Motivation:

Targeted activation of enzymes has broad applicability in industry and medicine. The fastest and most convenient way to get an activation signal to its destination is light waves. This is why light activation of enzymes is a relevant matter, but up to now no easy and widely applicable solution exists.
The method of photocaging is conceptually simple and in theory applicable for every protein. So far though it has only been used in intracellular studies for investigation of enzyme function [1-3]. But it has great potential to be used in enzyme production as well.
The purpose of this project is to demonstrate the versatility of photocaging through the following example.

In 2014, 595 000 tons of washing detergents were consumed solely within Germany [4]. Nowadays these contain a variety of enzymes to specifically and efficiently target common stains, which helps to significantly reduce the amount of surfactants needed.
0.5-1% of the washing detergents consist of proteases, whose long-time storage is especially problematic due to self-degradation. Later they even degrade other useful enzymes, in the final mixture. Thus they are inhibited by adding an equal percentage of boric acid. This chemical has been classified as a “substance of very high concern” by the European Chemicals Association as studies point towards its reproduction toxicity and teratogenicity [5]. Consequently, there have been many attempts to replace it, however none of the results have been as efficient as boric acid or had severe disadvantages regarding safety.
Boric acid actually naturally occurs in drinking water and the environment, but the usage of washing detergents makes the amount exceed the natural occurrence in waters. [6] It is obtained from mined, boron containing, minerals. Because of the continuously high amount of washing detergents consumed, avoiding usage of boric acid would be preferable to stop its influx into waters and render the handling of tons of chemicals for enzyme stabilization unnecessary.

Theory:

Photocaging allows expressing a protease, inhibited by only one covalently bound molecule. Directly prior to the washing process, it could be activated with light.
The key of this method is attaching a chemical protection group ta normally reactive molecule which is thereby shieled from reacting, which can be cleaved of when irradiated with the respective spectrum of light. By replacing an amino acid which is crucial to the function of the enzyme with the corresponding photocaged amino acid, a protected enzyme would be generated. Exposure to the right spectrum of light would lead to cleavage of the protection group and yield the original amino acid and therefore a functioning enzyme.
The introduction of this unnatural amino acid can be achieved via genetic code expansion. The host organism is provided with an orthogonal tRNA/synthetase pair that is capable of incorporating the unnatural amino acid of choice at a UAG codon. The UAG or “amber” codon is the least used stop codon in most organisms which makes it perfect for being reassigned. [1]
More and more orthogonal tRNA/synthetase pairs are being created, but researchers are limited to available ones, if time does not allow the development of the synthetase that meets one’s special needs. As this is the case for our project, the following strategies to achieve reversible inactivation have been developed.

Subtilisin is a protease from Bacillus subtilis whose derivatives are being used in many washing detergents, which makes it suitable for a proof of principle.
As a serine protease it catalyzes the cleavage of a peptide bond by means of a catalytic triad consisting of serine²²¹, aspartic acid³² and histidine⁶⁴. [7]

Previous studies proved, that replacing the serine with an alanine, which equals the loss of the reactive group, reduces the activity by factors of 2∙10⁶. As the protection group would shield the hydroxyl group the inhibition should be at least equally effective.
An orthogonal tRNA/synthetase pair which is able to introduce the photo-liable serine derivative O-(4,5-dimethoxy-2-nitrobenzyl)-L-serine (DMNB-serine) at an amber stop codon, exists in Saccharomyces cerevisiae. It was developed and successfully used by P. G. Schulz [2].
DMNB-serine is converted to serine by irradiation with low energy blue light. The first challenge is the production of subtilisin E in S. cerevisiae, for this an expression system has to be built. We chose to try three different approaches to compare efficiency of production: A plasmid with a constitutive promotor, one with an inducible and an inducible genome integration vector. Then by providing the yeast cells with the tRNA/synthetase pair for DMNB-serine and the respective unnatural amino acid can be incorporated.
Mutation of serine²²¹ to an amber stop codon in the subtilisin gene should lead to production of a photocaged and consequently inactive variant.
Also the naturally occurring secretion tag from B. subtilis, that allows relocation of the enzyme to the medium shall be exchanged with one working in S. cerevisiae, to make harvesting more convenient.
As previous efforts to express a protease from B. subtilis in S. cerevisiae failed due to hyperglycosylation, [8] which resulted in an inactive enzyme, we also pursued a second strategy. Thus the idea to develop a DMNBS-synthetase that works in Escherichia coli evolved.

Escherichia coli is widely used in synthetic biology. It offers the advantage of being a comparatively simple and well-understood model organism while being easy to handle in the laboratory environment. Also, an expansion of the genetic code has already been successfully implemented in E. coli multiple times [9-12] by introducing an orthogonal tRNA/synthetase pair.
Therefore, working in E. coli is an obvious choice.

Due to a limited range of tRNA/synthetase pairs for non-canonical amino acids in general and especially for those that act orthogonally in E. coli, photocaging serine in the active site of subtilisin E with DMNB-serine is currently not possible. Hence, another strategy is needed to produce temporarily inactive proteases. This part of the project focuses on utilizing the maturation process of subtilisin E.

Subtilisin E is an alkaline serine protease found in Bacillus subtilis that has to autoprocess itself to become functional. At first, the enzyme exists as a precursor, namely the pre-pro-subtilisin. The pre-sequence serves as a recognition sequence for secretion across the cytoplasmic membrane and is cleaved off in the course of the process. The pro-peptide acts as an intramolecular chaperone (IMC) and facilitates the folding of the protease. Folding is essential for the activity of an enzyme. Still, the maturation process of Subtilisin E is not completed, as the pro-peptide covers the substrate binding site and inhibits activity. However, enough proteolytic activity is achieved to autoprocess the IMC-domain and therefore cleave off the pro-peptide. Yet, the C-terminal end of the pro-peptide continues to block the substrate binding site. After the degradation of the pro-peptide, the substrate-binding site is cleared and the protease becomes effectively active. [13]

This mechanism can be used to implement a novel inactivation method.

O-(2-Nitrobenzyl)-L-tyrosine (ONBY) is a derivate of the canonical amino acid tyrosine. It carries a photo-labile protection group that can be cleaved off by irradiation with UV-light (365nm, [14]).
	By adding ONBY to the genetic code of E. coli and incorporating said amino acid in the pro-peptide cleavage-site of subtilisin E the maturation process is disturbed. Due to its size ONBY sterically hinders the protease [15]. The pro-peptide cannot be cleaved from the enzyme and subtilisin E is not able to achieve its full proteolytic activity. A temporarily inactive protease is produced.

After removal of the protection group the maturation process can be completed and subtilisin E acquires its full proteolytic activity.

Introduction

Photocleavable non-canonical amino acids offer the opportunity to control protein function on a non-invasive basis. Working with unnatural amino acids requires an additional, orthogonal pair of a tRNA and a cognate synthetase i.e. which does not crossreact with the endogenous tRNA/synthetase pairs [2].
A previously reported tRNA/synthetase pair for O-(4,5-dimethoxy-2-nitrobenzyl)-L-serine (DMNBS) which derived from Escherichia coli and was used in Saccharomyces cerevisiae [2] leads to the lack of a possibility to work with non-canonical amino acids replacing serine in E. coli by using a 21^st amino acid.

Based on computational modeling a synthetase for DMNBS in E. coli is designed, mainly because of three reasons:

1. Serine is a crucial amino acid for enzyme catalysis and therefore an obvious target for pathway analysis and control of protein activity via photocaging [2]
2. E. coli is the most used organism in the field of biotechnology.
3. DMNBS has been proven to work very well as a non-canonical amino acid in yeast and is superior to oNBY regarding the physico-chemical properties, e.g. a higher quantum yield for photocleavage and an absorption wavelength apart from the UV spectrum. [2]

For that purpose, the well described tyrosyl-tRNA/synthetase pair of Methanococcus janaschii is mutated to selectively incorporate DMNBS into proteins in E. coli.

The cognate tRNA's anticodon contains an amber stop anticodon. Hence, it is possible to incorporate an amino acid at a chosen position in a protein via amber codon suppression

Designing of DMNBS-RS

[picture tyrosine in binding pocket of wild type tyrosyl synthetase]

As the basis for designing a DMNB-S specific synthetase the wild type tyrosyl-synthetase of M. janaschii (Mj-Tyr-RS) is chosen due to its well characterized properties, its resemblance to the E. coli-leucyl-synthetase (Ec-Leu-RS) derived DMNBS-synthetase used in yeast and its availability through the iGEM-Team Austin, Texas, 2014.

It was also previously shown, that this pair is orthogonal to endogenous E. coli synthetases [16].

The determination of the alterations to be made for changing the amino acid specificity are based on the comparison of the crystal structures of Mj-tyrosyl- and Ec-Leu-tRNA/synthetase pairs. Highly conserved amino acids Y³², L⁶⁵ and D¹⁵⁸ are known to build H-bonds with the natural ligand tyrosine [17]. Furthermore, these sites resemble the mutated sites within E. coli-leucyl-synthetase which was already changed to a DMNBS-specificity and is successfully used in yeast to incorporate DMNBS [16,18]. Some more positions are chosen to be mutated due to their location within the 2Å-radius of bound DMNBS.

[picture DMNBS in binding pocket of designed DMNBS-tRNA/Synthetase]
Figure 1: Visualization of DMNBS binding to designed DMNBS-tRNA/synthetase

For the purpose of creating a mutation library, a semi rational method is applied by using partially randomized codons for site saturated mutagenesis, also with respect to the codon usage of E. coli.

Characterization of synthetase plasmids

[picture of YRS or DMNBSRS plasmid]
Figure 2: Tyrosyl-tRNA/synthetase pair or DMNBS-tRNA/synthetase pair plasmid

Via the synthetase plasmid the tRNA is encoded as well as the cognate synthetase, each with its own promotor. The tRNA's anticodon is changed to an amber-anticodon. A gentamycin resistance reporter is used for selection.

Characterization of screening plasmids

[Picture of pRXG]
Figure 3: Reporter plasmid pRXG

As the screening method for measuring the fidelity and incorporation efficiency of the created tRNA/synthetase pairs, a fluorescence-based method as established by the the iGEM-Team Austin, Texas, 2014 is used: a two-plasmid system, where the first plasmid (pRXG) contains an IPTG-inducible RFP-sfGFP reporter and the second plasmid the tRNA/synthetase pair.
The pRXG plasmid consists of a RFP domain which is connected through a linker sequence containing a recoded amber stop codon with a sfGFP domain. The expression of the plasmid gives either a red fluorescence, or - if the ncAA will be incorporated at the amber stop codon within the linker site - both a red and a green fluorescence.
By comparison of fluorescent levels it is possible to determine incorporation efficiency of the generated library.
A Kanamycin resistance reporter is used for selection of the reporter plasmid.
The reporter plasmid, as described above, has been created by iGEM Team Austin, Texas, 2014. A BioBrick has been built from it and characterized for further use: by iGEM-Team Austin, Texas, 2014

Screening

[Picture of pRXG]

For both, positive and negative screening, M9 minimal medium is used, because it has a lower optical density itself and contains buffer to stabilize the fluorescence proteins.
As a pretest, the excitation and emission wavelengths of the fluorescence proteins are measured via online screening with a BioLector, along with the OD. Thus the wavelengths for the screening are identified.
[Picture wavelengths]

To transform the mutation library into E. coli, the following strategy is applied: The reporter plasmid is transformed into BL21 DE3 gold from which subsequently chemical competent cells are made. The library then is transformed into those cells and onto M9 minimal solid medium with gentamycin and kanamycin for selection with a time of growth of 48 hours at 30°C.
A preculture is inoculated each well with one colony in 96-well microtiter plates in M9 minimal medium. Microtiter plates are sealed with microporous tape sheets, because fluorescence proteins mature under the influence of oxygen. Incubation at 30°C, 900rpm for 48 hours.

As an intermediate step the precultures are replicated onto a next microtiter plate for the purpose of homogenizing the cell density.

Thereof two separate screening microtiter plates are inoculated: Firstly, a positive screening plate with M9 minimal medium, addition of 2mM DMNBS as well as 100µM IPTG for expression. Secondly a negative screening plate without DMNBS. All microtiter plates are measured after the same amount of time. Incubation at 900rpm, 30°C, pH 7.6.

OD600 and both, red and green fluorescence levels are measured as an endpoint measurement. Thus the most efficient clones can be determined to selectively incorporate DMNB-S at an amber stop codon.

To obtain final results, the new-found, working DMNBS tRNA/synthetase pair clones undergo an online measurement rescreening with each nine replicates and the previously mentioned growth and screening conditions.

Results can be found here

[1]D. P. Nguyen, M. Mahesh, S. J. Elsässer, S. M. Hancock, C. Uttamapinant, and J. W. Chin, “Genetic Encoding of Photocaged Cysteine Allows Photoactivation of TEV Protease in Live Mammalian Cells,” J. Am. Chem. Soc., vol. 136, no. 6, pp. 2240–2243, Feb. 2014. [2]D. P. Nguyen, M. Mahesh, S. J. Elsässer, S. M. Hancock, C. Uttamapinant, and J. W. Chin, “Genetic Encoding of Photocaged Cysteine Allows Photoactivation of TEV Protease in Live Mammalian Cells,” J. Am. Chem. Soc., vol. 136, no. 6, pp. 2240–2243, Feb. 2014. [3]A. S. Baker and A. Deiters, “Optical Control of Protein Function through Unnatural Amino Acid Mutagenesis and Other Optogenetic Approaches,” ACS Chem. Biol., vol. 9, no. 7, pp. 1398–1407, Jul. 2014 [4]Industrieverband Körperpflege- und Waschmittel e. V., “Bericht Nachhaltigkeit in der Wasch-, Pflege- und Reinigungsmittelbranche in Deutschland 2013–2014”, 2014. [Online]. Available: http://www.ikw.org. [Accessed: 07-Oct-2016]. [5]European Chemicals Agency, “Member State Commitee Draft Support Document for Identification of Boric Acid as a Substance of Very High Concern Because of Its CMR Properties”, 2011. [Online]. Available: https://echa.europa.eu. [Accessed: 16-Oct-2016]. [6]Landesanstalt für Umwelt, Messungen und Naturschutz Baden-Württemberg, „Bor - Ableitung einer Geringfügigkeitsschwelle zur Beurteilung von Grundwasserverunreinigungen“, 2002. [Online]. Available: http://www.fachdokumente.lubw.baden-wuerttemberg.de. [Accessed 07-Oct-2016]. [7]P. Carter and J. A. Wells, “Dissecting the catalytic triad of a serine protease,” Nature, vol. 332, no. 6164, pp. 564–568, Apr. 1988. [8]L. F. Wang and R. J. Devenish, “Expression of Bacillus subtilis neutral protease gene (nprE) in Saccharomyces cerevisiae,” J. Gen. Microbiol., vol. 139, no. 2, pp. 343–347, Feb. 1993. [9]J. W. Chin, A. B. Martin, D. S. King, L. Wang, and P. G. Schultz, “Addition of a photocrosslinking amino acid to the genetic code of Escherichia coli,” Proc. Natl. Acad. Sci., vol. 99, no. 17, pp. 11020–11024, 2002. [10]J. W. Chin, “Reprogramming the genetic code,” EMBO J., vol. 30, no. 12, pp. 2312–2324, 2011. [11]R. A. Mehl et al., “Generation of a Bacterium with a 21 Amino Acid Genetic Code,” J. Am. Chem. Soc., vol. 125, no. 4, pp. 935–939, Jan. 2003. [12]L. Wang and P. G. Schultz, “Expanding the Genetic Code,” Angew. Chem. Int. Ed., vol. 44, no. 1, pp. 34–66, Jan. 2005. [13]X. Fu, M. Inouye, and U. Shinde, “Folding Pathway Mediated by an Intramolecular Chaperone,” J. Biol. Chem., vol. 275, no. 22, pp. 16871–16878, 2000. [14]M. S. Kim and S. L. Diamond, “Photocleavage of o-nitrobenzyl ether derivatives for rapid biomedical release applications,” Bioorg. Med. Chem. Lett., vol. 16, no. 15, pp. 4007–4010, Aug. 2006. [15]C. Chou, D. D. Young, and A. Deiters, “A Light-Activated DNA Polymerase,” Angew. Chem. Int. Ed., vol. 48, no. 32, pp. 5950–5953, Jul. 2009. [16]A. Deiters, D. Groff, Y. Ryu, J. Xie, and P. G. Schultz, “A Genetically Encoded Photocaged Tyrosine,” Angew. Chem. Int. Ed., vol. 45, no. 17, pp. 2728–2731, Apr. 2006. [17]A. Pavelka, E. Chovancova, and J. Damborsky, “HotSpot Wizard: a web server for identification of hot spots in protein engineering,” Nucleic Acids Res., vol. 37, no. Web Server issue, pp. W376-383, Jul. 2009. [18]N. Wu, A. Deiters, T. A. Cropp, D. King, and P. G. Schultz, “A Genetically Encoded Photocaged Amino Acid,” J. Am. Chem. Soc., vol. 126, no. 44, pp. 14306–14307, Nov. 2004.