Bacterial identification requires a series of experiments to accurately designate the taxonomic affiliation of a strain. Sequencing of 16S rRNA gene has been considered as one of the most important feature of bacteria taxonomy. If two bacterial strains show less than 97% 16S rRNA gene sequence similarity, they are separate species. To find out the sequence similarity reference sequences are critically required. There are a number of databases (NCBI, EZtaxon, RDP etc…) offering well curated sequences of 16S rRNA gene of validly published type strains, and allow microbiologists to designate taxonomic affiliation to his/her strains accurately. Sequencing of 16S rRNA gene for strain confirmation and validation is a routine for us at the Microbial Culture Collection, we use EZTaxon database for bacterial identifications. Yesterday, I was in between of confirming few type strains of Mesorhizobium acquired form DSMZ, when I searched the sequence of Mesorhiozbium loti DSM 2626(T) it showed 99.77% sequence similarity with Mesorhizobium qingshengii CCBAU 33460(T) (Fig. 1).
I thought, ok it’s fine and may happens some time when the strains have very similar 16S rRNA gene sequences, but to my surprise the type strain of M. loti USDA 3471(T) was placed at 22 position with a similarity of 97.97%. This is something very unusual, I never faced such situations where 16S rRNA gene sequence of same strain showed such a huge difference. I went back to my sequence contig and confirm the quality of sequence once again, which was absolutely alright. After this, I compared the newly generated sequence with the sequence accession mentioned in the strain description webpage of DSMZ, it showed 99.9% similarity, which confirmed that the strain is original. Now, several questions poured in to my mind making me restless:
Is the strain deposited at DSMZ is not the original type strain?
Where DSMZ got this strain in to their collection?
What could the possible reason for low sequence similarity?
Today, I started my day with the objective to find out answer for all my questions. The first thing, I did was to search the sequence in NCBI blast using options of exclude models and uncultured/environmental samples sequences, and limit to sequences from type material. The blast results gave answer to all my questions within few seconds as there were four sequences of M. loti type strains as top hits (Fig.2).
It was quite hilarious to find such answer, which I never expected, there were multiple sequences of the same strain with huge variation among them. Now, a new question arise that how is it possible?
I found answer to this question in a article by Willems et al. (2001) http://www.sciencedirect.com/science/article/pii/S0723202004700697, reporting the differences between subcultures of the Mesorhizobium loti type strains from different culture collections. After reading this article, I come to know that the strain with which I am dealing is the valide type strain of M. loti and the 16S rRNA gene sequence of M. loti USDA3471(T) curated by EZTaxon is not the valide type strain. This article not only cleared my confusion on the type strain of M. loti, but also attracted my attention towards the importance of quality control of reference material deposited at or exchanged between culture collection.