Protein modeling with α-N-acetylgalactosaminidase from E. meningoseptica

Our modeling aims to improve the activity of α-N-acetylgalactosaminidase from E. meningoseptica . There would be many practical advantages for a more efficient enzyme: For instance, conversion has to be completed as fast as possible for routine tasks in hospitals. An another example would be to compensate for a possible loss of activity, by being fixed onto a membrane. In order to achieve this we used the experimentally - determined structure of α-N-acetylgalactosaminidase from Liu et al. (2007).

As described in “Bacterial glycosidases for the production of universal red blood cells”, the active site consists of Tyr307, Tyr225, His228, Glu149, Tyr179 and Arg213. With this we tried to examine the interactions of the residues with the substrate. The precise distances, which would allow interactions, are listed in Table 1. The values refer to Schulz & Schirmer's „Principles of Protein Structure“.

Table 1: Different distances between residues which would allow interactions.

It was found that the substrate in the active site (Figure 1) is not stabilized at OH - 6 by residues (Substate Structure), as already mentioned in Liu et. al. (2007). It has been verified with the Protein Interactions Calculator that no interactions with OH - 6 were detected.

Figure 1: Active site of N-Agal with CASTp in PyMOL.

After that, we discussed with Prof. Dr. Thilo Stehle (Interfaculty Institute of Biochemistry) and decided that it would be possible to facilitate the nucleophile attack of the hydroxyle group from Tyr179. Furthermore we tried different mutations in Leu183 and Val186. We found that a mutation of Val186 to Asp186 should stabilize the substrate. The distance between OH - 6 and Asp186 - COOH is 2,9 Å and would be suitable for a hydrogen bond (Figure 2). Also the distance to other residues is over 5 Å.

It is important to note that such negative charge could have a negative effect on the activity. Due to that we tried Asn186 (Figure 3), but the distance with 3,03 Å is not optimal. We again consulted Prof. Dr. Stehle. He made the point, that prediction tools generally don’t always deliver very accurate results. When we simulated substrate binding in the mutated active site using CLC Drugdiscovery Workbench, we could not obtain an applicable result. However Prf. Dr. Stehle also noted, that it is very difficult to predict the activity for the mutated protein soley based on structure. It might strongly bind the substrate with the new hydrogen bond.

Figure 2: Mutation Asp186.

Figure 3: Mutation Asn186.

In conclusion, the best option is a mutation of Val186 to Asp186, where no collisions with other residues occur and a stabilizing effect at OH - 6 is achieved. The mutant looks to be promising, however no definite statement can be made about the mutants effect to the enzymatic activity. With this modeling, we managed to suggest a possible improvement for our enzyme, which can be put into practice and further compared to the enzyme in its native state


[1] Liu et al. (2007) : Bacterial glycosidases for the production of universal red blood cells

[2] G.E. Schulz and R.H Schirmer „Principles of Protein Structure“.