The tailspike protein gp17 of phage LSB-1 is not an endosialidase
David Schwarzer, Hannover Medical School
31 May 2011
The authors Chai et al. present a novel phage named LSB-1. By analyzing the genome they could identify this phage belonging to the T7-like phage family. The phylogenetic trees in Figures 5¿7 of their study nicely illustrate the close relationship to the other T7-like phages which were used for comparison. In these three phylogenetic trees, the phage LSB-1 was found to be closely related to coliphage K1F. This T7-like phage infects the pathogen Escherichia coli K1. These bacteria are encapsulated by a thick capsule composed of ¿2,8-linked sialic acid homopolymers (polySia). The tailspike proteins of K1-phages that are required to specifically bind and degrade the capsule of K1 strains are named endosialidases. Endosialidases are common to all known K1-phages so far and moreover K1-phages are the only known source of endosialidases.
However, the authors state that the tailspike protein gp17 is closely related to gp17 of K1F, which is the endosialidase endoNF. First of all, they showed that phage LSB-1 infects acapsular E. coli strains with the O124-antigen, which do not contain sialic acids or polySia. LSB-1 was NOT shown to infect K1 strains; therefore LSB-1 most likely does not comprise an endosialidase. The authors further wrote that gp17 of LSB-1 has an endosialidase activity, which was not shown in the study.
Second, the sequence of gp17 they have published shares high sequence similarity with endoNF of more than 90% in the N-terminal stretch of ~220 amino acids. The rest of both proteins share sequence identities of less than 10%, which is not enough to assign gp17 of LSB-1 an endosialidase function. The N-terminal portion that is highly conserved between gp17 of LSB-1 and endoNF is common to all T7-like phages as depicted in Figures 5¿8. This stretch has previously been identified as the head-binding domain that is required to attach the respective gp17 to the phage particle. The tree in Figure 8 of the study clearly shows that gp17 of K1F (endoNF) and of LSB-1 are distantly related. A common head-binding domain classifies a tailspike protein to a phage ¿ in this case the T7-like family. The function of the tailspike classifies the host specificity - which seems to be different for K1F and LSB-1.
Finally, the authors tried to model gp17 of LSB-1 into the 3D-structure of endoNF. This is simply not reasonable when both proteins share a sequence similarity of less than 20% and e-values of greater than 1! Further, the crystal structure of the N-terminal head-binding domain of endoNF - common to gp17 of LSB-1 - has not been solved as yet. The model of gp17 does not have anything in common to a sialidase or an endosialidase - e.g. the typical sialidase ß-propeller is not found. The ¿catalytic triad¿ in their model is not located in a pocket in which substrate may bind nor are these denoted amino acids conserved in an alignment of both sequences.
The LSB-1 gp17 is most likely a glycosidase with unknown substrate specificity, but not an endosialidase. To non-endosialidase experts this article might be misleading.
Virology Journal
The tailspike protein gp17 of phage LSB-1 is not an endosialidase
31 May 2011
The authors Chai et al. present a novel phage named LSB-1. By analyzing the genome they could identify this phage belonging to the T7-like phage family. The phylogenetic trees in Figures 5¿7 of their study nicely illustrate the close relationship to the other T7-like phages which were used for comparison. In these three phylogenetic trees, the phage LSB-1 was found to be closely related to coliphage K1F. This T7-like phage infects the pathogen Escherichia coli K1. These bacteria are encapsulated by a thick capsule composed of ¿2,8-linked sialic acid homopolymers (polySia). The tailspike proteins of K1-phages that are required to specifically bind and degrade the capsule of K1 strains are named endosialidases. Endosialidases are common to all known K1-phages so far and moreover K1-phages are the only known source of endosialidases.
However, the authors state that the tailspike protein gp17 is closely related to gp17 of K1F, which is the endosialidase endoNF. First of all, they showed that phage LSB-1 infects acapsular E. coli strains with the O124-antigen, which do not contain sialic acids or polySia. LSB-1 was NOT shown to infect K1 strains; therefore LSB-1 most likely does not comprise an endosialidase. The authors further wrote that gp17 of LSB-1 has an endosialidase activity, which was not shown in the study.
Second, the sequence of gp17 they have published shares high sequence similarity with endoNF of more than 90% in the N-terminal stretch of ~220 amino acids. The rest of both proteins share sequence identities of less than 10%, which is not enough to assign gp17 of LSB-1 an endosialidase function. The N-terminal portion that is highly conserved between gp17 of LSB-1 and endoNF is common to all T7-like phages as depicted in Figures 5¿8. This stretch has previously been identified as the head-binding domain that is required to attach the respective gp17 to the phage particle. The tree in Figure 8 of the study clearly shows that gp17 of K1F (endoNF) and of LSB-1 are distantly related. A common head-binding domain classifies a tailspike protein to a phage ¿ in this case the T7-like family. The function of the tailspike classifies the host specificity - which seems to be different for K1F and LSB-1.
Finally, the authors tried to model gp17 of LSB-1 into the 3D-structure of endoNF. This is simply not reasonable when both proteins share a sequence similarity of less than 20% and e-values of greater than 1! Further, the crystal structure of the N-terminal head-binding domain of endoNF - common to gp17 of LSB-1 - has not been solved as yet. The model of gp17 does not have anything in common to a sialidase or an endosialidase - e.g. the typical sialidase ß-propeller is not found. The ¿catalytic triad¿ in their model is not located in a pocket in which substrate may bind nor are these denoted amino acids conserved in an alignment of both sequences.
The LSB-1 gp17 is most likely a glycosidase with unknown substrate specificity, but not an endosialidase. To non-endosialidase experts this article might be misleading.
Competing interests
None declared