Lignocellulose-derived sugars represent the largest reserve of fermentable sugars in Nature
[1]. However, lignocellulose is difficult to deconstruct and utilize due to its recalcitrant structure and diverse and complex composition, with cellulose in crystalline and fibrous forms
[2][3]. Three dominant enzymatic systems are utilized by cellulolytic organisms to overcome the recalcitrant nature of lignocellulose, including the free cellulases derived from aerobic fungi and bacteria, the cellulosome system mainly produced by anaerobic bacteria, and multimodular glycoside hydrolases with carbohydrate-binding modules usually produced by
Caldicellulosiruptor species. One common feature of these various cellulolytic systems is the high expression of glycoside hydrolase family 48 (GH48) cellulases, which are considered to be the key component of cellulose degradation
[4][5][6].
Usually, only one or two (rarely three) GH48 gene(s) are present in the genomes of cellulolytic organisms. For the well-known cellulolytic bacteria
C. thermocellum, which is one of the most efficient cellulose degraders in nature, two GH48 cellulases—cellulosomal Cel48S and non-cellulosomal Cel48Y—exhibit hydrolysis activities on crystalline cellulose. These GH48 cellulases are upregulated during growth on crystalline cellulose. Meanwhile, the deletion of Cel48S and Cel48Y led to a significant decrease in performance but does not completely abolish cellulolytic activity
[6]. Moreover, three mutants of the cellulolytic bacterium
Ruminococcus albus which are impaired in the production of Cel9B and Cel48A were all reported to be defective in adhesion to and degradation of cellulose
[4]. Additionally, deletion of the CelA gene in
C. bescii seriously damaged its ability to grow on crystalline cellulose and abolished its growth on lignocellulosic biomass, with a 15-fold decrease in sugar release on crystalline cellulose compared with the parent and wild-type strains. Meanwhile, the loss of exoglucanase activity could not be compensated by other enzymes in the
C. bescii secretome
[7]. Except for the genetics perspective, the importance of GH48 cellulases for crystalline cellulose hydrolysis has also been addressed from the perspectives of metatranscriptomic
[8] and genomic analyses
[9][10]. GH48 cellulases are necessary for microorganisms to degrade crystalline cellulose, so GH48 genes were reported to be a suitable molecular marker for the characterization of truly cellulolytic bacteria, especially in anaerobic environments
[5][11]. The detection and quantification of GH48 genes also can be used to identify cellulolytic organisms
[12]. Moreover, Cel48A, the GH48 cellulase domain in CelA, exhibits catalytic promiscuity in hydrolyzing xylan with an unrevealed mechanism
[13].
2. Structures of Glycoside Hydrolase Family 48 Cellulases
As early as 1998, Parsiegla et al. reported the first crystal structure of GH48 cellulase, the catalytic domain of processive endocellulase Cel48F from
C. cellulolyticum in complex with a thiooligosaccharide inhibitor
[14]. Subsequently, they reported structures of native and mutated Cel48F in complex with the cello-oligosaccharides, hemithiocellooligosaccharide, and thio-oligosaccharide inhibitors (
Table 1)
[15][16]. Currently, 26 structures of 11 unique GH48 cellulases including Cel48F, CelS from
C. thermocellum, ExgS from
Clostridium cellulovorans, TfCel48A from
T. fusca, two GH48s from
Caldicellulosiruptor genus, three GH48s from
Bacillus genus, and HcheGH48 from
Hahella chejuensis. Most of them are exocellulases, except for Cel48F, which is a processive endocellulase.
Table 1. Characterized and structure-determined GH48 cellulases.
All these GH48 structures exhibit similar overall fold, with the C
α RMSD values ranging from about 0.5 Å to 0.7 Å by superimposing CelS to other GH48 cellulases. They share a typical (α/α)
6 barrel consisting of an inner core of six mutually parallel α-helixes (helices with even numbers) and an outer shell of six peripheral α-helixes (helices with odd numbers), and the N-terminus of each inner helix is connected by long loops, additional helices, or sheets to the C-terminus of one outer helix (
Figure 1A). The catalytic residues are located in the N-terminal region of two inner helices, while most substrate-binding residues are located on the additional elements which form a layer covering the barrel. The covering elements may play roles in modulating the function of GH48 enzymes. BpCel48 from
Bacillus pumilus exhibits eight longer loops compared to other GH48 structures. Structural overlay revealed that all three GH48 enzymes from
Bacillus sp. feature these extra loops (
Figure 1B). Molecular dynamics simulations indicated that BpCel48 loops near the tunnel exit do not affect product inhibition. However, these loops are speculated to be responsible for the lower thermostability of BpCel48 by being more exposed to solvent
[27][28]. A recent study indicates that two extra longer loops (loop2 and loop6) located at the exit of the active site in BlCel48B act as an extension of the catalytic pocket and form a platform for product anchoring at the exit from the open-cleft part of the active site
[26]. Among these GH48 structures, HcheGH48 is more special as it is from Proteobacteria, whose gene is believed to be obtained by horizontal gene transfer
[30]. Consistent with the gene transfer statement, the structure of HcheGH48 is almost identical to other GH48 structures. Besides, a structural element termed ω-loop located between residue Pro469 and Ala482 (as in Cel48F) in all cellulases is proposed to distinguish cellulases and non-cellulases from insects (i.e., chitinases) coded by horizontally transferred GH48 genes.
Figure 1. Structure and substrate sites of GH48 cellulase. (A) The overall structure of GH48 cellulase, CelS (PDB entry 5YJ6). The core (α/α)6 barrels are colored red and violet, and the additional secondary structures are gray. The helices of the inner barrel with even numbers are in red while the helices of the outer barrel with odd numbers are shown in violet. (B) Superimposition of CelS (PDB entry 5YJ6, orange) and three Bacillus GH48 cellulases (PDB entry 5WMA, marine; PDB entry 5BV9, cyan; PDB entry 7KW6, gray). The extra loops of Bacillus GH48s are indicated by red arrows. (C) The substrate sites of GH48 cellulase. The cellobiose in subsites +1 and +2 and cellohexaose in subsites −2 to −7 in CelS-cellohexaose are shown in deep purple (PDB entry1L2A). The thiocellodecaose located at subsites +2 to −7 in CelF E55Q/thiocellodecaose complex are colored in gray (PDB entry 2QNO). The figure was prepared using PyMOL (Schrödinger).
Structures of GH48 enzymes in complex with oligosaccharides revealed the active-site topology, generally featuring a tunnel-like substrate binding part (subsites named −7 to −1) and an open-cleft product binding part (subsites named +1 and +2) (
Figure 1C). In most complex structures, the tunnel-shaped active site is occupied by cello-oligosaccharides, and the cleft part is bounded by cellobiose in subsites +1 and +2 after the cleavage site. Nevertheless, another subsite +3 is surmised in the complex of the inactive mutant E55Q of CelF and cellohexaose or cellotetraose, indicating that there is sufficient sugar-binding potential at the tunnel exit
[15]. Residues constituting the tunnel are quite conserved. In the statistics of the five GH48 enzymes, including BpCel48, CbCel48A, TfCel48A, CelS, and Cel48F, 27 residues of 36 that represent the tunnel walls and contact with the substrate/product are universally conserved and most of the rest are highly conserved
[28]. A large content of the conserved residues is aromatic residues, as well as several charged residues, including Arg, Asp, and Glu residing along the tunnel exit
[27]. The aromatic residues interact and stabilize the cellulose chain along the tunnel length by stacking interactions with the sugar moieties. These aromatic residues are supposed to serve as lubricating agents to reduce the sliding barrier in the processive action
[15]. Further studies confirm their essential roles in the molecular recognition of insoluble cellulosic substrates as their mutants dramatically affect the enzyme hydrolysis rate and processivity
[27][29].
3. Strategies and Progress of Engineering Glycoside Hydrolase Family 48 Cellulases
Compared with other family cellulases, GH48 cellulases exhibit relatively low specific activity on cellulose in assays in vitro. Three main factors are speculated to be responsible for the low enzymatic activities of GH48s: the inefficient acquisition of cellulose by the tunnel entrance, the slow processivity of the cellulose substrate in the tunnel, and the end-product inhibition
[27]. A fundamental factor affecting the enzymatic activity of GH48 is the substrate properties. A comprehensive study of enzymatic properties of the processive BlCel48B cellulase from
B. licheniformis indicates that the heterogeneity and structural nature of cellulose substrates, including substrate size and morphology, impact the substrate affinity, cleavage patterns, processivity, and hydrolytic efficiency of BlCel48B
[26]. In addition, other cellulases are reported to be strongly influenced by the ratio between the average free path for cellulase processive dislocation after one catalytic step and its processivity, as well as by the physical and chemical structure of the substrate
[26][35][36]. The end-product cellobiose has been reported to strongly inhibit the activity of several GH48 cellulases, such as
C.
thermocellum CelS
[37] and
T. fusca Cel48A
[38].
Various factors are reported to influence the product binding affinity, such as the pH of the solution, the type of the product, and the enzymatic environment
[39][40][41]. The product inhibitory effect of four GH48s has been quantitatively evaluated, with CelS exhibiting the highest product inhibitory level, followed by BpCel48, CelF, and CbCel48A. A series of single mutants with theoretically reduced levels of product inhibition have also been proposed
[27]. For the well-studied Cel48F, a hydrogen bond rearrangement that reduces the sliding barrier and stimulates the product to move toward the exit is important for the product release progress. This provides clues and directions to the modification or the mutation of cellulase to enhance the catalytic activity
[42][43].
Effects on improving the enzyme secretion and stability have also been made to enhance GH48 enzymatic activity on cellulose. A PelB signal peptide mediating post-translational secretion has been attached to the N-terminal end of CelS (P-Cel48S) and allowed catalytically active Cel48S to be successfully produced in the culture medium of recombinant
Escherichia coli [44]. Meanwhile, recombinant Cel48S via the co-translational pathway (attached with a DsbA signal peptide) yielded a 2.2-times higher specific activity than that associated with P-Cel48S expression. A set of Cel48 chimeras created from the catalytic domains of three native Cel48 enzymes CelF, CelS, and CelY by structure-guided recombination have been evaluated and subsequent sequence-function analysis demonstrates a high degree of additivity in the sequence–stability relationship, and this will help to predict highly stable and active Cel48 enzymes
[45].
In addition to product inhibition, the sliding of the substrate into the active site is another crucial step in cellulose degradation. Recently, nonequilibrium molecular dynamics simulations are carried out to investigate the energetics and mechanism of the substrate dynamics and product expulsion in CelS. The results indicate that product removal is relatively easier and faster than the sliding of the substrate to the catalytic active site
[39]. Therefore, the details of the substrate passage in the processive action of GH48s will be another noteworthy entry point to the rationale design of enzymes with better yield and performance.