Introducing noncanonical amino acids for studying and engineering bacterial microcompartments

Abstract

Bacterial microcompartments (BMCs) with selectively permeable shells and encapsulated enzyme cores are well-suited candidates for nano-bioreactors because of their advantages of enhancing pathway flux and protection against toxic products. To better study and engineer protein-based BMCs, a series of protein chemistry approaches are adopted. As one of the most advanced techniques, genetic code expansion can introduce various noncanonical amino acids (ncAAs) with diverse functional groups into target proteins, thus providing powerful tools for protein studies and engineering. This review summarizes and proposes useful tools based on current development of the genetic code expansion technique towards challenges in BMC studies and engineering.

Introduction

Functioning analogously to eukaryotic organelles, bacterial microcompartments (BMCs) have been found in more than 20% of bacteria taxa [1–3]. BMCs consist of a thin layer of protein shell with sizes between 100 and 200 nm and an encapsulated enzyme core related to different metabolic pathways [4]. Carboxysomes (anabolic) and metabolosomes (catabolic) are the two basic types of BMCs [1]. The anabolic carboxysomes are the most well-studied BMCs due to wide distribution in cyanobacteria and importance in carbon fixation [5]. Differently, metabolosomes are more diverse and have been found to participate in the catabolism of various compounds such as 1,2-propandiol [6], ethanolamine [7], choline [8], and fucose [9]. Evolution has designed BMCs to isolate biochemical reactions with the purposes of increasing local concentrations of enzymes and substrates, enhancing pathway flux, confining volatile intermediates, and protecting cells from toxic products, thus making BMCs well-suited candidates for nano-bioreactor engineering [10]. Additionally, the short N-terminal targeting sequences of encapsulated enzymes are able to introduce heterologous proteins into BMCs [11,12], offering the possibility to encapsulate non-native proteins into BMCs for diverse purposes [13,14]. Furthermore, more efforts have been made to optimize BMC functions, such as modifying the selective permeability of BMCs by engineering shell proteins [15–17]. Indeed, as BMCs are exclusively composed of proteins, approaches for protein engineering will be key options for BMC studies and engineering.

One classic approach in protein engineering is to substitute amino acid residues at designed positions, thus modulating protein structures, functions, and properties for desired purposes. However, traditional methods only have a limited candidate pool of 20 canonical amino acids. To overcome this limitation and introduce more diversified non-canonical amino acids (ncAAs) into proteins, the genetic code expansion technique has been developed [18,19]. In this approach, an orthogonal pair of aminoacyl-tRNA synthetase (AARS) and tRNA (no crosstalk with endogenous AARSs and tRNAs in host cells) is used to incorporate an ncAA in response to an assigned codon (usually a stop codon) at the desired position in the target protein [20] (Figure 1). As the key part of genetic code expansion, a series of orthogonal AARS/tRNA pairs from archaeal or eukaryotic cells have been developed for bacterial expression systems [21]. For using Escherichia coli as host cells, the tyrosyl-tRNA synthetase (TyrRS) system from Methanocaldococcus jannaschii and the pyrrolysyl-tRNA synthetase (PylRS) system from Methanosarcinaceae are widely used [20,22]. In the last decade, a series of orthogonal translation systems (OTSs) have been evolved to introduce more than 300 ncAAs into proteins in living cells for different purposes, such as labeling proteins [23,24], studying protein posttranslational modifications [25,26], mapping protein-protein interactions [27,28], and modulating enzyme functions [29,30]. Such numbers of ncAAs harboring different functional groups with variable sizes, charges, and properties provide a much larger pool of tunable candidates than traditional approaches. Here we summarize and propose applications of genetic code expansion in BMC studies and engineering.

Figure 1.

The scheme of the genetic code expansion technique. The engineered aminoacyl-tRNA synthetase (AARS) charges its cognate tRNA with a specific noncanonical amino acid (ncAA). The ncAA-charged tRNA contains a specific anti-codon (CUA for example) to read the corresponding codon (usually a stop codon) in mRNA and, the ncAA is incorporated into protein during translation at a desired site.

Applications of genetic code expansion in BMC studies and engineering

In our recent studies, we were able to genetically incorporate ncAAs into Pdu BMCs in living Salmonella cells [31]. Developing specific OTSs is the prerequisite for applying the genetic code expansion technique in Salmonella cells. By changing promoters and optimizing gene sequences of introduced orthogonal pairs of AARSs and tRNAs as well as modifying Salmonella genes for controllable protein expression, three OTSs originally developed in E. coli (the TyrRS system, the PylRS system, and the phosphoseryl-tRNA synthetase system from methanogenic archaea) were successfully introduced into Salmonella [31]. With these customized OTSs specific for Salmonella cells, different tools based on ncAAs were developed for BMC studies. We further propose several potential applications of genetic code expansion in BMC engineering (Figure 2).

Figure 2.

Applications of genetic code expansion in BMC studies and engineering. P-azido-phenylalanine (pAzF) is used for site-specific labeling for imaging. Benzoyl-phenylalanine (Bpa) is a photocrosslinking amino acid to map protein-protein interactions. Naphthylalanine (Npa) provides more hydrophobicity, possibly increasing interactions between shell proteins and stabilizing BMCs. 2,3-diaminopropanoic acid (DAP) has positive charges in physiology conditions, thus modulating pore properties to facilitate uptake of specific molecules.

1. Labeling BMCs in living cells

Incorporation of ncAAs with specific click chemistry functionality into proteins enables unique biorthogonal chemistries which are inert to natural biological reactions, allowing protein labeling for different purposes such as imaging and proteomic studies [32]. Azide-alkyne cycloaddition is one of the most popular click chemistries with advantages of fast kinetics, low toxicity, and easy operation. The azido group of p-azido-phenylalanine (pAzF) can react with an alkyne-fluorescent dye by the click reaction and produce fluorescence signals for detections [33]. Moreover, due to the small size of pAzF, similar to canonical amino acids, it can be site-specifically incorporated into target proteins to replace selected residues, thus minimally affecting protein structures [34]. In our recent study, PduA, one of the major shell proteins of Salmonella Pdu BMCs [35,36], has been selected for the pAzF-derived labeling. Based on the crystal structure of PduA [37], N67 was chosen as the target position. After site-specifically mutating the original codon for N67 in pduA gene to an amber stop codon (UAG), the pAzF-containing PduA protein was expressed together with other Pdu proteins [31]. By the click reaction with a fluorescent dye harboring an alkyne group, Pdu BMCs were labeled for fluorescence imaging in living Salmonella cells.

2. Mapping BMC protein interactions by photocrosslinking

Protein-protein interactions between shell proteins or between shell proteins and encased enzymes are crucial for BMC assembly [38,39]. However, some of the interactions could be weak or transient due to different factors such as pH and ionic conditions. Thus, it could be challenging to map the protein-protein interaction network of BMC proteins completely. NcAAs with photocrosslinkable moieties can be easily irradiated with UV light to form covalent bonds between proximal molecules to detect transient and weak interactions, which have been widely used to study protein-protein interactions both in vitro and in vivo [40]. Benzoyl-phenylalanine (Bpa) is one of the most favorable photocrosslinking amino acids. It can be genetically incorporated into proteins in Salmonella cells by the optimized TyrRS systems [31]. Because such incorporation is site-specific by the genetic code expansion technique, this approach has the advantage over the classic cysteine-derived crosslinking which has concerns of multiple cysteine residues in a single protein. On the other hand, it is necessary to select proper Bpa incorporation sites before applying Bpa-derived photocrosslinking. Ideal positions should be at the outer surface of the target protein and avoid interface between its own subunits. It is achievable because the structures of most architectural subtypes (i.e. hexamers, trimers, and pentamers) have been solved, enabling homology modeling of most BMC shell proteins [41].

3. Stabilizing BMC shells

Studies have shown that environmental factors such as pH and ionic conditions could affect the stability and variability of BMC shell protein self-assembly [42]. The approach based on canonical amino acid substitutions (K28A and R78A mutations in BMC-H proteins) has been used to feasibly manipulate the self-assembly and dynamics of BMC shell proteins [43]. As ideal nano-bioreactors, it is important to stabilize BMC shells against extreme pH, high temperature, or high ionic strength for different applications. Due to diversified side chains, ncAAs could provide more intermolecular or intramolecular interactions between shell proteins for higher stabilities. Based on properties of key residues affecting shell stabilization, genetic code expansion could introduce corresponding ncAAs to increase BMC shells. For key residues with hydrophobic side chains such as valine, isoleucine and phenylalanine, ncAA analogs with more hydrophobicity could be introduced to provide more hydrophobic interactions. For key residues with longer side chains like glutamine and lysine which could form hydrogen bonds with protein backbones, ncAAs with additional amino groups or hydroxyl groups could offer extra hydrogen bonds. According to these criteria, several potential ncAA analogs are listed (Figure 3). Specific ncAA incorporation systems need to be evolved for those ncAAs, individually. The PylRS system will be a good choice as it has a high flexibility of substrate specificities from lysine derivatives to phenylalanine analogs [44].

Figure 3.

The structures of canonical amino acids and their corresponding ncAA analogs. Canonical amino acids are underlined.

4. Modulating Pdu BMC pore properties for metabolite selectivity

One unique property of BMCs as ideal nano-bioreactors is their selective pores for specific transport of substrates, products, and cofactors. Crystal structures of BMC shell proteins reveal that pores formed by different shell proteins diverge in steric/electrostatic properties, and BMCs could control metabolite movement based on various pore properties [45]. Thus, shell permeability can be tuned by modifying residues that surround the pores [15–17]. Recently, approaches of canonical amino acid substitutions have been used to modify key residues that affect the properties of shell pores. For instance, site-specific replacement of S40 of PduA could change the size and polarity of the opening pore. The substitution with histidine enables Pdu BMCs to have better uptake of glycerol than the native substrate 1,2-propandiol [17]. For another Pdu protein, PduT, different amino acid substitutions at its C38 were demonstrated with different influences on the shell permeability. PduT-C38S and PduT-C38A variants could increase the shell permeability for 1,2-propanediol, propionaldehyde, NAD⁺ and NADH, while PduT-C38I and PduT-C38W could not alter the permeability of Pdu BMCs [15]. Shell proteins have also been engineered for electron transport. By mutating S55 to cysteine, BMC-T1 was shown to bind a [4Fe-4S] cluster [46]. Addition of three histidine residues around the pore region of BMC-T1 generated a new binding site for copper [47]. Again, the larger pool of ncAAs could provide more candidates than the limited set of 20 canonical amino acids to modulate pore properties for desired reactions encapsulated in BMCs. Previously, pyruvate decarboxylase (PDC) and alcohol dehydrogenase (ADH) have been encapsulated into an ethanol reactor based on Pdu BMCs, however, the ethanol production was not significantly improved compared with non-compartmentalized reactions [48]. One possible reason could be that BMC pores have no optimal permeability to the substrate pyruvate. To increase the uptake of negatively charged pyruvate, 2,3-diaminopropanoic acid (DAP, structure in Figure 2) with one positive charge could be a potential substitution for Ser40 of PduA, and the genetic incorporation system for DAP has been established recently [49].

Conclusions and perspectives

The genetic code expansion strategy can offer various potential tools to study and engineer BMCs by incorporating a variety of ncAAs with expanded functional groups into BMC proteins. Our previous study has successfully labeled the PduA protein in Pdu BMCs with pAzF [31], and we also proposed several potential applications. Particularly, stabilizing and modifying pore selective permeability of BMCs are important for BMC engineering. Therefore, more future studies need to be conducted to develop these ncAA-based systems for stabilization and selective permeability of BMCs in experimental conditions. Additionally, more tools based on ncAA incorporations will be explored to facilitate BMC engineering due to the rapid development of genetic code expansion and the gradual enlargement of ncAAs library.

Highlights.

• Genetic incorporation of noncanonical amino acids (ncAAs) can facilitate BMC studies and engineering.

• NcAAs with specific click chemistry functionality can be used to label BMC proteins by bioorthogonal reactions.

• NcAAs with photocrosslinking properties can help study the protein-protein interactions between BMC proteins.

• NcAAs with varied functional groups can be used to stabilize BMCs and modulate selective permeability of BMC shells.