Glyco2D

Glyco2D is the software we developed to display glycogen structures based on different parameters. We built Glyco2D based on the mathematical model described by Meléndez-Hevia et al. ₁. This model described the structural properties of glycogen based on different parameters such as chain length, branching degree and the number of tiers. The model was able to demonstrate the optimal values of these parameters for maximizing the glucose stored in the smallest volume and the number of non-reducing ends. These optimal values were branching degree of 2, chain length of 13 glucose units and approximately 12 tiers ₁. The software we developed uses these properties to predict the structure of the molecule₂.

The software was created in C++ using openGL. The individual glucose molecules are represented by black squares and they are used as building blocks for the chains and branches.

To create the glycogen structures we made the following assumptions:

•The branching points on the chain are always the 5th and 9th glucose molecule on the chain.
• All chains are equal in length.
• The branching degree is 2 on each chain, except on the final tier.

Glyco2D Predictions

The software creates a pool of glucose units which are used to build the structure one tier at a time. Each time a new chain is about to be built, the software checks from a total of 24 possible directions and eliminates those that would grow toward the inner parts of the structure as it is not physically possible that the molecule is synthesized towards the core of the structure. The procedure to make this elimination is to discard the growth of any chain with a distance between its end and the centre of the molecule that is less than the distance of another chain belonging to two preceding tiers.

To make it easier to visualise, the software does not allow for chains to cross paths on the same plane, which does reduce the number of valid tiers compared to a three dimensional model. If there is more than one possible valid chain then the software will randomly select which valid chain to build.

Acylation simulation

Bacterial glycogen and plant starch consist of chains of glucose residues connected by alpha-1,4-glycosidic linkages with alpha-1,6-glycosidic linkages forming branch points. Our main aim was to produce acylated or butrylated starch. As plants are more difficult to work with, we initially expressed four putative acyltransferases in E. coli to see if we could modify bacterial glycogen. However, the activity on these enzymes is still unknown. We don't know at which position in the glucose molecule that the enzyme might add the acyl group. If the group is added to the free end available at a growing branch, it will compete with the glycogen synthase and disrupt the growth of the molecule.

Even though glycogen acylation has not yet been described, phosphorylation of glycogen has been studied previously in both muscle and bacterial glycogen. Tagliabracci V. S. et al. showed that the phosphate in glycogen is present as C2 and C3 phosphomonoesters ₃ . Phosphorylation of starch has been also characterized. According to Blennow A. et al., the phosphate groups bind at the free C6 and C3 hydroxyl groups of the glucose units. Both groups are located at the hydrophilic surface of the double helix, which might affect the stability of the molecule ₄.

We therefore needed to model the putative changes in glycogen structure depending on the location of the modification. Our aim was to produce carbohydrate molecules with 5-10% of the residues modified since this level of butrylation (achieved by chemical modification) has positive benefits to the colon of rats ₅ Our model indicates that this level of modification is only viable if the enzyme can modify any base - If it can only use the carbon-4 position it would impact on the growth of the molecule.

Glyco2D Works!

When we increase the initial percentage of random acylation, the total number of modified glucoses increased. We showed a great linear correlation between both (see Figure 3)

Acylation on Carbon 4

As we're using putative acyltransferases, we don't know the carbon in the glucose molecule that gets acylated. If it binds to C4, the acylation will restrict further growth of the glycogen molecule.

As we can see, by having a chance of acylation of 1% at C4, the glucose lost is around 50%. By 'glucose lost' we mean all the glucose molecules that could haven been added to the unmodified molecule (see Figure 4).

Therefore, if the acyl group always binds to C4, we won't have a viable glycogen molecule (see Figure 5).

Figure 4: Results from acylation simulation with acyl group added into carbon 4.