*Corresponding author. Present address: Department of Pharmacology, University of Illinois, Room 403, 835 S. Wolcott Avenue, Chicago, IL 60612-7344, USA. Tel.: +1 (312) 996-4318; Fax: +1 (312) 996-1225, E-mail address: email@example.com
Hemolysin II gene from Bacillus cereus VKM-B771 has been sequenced. The deduced primary translation product consists of 412 amino acid residues which corresponds to the protein with an Mr of 45.6 kDa. The predicted mature Hly-II protein (residues 32 to 412) is of 42.3 kDa, which is in close agreement with the mini-cell electrophoresis analysis. Hly-II deletion variant lacking 96 C-terminal residues still has hemolytic activity. The protein primary structure analysis revealed no homology with any known Bacillus cytolysins. Significant general homology (31–28% identity) was found between the hemolysin II and Staphylococcus aureus alpha-toxin, gamma-hemolysin (HlgB), and leukocidins (LukF, LukF-R, LukF-PV). The data suggest that hemolysin II belongs to the group of β-channel forming cytolysins.
Bacillus cereus produces several extracellular hemolysins considered as potential factors of virulence of this toxicogenic microorganism: cereolysin , sphingomyelinase , cereolysin AB , cereolysin-like hemolysin , hemolysin BL , hemolysin III , and hemolysin II [7,8]. Till recently, hemolysin II was one of the most controversial hemolysins of B. cereus. This partially purified novel hemolytic factor of B. cereus was described in 1978 . Although the properties of hemolysin II were quite well described, the information on further purification or cloning of its gene was not documented. The existence of hemolysin II as a novel hemolytic protein was doubtful [3,9]. Recently, we have succeeded in cloning the hemolysin II genetic determinant . The cloned 2.9 kb EcoRI fragment of B. cereus VKM-B771 genomic DNA was shown to be distinct from cereolysin AB  and hemolysin III genes . The hemolytic product encoded by the cloned DNA fragment possessed all the properties ascribed to hemolysin II and distinguishing it from cereolysin [7,8]. These data gave evidence that hemolysin II is an independent B. cereus hemolytic factor. Later we found the hemolysin II determinant in a number of B. cereus strains and in most of the Bacillus thuringiensis strains tested .
Here we present the complete nucleotide sequence of the hemolysin II gene and its deduced protein primary structure and deletion analyses.
2Materials and methods
2.1Plasmid constructs and strains
Plasmid pUJ1 containing the hemolysin II genetic determinant was constructed previously by cloning the 2.9 kb EcoRI fragment of the B. cereus VKM-B771 DNA into vector pUC19 . It was used for generation of a set of unidirectionally deleted derivatives by exonuclease III digestion. The Escherichia coli strain Z85  was used in hemolytic phenotype detection experiments and as a host for plasmid DNA preparation.
2.2Hemolytic phenotype detection
The hemolytic phenotype of the recombinant cells was tested by the appearance of hemolysis zones (clearance zones) around colonies grown on LB agar containing 1% suspension of human erythrocytes supplemented with ampicillin (100 μg ml−1). The cells were grown overnight at 37°C and thereafter incubated at 20°C for several more hours .
2.3DNA manipulations, sequencing and sequence analysis
DNA manipulations were performed by standard techniques. Digest of DNA by exonuclease III and Mung Bean nuclease (Promega) for generation of the deletion derivatives of pUJ1 was performed according to the manufacturer's instructions. The sequencing reactions were carried out by the dideoxy chain-terminating method with Klenow and Delta Taq Version 2.0 polymerases using Amersham and USB sequencing kits, respectively. Both DNA strands were sequenced using universal direct primer and walking primers. The walking primers were synthesized on a Gene Assembler automated DNA synthesizer (Pharmacia, Sweden) by the cyanoethyl-phosphoramidite method and purified on an OPC column (Applied Biosystems, USA). The nucleotide sequence obtained was analyzed with the PC/Gene software (Intelligenetics). The search for homologous sequences and alignment were made with the FASTA and the BLAST programs.
3Results and discussion
The hemolysin II genetic determinant was cloned previously within the 2.9 kb EcoRI fragment of genomic DNA of B. cereus VKM-B771  (Fig. 1, plasmid pUJ1). Deletion of a 0.5 kb EcoRI-BamHI sub-fragment at the left flank of the 2.9 kb EcoRI fragment resulted in a loss of the hemolytic phenotype, which indicated localization of hemolysin II gene at the left flank of this DNA fragment  (Fig. 1, plasmid p701-B). In this study, for the further localization of the gene we produced a nested set of unidirectional deletions from the right EcoRI site of the 2.9 kb fragment (Fig. 1). A minimal DNA fragment providing recombinant E. coli cells with the hemolytic phenotype was 1.2 kb (plasmid pUJ40). However, the hemolytic phenotype caused by the presence of this fragment was noticeably weaker than that of longer fragments (Fig. 2). E. coli cells bearing the plasmid pUJ30 with 1.4 kb insert had almost as strong hemolytic phenotype as that of cells with original 2.9 kb insert plasmid (Fig. 2). The 1.4 kb DNA fragment from plasmid pUJ30 was sequenced (Figs. 1 and 3). Sequence analysis has revealed the only ORF starting at position 252 from the left EcoRI flank. Potential initiation codons in this ORF are located at positions 258, 270, 321, 435. However, only the ATG codon at position 270 is preceded by a putative ribosome binding site, GAAGGAG, at a distance of 11 base pairs (bp), which suggests that translation of the hly-II gene initiates at this ATG. However, we did not find any termination codons in the reading frame within the 1.4 kb fragment sequenced. So, the sequence analysis was performed further and revealed in-frame TAA stop codon at position 1506 (Fig. 3). Thus, analysis of the sequence obtained (1669 bp) showed that the ORF consists of 1236 nucleotides. Potential transcription terminator, a region of dyad symmetry with estimated free energy of −13.3 kcal mol−1, was found downstream from the ORF (nt 1594 to 1618). Upstream of the ORF (nt 26 to 69) an additional palindrome was found, which could form a perfect 22 nt steam structure (ΔG of −27.8 kcal mol−1) and might serve as a transcription terminator of an upstream gene(s).
The length of the deduced primary translation product is 412 amino acid residues which corresponds to an Mr of 45.6 kDa. This protein appears to possess the only long stretch of hydrophobic amino acids at position 10–31. This N-terminal hydrophobic region is preceded by the amino acid sequence enriched in four positively charged lysine residues. Thus the N-terminal region of the deduced hemolysin II exhibits properties typical for known signal peptides . Assuming that the 31 N-terminal amino acids represent a signal peptide, the mature Hly-II protein is of 42.3 kDa. This calculated value is in good agreement with the mini-cell analysis of proteins produced by plasmid pUJ1  and pUJ1 derivative plasmid containing hly-II 5′-flanking and coding regions (nt 1 to 1508, Fig. 3) (Budarina, Zh.I., unpublished data). The deduced Hly-II protein primary structure analysis has revealed no homology with known hemolysins from genus Bacillus, which corroborates our conclusion made previously  that hemolysin II is an independent B. cereus hemolysin encoded by its own gene different from those described to date. However, significant general homology has been revealed between the deduced hemolysin II and Staphylococcus aureus alpha-toxin  (31% identity, Fig. 4). The homology with alpha-toxin from S. aureus is observed within the 317 N-terminal amino acids of the Hly-II deduced protein, while the Hly-II primary translation product is 93 amino acids longer than alpha-preprotoxin, whose deduced length is 319 amino acid residues. As mentioned above, the 1.2 kb derivative of the original 2.9 kb insert was still capable of producing hemolytic activity (Fig. 2). We sequenced the 1.2 kb insert. Its precise length is 1218 bp and it encodes protein of 316 amino acid residues. The size is surprisingly close to the size of alpha-toxin. As shown by blood agar tests, 1.4 kb insert encodes truncated protein which appears to be as hemolytic as native Hly-II protein (Fig. 2). The precise length of the insert is 1405 bp, and the length of the corresponding protein is 378 amino acid residues, which is 33 residues shorter than native Hly-II. Thus, 316 N-terminal amino acids (including potential signal peptide) are sufficient for reduced but detectable hemolytic activity of Hly-II; next 63 amino acids (from 317 to 379) contribute to the hemolytic activity; contribution of the 33 C-terminal amino acids in the Hly-II hemolytic activity is barely detectable on human blood agar.
Alpha-toxin is a hemolysin produced by most pathogenic strains of S. aureus. It causes lysis of all types of mammalian cells  by forming heptamers which insert into the cell membrane and generate hydrophilic pores . Taking into account the found homology, it is interesting to note a number of common features between the alpha-toxin and hemolysin II. Both these proteins exhibit a pronounced Arrenius effect (the ability to restore hemolytic activity after short-term heating at high temperature) [8,16], show similar dependence of the hemolytic activities on temperature [8,17], and possess increased specificity to rabbit erythrocytes in comparison with human ones (, Budarina, Zh.I., unpublished data).
Besides alpha-toxin of S. aureus, several cytolytic proteins from this microorganism showed homology with hemolysin II: leukocidin R, component F (29% identity) , gamma-hemolysin, component B (29% identity) , leukocidin, component F (28%) , Panton-Valentine leukocidin, component F (28% identity)  (alignments not shown). All these proteins belong to the family of β-channel forming toxins . The data suggest that hemolysin II is a member of this group of cytolytic proteins. The Hly-II polypeptide is remarkably longer than other cytolysins of the group. However, deletion of about one fourth of its length from its C-terminus results in still hemolytic protein whose size is extremely close to the sizes of its homologs. Influence of deletion of 33 C-terminal amino acids of the hemolysin is almost undetectable on the blood agar. The facts raise question about functional role of C-terminus of hemolysin II. Taking into account limitations of the blood agar test, we consider quantitative analysis of hemolytic activities of the hemolysin II and its truncations as the next step in our investigation of this hemolysin.
The authors wish to thank Dr. M.G. Shlyapnikov for the synthesis of oligonucleotides, Dr. M.A. Sinev and Dr. A.Yu. Tomashevski for critical discussions, and Lyuba Ulyanova for excellent technical assistance.