesterol slightly affects the efficiency of oligomerization
[1]. The activation of PS2 induced by proteinase K
[25] leads to the exposure of specific regions that bind to the receptor
[18,25][23][25].
PS3 acts as a pore former in cancer cells, thereby increasing cellular permeability
[12,40][12][26]. Although PS3 is structurally similar to the Cry proteins, containing the five conserved blocks that characterize Cry
[7[7][26],
40], the PS3 and Cry proteins are fundamentally different due to a castor domain
[7,40[7][26][27],
41], which is present in many unrelated proteins and is presumed to enhance/induce carbohydrate-binding capacity
[40][26]. Similar to the above-described PS, the mechanisms of action of PS3 remain largely unknown. Krishnan et al. suggested that PS3 is most likely pore forming
[16][28], which leads to an imbalance in ATP, increased cell size, and membrane damage
[40,41][26][27]. Its cytotoxic activity was evident in the HL-60 and HepG2 cell lines
[7[7][26][27],
40,41], but it did not affect HeLa cells
[41][27].
Studies on PS3, PS4, PS5, and PS6 are limited compared to those on PS1 and PS2, and many action modes remain undetermined. PS4 shows homology with both Cry and pore-forming β-type aerolysin. It has been reported to be cytotoxic to the Caco-2, Sawano, and MOLT cell lines
[5,41][5][27]. Its structure mainly comprises β-sheet domains, and its pore-forming activity is not dependent on cholesterol
[21,41][16][27]. Cells treated with this protein show an increase in size due to an increase in the cytoplasmic compartment and shrinkage of the nucleus, leading to the rupture of the cytoplasmic membrane and cell death
[42][29]. PS5 and PS6 are the most recently discovered PS proteins. They have three domains, similar to PS1 and PS2, and presumably have pore-forming activity. They have been reported to show cytotoxic activity in liver and cervical cancer cell lines. However, there is no further information on their mechanisms of action
[14][17].
4. Perspectives on the Improvement of Bt Parasporins as an Innovative Strategy for Controlling Cancer Cells
By deciphering structure-function relationships, proteins with improved properties, e.g., desired thermal activity, selectivity, specificity, or folding, can be designed
[43][30]. For example, engineered proteins with various substitutions of amino acids are used in receptor- and channel-protein-binding studies
[44][31]. Protein engineering is called the synthesis of proteins with enhanced functionality in vitro and in vivo due to altered physical, chemical, or biological properties through genomic and post-genomic strategies. Genetic improvement is closely linked to complementary computational methods, which aim to optimize the generation of mutant libraries by simulating the experimental conditions of directed mutagenesis techniques
[45,46,47,48][32][33][34][35]. In addition, other computational methods are oriented toward predicting protein structures and designing models that allow the prediction of molecular interactions and pinpoint amino-acid residues or regions at crucial positions in natural and mutant proteins
[43,49][30][36]. The computational technique most widely used for studying the possible interactions of
Bt Cry toxins with insect receptors is molecular docking, followed by molecular dynamics, which has proven to help predict the stability of the interactions and analyze the molecular mechanisms of action.
Florez et al.
[50][37] obtained five Cry11 variants by DNA shuffling and showed the toxic activity against
Aedes aegypti and
Culex quinquefasciatus for three of them. Molecular docking simulations were performed for these three variants, and the amino acids with possible interactions were identified. BenFarhat-Touzri et al.
[51][38] cloned and sequenced the Cry1D-133 toxin and determined its toxicity against
S. littoralis larvae. Molecular docking simulations were performed to explain the enhanced toxicity of this toxin and showed that the number of toxin–receptor interactions was higher than that of the interactions exhibited by the Cry1D toxin.
The use of computational techniques based on molecular dynamics has enabled researchers to study the mechanisms of action of Cry toxins. The study of molecular dynamics has provided novel insights into the oligomerization of Cry toxins at a molecular level. Sriwimol et al. simulated the Cry4Ba structure with a three-dimensional reconstructed map for trimeric protein states. For the first time, they showed the need for membrane-induced conformational changes in Cry4Ba toxin monomers to allow the molecular assembly of a pre-pore trimer, which can be inserted into the target membranes to generate a lytic pore
[52][39].
Other molecular dynamic studies have been applied to investigate the residue interactions relevant to the toxicity of the
Bt Cry toxin family. Pacheco et al. discovered the importance of salt-bridge formation between α-helix residues from adjacent monomers for the toxicity and oligomerization of the Cry1Ab and Cry5Ba toxins by molecular dynamics’ simulations
[53][40]. They showed a critical role for the salt bridge between the E101 and R99 residues of Cry1Ab
[54][41]. Site-directed mutagenesis experiments confirmed decreased oligomerization and toxicity potential for Cry1Ab-E101K and Cry1Ab-R99E mutants.
Interestingly, the R99–E101 salt bridge is not fully conserved in Cry proteins, with both or one of the residues being different in Cry5Ba. However, Pacheco et al. showed that additional salt bridges with similar structural functions could also be formed in these Cry proteins. In conclusion, the computational analysis highlighted the importance of salt-bridge formation between the α-3 helices of adjacent monomers for inducing/facilitating a conformational change crucial for Cry toxicity
[53][40].