Phenylalanine ammonia-lyase (PAL, EC 4.3.1.26) serves as the pivotal enzyme in the phenylalanine pathway within plants, primarily responsible for catalyzing the hydrolysis of L-phenylalanine (L-Phe) to generate trans-cinnamic acid[1]. PAL plays a crucial role in plant defense and bridges secondary and primary metabolism[2]. Additionally, AvPAL*, found in Anabaena variabilis, is employed for treating phenylketonuria and uremia[3]. PAL is predominantly present in plants and microorganisms such as Camellia sinensis, Petroselinum crispum, Bambusa oldhamii, Rhodotorula glutinis and Photorhabdus luminescens [4], [5], [6], [7], [8]. In monocotyledonous plants like Zea mays and Sorghum bicolor, PAL exhibits tyrosine ammonia-lyase (TAL) activity alongside its function of converting L-tyrosine (L-Tyr) into p-coumaric acid[9].

As the principal product resulting from PAL activity, trans-cinnamic acid possesses notable antioxidant and antimicrobial properties highly sought after by the food and pharmaceutical industries[10]. On the other hand, p-coumaric acid holds significant economic value as a chemical constituent involved in flavonoid production along with polyphenols, coumarins, stilbenes, lignans, and polymeric materials while also exhibiting protective effects against cancer, atherosclerosis, oxidative cardiac damage, neuronal damage, and anti-inflammatory activity[11], [12].

Currently, trans-cinnamic acid and p-coumaric acid are primarily obtained through chemical synthesis and plant extraction methods, but these approaches have certain limitations. Chemical synthesis methods such as the Perkin reaction[13], the Heck reaction[14], [15], and the Doebner condensation method[16] necessitate high temperatures and large quantities of chemical solvents, resulting in significant costs and environmental concerns[17]. Plant extraction methods fail to yield sufficient quantities to meet industrial demands due to the low initial content present in plants[18]. In contrast, biosynthesis offers a greener and more environmentally friendly alternative compared to chemical synthesis or plant extraction.

Bio-synthesis of p-Coumaric acid can be carried out by hydroxylation of trans-cinnamic acid by 4-cinnamic acid hydroxylase (C4H), or by tyrosine ammonia-lyase (TAL) catalyzing L-tyrosine (L-Tyr). However, the limited expression and activity of C4H during the hydroxylation process restrict p-coumaric acid synthesis[19]. On the other hand, PAL and TAL belong to the 5-methylene-3,5-dihydro-4H-imidazol-4-one (MIO)-containing class I lyase-like enzyme family (MIO-enzymes), and it was reported that the special 4-methylnitrosidazole-5-one (MIO) structure of PAL contributes to the deamination of PAL to L-Phe [20]. The generation of trans-cinnamic acid can be through phenylalanine deamination with PAL catalysis, while some plant-derived PAL such as ZmPAL2, which derived from Zea mays also has a deamination effect on L-Tyr. Thus, PAL without cofactor regeneration has the potential advantage of industrial production of trans-cinnamic acid and p-coumaric acid (Fig. 1). However, the currently reported conversion rates of different sources of PAL usually decreased along with increasing substrate concentrations (Table 1), among which ZmPAL2 could catalyze 55 mM L-Phe（9.1 g/L）with the conversion rate was less than 50% [9], [21], [22], [23]. The conversion rates of different sources of TAL reported so far at 0-6 mM L-Tyr concentrations also decreased with increasing substrate concentrations, among which TAL from Rhodotorula glutinis could catalyze 6 mM L-Tyr, with a conversion rate of less than 58%[19], [24]. Therefore, it is necessary to increase the activity of PAL or TAL in the biosynthesis of trans-cinnamic acid and p-coumaric acid.

In recent years, the combination of semi-rational and rational design has expedited enzyme evolution by constructing concise mutation libraries through saturation mutagenesis targeting specific amino acid residues, followed by screening to identify superior mutants[25]. Computer-assisted tools facilitate faster elimination of non-beneficial mutations, enabling the selection of a few specific amino acid residues for library construction based on structure-function relationships, ultimately leading to the discovery of activity-enhancing mutants [26], [27]. Enzymatic channels play crucial roles in substrate regulation and small molecule entry, CAVER is a widely utilized software tool that automatically analyzes tunnels and channels in static macromolecular structures to identify and characterize transport pathways[28]. Additionally, pSUFER (protein Strain, Unsatisfactoriness, and Frustration findER), another computational tool similar to localized frustration computation in proteins, is employed[29]. The present approach enables the identification of residues detrimental to protein folding, categorizing them as suboptimal, a feature often associated with protein activity. By employing the output Rosetta energy score, mutant residues can be effectively screened, thereby accelerating the process.

To obtain a highly efficient biocatalyst capable of catalyzing L-Phe and L-Tyr, we evolved the ZmPAL2, a Zea mays-derived PAL known for its tolerance to higher substrate concentrations compared to other PAL[9], a strategy combining saturation mutagenesis and computational design was employed to remodel the substrate-binding pocket and adjacent loop region of ZmPAL2 to enhance enzyme activity. As a result, we successfully obtained a mutant variant exhibiting significantly improved activity towards both L-Phe and L-Tyr. This research provides valuable engineered PAL for efficiently producing trans-cinnamic acid and p-coumaric acid.