Tools for the discovery of biopolymer producing cysteine relays

Abstract

Cysteine relays, where a protein or small molecule is transferred multiple times via transthiolation, are central to the production of biological polymers. Enzymes that utilise relay mechanisms display broad substrate specificity and are readily engineered to produce new polymers. In this review, I discuss recent advances in the discovery, engineering and biophysical characterisation of cysteine relays. I will focus on eukaryotic ubiquitin (Ub) cascades and prokaryotic polyhydroxyalkanoate (PHA) synthesis. These evolutionarily distinct processes employ similar chemistry and are readily modified for biotechnological applications. Both processes have been studied intensively for decades, yet recent studies suggest we do not fully understand their mechanistic diversity or plasticity. I will discuss the important role that activity-based probes (ABPs) and other chemical tools have had in identifying and delineating Ub cysteine-relays and the potential for ABPs to be applied to PHA synthases. Finally, I will offer a personal perspective on the potential of engineering cysteine-relays for non-native polymer production.

The ubiquitylation cascade is a promiscuous cysteine relay

Transthiolation is the process of exchanging a thiol group from one molecule to another. In a biological context, this usually involves exchanging a molecule between cysteines or between a cysteine and a cysteine-derived molecule (such as Coenzyme A (CoA) or glutathione). Numerous unrelated biological polymers are produced as the end point of cysteine relays involving multiple sequential transthiolation events. Protein ubiquitylation is the most well characterised of these cysteine relays (Fig. 1). In the classical ubiquitylation cascade, Ub is activated by an ATP-dependent Ub activating enzyme (E1). The carboxy terminus of Ub is attached via a thioester bond to the catalytic cysteine of the E1. A ubiquitin-conjugating enzyme (E2) then binds to the E1, where its catalytic cysteine is juxtaposed with that of the E1, facilitating transthiolation (Olsen and Lima 2013). The E2 ubiquitin conjugate (E2~Ub) can then bind one of many (~ 600 in humans) RING (Really Interesting New Gene) type ubiquitin E3 ligases (E3) which allosterically activate the E2~Ub conjugate for Ub discharge to the epsilon-amino group of lysine within substrate proteins (Deshaies and Joazeiro 2009; Plechanovova et al. 2012). This ‘canonical’ pathway with Ub transfer from E1 catalytic cysteine to E2 catalytic cysteine and finally to a lysine residue within a substrate accounts for the majority of ubiquitylation events within the cell (Deshaies and Joazeiro 2009). However, over 35 years ago, it was noted that ubiquitylating enzymes are promiscuous and can utilise ‘non-native’ substrates (Rose and Warms 1983); this feature of the pathway is frequently exploited in structural and mechanistic studies (Burchak et al. 2006; Eddins et al. 2006; Plechanovova et al. 2012) (Table 1). Several Ub-like proteins with dedicated E1s and E2s have been discovered and are reviewed elsewhere (Cappadocia and Lima 2018). Of these Ub-like proteins, the small ubiquitin-like modifier (SUMO) and its associated conjugation and deconjugation machinery have proven to be the most biotechnologically tractable (Hofmann et al. 2020; Lau et al. 2018; Malakhov et al. 2004).

Fig. 1.

The ubiquitylation cascade involves multiple cysteine relays. a Ubiquitin (Ub) E1 catalyses adenylation of the Ub C-terminus. Following adenylation, the Ub C-terminus is attached via a thioester bond to the catalytic cysteine of E1. E1 then transfers Ub to the catalytic cysteine of a Ub-conjugating enzyme (E2). b E2s catalyse lysine aminolysis (transfer of Ub to the epsilon-amino group of lysine), many E2s are allosterically activated by RING domain containing Ub E3 ligases (E3s). c E2s transfer Ub to the catalytic cysteine of HECT and RBR E3s, the HECT or RBR then catalyses lysine aminolysis. d, E2s transfer Ub to the upstream catalytic cysteine of RCR E3s; Ub is subsequently intramolecularly transferred to the downstream cysteine. The downstream cysteine catalyses transfer of Ub to threonine within substrate proteins

Table 1.

The ubiquitylation cascade is a well-characterised cysteine relay. Most enzymes in the cascade are promiscuous and can transfer Ub to ‘non-native’ substrates

Enzyme class	Native activity	Non-native activities observed or engineered in the laboratory	Key references describing non-native activities
E1	Adenylation of the Ub carboxyl terminus. Transfer of Ub to the active site cysteine of E2s.	Transfer of Ub to the active site serine or lysine of E2s. Transfer of Ub to small molecule thiols. Transfer of chemically modified Ub to E2s. Transfer of Ub-like peptides to E2s.	(Burchak et al. 2006; Eddins et al. 2006; El Oualid et al. 2010; Mulder et al. 2016; Plechanovova et al. 2012; Rose and Warms 1983).
E2	Transfer of Ub to the epsilon amino-group of lysine and the amino-terminus of proteins. Transfer of Ub to active site cysteines in HECT, RBR and RCR E3s.	Transfer of Ub to serine, threonine and hydroxylamine. Transfer of Ub-like peptides to the epsilon amino-group of lysine. Transfer of chemically modified Ub to E3s. Transfer of Ub to lysine analogues (L-2,3-diaminopropionic acid, Nα-acetyl-L-ornithine and L-homolysine).	(Hofmann et al. 2020; Liwocha et al. 2020; Mulder et al. 2016; Saha and Deshaies 2008; Wang et al. 2009; Wenzel et al. 2011).
HECT	Transfer of Ub to the epsilon amino group of lysine.	Transfer of Ub to serine and/or threonine. Transfer of Ub to lysine analogues (L-2,3-diaminopropionic acid, Nα-acetyl-L-ornithine and L-homolysine).	(Liwocha et al. 2020; Xu et al. 2018).
RBR	Transfer of Ub to the epsilon amino group of lysine and the amino terminus of proteins.	Transfer of Ub to serine and/or threonine.	(Kelsall et al. 2019).
RCR	Transfer of Ub to threonine.	Transfer of Ub to serine, glycerol and tris(hydroxymethyl)aminomethane (Tris)	(Mabbitt et al. 2020; Pao et al. 2018).
DUB	Cleavage of isopeptide and peptide linked Ub.	Cleavage of Ub linked to serine, threonine and fluorophores.	(Dang et al. 1998; De Cesare et al. 2021).

Apart from RING family E3s, three classes of cysteine-dependent E3 have been identified. Given the promiscuity of E2 binding and transthiolation, it is plausible that many other classes remain to be described (Ahel et al. 2020). These cysteine-dependent E3s form a thioester adduct with Ub (E3~Ub) and subsequently transfer Ub to substrate proteins. The three established classes are HECT (homologous to E6AP) (Scheffner et al. 1995), RBR (RING-Between-RING) (Marin et al. 2004; Wenzel et al. 2011) and RCR (RING-Cys-Relay) E3 ligases (Mabbitt et al. 2020; Pao et al. 2018). The HECT and RBR E3s bind E2~Ub and Ub is transferred to the active site cysteine to produce E3~Ub. Ub transfer is facilitated by cysteine proximity and does not require allosteric activation of the E2 (Dove et al. 2017; Dove et al. 2016; Kamadurai et al. 2009; Lechtenberg et al. 2016; Yuan et al. 2017). In contrast, the RCR E3 ligase provides a degree of allosteric activation to E2~Ub (Mabbitt et al. 2020; Pao et al. 2018). The RCR is unique in having two catalytic cysteines. The upstream cysteine accepts Ub from the E2 and subsequently relays Ub to the downstream cysteine. Finally, the downstream cysteine transfers Ub to threonine residues within protein substrates (Mabbitt et al. 2020; Pao et al. 2018).

A feature of many E2s and E3s is that they catalyse the formation of Ub polymers. Given Ub has 7 lysines, there are many possible linear and branched Ub chains; the physiological roles of these polymers has been reviewed extensively elsewhere (Haakonsen and Rape 2019; Stanley and Virdee 2016; Yau and Rape 2016). A recent report of ester-linked Ub polymers (Ub chains linked at threonine or serine) in vitro suggests that ester-linked Ub chains may also function in cell signalling (Kelsall et al. 2019). If this proves to be the case, serine and threonine would provide 10 potential branch points to Ub chains. Further complicating the Ub chain topology, Ub is cleaved from substrate proteins by one of many (~95 in humans) deubiquitylating enzymes (DUBs). Deubiquitylating enzymes have been categorised into several families; however, mechanistically, they are all either cysteine proteases or zinc proteases (Nijman et al. 2005). Like the other members of the ubiquitylation cascade, DUBs display varying degrees of promiscuous activity (Table 1).

Activity-based probes for the ubiquitylation cascade

Chemical tools which directly detect an enzyme’s active site and subsequently covalently modify the enzyme are frequently referred to as activity-based probes (ABPs). Activity-based probes are a logical extension of dye or fluorophore-linked chemical probes that specifically modify thiols or amines. Compounds like Ellman’s reagent (5,5′-dithio-bis-(2-nitrobenzoic acid) and Sanger’s reagent (1-fluoro-2,4-dinitrobenzene) are useful tools for quantifying thiols and amines, respectively, but they lack the specificity required to provide a qualitative readout of enzyme activity. Activity-based probes typically consist of three parts: warhead, recognition element and handle. The warhead (reactive group) is activated by the enzyme and subsequently covalently modifies the enzyme. The warhead is connected to a recognition element which provides specificity to the ABP. Finally, a handle such as a fluorophore or affinity tag facilitates detection of the ABP-enzyme complex (Heal et al. 2011; Liu et al. 1999; Speers and Cravatt 2004). The principal advantages of ABPs over other enzyme detection systems are as follows: firstly, that they report changes in activity even when expression levels are constant and secondly, that they detect enzyme mechanism irrespective of sequence (Liu et al. 1999). Coupling ABP labelling with mass spectrometry has facilitated the discovery of ubiquitinylating enzymes with modest sequence homology to known members of the pathway (Borodovsky et al. 2002; Pao et al. 2018).

Before discussing ABPs for the ubiquitylation cascade, it is important to distinguish between unimolecular (SN1) and bimolecular (SN2) nucleophilic substitutions. In the first step of SN1 reactions, loss of the leaving group results in a carbocation intermediate. The activation energy of the rate-determining step is inversely proportional to the stability of the carbocation or partial carbocation intermediate. In contrast, in SN2 reactions loss of the leaving group and bond formation are simultaneous. All ubiquitylating enzymes utilise a thiolate (S^-) nucleophile that undergoes concerted (SN2) nucleophilic substitutions with the alkyl halide electrophiles employed in many ABPs. In SN2 reactions, the rate constant is determined by both the concentration of ABP and target enzyme. The effective concentration of the ABP-enzyme complex and hence the rate of reaction can be increased by incorporating a recognition element into the ABP. The use of recognition elements in ABPs is an extension of the organic chemistry principle of template-assembled synthesis where non-covalent interactions are used to promote covalent bond formation.

Like other cysteine proteases, DUBs that employ a cysteine protease mechanism are excellent targets for ABPs (Borodovsky et al. 2002; Faleiro et al. 1997; Greenbaum et al. 2000; Otto and Schirmeister 1997). In the earliest example of a DUB probe, radiolabelled Ub modified with a nitrile warhead served as an ABP for proteasome-associated DUBs (Lam et al. 1997). A later study uncovered a family of DUBs using ABPs with one of seven thiol reactive groups as the warhead (Borodovsky et al. 2002) (Fig. 2). Ideally, the warhead of an ABP will have limited off-target reactivity. This is particularly the case with cysteine reactive warheads, which at worst, are highly non-specific (Muller and Winter 2017). Of the available warheads, terminal alkynes appear to provide the best selectivity for cysteine protease DUBs over free thiols (Ekkebus et al. 2013). Deubiquitylating enzymes differ markedly in their specificity for isopeptide linkages within Ub dimers. Activity-based probes in which the warhead is placed between two Ub molecules (mimicking the isopeptide bond) allow the specificity of cysteine DUBs for Ub linkages to be probed (Li et al. 2014; McGouran et al. 2013) (Fig. 2). Metalloprotease DUBs have received little attention from the ABP community. Recently, an ABP with a zinc chelator (8-mercaptoquinoline) warhead was reported (Hameed et al. 2019). It is not yet clear if this warhead will be sufficiently specific to probe endogenously expressed metalloprotease DUBs.

Fig. 2.

Warheads employed in activity-based probes for deubiquitylating enzymes (DUBs). a Cysteine DUBs cleave the isopeptide bond between Ub and a substrate protein. b The C-terminal glycine of Ub can be replaced by a thiol reactive group or a metal ion chelator (boxed) (Borodovsky et al. 2002; Ekkebus et al. 2013; Hameed et al. 2019; Lam et al. 1997). c Isopeptide-linked Ub dimers (depicted top left) are preferentially cleaved by many DUBs. Warheads can be placed between the Ub C-terminus and any of the 7 lysine positions in a second Ub. Construction of some DUB probes requires replacement of lysine with cysteine or the unnatural amino acids azidohomoalanine (Aha) or diaminobutyric acid (Dab) (Haj-Yahya et al. 2014; Li et al. 2014; McGouran et al. 2013; Mulder et al. 2014)

Activity-based probes for E1s, like cysteine-DUB probes, take advantage of the enzyme’s affinity for Ub and the reactivity of the active site cysteine. E1s adenylate the carboxyl terminus of Ub, transfer Ub to a conserved cysteine and subsequently transfer Ub to E2s. An E1-specific ABP was made by placing an electrophilic warhead (activated vinyl–sulfone) between Ub and an AMP mimic (Lu et al. 2010) (Fig. 3). Using this probe, E1–Ub complexes were isolated for structural studies (Hann et al. 2019; Lu et al. 2010; Olsen et al. 2010).

Fig. 3.

A warhead to capture the tetrahedral E1~Ub thioester intermediate. a E1 catalyses adenylation of Ub (not shown); subsequently. the catalytic cysteine attacks the adenylate resulting in a thioester bond between the Ub C-terminus and the E1. b An AVSN (5′-(vinylsulfonylaminodeoxy)adenosine) warhead attached to the modified C-terminus of Ub was used to trap a covalent E1-Ub-AVSN complex (Hann et al. 2019; Lu et al. 2010)

A second class of E1 ABP took advantage of the affinity of E2s for the E1. In this ABP, an E2 was used as the recognition element and the warhead was attached to the E2’s catalytic cysteine (Stanley et al. 2015). This ABP also reacted specifically with the active site cysteine of a bacterial HECT-like protein and an RBR E3 ligase in vitro (Stanley et al. 2015) (Fig. 4). Reactivity with HECT and RBR E3s was not surprising, given that structural studies suggest that a relatively small number of interactions are sufficient for E2~Ub binding and Ub transfer (Dove et al. 2017; Dove et al. 2016; Kamadurai et al. 2009; Lechtenberg et al. 2016; Yuan et al. 2017). The Virdee group further developed their E1 probe (Stanley et al. 2015), so that it had increased reactivity with cysteine E3s (Pao et al. 2016). In this ABP, an activated vinyl–sulphide warhead is placed between the C-terminus of Ub and the catalytic cysteine of the E2 (Pao et al. 2016) (Fig. 5). Addition of Ub to the ABP increased its affinity/reactivity with RBR E3s and allowed for profiling of an RBR E3’s activity in cell lysates (Pao et al. 2016). This ABP was subsequently used to profile cysteine-dependent E3s in cell lysates; the probe detected ~80 % of known human HECT and RCR E3s as well as E3s whose mechanism had not been fully described (Pao et al. 2016). One of these E3s was a 500-kDa protein with a RING domain at the C-terminus (known as MYCBP2), Pao and colleagues found that the 30-kDA C-terminal domain of MYCBP2 contains two catalytic cysteines. The upstream cysteine receives Ub from an E2 and subsequently relays Ub to the downstream cysteine. The downstream cysteine is positioned in a well-defined active site (similar to a canonical catalytic cysteine, histidine and glutamic acid catalytic triad). This second site is responsible for transfer of Ub to ester substrates (Pao et al. 2016). The MYCBP2 mechanism was dubbed RING-Cys-Relay (RCR). Subsequently, the X-ray crystal structure of the ABP-MYCBP2 complex was solved. This structure mimics the transthiolation intermediate where Ub is transferred from the catalytic cysteine of the E2 to the upstream catalytic cysteine of MYCBP2 (Mabbitt et al. 2020). It is likely that ABPs will continue to be improved and used to survey the catalytic diversity of cysteine-dependent E3s. Mechanistically, unique cysteine E3s, like MYCBP2, may be useful drug targets (Mabbitt et al. 2020).

Fig. 4.

ABP to capture E1-E2 transthiolation intermediate. a Ub is transferred from the active site cysteine of E1 to the active site cysteine of an E2 via transthiolation. b Modification of the E2 active site cysteine with an activated vinylsulfide warhead results in an ABP for E1 to E2 transthiolation (Stanley et al. 2015). c The ABP also selectively modified the active site cysteine of a small cohort of RBR and HECT E3s

Fig. 5.

ABP to capture E2-E3 transthiolation. a Ub is transferred from the active site cysteine of E2 to the active site cysteine of a HECT or RBR via transthiolation. a An ABP for HECT and RBR E3s. Both E2 and Ub (residues 1-73) serve as the recognition element and the activated vinyl sulphide warhead reacts with the catalytic cysteine of HECT and RBR E3s (Pao et al. 2016). The ABP also reacts specifically with E1 and RCR E3s (Mabbitt et al. 2020; Pao et al. 2018)

An elegant ABP which takes advantage of the promiscuous nature of the ubiquitylation cascade was presented by Mulder and colleagues (Mulder et al. 2016) (Fig. 6). They replaced the C-terminal glycine of Ub with dehydroalanine (Dha). Like Ub, Ub-Dha is activated by the E1 and is subsequently relayed to E2s and cysteine-dependent E3s. However, at each step, the activated methylene group of Dha can attack the enzymes catalytic cysteine to form a stable thioether bond (Mulder et al. 2016). In principle, this probe could facilitate the discovery of new E2s and cysteine-dependent E3s. However, at present, no new E2 or E3 has been discovered using Ub-Dha.

Fig. 6.

A cascading probe for the ubiquitin cysteine relay. Substitution of Gly76 with dehydroalanine (Dha) results in a cascading ABP. At each step in the cysteine relay, Ub-Dha can either be transferred or form a stable thioether adduct (Mulder et al. 2016)

Polyhydroxyalkanoate synthesis

Polyhydroxyalkanoates (PHAs) are produced as storage polymers by many bacterial species. A recent review provides a historical overview of the discovery of PHAs (Choi et al. 2020). Polyhydroxyalkanoate granules are closely associated with a group of regulatory, biosynthetic and DNA binding proteins resulting in an organelle-like structure(Jendrossek 2009). In nature, alkanoic acids incorporated into PHAs are derived from fatty acid biosynthesis and β oxidation. At least ten alternate pathways have been engineered to produce PHA precursors in the laboratory (Meng et al. 2014). Irrespective of origin, the alkanoic acid precursors are attached via a thioester bond to CoA. The hydroxyalkanoate–CoA thioester is then polymerised by PhaC (polyhydroxyalkanoate synthase). PhaC are stereospecific for (R)-n-hydroxy fatty acids, with the prototypical substrate being (R)-3-hydroxybutyric acid. Over 100 alternate substrates have been identified; however, it should be noted that these alternate substrates are often incorporated into random copolymers rather than pure homopolymers (Table 2).

Table 2.

Polyhydroxyalkanoate synthases are highly promiscuous and can utilise at least 100 distinct substrates

Enzyme class	Observed activities	Key references describing substrates
PhaC	Polymerisation of hydroxyalkanoates with between 2 and 6 carbons in the backbone of the polymer. The R-pendant (group not in the backbone of the polymer) is of variable length. R-pendants can contain epoxide groups, double bonds and variable terminal moieties. Terminal groups including alkyne, aromatic, aryloxy, cyano, halogen and nitrosyl groups.	(Choi et al. 2016; Curley et al. 1996; Hazer et al. 1996; Kim et al. 1998; Steinbüchel and Valentin 1995; Taguchi et al. 2008)
PhaZ	Cleave of polyhydroxyalkanoates	(de Eugenio et al. 2007; Jendrossek and Handrick 2002)

PhaC have been separated into four classes based on their substrate preference and subunit composition. Class I PhaC preferentially utilise (R)-3-hydroxy fatty acids with between 3 and 5 carbons. Several X-ray crystal structures of the C-terminal catalytic domain of class I PhaC have been solved (Chek et al. 2017; Chek et al. 2020; Kim et al. 2017a; Wittenborn et al. 2016). The catalytic domain adopts an α/β-hydrolase fold with a cysteine, histidine and aspartic acid catalytic triad. A mobile subdomain protrudes from the canonical core of the α/β-hydrolase and appears to contribute to homodimerisation as well as providing a lid to the catalytic machinery (Chek et al. 2017; Chek et al. 2020; Kim et al. 2017a; Wittenborn et al. 2016). The N-terminal domain, missing from the crystallisation construct, is required for PhaC localisation to the PHA granule. Constructs lacking the N-terminal domain have negligible in vivo activity (Kim et al. 2017b).

Two catalytic mechanisms have been proposed for PhaC. The first is a processive model requiring one active site and the second is a pingpong model requiring the sequential action of two active sites. Following publication of the X-ray crystal structure of Class I PhaC, the processive model has found favour. This is due to the distance between the active sites of monomers within the observed PhaC dimers being too large for the pingpong mechanism to take place (Chek et al. 2017; Chek et al. 2020; Kim et al. 2017a; Wittenborn et al. 2016). In the processive model, histidine deprotonates the catalytic cysteine, facilitating nucleophilic attack of the carbonyl carbon of the hydroxyalkanoate–CoA substrate. This results in a covalent hydroxyalkanoate thioester intermediate. A second hydroxyalkanoate–CoA then enters the active site and the terminal hydroxyl group is deprotonated by the aspartate–histidine diad (Kim et al. 2017a; Wittenborn et al. 2016) or directly by aspartate (Chek et al. 2020). The terminal alkoxide then attacks the carbonyl carbon resulting in growth of the PHA polymer (Chek et al. 2017; Chek et al. 2020; Kim et al. 2017a; Wittenborn et al. 2016) (Fig. 7).

Fig. 7.

Reaction scheme for PhaC catalysed 3-hydroxybutyrate polymerisation as proposed by Wittenborn et al. 2016. Histidine deprotonates the catalytic cysteine facilitating nucleophilic attack of the hydroxyalkanoate-CoA (3-hydroxybutyrate-CoA). A second hydroxyalkanoate-CoA binds and the hydroxyl group is deprotonated by histidine. The alkoxide then attacks the hydroxyalkanoate thioester resulting in a non-covalent intermediate. The hydroxyalkanoate chain is then transferred back to the catalytic cysteine. A modified reaction scheme, based on a CoA bound PhaC structure, has also been proposed. In this scheme, the catalytic histidine moves out of the active site following formation of the first covalent intermediate (Chek et al. 2020)

Class II PhaC preferentially polymerise (R)-3-hydroxy fatty acids with between 6 and14 carbons. Based on sequence similarity and structural predictions, Class II PhaCs are likely to adopt an α/β-hydrolase fold and utilise the same catalytic mechanism as their Class I counterparts (Amara and Rehm 2003). Class III PhaC form a heteromer with a second protein, PhaE, and this association is required for in vivo activity (Liebergesell et al. 1992). Similarly, Class IV PhaC form a heteromer with PhaR (McCool and Cannon 2001). The substrate specificity of Class III is similar to Class II, whilst the specificity of Class IV is variable (Liebergesell et al. 1993). Like other members of the α/β-hydrolase superfamily substrate, specificity is likely determined by the size and shape of the substrate binding site. However, in the case of PhaC, affinity for the product (PHA granules) driven by accessory proteins (PhaE or PhaR) or the N-terminal domain may contribute to specificity in vivo.

Polyhydroxyalkanoate depolymerases (PhaZ)

Both intracellular and extracellular serine–hydrolases catalyse the depolymerisation of PHA. Intracellular PhaZ depolymerases mobilise hydroxyalkanoates when cells are subject to carbon limiting conditions. In contrast, extracellular PhaZ scavenge hydroxyalkanoates from the environment (Jendrossek and Handrick 2002). All PhaZ with known or predicted structures are α/β-hydrolases with a canonical serine, histidine and aspartate catalytic triad (Hisano et al. 2006; Jendrossek and Handrick 2002; Jendrossek et al. 2013; Liu et al. 2015). The key differences between depolymerases are the presence or absence of a secondary domain that binds PHA and the activity of the PhaZ with crystalline substrates as opposed to native amorphous substrates (Jendrossek and Handrick 2002).

Probes for CoA-dependent enzymes

Probes that label either the substrate or active site nucleophile in CoA-dependent reactions have been developed (Hwang et al. 2007; Yang et al. 2010; Yu et al. 2006) (Fig. 8). In the first example, chloroacetyl–CoA was used to identify the substrate of an N-acetyl transferase. Transfer of the chloroacetyl group resulted in a chloroacetylated protein that could subsequently be detected using a thiol reactive dye or affinity tag (Yu et al. 2006). This class of probe was improved by replacing the terminal chloride with an alkyne; the alkyne served as a handle for Cu(I)-catalysed azide-alkyne cycloaddition of an affinity tag (Yang et al. 2010). A second class of probe utilises a thiocarbamate sulfoxide warhead that carbamylates the nucleophile of CoA-dependent enzymes (Hwang et al. 2007). The probe designed by Hwang and colleagues was successfully utilised with purified N-acetyl transferases and N-acetyl transferases in cell lysates. However, the specificity of the probe in lysates is not clear.

Fig. 8.

Tools for probing coenzyme A (CoA)-dependent enzymes. a Chloroacetyl-CoA is transferred to substrates by N-acetyl transferases. The chloroacetlyated protein can then be detected using thiol reactive fluorophores or biotin(Yu et al. 2006). b Pentynoyl CoA is transferred to substrates by N-acetyl transferases. The modified protein can then be detected by Cu(I)-catalysed azide-alkyne cycloaddition (CuAAC) of biotin or a fluorophore (Yang et al. 2010). c A sulfocarbamate warhead between CoA and desthiobiotin was employed to label the catalytic cysteine of an N-acetyl transferase (Hwang et al. 2007)

Chemical tools for probing PhaC

Activity-based probes could be used to isolate covalent PhaC-substrate intermediates for structural studies or to identify novel PhaC-like enzymes. For structural studies, ABPs can trap mechanistically informative intermediates. The covalent PhaC–hydroxyalkanoate–CoA transition state is analogous to the covalent E1–E2–Ub transition state. Activity-based probes similar to those employed with the ubiquitylation machinery could be used to trap PhaC–hydroxyalkanoate–CoA complex (Lu et al. 2010; Stanley et al. 2015). This class of ABP could also be used to identify PhaC-like enzymes. The caveat here is that the CoA moiety would provide the majority of the recognition element and the probe would be expected to react with non-target CoA-dependent enzymes.

An outstanding question in PhaC mechanism is what the steric constraints are on the side group (R-pendant) of the hydroxyalkanoate and how these constraints differ between PhaC classes. A substrate mimic that alkylates the active site cysteine of PhaC (i.e., a PhaC inhibitor) could be used to isolate covalent PhaC hydroxyalkanoate complexes for structural studies. Similar probes have been instrumental in elucidating the mechanisms of cysteine proteases (Otto and Schirmeister 1997).

Activity-based probes for PhaZ

Probes with a flurophosphonate warhead have been successfully employed with recombinant PhaZ (Huang et al. 2012). In order to improve the specificity of this class of broad-spectrum serine–hydrolase probe, a hydroxyalkanoate recognition element could be incorporated between the warhead and handle; many analogous probes with short-peptide recognition elements have been developed (Faucher et al. 2020; Liu et al. 1999).

Engineering cysteine relays

Given the promiscuity of cysteine relays, it would appear a simple task to engineer novel polymers. In many instances, the difficulty is not finding a promiscuous activity, rather it is that the yield of product is sufficient for biochemical assays but insufficient for large-scale production. For example, only trace amounts of Ub linked by an ester bond to serine could be produced using the RCR E3 ligase, whereas large amounts of threonine linked-Ub can be produced under the same reaction conditions (De Cesare et al. 2021). Similarly, only small amounts of ester-linked Ub dimers can be produced in vitro using the RBR E3 ligase HOIL-1 (Kelsall et al. 2019).There is potential to improve both E3s by computational design or laboratory-directed evolution; however, this would be a major undertaking given the lack of a high-throughput assay and the complexity of the RBR and RCR catalytic mechanisms. It is more likely that better catalysts will be obtained by screening the thousands of naturally occurring E2s and E3s using conventional enzyme assays or ABPs.

The recent PhaC structures should allow for computational design of PhaC variants with altered substrate specificity. A hydroxyalkanoate bound structure and a structure of a Class II PhaC would further facilitate the design process. In the absence of a strong selective pressure or a high-throughput assay, laboratory-directed evolution of PhaC would be difficult but not impossible (Kichise et al. 2002). It is possible that PhaC-like proteins that lack sequence homology with the established classes exist in nature and that these could be identified using ABPs. Similar studies have uncovered new serine hydrolases (Lentz et al. 2018; Ortega et al. 2016). Ideally, multiple microbial species would be profiled simultaneously. The challenge here is that identification of labelled proteins by mass spectrometry relies on database searching and a robust method to search proteomes of multiple related species would need to be implemented (Cheng et al. 2018; Schiebenhoefer et al. 2019). Whilst it has been demonstrated that microbial communities can be probe labelled, separated and subsequently identified by 16S rRNA sequencing (Whidbey et al. 2019), to my knowledge, mass spectrometry-based ABP-profiling has not been achieved with microbial communities.

Conclusions

When engineering a cysteine relay to use a novel substrate, we choose the starting point; either we improve the latent promiscuous activity of well a characterised ligase or screen for a new ligase. Activity-based probe profiling widens the screening net to include mechanistically similar but not necessarily evolutionarily related enzymes. It is likely that ABP profiling will continue to uncover ligases with unusual substrate specificity and that these enzymes will be the equal or better of those engineered in the laboratory.

Funding

I am grateful for the financial support provided by Ministry of Business, Innovation and Employment New Zealand via Scion’s Strategic Science Investment Fund contract C04X1703.

Data Availability

All underlying data are available from the author upon request.

Code availability

Not applicable.

Declarations

Ethics approval

Not applicable.

Consent to participate

Not applicable.

Consent for publication

Not applicable.

Conflicts of interest

The author declares no competing interests.