Team:Aachen/Lab/Glycogen/Characterization

How We Approached Glycogen Characterization

After building our mono- and polycistronic constructs, their characterization was the next and possibly most important step. In this process, we investigated both cell growth and glycogen formation to prove the accumulation of glycogen due to the overexpression of glycogen building genes and the knockout of glycogen degradation genes. In our approaches, we wanted to examine the function of glgCAB, draw conclusions on fermentation conditions and estimate the increase in glycogen. Therefore, we did growth experiments with Escherichia coli BL21 Gold (DE3) and analysed respective samples via iodine staining and dinitrosalicylic acid staining. Furthermore, we did glycogen purifications to apply a glycogen kit.


Achievements

Part Gene Result
K1585321 glgCAB characterized by iodine staining and therefore improved BBa_K118016
K1585320 glgAB characterized polycistronic BioBrick by iodine staining
K1585310 glgA characterized monocistronic BioBrick by iodine staining
K1585311 glgB characterized monocistronic BioBrick by dinitrosalicylic acid staining
K1585351 ΔglgP characterized gene knockout by iodine staining


a

Iodine staining

Due to the great number of constructs that we aimed to screen, there was a need for a simple and rapid characterization method. Since iodine staining meets these criteria, we chose the procedure for first characterzations. Glycogen accumulation is highest in the stationary phase of bacterial growth [1]. We could also prove this by staining samples from different growth phases. Therefore, the following overnight cultures were grown to this level. To ensure the best comparison between different strains, the samples were adjusted to the same OD before staining.

Results from iodine staining

Our first BioBrick that was successfully characterized by iodine staining is the glycogen synthase, GlgA. The cell pellet of GlgA turned dark brown when resuspened in iodine solution while the wild type cells changed to a light yellow. The iodine molecules interact with the helical structure in the glycogen and thereby change their absorbance and appear dark blue or black/brown. Therefore the dark color shows that more glycogen is present in the BL21 Gold (DE3) + glgA cells compared to the wild type.


Aachen 15-09-02 WT, glgA v2.png
Iodine staining GlgA vs. wild type
Cultivated in LB + 20 mM glucose, BL21 Gold (DE3) wild type + glgA stained distinctly darker than the wild type.


The glgC BioBrick, based on BBa_K118016 by Team Edinburg 2008, was fused with the well-characterized RBS B0034. The dark color shows that the strong effect of GlgC on glycogen accumulation was preserved.


Aachen 15-08-31 glgC characterization LB.png
Iodine staining BL21 Gold (DE3) + glgC vs. BL21 Gold (DE3) wild type
Cultivated in LB and on LB + 40 mM glucose, BL21 Gold (DE3) wild type + glgC stained considerably darker than the wild type.


Iodine staining was also tested on cells cultivated in minimal medium M9. The result of the single knockout ΔglgP was that it was clearly darker than the wild type which proves a higher accumulation of glycogen. This shows that the absence of a glycogen phosphorylase has a strong effect on the glycogen production.


Aachen 15-09-05 M9 WT, delta P v2.png
Iodine staining ΔglgP vs. wild type
Cultivated in M9, BL21 Gold ΔglgP stained distinctly darker than the BL21 Gold (DE3) wild type.


Finally, after our polycistronic construct glgCAB was assembled in an expression vector, it was also cultivated on LB with 20 mM glucose and stained. In several trials, glgCAB was darker than the wild type. Compared to the already existing BioBrick (BBa_K118016) glgC, our polycistronic construct was colored more brown than blue. This results from a higher braching frequency in the glycogen structure.[2]This indicates that the branching enzyme in our glgCAB is, in spite of a point mutation, functionally expressed and influences the glycogen structure.

Aachen glgCAB , WT v2.png
Iodine staining BL21 Gold (DE3) + glgCAB vs. wild type
Cultivated in LB + 20 mM glucose, BL21 Gold (DE3) + glgCAB stained distinctly darker than the BL21 Gold (DE3) wild type.


Aachen 15-09-09 WT, glgCAB1, glgCAB5, glgC v2.png
Iodine staining BL21 Gold (DE3) + glgCAB vs. wild type vs. glgC
Cultivated in LB + 20 mM glucose, BL21 Gold (DE3) + glgCAB and BL21 Gold (DE3) + glgC stained distinctly darker than the BL21 Gold (DE3) wild type. It can be observed that the strain expressing glgC is more blue which indicates more linear glycogen. The glgCAB #1 strain is very dark and more brownish which suggests a higher frequency of branches. This could result from the influence of the branching enzyme.


One of our final experiments was the combination of two of our characterized glycogen constructs. We transformed glgCAB into our E. coli BL21 Gold (DE3) ΔglgP to achieve the highest glycogen accumulation. After staining both samples with iodine, we observed, that the cells with a combination of knockout and polycistronic plasmid, were even darker than the E. coli BL21 Gold (DE3) ΔglgP cells.

Aachen 15-09-17 deltaP +glgCAB vs deltaP.jpg
Iodine staining BL21 Gold (DE3) ΔglgP + glgCAB vs. BL21 Gold (DE3) ΔglgP
Cultivated in LB + 20 mM glucose, BL21 Gold (DE3) ΔglgP + glgCAB stained distinctly darker than BL21 Gold (DE3) ΔglgP. It shows that even higher glycogen accumulation can be achieved by combining overexpression of all three synthesis enzymes and the glgP knockout.

Dinitrosalicylic Acid Staining

The iodine staining of strains containing glgB (branching enzyme) was not quite distinct. We still wanted to investigate the functionality of the strains. Therefore, we used dinitrosalicylic acid staining (DNS) for detection of reducing ends [3] which should correspond to the branching frequency. All samples of glgB strains were compared to wild type samples. For best comparison the samples were grown to stationary phase and adjusted to the same OD before staining. Since every sugar or alkyl would have reacted with 3,5-dinitrosalicylic acid, we purified our samples before the staining. In order to identify the branching frequency, we analyzed the absorbance values of the purified samples compared to the absorbance values of hydrolyzed samples. By calculating the absorbance ratio of the non-hydrolyzed divided by the hydrolyzed samples, we aimed for information about the branches per glycogen unit. The reaction principle can be described as follows.

Aachen GlycogenDNS assay reaction scheme.png
reaction principle

a

3,5-Dinitrosalicylic acid reacts with the reducing ends to 3-Amino-5-Nitrosalicylic acid, resulting in a changed π-system and therefore a change in absorbance. During the reaction, the reducing ends are oxidized whereas one nitro group of 3,5-Dinitrosalicylic acid is reduced to an amino group of 3-Amino-5-Nitrosalicylic acid. The more free reducing ends are present the more 3-Amino-5-Nitrosalicylic acid will be formed. Thus, the absorbance will increase.


Results from Dinitrosalicylic Acid Staining

The absorbance of non-hydrolyzed samples of the glgB strain is higher than the absorbance of the wild type (see figure DNS staining of hydrolyzed samples of the glgB strain and wild type). To identify the amount of glycogen in the samples, we hydrolyzed the samples and applied our staining. Our results show that the number of branches in glycogen is higher in the glgB strain compared to the wild type (DNS staining of non-hydrolyzed samples of the glgB strain and wild type).


Aachen GlycogenDNSNotHydrolyzed.png
DNS staining of non-hydrolyzed samples of the glgB strain and wild type
The non-hydrolyzed samples of the glgB strain and wild type show that the overexpression of glgB leads to glycogen molecules with a higher number of branches compared to the wild type. Error bars show propagation of uncertainty.
Aachen GlycogenDNSHydrolyzed.png
DNS staining of hydrolyzed samples of the glgB strain and wild type
The hydrolyzed samples indicate that the glycogen amount is higher in the wild type compared to the glgB overexpressing strain. Error bars show propagation of uncertainty. Since cell fragments are still left in the samples, the blank also shows an absorbance. Therefore, the line does not go though the origin.

a

The non-hydrolyzed values divided by the hydrolyzed values generate a ratio. Based on this ratio we can identify the number of branches per glycogen unit. It indicates that the number of branches per unit is higher in the glgB strain compared to the wild type.


Aachen GlycogenDNSRatio.png
ratio of non-hydrolyzed and hydrolyzed DNS staining values of the glgB strain and wild type
Although the amount of glycogen is higher in the wild type (see figure above), the number of branches per glycogen unit is higher in the glgB overexpressing strain. Therefore the overexpression of glgB successfully increased the number of branches. Error bars show propagation of uncertainty.
Aachen GlycogenDNSCalibration.png
glycogen calibration curve
With the calibration curve we are able to identify the glycogen concentration of our samples. Error bars show propagation of uncertainty.

a


The results of our experiments show that the overexpression of glgB leads to glycogen molecules with a higher number of branches compared to the wild type. Therefore, we proved the functionality of our glgB BioBrick.

a

Growth experiments

In order to characterize the growth performance of our knockout strains and transformants, we did several growth experiments in shake flasks. By combining these experiments with iodine staining and HPLC analysis of respective samples, we aimed to draw conclusions on possible correlations between the accumulation of glycogen and growth rate. In general, the growth was observed by taking regular OD samples manually or executing the experiment in the Cell Growth Quantifier (CGQ, Aquila Biolabs) for online measurement.


The first growth experiment was performed with electrocompetent BL21 Gold (DE3) single knockouts ∆glgP and ∆glgX in which we transformed glgC in pSB1A30. We compared these strains to the BL21 Gold (DE3) wild type in M9 medium by taking samples manually. As you can see below, the strains without a debranching enzyme enter the stationary phase earlier. This might be because glycogen cannot be used for cell growth. Strains additionally expressing GlgC have an overall lower OD, which indicates a metabolic burden due to the overexpression of glgC.

Aachen Gly BL21vsKnockoutsGrowth.png
Growth of single knockout strains
The growth perfomance was observed in 250 mL shake flasks in M9 medium with 6.6 mM NH4Cl and 40 mM glucose. One could observe that ∆glgX+ glgC had a shorter lag-phase but entered the stationary phase earlier. The growth curve of the wild type and ∆glgP + glgC was relatively similar.
Aachen Gly OnlineGrowth1.png
Growth of single knockout strains containing glgC
The conditions were the same as for the growth experiments on the left.



When we examined all three cultures via iodine staining, ∆glgP + glgC was darker than the wild type. In HPLC analysis, we found out that we did not have nitrogen limitation in the stationary phase, which is necessary for glycogen accumulation.[4] After another experiment with the single knockouts in BL21 Gold (DE3) where the nitrogen concentration was too low, we optimized our M9 medium. Using this M9 medium (6,6 mM NH4Cl, 40 mM glucose) and doing online measurements in the CGQ, we were able to reproduce the results for the single knockout strains and the single knockout strains with glgC. Via iodine staining of samples in different growth phases, we could verify that most glycogen accumulation occured in stationary phase.


When we cultivated glgC, glgCAB and the BL21 Gold (DE3) wild type in LB with 20 mM glucose, we observed that BL21 Gold (DE3) + glgCAB was able to grow in spite of expressing three additional genes. This is a promising precondition for high future glycogen production.


Aachen growth glgC, glgCAB, WT.png
Growth of BL21 Gold (DE3) containing glgCAB
Despite of additional metabolic burden, the strain containing glgCAB reached the highest end OD on LB with 20 mM glucose
Aachen iodine staining of glgC, glgCAB.png
Iodine staining after the growth experiment
The darker iodine stain of this strain suggests that the strain expressed glgCAB and accumulated more glycogen in the stationary phase than the wild type and the strain with glgC

a

Outlook

Feeding experiments

To prove the hypothesis that other bioprocesses can be built upon bacterially produced glycogen, the following experiment is a practical approach. The expectation is that feeding a wild type culture in stationary phase with cooked cells containing glgCAB in an expression vector should initiate subsequent growth due to the additional glucose in the feed.


Procedure

  • inoculate three 25 ml maincultures of BL21 Gold (DE3) wild type in C-lim M9 medium from LB precultures
  • inoculate two 10 ml LB + 20 mM glucose cultures (one of wild type culture and one containing glgCAB in psB1A30), induce the strain containing glgCAB with IPTG
  • adjust the two feed cultures to the same OD, centrifuge, resolve in water, cell lyse and hydrolization
  • feed WT1 with WT, WT2 with CAB, WT3 with water (500 µl)
  • track growth curves

HPLC analysis

Another approach to measure the glycogen accumulation quantitatively would be the analysis of samples by HPLC. With this method purified glycogen can be directly measured or the glucose levels can be observed after glycogen hydrolysis. We worked on this approach but the main problem was that we did not have a reliable glycogen purification protocol. Hence, our samples could not be analyzed properly. Nevertheless, by optimizing this aspect the glycogen content of cells should be clearly visible as we have already seen correct peaks for the synthesis operon in the glgP knockout strain.

a

Combining glycogen synthesis and MCC pathway

MCC operon in Bl21 Gold (DE3) ∆glgP


Over the course of our project, we initially pursued working in the subgroups “methanol” and “glycogen”. As our work progressed and first milestones were reached, we took a step forward and thought of ways to combine both aspects –methanol assimilation and glycogen accumulation- in one organism. Since the glgP single knockout in Bl21 Gold (DE3) gave good results regarding its growth performance and glycogen accumulation, we decided to transform the polycistronic construct of methanol assimilating genes (mdh,hps, phi, xpk) into this strain. We saw that this strain was able to grow on M9 with 0.522 M methanol. The most effective way of further optimizing its growth and methanol assmilitation is to apply promoters with different strengths.

Laboratory Notebook

References

  1. Romeo T, Preiss J. Genetic regulation of glycogen biosynthesis in Escherichia coli: in vitro effects of cyclic AMP and guanosine 5'-diphosphate 3'-diphosphate and analysis of in vivo transcripts. J Bacteriol. 1989 May;171(5):2773–2782
  2. Lengeler J., Drews G., Schlegel H. Biology of the Prokaryotes, page 194, 9.5.2
  3. S. K. Meur, V. Sitakara Rao, and K. B. De. Spectrophotometric Estimation of Reducing Sugars by Variation of pH. Z. Anal. Chem. 283, 195-197 (1977)
  4. Fung T. et. al. 2013. Residual Glycogen Metabolism in Escherichia coli is Specific to the Limiting Macronutrient and Varies During Stationary Phase. Journal of Experimental Microbiology and Immunology (JEMI)