Engineering Caldicellulosiruptor bescii with Surface Layer Homology Domain-Linked Glycoside Hydrolases Improves Plant Biomass Solubilization

ABSTRACT

Extremely thermophilic Caldicellulosiruptor species solubilize carbohydrates from lignocellulose through glycoside hydrolases (GHs) that can be extracellular, intracellular, or cell surface layer (S-layer) associated. Caldicellulosiruptor genomes sequenced so far encode at least one surface layer homology domain glycoside hydrolase (SLH-GH), representing six different classes of these enzymes; these can have multiple binding and catalytic domains. Biochemical characterization of a representative from each class was done to determine their biocatalytic features: four SLH-GHs from Caldicellulosiruptor kronotskyensis (Calkro_0111, Calkro_0402, Calkro_0072, and Calkro_2036) and two from Caldicellulosiruptor hydrothermalis (Calhy_1629 and Calhy_2383). Calkro_0111, Calkro_0072, and Calhy_2383 exhibited β-1,3-glucanase activity, Calkro_0402 was active on both β-1,3/1,4-glucan and β-1,4-xylan, Calkro_2036 exhibited activity on both β-1,3/1,4-glucan and β-1,4-glucan, and Calhy_1629 was active only on arabinan. Caldicellulosiruptor bescii, the only species with molecular genetic tools as well as already a strong cellulose degrader, contains only one SLH-GH, Athe_0594, a glucanase that is a homolog of Calkro_2036; the other 5 classes of SLH-GHs are absent in C. bescii. The C. bescii secretome, supplemented with individual enzymes or cocktails of SLH-GHs, increased in vitro sugar release from sugar cane bagasse and poplar. Expression of non-native SLH-GHs in vivo, either associated with the S-layer or as freely secreted enzymes, improved total carbohydrate solubilization of sugar cane bagasse and poplar by up to 45% and 23%, respectively. Most notably, expression of Calkro_0402, a xylanase/glucanase, improved xylose solubilization from poplar and bagasse by over 70% by C. bescii. While Caldicellulosiruptor species are already prolific lignocellulose degraders, they can be further improved by the strategy described here.

IMPORTANCE Caldicellulosiruptor species hold promise as microorganisms that can solubilize the carbohydrate portion of lignocellulose and subsequently convert fermentable sugars into bio-based chemicals and fuels. Members of the genus have surface layer (S-layer) homology domain-associated glycoside hydrolases (SLH-GHs) that mediate attachment to biomass as well as hydrolysis of carbohydrates. Caldicellulosiruptor bescii, the most studied member of the genus, has only one SLH-GH. Expression of SLH-GHs from other Caldicellulosiruptor species in C. bescii significantly improved degradation of sugar cane bagasse and poplar. This suggests that this extremely thermophilic bacterium can be engineered to further improve its ability to degrade specific plant biomasses by inserting genes encoding SLH-GHs recruited from other Caldicellulosiruptor species.

INTRODUCTION

Lignocellulose has significant potential as a renewable resource for biofuel and bio-based chemical production due to its abundance and cost (1). The biological conversion of lignocellulose into industrial chemicals requires synergistic action of cellulases, hemicellulases, and other plant biomass-degrading enzymes to overcome biomass recalcitrance (2, 3). The simultaneous presence of these enzymes and the subsequent ability to ferment both C₅ and C₆ sugars without significant catabolite repression are rare among microorganisms but do occur in several extremely thermophilic Caldicellulosiruptor species (4). These bacteria (optimal temperature [T_opt] of >70°C) have drawn considerable attention for use in consolidated bioprocessing (CBP), where lignocellulose solubilization and conversion are achieved with a single microorganism (5). Members of this genus are extremely thermophilic, Gram-positive, fermentative anaerobic bacteria that can grow on glucans and hemicelluloses; some species can also grow on microcrystalline cellulose (6). A genetic system developed for Caldicellulosiruptor bescii (7) has been improved significantly (8–11), thereby creating a metabolic engineering platform for bio-based chemical production (12). Recent developments, both experimentally and computationally, have further improved prospects for C. bescii (13–16).

Six large multidomain glycoside hydrolases (GHs), CelA (Athe_1867), CelB (Athe_1859), CelC (Athe_1857), CelD (Athe_1866), CelE (Athe_1865), and CelF (Athe_1860), present in the glucan degradation locus (GDL) of the C. bescii genome, play a key role in microcrystalline cellulose hydrolysis (17, 18). These GDL enzymes are naturally optimized for synergistic cellulose degradation (19). Caldicellulosiruptor species also use unique carbohydrate binding proteins, called “tāpirin,” and surface layer (S-layer) homology (SLH)-associated GHs (SLH-GHs) to attach specifically to lignocellulosic plant biomass (20, 21). Intracellular and extracellular GHs across the genus Caldicellulosiruptor have been examined in detail, such as the cellulase CelA (22–24) and the Caldicellulosiruptor kronotskyensis xylanase (Xyn10A) (25), but a comprehensive study of the SLH-GHs has not been done.

The surface layer (S-layer) in microorganisms lies on the outside of the cell envelope, and it is noncovalently associated at either the N or C terminus of the S-layer proteins (26–28). Bacterial SLH domains in S-layer-associated proteins are found in several microorganisms and mainly function as cell wall anchors (29–31). In the genus Caldicellulosiruptor, SLH-GHs are displayed on the cell surface and help deconstruct certain types of lignocellulosic plant biomass. For example, a Caldicellulosiruptor saccharolyticus SLH endoglucanase (Csac_0678) attaches to and hydrolyzes glucans in switchgrass (32). An SLH-GH in Caldicellulosiruptor kronotskyensis (Calkro_0402) was characterized as a xylanase, and a knock-in of Calkro_0402 into a C. bescii strain improved substrate attachment and degradation of certain xylans compared to wild-type C. bescii. Calkro_0402 likely plays a role in helping the cell attach to xylan substrates (20). Furthermore, Caldicellulosiruptor kristjanssonii Xyn10C-GE15A, a homolog of Calkro_0402, exhibited thermostable xylanase-glucuronoyl esterase activity; it specifically hydrolyzed glucuronoarabinoxylan, a type of xylan found in corn biomass, made possible by its two different catalytic domains (33). While two SLH-GHs (Calkro_0402 and Calkro_2036) and their homologs have been studied, representatives of the other four SLH-GHs have not (20, 32, 34, 35). Here, we consider representatives of all six classes of SLH-GHs from Caldicellulosiruptor species: Calkro_0111, Calkro_0402, Calkro_0072, and Calkro_2036 from C. kronotskyensis and Calhy_1629 and Calhy_2383 from Caldicellulosiruptor hydrothermalis (Table 1 and Fig. 1). The SLH-GHs were produced recombinantly in either C. bescii or Escherichia coli and characterized biochemically. Then, in vitro lignocellulosic plant biomass hydrolysis was done with cocktails of SLH-GHs. Finally, six strains of C. bescii were created to express each of the non-native SLH-GHs and then evaluated for potential contributions to C. bescii biomass solubilization.

TABLE 1.

Caldicellulosiruptor sp. surface layer homology domain glycoside hydrolases

Domainsa	Locus	Mr (truncated) (kDa)	pHoptb	Topt (°C)	Homolog(s)
SLH-SLH-SLH-CBM54-FN3-GH16-ACL-FN3-FN3-ACL-FN3-FN3-GH55-CBM32-CBM32	Calkro_0111	240	5	75
CBM22-CBM22-CBM22-GH10-CBM9-CBM9-Cad-SLH-SLH	Calkro_0402	160	6	75	Calow_1924, Calla_0206, Calkr_2245
SLH-SLH-SLH-CBM54-GH16-CBM4-CBM4-CBM6-CBM4-CBM4	Calkro_0072	170	7	70	Calhy_0060
GH5-CBM28-SLH-SLH-SLH	Calkro_2036	60	5	75	Athe_0594, Calhy_2064, Csac_0678, Calkr_2007, Calow_0352
SLH-SLH-SLH-GH43-CBM54	Calhy_1629	110	6	80
CBM35-GH87-FN3-SLH-SLH-SLH	Calhy_2383	200	5.5	75

^aSLH, surface layer homology; CBM, carbohydrate binding module; GH, glycoside hydrolase; ACL, actin cross-linking-like; Cad, cadherin-like domain; FN3, fibronectin type III. Boldface indicates catalytic GH domain.

^bpH_opt, optimal pH.

FIG 1.

Caldicellulosiruptor species and substrate specificity of their SLH-GHs. Lanes: 1, birchwood xylan (β-1,4-xylose); 2, oat spelt xylan (β-1,4-xylose with arabinose substitution); 3, curdlan (β-1,3-glucose); 4, laminarin (β-1,3-glucose with β-1,6 branches); 5, lichenan (β-1,3- and β-1,4-glucose [ratio of 1:2]); 6, carboxymethyl cellulose (β-1,4-glucose); 7, pustulan (β-1,6-glucose); 8, tamarind xyloglucan (β-1,4 and β-1,6 linkages and α-1,6-xylose side chain); 9, wheat arabinoxylan (β-1,4-xylose-α-1,2/3-arabinose side chain); 10, barley-β-glucan (β-1,3- and β-1,4-glucose); 11, arabinan (α-1,5-arabinose with α-1,2-l-arabinofuranosyl); 12, debranched arabinan (α-1,5-arabinose). Surface layer homology domains (SLH), catalytic domains (glycoside hydrolase [GH]), carbohydrate binding domains (carbohydrate binding module [CBM]), and domains with unknown function (e.g., actin cross-linking-like [ACL], cadherin-like domain [Cad], fibronectin type III [FN3]). The asterisk indicates the difference between sample and negative control.

RESULTS AND DISCUSSION

Recombinant production of Caldicellulosiruptor SLH-GHs.

Six different classes of SLH-GHs are represented in Caldicellulosiruptor species, as shown in Table 1. These SLH-GHs are found across the genus and have different domain architecture and molecular masses. Each SLH-GH consists of at least one glycoside hydrolase (GH) domain, one or more carbohydrate binding modules (CBMs), and domains of unknown function and has repetitive SLH domains at either the N or C terminus. Because of the size and complexity of the SLH-GHs, they have been difficult to produce recombinantly (20). However, successful expression of representative SLH-GHs from the six classes was accomplished using either Escherichia coli or C. bescii as the host (Fig. 2); SLH domains were removed to facilitate protein expression. In C. bescii, this resulted in secreted protein expression, thereby aiding the purification process. The SLH-GHs without their SLH domain were expressed with a C-terminal histidine tag. To determine whether anchoring to the S-layer was important for biomass degradation, SLH-GHs were also expressed in C. bescii with their SLH domain intact; in these cases, the expression strain and supernatant were probed with a His-tagged antibody to confirm the expression on the S-layer and in the culture supernatant (see Fig. S1 in the supplemental material). The predominant band (more than 75% of total protein in the lane on the SDS-PAGE gel in Fig. 2) indicated that the M_rs of the purified, truncated proteins, Calkro_0111, Calkro_0402, Calkro_0072, Calkro_2036, Calhy_1629, and Calhy_2383, were approximately 240 kDa, 160 kDa, 170 kDa, 60 kDa, 110 kDa, and 200 kDa, respectively, consistent with the M_rs without the SLH domain determined from amino acid sequence.

FIG 2.

SDS-PAGE gel of purified recombinant SLH-GHs produced in E. coli or C. bescii host. One hundred micrograms of each purified truncated SLH-GH was resolved in 4 to 20% mini-Protean TGX stain-free gels. Lanes 1 and 8, benchmark protein ladder; lane 2, Calkro_0111_ΔSLH (240 kDa); lane 3, Calkro_0402_ΔSLH (160 kDa); lane 4, Calkro_0072_ΔSLH (170 kDa); lane 5, Calkro_2036_ΔSLH (60 kDa); lane 6, Calhy_1629_ΔSLH (110 kDa); lane 7, Calhy_2383_ΔSLH (200 kDa).

Biochemical characterization of SLH-GH enzymes.

The optimal pH and temperature of each SLH-GH were determined using the substrates on which the enzyme was most highly active (Tables 1 and 2). The optimal pH of these SLH-GHs was in the range of 5 to 7 (see Fig. S2), while their optimal temperatures were between 70 and 80°C (see Fig. S3); these optima correspond to the optimal growth temperatures of Caldicellulosiruptor species from which the encoding genes were recruited. Figure 3 provides information on thermostability of these enzymes. All SLH-GHs exhibited high stability at 70°C, and as expected, stability decreased with increasing temperature. At 75°C, all SLH-GHs, with the exception of Calkro_0072 and Calhy_2383, retained more than 50% of their original activity after 5 h of incubation, with half-lives (t_1/2) of at least 5 h. The half-lives of Calkro_0072 and Calhy_2383 at 75°C were approximately 2 to 3 h. At 80°C, Calkro_0111, Calhy_1629, and Calhy_2383 exhibited high thermostability with half-lives of ~4 h. However, the other SLH-GHs lost 50% of their original activity within 30 min of incubation at 80°C. At 85°C, most SLH-GHs had ~30-min half-lives, except for Calhy_1629 and Calhy_2383, which had t_1/2 of ~1.5 h and ~2.5 h, respectively.

TABLE 2.

Specific activity of SLH-GHs on selected substrates

Enzyme	Substrate 1	Substrate 2	Substrate 3	Substrate 4	Substrate 5
Calkro_0111
Substrate	Laminarin	Barley-β-glucan
Sp acta	2,324 ± 187	878 ± 35
Calkro_0402
Substrate	Barley-β-glucan	Wheat arabinoxylan	Birchwood xylan	Oat spelt xylan	Lichenan
Sp act	4,048 ± 769	3,920 ± 305	3,846 ± 254	3,810 ± 455	2,399 ± 187
Calkro_0072
Substrate	Laminarin	Lichenan	Curdlan	Barley-β-glucan
Sp act	4,291 ± 221	4,241 ± 260	3,494 ± 260	2,857 ± 187
Calkro_2036
Substrate	Lichenan	CMC	Barley-β-glucan
Sp act	4,310 ± 218	3,231 ± 195	2,317 ± 293
Calhy_1629
Substrate	Arabinan	Debranched arabinan
Sp act	1,585 ± 38	3,214 ± 158
Calhy_2383
Substrate	Barley-β-glucan	Lichenan	Laminarin	CMC
Sp act	1,482 ± 48	1,470 ± 53	1,219 ± 48	1,078 ± 20

^aSpecific activity is shown as micromoles of reducing sugar per minute per micromole of protein.

FIG 3.

Thermostability analysis of SLH-GHs at different temperatures: 75°C, 80°C, and 85°C. Enzymatic activity was measured at 0, 0.5, 1, 2, 3, 4, and 5 h. The residual activities are presented in percentages, based on their initial activities.

Substrate specificity of Caldicellulosiruptor SLH-GHs.

A variety of polysaccharides were incubated with each SLH-GH to investigate substrate range. As seen in Fig. 1, Caldicellulosiruptor SLH-GHs consist of catalytic domains (glycoside hydrolase [GH]), carbohydrate binding domains (carbohydrate binding module [CBM]), and domains of unknown function (e.g., actin cross-linking-like [ACL], cadherin-like domain [Cad], and fibronectin type III [FN3]). Each SLH-GH was active on multiple substrates, reflecting the differences in sugar backbones and side chains.

Calkro_0111, consisting of GH16 and GH55 catalytic domains, was most active on substrates containing β-1,3 linkages in their backbones, with either β-1,4 or β-1,6 side chains (laminarin and barley-β-glucan, Fig. 1). According to the Carbohydrate Active enZYmes (CAZY) database (36), the GH16 family consists of several subgroups, including exo- and endo-β-1,3-glucanases. Moreover, a truncated Calkro_0111, to separate the GH domains, displayed endo- or exoglucanase activity on laminarin, reflecting the GH16 and GH55 domains, respectively (20). GH16 domains can release oligosaccharides from laminarin (20, 37), and GH55 domains can release glucose and laminaribiose from laminarin (38, 39). Therefore, Calkro_0111 can be classified as a laminarinase.

Calkro_0402, which has a GH10 domain, exhibited high activity on substrates containing β-1,4-xylose linkages, including birchwood xylan, oat spelt xylan, lichenan, wheat arabinoxylan, and barley-β-glucan (Fig. 1). Calkro_0402 and its homologs have been previously described as endo-β-1,4-xylanases (20, 33), representative of the high activity on xylan and wheat arabinoxylan. The results here also show hydrolysis of β-1,3/1,4-glucans, as evidenced by the activity on lichenan and barley-β-glucan, most likely indicating endo-β-1,3/1,4-glucanase (lichenase) activity. Calkro_0402 was not active on substrates with only β-1,3-glucans or only β-1,4-glucans.

Calkro_0072, which contains a GH16 domain, was highly active on curdlan, laminarin, lichenan, and barley-β-glucan, all of which contain β-1,3-glucans (Fig. 1). Calkro_0072 can be described as a β-1,3-glucanase, as observed for similar enzymes from other microorganisms (40, 41).

Calkro_2036 is the smallest enzyme among Caldicellulosiruptor SLH-GHs. Its catalytic domain belongs to family GH5, and its homolog in C. bescii, Athe_0594, has been previously described as a cellulase and a lichenase (34). Here, Calkro_2036 was highly active on the substrates lichenan, carboxymethyl cellulose (CMC), and barley-β-glucan (Fig. 1). Trace activity on oat spelt xylan was detected. This indicated both cellulase and lichenase activity, with additional activity on oat spelt xylan, likely due to glucose side chain hydrolysis.

Calhy_1629 belongs to family GH43. It was mostly active on arabinan substrates that have α-1,5-arabinan as their backbone (arabinan and debranched arabinan). GH43 domains are typically found in arabinanases, with homologs in other microorganisms, including Thermotoga thermarum (42) and Bacillus subtilis (43). Thus, the primary activity of Calhy_1629 is as an arabinanase.

Lastly, Calhy_2383 contains a GH87 catalytic domain, commonly found in fungi and usually part of a multidomain enzyme (44, 45). Enzymes containing GH87 domains are mainly active on α-1,3-glucan substrates (46). Here, Calhy_2383 was active on lichenan, laminarin, and barley-β-glucan, which contain β-1,3-glucan mixed with β-1,4 or β-1,6 glucan linkages. Lower activity was detected on curdlan (β-1,3-glucose), and CMC (β-1,4-glucose). No activity was seen on any of the xylose-containing substrates (Fig. 1). This suggests a primary activity as a β-1,3-glucanase, with additional secondary activities as a β-1,4-glucanase. Calhy_2383 activity on CMC is particularly interesting as Caldicellulosiruptor hydrothermalis is unable to effectively degrade cellulose, indicating that the β-1,4-glucanase activity seen here is not sufficient for C. hydrothermalis to degrade crystalline cellulose.

Engineering C. bescii’s S-layer to improve solubilization of plant biomass.

C. bescii, the only Caldicellulosiruptor species for which molecular genetic tools are available, has only one SLH-GH (Athe_0594). Given its lack of five of the six SLH-GH types in the genus Caldicellulosiruptor, the objective was to determine if C. bescii could be a better biomass degrader with additional SLH-GHs recruited from other species. Previously, the function of SLH domains has been examined in other microorganisms (30, 47, 48). The SLH domains in Gram-positive bacteria are involved in surface attachment using the interaction between peptidoglycan-linked secondary cell wall polymers and SLH domains (49). This is likely the case for the Caldicellulosiruptor SLH-GHs studied here. The SLH-GHs were heterologously expressed in C. bescii with and without their S-layer domain intact and in all cases with C-terminal 6×His tags. To confirm that the SLH domain-containing enzymes were localized to the cell surface, cells were incubated with Alexa Fluor 488-conjugated nickel-nitrilotriacetic acid (Ni-NTA) beads and imaged with confocal microscopy. As seen in Fig. 4, these expression strains displayed the SLH-GHs on their cell surface, as indicated by the green fluorescence observed. Meanwhile, strains with deleted SLH domains did not exhibit green fluorescence on the cell surface.

FIG 4.

Fluorescence microscopy showing surface-bound GHs via SLH domains in different SLH-GH strains revealing that SLH domains in Caldicellulosiruptor SLH-GHs function as S-layer binding domains to associate the rest of the domains on the cell surface. The expression strains have His tags on the expressed protein, which can specifically bind to Ni-NTA fluorescent dye. All cells are labeled with Ni-NTA fluorescent dye (green) and are counterstained with DAPI (blue). Bars, 2 μm.

Lignocellulosic plant biomass solubilization by C. bescii secretomes supplemented with SLH-GH cocktails.

The addition of SLH-GH cocktails to the parent strain’s secretome, in vitro, was done to see if this enhanced reducing sugar release from lignocellulosic plant biomasses (Table 3). Sugar cane bagasse and poplar, which have different carbohydrate profiles, were used as growth substrates to generate the secretomes. These biomasses each contain more cellulose than hemicellulose, but they differ in sugar composition (glucose, xylose, galactose, arabinose, and mannose) (Fig. 5). Some degree of synergism occurred in all SLH combinations with the parent strain’s secretome in vitro. The best case was a mixture of 75% secretome and 25% SLH-GHs for both biomasses. For sugar cane bagasse hydrolysis, the mixture of 75% secretome and 25% SLH-GHs released 79% more reducing sugar than did secretome only. Other combinations, 50% secretome or 25% secretome with the balance being SLH-GHs, were able to release 12 and 32%, respectively, more reducing sugars than those with using secretome only (Table 3).

TABLE 3.

In vitro biomass hydrolysis after 24 h of incubation

SLH-GH cocktail	Normalized sugar release (μg sugar/100 μg enzyme)	% improvement vs secretome only
Sugar cane bagasse
Secretome only	15.3
25% secretome–75% SLH-GHs	17.1	12.1
50% secretome–50% SLH-GHs	20.2	32.4
75% secretome–25% SLH-GHs	27.3	78.8
SLH-GHs only	12.1	−20.9
Poplar
Secretome only	4.2
25% secretome–75% SLH-GHs	6.1	47.6
50% secretome–50% SLH-GHs	8.6	107
75% secretome–25% SLH-GHs	17.5	321
SLH-GHs only	2.2	−47.6

FIG 5.

Carbohydrate solubilization of different SLH-GH strains compared with their parent strains. (A) Total carbohydrate solubilization. (B) Different sugar compositions in sugar cane bagasse and poplar. (C) Detailed carbohydrate solubilization in sugar cane bagasse. (D) Detailed carbohydrate solubilization in poplar.

Similar trends were observed with poplar biomass hydrolysis, but the SLH-GH addition was more impactful. Incubation with 75% secretome and 25% SLH-GHs improved reducing sugar release from poplar more than 321% compared to secretome only (Table 3). Other combinations of secretome and SLH-GHs also showed significant improvements for poplar. Synergistic enzymatic activity between two enzymes has been shown to exhibit cooperativity between the different domains. Adding glucuronoyl esterase (CkGE15) to a commercial xylanase, at its optimal temperature, resulted in increased xylan liberation from lignin in wheat straw (33). Enzymatic synergism between cellulases and hemicellulases has also been reported to improve lignocellulosic plant biomass deconstruction (50–52). Natively, C. bescii has only a single SLH-GH, but it has a plethora of other secreted cellulases and hemicellulases that are capable of deconstructing significant portions of plant biomasses (35). However, C. bescii is unable to solubilize all polysaccharides contained in certain lignocelluloses, especially xylans and arabinans (14, 16). Supplementing with SLH-GH cocktails could overcome this limitation.

Lignocellulosic plant biomass solubilization by C. bescii SLH-GH expression strains.

To further evaluate the role of each SLH-GH in lignocellulosic plant biomass deconstruction, C. bescii expression strains for each SLH-GH, with and without the SLH domain intact, were compared against the parent strain on sugar cane bagasse and poplar. Recall that Calkro_0111, Calkro_0072, and Calhy_2383 were most active on β-1,3-glucan-containing substrates; Calkro_2036, Calkro_0402, and Calhy_1629 were most active on lichenan, β-1,4-xylan, and α-1,5-arabinan, respectively. There was no significant difference in lignocellulosic plant biomass solubilization between strains with and those without SLH domains. Therefore, the strains with SLH domains were used for biomass solubilization. In natural environments where carbohydrate substrates are scarce, the intact SLH domain is likely important for scavenging nutrients.

As seen in Tables 4 and 5 and Fig. 5, the addition of Calkro_0402 (β-1,4-xylanase/lichenase) to the parent strain solubilized 44.5% and 23.0% more carbohydrate than the parent strain for bagasse and poplar, respectively. This strain had significantly increased xylose solubilization compared to all other strains (78.1% and 71.9%, for bagasse and poplar, respectively), representative of the enzyme’s xylanase activity. All strains were capable of improving glucose solubilization compared to the parent strain (Tables 4 and 5 and Fig. 5); the most significant solubilization was 48.5% and 35.2% for Calkro_2036 for bagasse and poplar, respectively. Calkro_2036 is homologous to Athe_0594 in C. bescii, but the higher expression level in the expression strain indicates that Athe_0594 needs to be at higher levels to be effective against specific biomasses. All of these enzymes have glucanase activity of various types, including β-1,3-glucanase (Calkro_0072 and Calhy_2383), β-1,4-glucanase (Calkro_2036 and Calhy_2383), lichenase (Calkro_0402, Calkro_2036, Calhy_1629 [trace], and Calhy_2383), and laminarinase (Calkro_0111). These glucanase activities support the increased in vivo glucose solubilization from plant biomasses (Fig. 5). Enzymes with the highest activities on lichenan and barley-β-glucan also had the highest glucan solubilization (Fig. 1 and 5 and Tables 4 and 5). Alternatively, several SLH-GHs (Calkro_0111, Calkro_0072, Calkro_2036, and Calhy_2383), which contributed significantly to glucose release, contain CBM domains distinct from those of C. bescii glucanases and may be responsible for enhancing glucan solubilization. Additionally, these SLH-GHs also increased mannose solubilization in poplar, indicating that these glucanases may also have activity on mannan; sugar cane bagasse has no significant mannose content (Fig. 5).

TABLE 4.

Improvement in sugar cane bagasse carbohydrate solubilization of SLH-GH strains

Strain	Carbohydrate			Glucose			Xylose			Arabinose
	Amt of sugar solubilized (mg)	%		Amt of sugar solubilized (mg)	%		Amt of sugar solubilized (mg)	%		Amt of sugar solubilized (mg)	%
		Solubilized	Improved		Solubilized	Improved		Solubilized	Improved		Solubilized	Improved
MACB1018	76.2	29.5		42.1	27.1		24.7	30.1		2.955	39.1
Calkro_0111	101	39.5	33.6	58.8	37.8	39.6	32.1	39.1	29.6	3.55	47.0	20.4
Calkro_0402	109	42.7	44.5	54.6	35.1	29.7	44.1	53.7	78.1	3.82	50.6	29.6
Calkro_0072	87.1	33.9	14.7	50.1	32.2	19.0	26.9	32.8	8.9	3.32	44.0	12.6
Calkro_2036	99.2	38.6	30.6	62.5	40.2	48.5	27.0	32.9	9.0	3.23	42.7	9.4
Calhy_1629	94.1	36.6	24.0	54.4	35.0	29.3	27.9	34.0	12.9	5.06	66.9	71.3
Calhy_2383	97.6	38.1	29.1	58.3	37.5	38.4	29.7	36.2	20.0	3.09	40.9	4.8

TABLE 5.

Improvement in poplar carbohydrate solubilization of SLH-GH strains

Strain	Carbohydrate			Glucose			Xylose			Mannose
	Amt of sugar solubilized (mg)	%		Amt of sugar solubilized (mg)	%		Amt of sugar solubilized (mg)	%		Amt of sugar solubilized (mg)	%
		Solubilized	Improved		Solubilized	Improved		Solubilized	Improved		Solubilized	Improved
MACB1018	78.6	31.4		47.6	30.0		15.8	26.1		4.2	29.4
Calkro_0111	91.0	32.4	15.6	55.9	35.2	17.5	18.7	30.9	18.4	5.5	38.3	30.3
Calkro_0402	96.8	34.3	23.0	53.7	33.8	12.8	27.2	44.9	71.9	5.1	35.3	20.2
Calkro_0072	90.8	36.3	15.4	58.7	36.9	23.3	16.6	27.4	4.7	4.8	33.7	14.7
Calkro_2036	95.2	38.0	20.9	64.2	40.4	34.8	14.8	24.4	−6.4	5.4	37.4	27.3
Calhy_1629	81.2	32.4	3.1	49.7	31.3	4.4	16.5	27.2	4.1	4.2	28.9	−1.5
Calhy_2383	86.0	34.3	9.3	52.3	32.9	9.9	17.8	29.3	12.3	4.8	33.6	14.3

The Calhy_1629 strain increased arabinose solubilization by 71.3% from sugar cane bagasse, the highest among strains; however, arabinose is not a major component in sugar cane bagasse, thus not contributing much to overall carbohydrate solubilization. It is noted that there was no significant difference in galactose solubilization between strains (all 64 to 66% solubilization), although galactose is a minor component in these biomasses (Fig. 5). The results shown here indicate that the surface-associated enzymes added to the parent strain of C. bescii cooperate with secreted carbohydrate active enzymes in the secretome to better deconstruct lignocellulosic plant biomasses, resulting in more fermentable sugar release for subsequent use in metabolic pathways.

The importance of surface layer association of the enzymes is likely not the primary cause of increased solubilization, as no difference was seen for Calkro_0402 and Calkro_2036, with or without the SLH domain, expressed in the C. bescii parent strain on sugar cane bagasse (data not shown). It is more likely that the unique CBM domains, catalytic activity, and/or increased expression level (in the case of Calkro_2036, the Athe_0594 homolog) is responsible for improved biomass degradation. The shared glucanase activity of these enzymes suggests that C. bescii might benefit from additional β-1,3/1,4-glucanase activity, at least for poplar and sugar cane bagasse. In other studies of C. bescii, coexpression of Thermotoga maritima β-d-xylosidase and Acidothermus cellulolyticus xylanase in C. bescii substantially improves exoproteome activity toward xylan and cell growth on cellulose (53). Additionally, coexpression of β-glucanases and a cellobiose phosphorylase in C. bescii increased cellulolytic activity of C. bescii exoproteome (54). These are evidence that synergism between the C. bescii secretome and non-native enzymes might be key to improving its ability to hydrolyze plant biomass. Future work could further explore the importance of native and non-native β-1,3/1,4-glucanases in C. bescii. In any case, these results show that there is room to improve C. bescii’s native ability to deconstruct plant biomass by adding or enhancing the level of carbohydrate active enzymes. While previous work has shown that C. bescii has fully optimized its native enzymes for crystalline cellulose degradation (19), non-native enzymes can still make significant improvements in solubilization of certain polysaccharides. Addition of these enzymes into C. bescii can help make it a better candidate in CBP where plant biomass deconstruction still remains one of the major bottlenecks.

MATERIALS AND METHODS

Bacterial strains, plasmids, and reagents.

Plasmid construction and expression of recombinant proteins were performed using the following Escherichia coli strains: NEB 5-alpha and NEB 10-beta (New England Biolabs [NEB], Ipswich, MA), Rosetta 2(DE3)pLysS (Fisher Scientific), and Arctic Express(DE3) RIL E. coli (Agilent Technologies). Overexpression of full-length and truncated SLH proteins in C. bescii was performed using strain C. bescii MACB1018 (8, 10). Genes of interest were amplified from Caldicellulosiruptor species genomic DNA by using the NEB Monarch kit and PCR (NEB Phusion High Fidelity DNA polymerase). Plasmid construction was done using the pET46 Ek/LIC vector kit (EMD Millipore) for E. coli protein expression or insertion into the C. bescii replicating vector pJMC046 containing a 6×His tag on the C terminus for C. bescii protein expression (19) using NEB HiFi Builder (New England Biolabs, Ipswich, MA). Plasmid sequences were confirmed by Sanger sequencing (Azenta Life Sciences, Morrisville, NC). Plasmids were maintained at −80°C in LB medium plus 15% glycerol and 50 μg/mL kanamycin, 50 μg/mL ampicillin, and/or 34 μg/mL chloramphenicol, as appropriate.

Genetic manipulation of C. bescii.

Competent cells of the MACB1018 strain and methylated plasmids were prepared using a previously described protocol (8, 18, 21). For each electroporation, 1 μg of methylated plasmids was added directly to 50 μL of competent cells at room temperature. The mixture was then transferred to a 1-mm-gap glass cuvette (USA Scientific) and electroporated using a Gene Pulser II system with a Pulse Controller Plus module (Bio-Rad). The electroporation was initiated at 2.0 kV, 400 Ω, and 25 microfarads. The voltage and resistance were increased as the size of plasmids increased. After the electroporation, the cells were immediately transferred to low-osmolality complex (LOC) medium (55) and incubated at 70°C for 1 h for recovery. After recovery, the transformants were passaged into defined modified medium DSM 516 (18) with 50 μg/mL kanamycin selection and 24 mM NH₄Cl. After 3 to 4 days, the culture was plated and embedded in DSM 516 medium with 1.5% agar and grown anaerobically at 70°C. Single colonies were picked and grown in liquid medium. The transformed DNA was isolated from liquid culture and sequenced. Primers used are listed in Table 6.

TABLE 6.

Primers used in this study

Primer	Sequencea	Use
ECalkro_0402 F	GACGACGACAAGATTCAAAGCAGCCAAACTAA	pET46 cloning
ECalkro_0402 R	GAGGAGAAGCCCGGTTACTTTTTGAGATTGTTAAGTGCTCTG	pET46 cloning
ECalkro_0072 F	GACGACGACAAGATAAATTCAGTTCAAATCTTCAAGGACGTT	pET46 cloning
ECalkro_0072 R	GAGGAGAAGCCCGGTTACTGGTCACTCACAACAGTCA	pET46 cloning
ECalkro_2036 F	GACGACGACAAGATAGGAAAATTATTTTAAAGTTTTGTGCAC	pET46 cloning
ECalkro_2036 R	GAGGAGAAGCCCGGTTATAGCTTGCCTGCTAAGTTCAAGAG	pET46 cloning
ECalhy_1629 F	GACGACGACAAGATGTATGTAGATAATGAAAG	pET46 cloning
ECalhy_1629 R	GAGGAGAAGCCCGGTTATTTTGTTTGAAACAATTTT	pET46 cloning
ECalhy_2383 F	GACGACGACAAGATGGAAGAACCCAATAGCAA	pET46 cloning
ECalhy_2383 R	GAGGAGAAGCCCGGTTAATTAGCTAAATTAGCA	pET46 cloning
Calkro_0111 F	AATCATACAAGGAGGTTTGGTGAGTAGTTATGAAACGAGGGGGAGCA	Cloning into pJMC046
Calkro_0111 R	CCTCCTTTAATGGTGGTGGTGGTGATGGTTCTTGCTCCCATACACTTCAAACTC	Cloning into pJMC046
Calkro_0402 F	AATCATACAAGGAGGTTTGGTGAGTAGTTATGAAAAAGCGACTTATAGCATTTCTTG	Cloning into pJMC046
Calkro_0402 R	CCTCCTTTAATGGTGGTGGTGGTGATGGTTCTTTTTGAGATTGTTAAGTGCTCTGAATAA	Cloning into pJMC046
Calkro_0072 F	AATCATACAAGGAGGTTTGGTGAGTAGTTATGCGTGTAAAAAATCTTGTTGC	Cloning into pJMC046
Calkro_0072 R	CCTCCTTTAATGGTGGTGGTGGTGATGGTTCTGGTCACTCACAACAGTC	Cloning into pJMC046
Calkro_2036 F	CAAGGAGGTTTGGTGAGTAGTTATGAGGAAAATTATTTTAAAGTTTTGTGC	Cloning into pJMC046
Calkro_2036 R	GTCCTCCTTTAATGGTGGTGGTGGTGATGGTTTAGCTTGCCTGCTAAGTTCAAG	Cloning into pJMC046
Calhy_1629 F	AATCATACAAGGAGGTTTGGTGAGTAGTTATGACCGAGGAAAGTGTTTATTTG	Cloning into pJMC046
Calhy_1629 R	CCTCCTTTAATGGTGGTGGTGGTGATGGTTTGCCAAAACTTTAATTTCATC	Cloning into pJMC046
Calhy_2383 F	AATCATACAAGGAGGTTTGGTGAGTAGTTATGGTCAGAAGAATTTATAAAAAG	Cloning into pJMC046
Calhy_2383 R	CCTCCTTTAATGGTGGTGGTGGTGATGGTTATTAGCTAAATTAGCAATAG	Cloning into pJMC046
Calkro_0111 F	AGTGGTTTGGATACAAGTAAGTGG	SLH domain deletion
SCalkro_0111 R	TCCTGGTGTGCCGTTAAAAAC	SLH domain deletion
SCalkro_0402 F	CAGAGCACTTAACAATCTCAAAAAG	SLH domain deletion
SCalkro_0402 R	AATGTTAACAGCTACAGGTGTTTT	SLH domain deletion
SCalkro_0072 F	GGCGGTGGAGAAAATTCTGTC	SLH domain deletion
SCalkro_0072 R	CATAACTACTCACCAAACCTCCTTG	SLH domain deletion
SCalkro_2036 F	CTAAACCATCACCACCACCAC	SLH domain deletion
SCalkro_2036 R	CATCAGCAATTGGAACATCGTC	SLH domain deletion
SCalhy_1629 F	GATTCTGGGGTTCGTAACGATAAT	SLH domain deletion
SCalhy_1629 R	CATAACTACTCACCAAACCTCCTTG	SLH domain deletion
SCalhy_2383 F	GAGAATTTGGCAAGGTTTGCT	SLH domain deletion
SCalhy_2383 R	TGCTCTTGTTATGTTCCTTTGAGG	SLH domain deletion

^aUnderlined portions identify Gibson assembly overlap regions.

Protein expression and purification.

Protein expression from E. coli strains was performed in 2× YT medium (10 g/L yeast extract, 16 g/L tryptone, 5 g/L sodium chloride) with the appropriate antibiotics (50 μg/mL kanamycin or 50 μg/mL ampicillin plus 34 μg/mL chloramphenicol for Rosetta; 50 μg/mL kanamycin and 50 μg/mL gentamicin for Arctic Express) at 37°C with 250-rpm agitation. Once the optical density at 600 nm (OD₆₀₀) reached 0.8, cultures were cooled and then induced with 1 mM isopropyl-β-d-1-thiogalactopyranoside (IPTG) at 20°C for 18 h. Cells were harvested at 6,000 × g for 10 min. Cell pellets were kept at −20°C prior to lysis and purification. Cell pellets were thawed and then resuspended in 20 mM sodium phosphate, 500 mM sodium chloride, pH 8.1. The cell suspension was lysed twice in a French press (Sim-Aminco) at ~15,000 lb/in². Because the target proteins are expected to be thermostable, cell lysate was heat treated at 65°C for 20 min to remove heat-labile proteins from E. coli. The cell lysate was centrifuged at 25,000 × g for 30 min at 4°C, and the supernatant was passed through a 0.22-μm filter to obtain cell extract.

For Calkro_0111, the C. bescii overexpression strain was grown in complex modified DSM 516 medium (18) plus 50 μg/mL kanamycin for 24 to 36 h at 75°C and the supernatant was harvested by centrifuging the culture at 6,000 × g for 10 min. The supernatant was filter sterilized and was kept at 4°C before purification.

Proteins were purified using 5-mL HisTrap HP Ni-Sepharose (Cytiva) immobilized metal affinity chromatography (IMAC) columns, operated according to the manufacturer’s instructions. Size exclusion chromatography was used after the IMAC column on a HiLoad 26/600 Superdex 200-pg column (Cytiva) using buffer containing 20 mM sodium phosphate, 500 mM sodium chloride, pH 8.1. All chromatography steps were performed on a Biologic DuoFlow fast protein liquid chromatograph (FPLC) (Bio-Rad). The purity of purified proteins was assessed via SDS-PAGE using 4 to 15% mini-Protean TGX stain-free gels (Bio-Rad) using Laemmli’s method with a Benchmark protein ladder (Life Technologies). The gel images were also analyzed using densitometry (ImageJ). The concentration of purified proteins was determined using the Pierce bicinchoninic acid (BCA) protein assay (Thermo Fisher) with bovine serum albumin (BSA) as a protein standard.

Polysaccharide substrate specificity.

Proteins (0.5 μM) were incubated with 1% (wt/vol) of the following substrates in a final assay volume of 100 μL: birchwood xylan, oat spelt xylan (MilliporeSigma, St. Louis, MO), curdlan, laminarin, lichenan, carboxymethyl cellulose, pustulan (Carbosynth), tamarind xyloglucan, wheat arabinoxylan, barley-β-glucan, arabinan, and debranched arabinan (Megazyme, Wicklow, Ireland). Assays were performed at 70°C for 20 min in a thermocycler. Released reducing sugars were measured using 3,5-dinitrosalicylic acid (DNS) assay, as previously described (19). The DNS reagent used contained 1.6% (wt/vol) sodium hydroxide, 30% (wt/vol) sodium potassium nitrate, and 0.1% (wt/vol) 3,5-dinitrosalicylic acid. Briefly, 25 μL of sample was incubated with 50 μL of DNS reagent and incubated in PCR tubes using the following program: 95°C for 5 min, 48°C for 1 min, and hold at 20°C. Then, 36 μL of DNS reaction mixture was diluted with 160 μL of distilled water, and absorbance was measured at 540 nm. Xylose or glucose was used as a standard for released xylooligosaccharide and gluco-oligosaccharide.

pH and temperature optima.

pH optima were determined at 70°C for 20 min in buffer pHs ranging from 3 to 10 (50 mM citrate buffer for pH 3 to 6, 50 mM sodium phosphate buffer for pH 6.5 to 8, and sodium bicarbonate buffer for pH 9 to 10). The temperature optima were evaluated at the optimal pH of each enzyme at 30, 40, 50, 60, 65, 70, 75, 80, 90, and 95°C. All reactions were done in triplicate with 0.5 μM SLH-GH in 100 μL of 1% (wt/vol) substrate, and no-enzyme controls were used. The DNS assay was used to assess the amount of released sugars from each reaction, as described above.

Enzyme thermostability.

Each SLH-GH was incubated at 75°C, 80°C, and 85°C for 5 h in its storage buffer. At different time points (0, 0.5, 1, 2, 3, 4, and 5 h), 0.5 μM preheated sample was taken and the residual activities were measured by incubating the enzyme-substrate mixture at 70°C for 20 min. Ten milligrams per milliliter of birchwood xylan, curdlan, CMC, arabinan, and lichenan was used as the substrates for Calkro_0402, Calkro_0072, Calkro_2036, Calhy_1629, and Calhy_2383, respectively. DNS assay was used to determine the amount of released sugar from each reaction, and the amount was compared with nondenatured enzyme.

SLH-GH specific activity.

Enzyme activity was determined by incubating each enzyme with a 1% polysaccharide substrate at the optimal temperature for 20 min in its optimal pH buffer. Substrates used in these assays were picked based on their highest activity found in the substrate specificity. DNS assay was used to measure released sugars. The enzyme activity was reported as micromoles of reducing sugar equivalents per minute.

Confocal microscopy.

The procedure was adapted from reference 56. All centrifugation steps were performed at 6,000 × g for 10 min at room temperature. Twenty-five milliliters of cell culture was drawn and pelleted. Cell pellets were washed three times with 1× phosphate-buffered saline (PBS) and were resuspended in 3.6 μM 2-[4-(aminoiminomethyl)phenyl]-1H-indole-6-carboximidamide hydrochloride (DAPI) (Thermo Fisher) in PBS and were left for 12 h at 4°C. After staining with DAPI, cells were pelleted and washed three times with 1× PBS. The cells were incubated with 300 μg/mL of Ni-NTA fluorescent dye (MilliporeSigma, St. Louis, MO) for another 12 h at 4°C. The cells were pelleted and washed three times with 1× PBS and were resuspended in 1× PBS prior to imaging. The images were taken on a Zeiss LSM880 confocal microscope (Carl Zeiss, Germany) operating in Airyscan fast mode (North Carolina State University Cellular and Molecular Imaging Facility [CMIF], Raleigh, NC). Images were processed by Zen Black 2.3 software.

Lignocellulosic plant biomass preparation.

All biomass substrates were prepared by milling and sieving, as previously described (16). The 40/80 sieve fraction was collected and was washed three times with 75°C distilled water to remove any soluble sugars. The biomass was then dried at 65°C overnight.

Lignocellulosic plant biomass hydrolysis with SLH-GH cocktails.

Enzyme synergism was investigated on sugar cane bagasse and wild-type poplar. Briefly, >80 sieve fractions, as described above, of each lignocellulosic plant biomass were collected and used in the experiment. Different combinations of SLH-GHs and secretome of MACB1018 cells grown on the respective lignocellulosic plant biomasses were incubated at 70°C with 500-rpm agitation. Sugar release was monitored at different time points (0, 2, 4, 8, 12, and 24 h) by DNS assay. All assays were performed in triplicate.

Biomass solubilization.

To assess biomass solubilization, 5 g/L of prepared 40/80 sieve fraction biomass was used in each culture. Caldicellulosiruptor strains were passaged three times on modified 671d medium, as described previously (57), on each biomass substrate, supplemented with 250 mM uracil and 50 μg/mL kanamycin. Solubilization cultures were prepared by inoculating 5 × 10⁸ cells/mL and grown at 75°C with 250 rpm for 7 days. Experiments were done in triplicate. After the fermentation, each culture was cooled to room temperature and harvested by centrifugation at 5,000 rpm for 10 min at room temperature. Soluble fermentation products were determined using a Waters Arc3 high-pressure liquid chromatograph (HPLC), as described in reference 14, using a refractive index detector and Rezex-ROA column with a 5 mM sulfuric acid mobile phase flowing at 0.6 mL/min at 50°C (model 2414 [Phenomenex]; Waters Corp.). Residual substrate biomass was washed three times and dried for 24 h at 75°C. The final dried plant biomass weight was recorded to calculate biomass solubilization.

Quantitative saccharification.

Composition of lignocellulosic plant biomasses was evaluated using the Klason method, as described in reference 16. The amount of sugar was measured using a Waters Arc3 HPLC using a Shodex SP0810 sugar column for separation with a water mobile phase at 0.6 mL/min at 85°C and a Waters model 2414 refractive index detector for detection.