From only one bacterial colony, THN1, a potential mlrA gene was amplified and sequenced. blast analysis showed a 98.5% identity between this sequence and the mlrA gene sequence LBH589 of Sphingomonas sp. ACM-3962. The 16S rRNA gene of this bacterial strain was also sequenced, and a homologous search by blastn showed a maximum identity (99%) to Novosphingobium aromaticivorans DSM 12444 (GenBank no. CP000248). Therefore, this bacterial strain was identified as Novosphingobium sp. THN1 belonging to the family Sphingomonadaceae. Removal of microcystin LR in the THN1 culture was observed following analysis of the remaining microcystin LR (Fig. 1). There was a sharp decline during the first 12 h
and 91.2% of the toxin was eliminated in this period. Because microcystin PF-02341066 mw LR could not be detected in the culture after 60 h, complete degradation was concluded.
No decrease in the toxin occurred in the negative control (data not shown). A potential mlr gene cluster with four genes mlrA, mlrB*, mlrC and mlrD was successfully cloned from THN1. All the gene sequences were confirmed to be mlr by aligning with the corresponding genes found in GenBank. The coverage of each mlr sequence from GenBank and their similarity to mlr of THN1 was calculated using bioedit V5.0.6 (Table 2). THN1 had maximum identities with different strains for each gene including mlrA (MD-1, 99.7%), mlrB* (C-1, 96%), mlrC (C-1, 91.7%) and mlrD (ACM-3962, 95.7%). A particularly low similarity (83.7%) of mlrA was found between THN1 and Y2 (Saito et al., 2003), indicating that the Y2 strain has experienced more variation. The two mlr clusters of THN1 and ACM-3962 had a similarity of 95.6%. Relative locations and directions of transcription for each mlr gene of THN1 were the same with ACM-3962. Because the only available mlrC gene sequence (1521 bps) from ACM-3962 does not contain a stop codon, the mlrC (1536 bps) coding 511 amino acid residues, found in this study, was the first reported complete ORF for this gene. Alignment of
mlrB* sequences diglyceride for THN1 and ACM-3962 showed three base insertions (Fig. 2a) at positions 30(C), 44(C) and 1176(G). Apparently, the insert mutations caused a frameshift and eight stop codons (Fig. 2b) within the gene sequence. In an attempt to determine whether mlrB* was transcribed into mRNA in the THN1 cells, we tried to amplify mlrB* from the total cDNA. As displayed in the gel image (Fig. 3), high-quality total RNA was extracted from THN1 cells and no genomic DNA could be detected in the RNA extracts after digesting with DNase. In PCR reactions using total cDNA, the mlrA amplicon was obvious, but no mlrB* product could be detected. In other words, no mRNA of mlrB* gene existed in the complete RNA for the THN1 cells. Upregulated expression of mlrA gene was detected upon exposure to microcystin LR (Fig. 4).