iGem Tübingen 2017

Team Inspiration Results Human Practice Lab Attribution

Biochemistry

Introduction

Streptomyces coelicolor
S.coelicolor is a bacterium that belongs to the genus of Streptomyces. Over 500 species of Streptomyces are known which are mainly found in soil and decaying vegetation where they grow in mycelial manner and produce spores. One characteristic is their earthy smell which is caused by the production of a volatile metabolite, Geosmin. Streptomyces are aerobic, gram-positive and have a genome with high GC-content. These little bacteria are most famous for their bioactive secondary metabolites that can be antifungal, antiviral, antitumoral, anti-hypertensive, immunosuppressant, and antibiotic.
Application in Biotechnology
Since Streptomyces are able to produce complex chemical components, they can be used for production of useful substances such as antibiotics. Furthermore, Streptomyces are used for heterologous protein expression as they have a few advantages compared to E.coli. Some proteins refuse to fold properly in E.coli, while they do in Streptomyces. Another advantage is their good secretion system which simplifies purification.

Production of antibiotics by Streptomyces
One class of antibiotics produced by Streptomyces are aminocoumarins which inhibit bacterial DNA gyrase enzymes by binding tightly to subunit B competing with ATP for the ATP binding site and thereby blocking supercoiling. The therapeutic index is quite small because aminocoumarins also bind to ATP binding sites of human topoisomerases type II.
All aminocoumarins have a similar structure of three rings (A to C) and are produced in a similar fashion. The prenylated ring A is the first part on which ring B and C are attached.

Our goal was to render the aminocoumarins specific to β-Lactam antibiotics resistant bacteria such as the dangerous Methicillin resistant Staphylococcus aureus (MRSA) which leads to many deaths every year. The new antibiotic should only target β-Lactam resistant bacteria for which reason one side important for the protein ligand interaction was masked by a β-Lactam ring. The resistant bacterium will cleave the β-lactam ring, activate the antibiotic and die. Therefore, our warhead, a β-lactam, was installed at the prenyl tail of Ring A. The bioinformatic analysis proofed our theoretical assumption that the binding of the closed β-Lactam ring to GyrB and human topoisomerase type II (LINK TO Bioinfo) is weaker than the binding of clorobiocin. Furthermore it showed that the cleaved β-Lactam ring binds stronger to the target than clorobiocin. The modification does not only decrease the binding strength to its targets in the uncleaved state, but also increases the binding strength to its target as soon as it is cleaved by a β-Lactamase. Our modified aminocoumarin will be less toxic for every organism that is not resistant to β-Lactamase , but more toxic for the ones with resistance.

Theoretical Background

Key features
In order to produce our new antibiotic we needed a production strain that lacks the biosynthetic pathway for Ring A and an amide synthase which accepts our substrate. Furthermore, we have to close the β-Lactam-ring using a β-Lactam synthetase (BLS).
RingA
The key enzyme for the production of Ring A of clorobiocin is CloQ a prenyltransferase (Heide et al. 2003). A cloQ defective S.coelicolor M512 strain (Clo/SA02) with a recombinant biosynthetic cluster was made available to us by the group of Prof. Heide at University of Tuebingen.
Amid synthase
As the structure of the amide synthase (CloL) from the clorobiocin biosynthetic pathway is not available, we decided to use amid synthases from the related pathways of novobiocin (NovL) and coumermycin A1 (CouL) and co-express them in our production strain to increase the chance of successful incorporation. As shwon by Anderle (2007) CloL is the most promiscuous of the three enzymes (Anderle et al. 2007). Introduction of genes in Streptomyces is possible by conjugation, which means one bacteria transfers DNA over a dedicated transportation complex to another bacteria.
Installation of the β-Lactam ring / warhead
Although, many possibilities for formation of a β-Lactam ring by chemical synthesis are available, we chose a path via enzymatical synthesis. This bears the advantage that the enzyme can be co-expressed in Streptomyces. The enzyme β-Lactam synthetase (Bachmann et al. 1998) seemed to be useful as it uses a β-amino acid to form the β-Lactam ring. The synthesis of the needed β-amino acid was easily possible. Based on available crystal structures of the BLS from S. clavuligerus (1mbz,1mc1, 1mb9,1m1z,1jgt), we decided to change two amino acids within the active site to increase the probability that the enzyme accepts our substrate (based on the structure with the trapped intermediate (1MBZ)).

Procedure

Amid synthases
For conjugation of amid synthase genes (CloL, CouL, SimL), the genes were cloned into a special plasmid pUWL-Apra-oriT which was supplied by the laboratory of Dr. Gust. The pUWL vector has origins of replication for propagation in E.coli and S.coelicolor and additional an origin of transfer (oriT), the startpoint of replication for conjugation and and a constitutive Streptomyces promoter (ermE up). In contrast to E.coli Streptomyces do not need a RBS. Streptomyces have a strong methylation based restriction system, hence, all DNA passed a methylation (dam) defective E.coli strain like ET12567. This strain contains pUZ8002, which supplies the transfer function to oriT containing plasmids by expressing Tra which allows Tra defective vectors in trans mobilisation. pUZ8002 will not be transferred (Paget et al. 1999)OpenWetware.
Germinated Streptomyces recipient spores and ET12567 donor cells were plated together on Mannitol-Soya Flour (MS)-agar plates with 10mM MgCl2. On the next day, the plate was overlaid with nalidixic acid (bacteriostatic to E.coli) and apramycin (plasmid selection). After three days, the first exconjugants appeared and a replica were prepared. Three colonies were picked and and diluted to gain single colonies. This step was repeated. After colonies appeared, one was picked and spread on a new plate. As soon as the plate was confluent grown two more plates were spread with each clon. When all three plates reached confluency, spores were harvested. Installation of the β-Lactam ring / warhead
Both β-Lactam synthetases (BLS, wildtype (WT) and double mutant (DM)) were cloned into pSB1C3 (Wt: BBa_K2372002, DM: BBa_K2372003) and pSB1C3-pRha-RBS (Wt: BBa_K2372009, DM in pSB1K3-pRha-RBS: not submitted). The proteins were over-expressed in E.coli and purified using Nickel immobilized metal ion chromatography (Ni-IMAC). The synthesized substrate was supplied together with Mg2+ and ATP to the reaction. The reaction is followed by using LC-MS, expecting a mass shift of 18.
Feeding experiment
A preculture in tryptic-soy-broth (TSB, 50mL) with antibiotics was prepared with spores. After 2-3 days the preculture was used for inoculation of chemically-defined medium (CDM).The preculture was checked for contaminants under the microscope, spun down and resuspended in CDM. The experiments were done in a 24 well bioreactor where each well contained 3 ml of Streptomyces culture. Different substrate concentrations and Streptomyces were put in the reactor and left at 30°C, 200 rpm for 7 days. 1 mL per well was spun down at 17000g and the supernatant was cleaned by centrifugation. This is the aqueous phase sample. A second 1 mL per well was extracted using two times 1 mL ethyl acetate, which was dried down and resuspended in Methanol. This is the organic phase sample. The product formation was analyzed and confirmed by LC-MS.

Figure 2: Workflow of the biochemistry part.

Results

Feeding
Aqueous and organic phase samples were analyzed by LC-MS. The MS chromatograms were searched for expected masses indicated at the left top corner. The black trace indicates the mean ion count (MIC) in positive, the pink trace in negative mode.
In order to examine if the correct culture conditions were used the clorobiocin producing strain Clo-BG1 and the clorobiocin standard were analyzed.
Clorobiocin standard (2.5µg):

Figure 3: Chromatogram at 280 nm of clorobiocin standard.

Figure 4: UV chromatogram at 280 nm of clorobiocin standard.

Figure 5: MS chromatogram of clorobiocin standard.

Figure 6: MS spectrum at minute 13 (positive mode)

Figure 7: MS spectrum at minute 17 (positive mode)

The most abundant ion is highlighted in cyan with a m/z of 391.2. This ion can be found in the organic phase of the non-fed Clo-BG1 strain:

Figure 8: UV chromatogram of CloBG1 at 280nm.

Figure 9: UV chromatogram of CloBG1 at 280nm.

Figure 11: MS spectrum of CloBG1 at 15min.

As expected we did not find m/z signal of 391.2 in the CloQ-defective strain (organic phase, non-fed):

Figure 12: UV chromatogram of SA02 at 280nm.

Figure 13: UV chromatogram of SA02 at 260nm.

Figure 15: positive mode MS spectrum of SA02 at 15min.

Figure 16: negative mode MS spectrum of SA02 at 15min.

The results indicate that the culture conditions were good. More ions fly with higher abundance in positive mode. Ions can be explained by fragmentation in the ionization process:

We find this ion (m/z = 391,1) again in the organic phase of fed samples (red trace):

But not in non-fed samples:

The four peaks (minutes 8.0; 10.5; 11.25 and 12.0) have to be major metabolites [M+H]+ ions, because the respective [M-H]- ions are also found. In the fed sample the first peak is much broader corresponding to the substrate and its methyl-esters:

Figure 21: substrate and methylesters)

The substrate can be detected much easier in aqueous phase samples:

Figure 22: Chromatogram at 280nm of the aqueous phase of fed sample.

Figure 23: MS Chromatogram of the aqueous phase of fed sample

Figure 24: MS spectrum of the aqueous phase of fed sample at minute 5.5.

[M+H]+-ion abundance was above detection limit, indicated by the overlaid red trace. The [methyl-M+H]⁺ ion is shown as the green trace in the MS chromatogram.
We expect our product at m/z = 743.21 ([M+H]⁺) in the organic phase. Looking for it we found the corresponding mass:

Figure 25: Structure of our new antibiotic with the open lactam ring.

Figure 26: MS chromatogram of organic phase of the fed, CouL conjugation.

Figure 27: MS chromatogram of organic phase of the fed, CouL conjugation zoomed in.

Figure 28: MS spectrum of organic phase of the fed, CouL Conjugationat minute 10.88.

Our two genes, β-Lactamase and GyrB, are under control of L-rhamnose and L-arabinose dependent promoters respectively. In this way we reduce leaky expression and can start expression at a defined time point. We used Kanamycin as positive control, Carbenicillin was not suitable because the GyrB expression plasmid encodes a Carbenicillin resistance. As negative controls we used the respective solvents and pSB1C3-SHV (BBa_K2372004) and pSB1C3-pRha-RBS (BBa_K2372007) which do not express protein.

Figure 2: TC: Tetracyclin (XL1blue are resistant, 250µg)), Kan: Kanamycin (500µg).

Kanamycin always shows a z.o.i. between 4 and 5.5 mm. XL1blue should be resistant to Tetracycline. The z.o.i. might be a result of the small solubility in water resulting in a steep concentration gradient instead of a more homogenous, thus lower concentration.

Figure 3: Carb: Carbenicillin (1000µg), Kan: Kanamycin (500µg), SHV control: BBa_K2372004, pRha-RBS-SHV: BBa_K2372008. 0.1% (w/v) L-rhamnose⋅H2O was used for induction.

The z.o.i. is drastically reduced in XL1blue samples but not that much in the TolC samples. TolC might not be a good expression host.

Figure 4: Clorobiocin (10µg), NDX: nalidixic acid (250µg), Kan: Kanamycin (500µg). 2.5µM of L-arabinose was used for induction.

Nalidixic acid (NDX) is an inhibitor of GyrA, while clorobiocin targets the ATP binding site in GyrB. Therefore NDX can be used as a control. TolC are more sensitive to clorobiocin and NDX than XL1blue. The radius of the z.o.i. in the GyrB expressing samples is reduced in TolC samples. Because the initially used L-rhamnose concentration seemed too low we used different concentrations to find out the most suitable.

Figure 5: 1000µg Carbenicillin in each sample.

The size of the z.o.i. does not decrease directly with the increase in L-rhamnose concentration. TolC seem not expressing protein in response to L-rhamnose.

Discussion

Discussion The measurement of the zone of inhibition in the agar-diffusion assay in E.coli with clorobiocin is challenging because E.coli export aminocoumarins via the outer membrane protein TolC. Furthermore clorobiocin has a low solubility in the water-based LB-agar. The first problem is circumvented by using a TolC knockout variant, the second has to be kept in mind while interpreting the data. Tetracycline has a low solubility in water too and led to small z.o.i. even with the resistant XL1 blue, possibly because of its accumulation and therefore high concentration around the wafer. Very small radii of the z.o.i. (<2mm) as seen in the clorobiocin samples and the discrimination between e.g. 4 and 4.5 mm are hard to measure using a conventional ruler. Sometimes the border of the z.o.i. is diffuse or there is more than one z.o.i. (concentric rings). An automated image analysis system might be better or the determination of “mean inhibitory concentrations” in liquid culture. After all the test-system is working. We were able to make E.coli more resilient against clorobiocin by expressing GyrBR. If we had more time an increasing amount of GyrA and GyrBR from Staphylococcus aureus would have been expressed and validated. The SHV-1 confers resistance to carbenicillin in an inducer concentration dependent manner. The next step would be to design a polycistronic gene structure with a constitutive promoter in order to only need one plasmid without the necessity for external induction. References: Anderle, C., Hennig, S., Kammerer, B., Li, S. M., Wessjohann, L., Gust, B., & Heide, L. (2007). Improved mutasynthetic approaches for the production of modified aminocoumarin antibiotics. Chem Biol, 14(8), 955-967. doi:10.1016/j.chembiol.2007.07.014 Bachmann, B. O., Li, R., & Townsend, C. A. (1998). beta-Lactam synthetase: a new biosynthetic enzyme. Proc Natl Acad Sci U S A, 95(16), 9082-9086. Paget, M. S., Chamberlin, L., Atrih, A., Foster, S. J., & Buttner, M. J. (1999). Evidence that the extracytoplasmic function sigma factor sigmaE is required for normal cell wall structure in Streptomyces coelicolor A3(2). J Bacteriol, 181(1), 204-211.