Discovery and design of novel glucoamylases suitable for raw starch hydrolysis at moderate temperatures via an integrated bioinformatics-assisted strategy

Abstract

Glucoamylases capable of hydrolyzing raw starch at moderate temperatures exhibit significant potential for the direct and efficient hydrolysis of raw starch. Using bioinformatics-assisted mining strategies, four candidate glucoamylases were initially identified from the GenBank database. Among these candidates, SeGA demonstrated high specific activities of 412.94 U/mg toward soluble starch and 143.51 U/mg toward raw corn starch. The half-life of SeGA was 192 h at 40 °C. Using SeGA as the initial enzyme, FuncLib design was performed to enhance its activity. The mutant SeGA-21 exhibited higher specific activity of 466.22 U/mg toward soluble starch and 173.98 U/mg toward raw corn starch, representing increases of 1.13-fold and 1.21-fold compared to SeGA, respectively. The half-life of SeGA-21 at 40 °C increased to 240 h, representing a 25% increase compared with SeGA. SeGA-21 can efficiently hydrolyze 30% raw corn starch, achieving a hydrolysis rate of approximately 31% after a 9-h reaction. When working synergistically with amylolytic enzymes, a hydrolysis rate of 71.5% was achieved for 30% raw corn starch at 40 °C within 15 h of reaction. Due to its ability to hydrolyze raw starch, moderate reaction temperature requirements, and excellent thermostability, SeGA-21 can be regarded as an efficient enzyme for the cold hydrolysis of raw starch and starch modification.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13068-025-02725-3.

Introduction

Starch, the second most abundant renewable and biodegradable natural polymer, has been widely utilized as a cost-effective and sustainable raw material in the food, pharmaceutical, and biofuel industries due to its unique physicochemical properties [8]. In the conventional process of starch–sugar production, a high temperature—typically around 100 °C—is required to disrupt the complex crystalline structure of starch. The subsequent enzymatic hydrolysis process involves α-amylase-mediated liquefication at approximately 90 °C and glucoamylase-catalyzed saccharification at approximately 60 °C [28]. Evidently, the current starch processing techniques are highly energy-intensive [8, 40]. In recent years, the processing of raw starch granules at moderate temperatures without prior treatment has attracted considerable attention. This approach aligns with the increasing demand for environmental preservation and sustainable practices in industrial processing.

Raw-starch amylolytic enzymes, which can digest raw starch at moderate temperatures, serve as the key enzymes for achieving the direct hydrolysis of raw starch and have emerged as a popular choice for the production of starch sugar [28]. Raw-starch amylolytic enzymes are primarily categorized into two types according to their specific roles in starch hydrolysis: raw-starch-degrading α-amylase and raw-starch-degrading glucoamylase. Among the amylolytic enzymes required for raw starch hydrolysis, raw-starch-digesting glucoamylase (RSDG) plays a crucial role in the hydrolysis of starch granules, as it can directly hydrolyze raw starch granules into glucose at moderate temperatures. The majority of glucoamylases are derived from fungi [28], including species from Aspergillus [12, 28], Penicillium (Sun et al., 2006, [38]), and Tetracladium [7]. In contrast, only a small proportion of glucoamylases are of bacterial origin, such as those from Saccharophagus degradans [34] and Caulobacter crescentus [22], Z. [41]. Among fungal glucoamylases, a few enzymes can directly hydrolyze raw starch granules into glucose, such as those from Aspergillus fumigatus [27, 28], Aspergillus clavatus [16], and Penicillium oxalicum [38]. These RSDGs are capable of degrading various types of raw starch with differing levels of efficiency; however, their specific activity toward raw starch is relatively low at moderate temperatures, such as an application temperature around 40 °C [27, 38]. Furthermore, the concentration of raw starch used for hydrolysis in most studies ranges from 10% to 15%, which is lower than that commonly employed in industrial processing [28]. The synergistic cold processing of highly concentrated raw starch, with concentrations ranging from 20% to 30%, has received limited research attention to date [8, 15, 25, 26]. This is primarily attributed to the limited number of α-amylases or glucoamylases reported to date that are capable of effectively hydrolyzing highly concentrated raw starch suspensions with concentrations ranging from 20% to 30% [25, 26].

In the post-genomic era, an efficient approach for obtaining novel catalysts is to exploit the vast amount of information available in genome databases, thereby eliminating the need for environmental sample collection [23]. The genomic mining strategy, which involves molecular modeling and structural bioinformatics, has enabled the discovery of enzymes that catalyze targeted reaction in a synthetic pathway [31]. The primary advantage of this approach lies in its effective sampling of specific enzymatic activities from sequence databases [31]. Here, an integrated strategy encompassing bioinformatics-assisted mining and semi-rational design was proposed to obtain novel raw starch digesting-glucoamylase suitable for cold starch hydrolysis. A glucoamylase-encoding gene, designated as SeGA, was identified from the database, and the recombinant enzyme SeGA exhibited high specific activity toward raw starch at moderate reaction temperatures. To further enhance the catalytic property of SeGA, FuncLib design was implemented. One FuncLib-design enzyme, SeGA-21, was screened out, exhibiting higher specific activity and excellent thermostability. To evaluate the application potential of SeGA and its mutant in raw corn starch hydrolysis, liquefaction and saccharification experiments were performed on high-concentration raw corn flour under moderate temperature conditions.

Materials and methods

Materials and culture medium

Escherichia coli JM109 was acquired from TransGen (Beijing, China) and used for gene cloning. Pichia pastoris GS115 and the plasmid pPIC9K were purchased from Invitrogen (Carlsbad, CA, USA). Soluble starch, amylose, and amylopectin were obtained from Sigma Chemical Co. (St. Louis, MO, USA). Rice, corn, bean, and wheat were purchased from the local supermarket (RT-Mart, Hefei, China). The samples were crushed using a food grinder and further ground with a mortar. All other chemicals and reagents were of analytical grade unless otherwise stated.

MD medium contained 13.4 g/L YNB, 0.4 mg/L biotin, and 20 g/L glucose. BMM medium contained 13.4 g/L YNB, 0.4 mg/L biotin, 2% (w/v) soluble starch, 0.5% (v/v) methanol, and 100 mM potassium phosphate buffer (pH 6.0). MD and BMM agar plates were prepared by supplementing the medium with 1.5% (w/v) agar. BMGY medium contained 10 g/L yeast extract, 20 g/L peptone, 13.4 g/L YNB, 0.4 mg/L biotin, 1% (v/v) glycerol, and 100 mM potassium phosphate buffer (pH 6.0). BMMY medium contained 13.4 g/L YNB, 0.4 mg/L biotin, 0.5% (v/v) methanol and 100 mM potassium phosphate buffer (pH 6.0).

Bioinformatics-assisted mining novel glucoamylase candidates

Two experimentally cold-adapted fungal glucoamylases were initially identified and retrieved from the GenBank database [36, 41]. They were employed as template for homologous search in NCBI databases (https://www.ncbi.nlm.nih.gov/). To eliminate redundancy, the CD-HIT online server was utilized to filter out sequences with a sequence identity cutoff of 0.9. Based on the average amino acid length of glucoamylases, proteins within the length range of 400–800 amino acids were retained, as this range corresponds to the typical length distribution of glucoamylases [24]. The sequences explicitly annotated as glucoamylase were retained for further analysis. A Hidden Markov Model (HMM) was then constructed using HMMER3.3 software and employed to search for candidates within the remaining sequences. The top 50 sequences with the highest scores, as assessed by HMMER3.3, were retained for further analysis.

Subsequently, the candidate sequences were analyzed for their physicochemical parameters using the ProtParam tool. Homology models were generated using the SWISS–MODEL online server, and their quality was assessed using the SAVES v6.0 online server. The structure of pentose was determined based on previously published literature [21]. Three-dimensional structure of pentaose, comprising α-1,4-glycosidic bonds, was obtained from PubChem (https://pubchem.ncbi.nlm.nih.gov/) and constructed using Chem3D software. AutoDock Vina was employed to dock the 3D structure of pentaose onto the catalytic sites of the candidate glucoamylases. The candidate enzymes were selected based on their affinity with the ligand and root mean square deviation (RMSD) values. Following the removal of sequences with more than 90% identity, codon optimization and sequence synthesis were performed for the remaining candidate sequences.

Expression and purification of the candidate glucoamylases

The plasmid pPIC9K harboring the synthetic gene was transformed into E. coli JM109 competent cells. The plasmids from positive clones were extracted and linearized using SacI. The linearized DNA fragments were transformed into P. pastoris GS115 by electrotransformation. The electroporated cells were plated onto MD agar plates and incubated at 28 °C for the selection of His⁺ transformants. Subsequently, the selected clones were cultured on BMM plates containing 2% soluble starch at 28 °C. A volume of 100 μL of methanol was carefully applied to the inner surface of the inverted plate lid every 24 h to induce expression. The positive clones were identified by the presence of a clear zone surrounding the transformant colonies.

Positive clones were selected and cultivated in 5 mL BMGY medium at 28 °C and 200 rpm. When OD₆₀₀ of cultures reached 2.0, cells were harvested by centrifugation at 3000×g and 4 °C for 10 min, and subsequently resuspended in 400 mL BMMY medium in 1 L flasks to a final cell density of 1.0. To induce enzyme expression, the cultures were incubated at 28 °C and 200 rpm, with the daily addition of 0.5% (v/v) methanol. After a 7-day incubation period, the aqueous culture broth was centrifuged at 6000×g for 30 min at 4 °C. The supernatant was dialyzed against a 50 mM citrate–phosphate buffer (pH 5.5) overnight and then centrifuged at 12,000×g for 20 min. Enzyme homogeneity was assessed by sodium dodecyl sulfate sulfate–polyacrylamide gel electrophoresis (SDS–PAGE) using a 12% polyacrylamide gel.

Enzyme assay

Enzyme activity was determined by measuring the reducing sugars released from the hydrolysis of raw corn flour or soluble starch using the dinitrosalicylic acid (DNS) assay [8, 12]. The reaction mixture contained 30 μL of enzyme solution and 570 μL of citrate–Na₂HPO₄ buffer (50 mM, pH 5.5) supplemented with 2% raw corn flour or 2% soluble starch. The mixture was incubated at 40 °C or 60  C for 10 min, and the reaction was terminated by adding 300 μL of DNS reagent. After heating in boiling water for 15 min, the amount of reducing sugars released was measured at 540 nm. One unit of glucoamylase activity was defined as the amount of enzyme required to release 1 μmol of reducing sugars, expressed as glucose, per minute under the standard assay conditions described above.

Enzymatic characteristics of the candidates

The effect of temperature on enzyme activity was determined over a range of 30 °C to 65 °C using 2% raw corn flour or soluble starch as substrates. The optimal pH of the candidates was measured in 50 mM citrate–Na₂HPO₄ buffer ranging from pH 3.5 to 6.5. Thermostability was determined by incubating the enzyme in citrate–Na₂HPO₄ buffer (50 mM, pH 5.5) at 40 °C and 50 °C. At appropriate intervals, the residual activity was measured using the DNS method as described above. pH stability was assessed by dispersing the enzyme in citrate–Na₂HPO₄ buffer (50 mM, pH 5.0–6.0), and the residual activity was subsequently measured. All experiments were conducted in triplicate, and values represent the mean ± standard deviation.

The effects of metal ions on enzyme activity were assessed under the standard assay conditions by supplementing the reaction mixture with Na⁺, K⁺, Li⁺, Ca²⁺, and Fe²⁺ at concentrations of 1.0 mM and 5.0 mM, respectively. The enzyme activity in the absence of metal ions addition was defined as 100%. In the reaction mixture containing additional effectors, the enzyme activity was determined. Relative activities were calculated as the ratio of enzyme activity in the presence of metal ions compared to that in their absence.

Kinetic constants and substrate specificity of the candidate glucoamylases

The kinetic constants, including K_m and k_cat, were determined using raw corn flour and soluble starch as the substrate based on the DNS assay. The reaction was performed by incubating the enzyme in citrate–Na₂HPO₄ buffer (50 mM, pH 5.5) supplemented with varying concentrations of raw corn flour or soluble starch (10–200 mg/mL), respectively. The reaction was carried out at 60 °C for 10 min. Kinetic parameters were determined by fitting the experimental data to the Lineweaver–Burk plot derived from the Michaelis–Menten equation using Origin 8.0 software.

The substrate specificity was determined under standard assay conditions using various substrates at a concentration of 2% (w/v), including amylose, amylopectin, raw rice flour, raw bean flour, raw corn flour and raw wheat flour. Enzyme activity was determined using the DNS method as described above.

Scanning electron microscopy

The morphology of raw corn starch before and after hydrolysis was visualized using a Scanning Electron Microscope (SEM, HITACHI S4800, Japan). The reaction was conducted using the candidate glucoamylase of 0.1 U/mg raw corn flour (U/mg RCF) and 30% raw corn flour. Following incubation at 40 °C for 3–12 h, the mixture was centrifuged at 8,000×g to recover the pellets. Then, the pellets were washed three times with ethanol and subsequently dried at 35 °C until a constant weight was achieved. The samples were fixed on a specimen holder using a silver plate and coated with Pt using Ion Sputter E-1010 at 5.0 kV and 20 mA for 40 s. The specimens were then viewed with SEM.

FuncLib design

The glucoamylase sequences from fungi, classified under EC number 3.2.1.3, were retrieved from the Carbohydrate-Active enZYmes database (CAZy, http://www.cazy.org/). After performing multiple sequence alignment using ClustalX 2.0, the sequence logo was generated with WebLogo3 (http://weblogo.threeplusone.com/create.cgi). Less-conserved sites within 6 Å of the active sites were targeted for FuncLib design using an automated, web-accessible server (http://FuncLib.weizmann.ac.il). Each mutant was designed to contain 3 to 5 mutation sites. Based on FuncLib calculations, the 50 designs with the lowest free energy were selected for subsequent docking analysis. The final selection of the top 5 designs was based on their binding affinity for pentaose as the ligand and their RMSD values. Subsequently, the candidate designs were synthesized commercially and expressed in P. pastoris, as described in the preceding sections.

Hydrolysis of raw corn starch by amylolytic enzymes at moderate temperatures

To evaluate the applicability of candidate glucoamylases in raw starch hydrolysis at moderate temperatures, the hydrolysis of high-concentration raw corn flour was performed using the candidate glucoamylases in combination with α-amylase AmyZ1:ΔTG and a commercial pullulanase (Shanghai Macklin Biochemical Technology Co., Ltd). AmyZ1 is an α-amylase with high specific activity toward raw corn starch, initially characterized by our research group in 2019 [8]. After modification, the mutant AmyZ1:ΔTG exhibited enhanced thermostability [6]. In this study, AmyZ1:ΔTG was employed in combination with the candidate glucoamylases to hydrolyze raw corn flour.

Crushed corn kernels were ground to a fine powder using a mortar and subsequently sieved through an 80-mesh screen. Raw corn flour at concentration of 30%, without any pretreatment, such as gelatinization, was utilized as the substrate. The purified AmyZ1:ΔTG and a commercial pullulanase were added at a final concentration of 4 U/mg raw corn flour (U/mg RCF) and 0.5 U/mg RCF, respectively [6, 8, 33]. The dosage of glucoamylase candidates was optimized, with the enzyme loading ranging from 1 to 4 mg/g RCF. The reaction mixture was incubated at 40 °C and 150 rpm for 18 h. After optimizing the enzyme dosage, the reaction time was optimized. Reaction mixtures containing either the candidate glucoamylases or AmyZ1:ΔTG were used as controls. At appropriate intervals, the samples were withdrawn and reducing sugars were measured by the DNS method. The rates of hydrolysis were calculated according to the formula described below, where the factor 0.9 represented the conversion factor resulting from the hydrolysis reaction.

$$\text{Hydrolysis rate}\left(\text{\%}\right)=[\text{Reducing sugars after hydrolysis}(\text{mg}/\text{mL})÷\text{Initial weight of raw corn flour}(\text{mg}/\text{mL})]\times 0.9\times 100$$

Statistical analysis

All the experiments were performed independently in triplicate. Data were represented as mean values with standard deviations (Mean±SD). SPSS Statistic 27 (SPSS Inc., Chicago, Illinois, USA) was used for statistical analysis. All data were analyzed using one-way ANOVA followed by the Welch–Duncan procedure between the means with SPSS software. p < 0.05 was regarded as statistically significant.

Results and discussion

Bioinformatic screening for novel glucoamylases

The majority of known glucoamylases exhibit an optimal reaction temperature of approximately 60 °C [41]. Only the mesophilic fungal glucoamylases from Tetracladium sp. [7] and Aspergillus tritici WZ99 [36] exhibited optimal reaction temperatures of 30 °C and 45 °C, respectively. Consequently, two experimentally characterized fungal glucoamylases with low-temperature catalytic properties, AmyT1 from Tetracladium sp. [7] and AtriGA15A from Aspergillus tritici [36], were selected as templates (Table S1). Employing the selected templates, 6405 putative glucoamylase sequences were identified from the NCBI database (Fig. 1). After non-redundant screening, a total of 2086 sequences were selected from the initial set. Based on the average amino acid length of glucoamylases, 951 sequences were retained within the length range of 400 to 800 amino acids. One hundred and eighty-five sequences were selected as they were explicitly annotated as glucoamylases. By utilizing the HMM constructed from two template sequences, the top 60 sequences with the highest scores as evaluated by HMMER 3.3 were selected for further analysis. This cutoff was selected, because the sequence scores declined more rapidly after rank 60 (Fig. S1a). After the prediction of physicochemical properties, a final set of 28 candidates were selected for further analysis. The glucoamylase from Penicillium oxalicum (PDB ID: 6FHV) was selected as the template for homology modeling [21], given its sequence identity with the candidate sequences exceeding 40%. Twenty-three candidates passed the quality evaluation conducted using the SAVES v6.0 online server. AutoDock Vina was employed to dock the three-dimensional structures of glucoamylases with a pentaose ligand consisting of α-1,4-glycosidic bonds (Fig. S1b). Six sequences possessed higher affinity with the ligand and greater reproducibility in docking analysis (Fig. S2). After multiple sequence alignment, four final candidate sequences were selected for synthesis and subsequent characterization (Fig. 1 and Table S2). Among the four candidate enzymes, those derived from Septoria linicola and Mytilinidion resinicola were found to contain carbohydrate-binding modules (CBMs), whereas the enzymes from Phaeomoniella chlamydospora and Botrytis cinerea did not possess such modules (Fig. S3).

Fig. 1.

Workflow for the bioinformatics-assisted mining of novel glucoamylase candidates. NH number of hits

To date, numerous isoforms of fungal glucoamylases exhibiting diverse biochemical characteristics have been reported in the literature [12]. The majority of glucoamylases have been obtained through traditional screening methods, which involve identifying bacterial or fungal strains with amylolytic activity, analyzing the genomes of these strains to identify potential encoding genes, and subsequently expressing and characterizing the novel candidate enzymes [7, 12, 28, 38]. In general, certain fungal species, such as Aspergillus and Penicillium spp., have frequently been screened out as strains exhibiting amylolytic activity. Consequently, the majority of glucoamylases have been derived from specific fungal species, such as Aspergillus and Penicillium spp. Currently, bioinformatic screening from extensive sequence databases has sufficient capability to identify promising enzyme genes with favorable biochemical characteristics, thereby reducing the necessity for direct environmental sampling [30, 31]. By adopting bioinformatic approaches, four candidate glucoamylase genes were identified in this study, originating from fungal species other than Aspergillus and Penicillium spp. For example, the candidate enzymes were derived from Septoria linicola (the causative agent of pasmo disease in flax) [14], Phaeomoniella chlamydospora (responsible for grapevine trunk diseases) [1], and Botrytis cinerea (a pathogen causing gray mold disease) [5]. In general, such pathogenic micro-organisms have not been considered as viable candidates for enzyme discovery due to their inherent pathogenic properties. For genes originating from pathogenic fungal species, their heterologous expression in a biosafe host, such as Pichia pastoris, can effectively mitigate biosafety risks. Higher expression levels can be achieved by optimizing the host expression system. Using this strategy, several genes encoding potential industrial biocatalysts have been cloned from Beauveria bassiana and successfully expressed in various host systems, including E. coli and Pichia pastoris [2]. Similarly, given their fungal origin, the candidate glucoamylase genes in this study will be expressed in Pichia pastoris with the aim of reducing biosafety risks and enhancing production efficiency.

Therefore, the primary advantage of bioinformatic-assisted screening lies in its capacity to efficiently identify specific enzymatic activities from sequence databases, irrespective of the microbial origin of the enzyme.

Expression and comparison of candidate glucoamylases

Codon-optimized genes encoding four candidate enzymes were cloned into the pPIC9K vector and recombinantly expressed in P. pastoris GS115. Two of the enzymes failed to be successfully expressed in P. pastoris, whereas the other two were purified and designated as PcGA (Fig. S4) and SeGA (Fig. 2a). Given their fungal origin, the candidate glucoamylase genes in this study have been expressed in Pichia pastoris. SeGA exhibited a relatively high protein expression level, whereas PcGA showed a comparatively low expression level. Regarding PcGA, the use of alternative expression hosts, such as glycoengineered strains of Komagataella phaffii, could enhance the protein expression level.

Fig. 2.

SDS–PAGE analysis of SeGA and its mutants. a SDS–PAGE analysis of SeGA and SeGA-21. Line 1: SeGA after purification; Line 2: SeGA-21 after purification; b SDS–PAGE analysis of other mutants. Line 1: SeGA-4 after purification; Line 2: SeGA-5 after purification; 3: SeGA-17 after purification; 4: SeGA-35 after purification. M: protein molecular weight marker

The glucoamylase PcGA exhibited an optimum temperature of 45 °C and retained approximately 60% residual activity at 30 °C (Fig. S5a). PcGA exhibited the optimum activity at pH 4.5 (Fig. S5b), which was comparable to the acidic reaction conditions of most glucoamylases [12]. However, the specific activity of PcGA was relatively low, with only 73.51 U/mg against soluble starch, and it exhibited no detectable hydrolytic activity toward raw starch substrates, including corn, rice, wheat, and bean flour (Table S2). The other one, SeGA, had an optimal temperature of 60 °C and maintained approximately 40% residual activity at around 40 °C for soluble starch and raw corn starch (Fig. 3a). SeGA exhibited an optimal reaction pH of 4.5 for soluble starch and 5.5 for raw corn flour (Fig. 3b). SeGA remained stable within the temperature range of 40 to 50 °C (Fig. 3c). After incubation at 40 °C for 192 h, approximately 50% of the original activity was retained (Fig. 3c). The half-life was approximately 24 h when incubated at 50 °C (Fig. 3c). SeGA exhibited excellent stability, with greater stability at pH 5.5 compared to pH 5.0 and 6.0 (Fig. 3d).

Fig. 3.

Effects of temperature and pH on activity and stability of SeGA. a Optimal temperature for SeGA activity using raw corn flour and soluble starch as substrates. The assays were measured at temperatures from 40 °C to 65 °C in citrate–Na₂HPO₄ buffer (50 mM); b optimal pH for SeGA activity using raw corn flour and soluble starch as substrates. The optimum pH was determined in citrate–Na₂HPO₄ buffer (50 mM, pH 4.0–6.5); c effect of temperature on enzyme stability. The purified enzyme was incubated in citrate–Na₂HPO₄ buffer (50 mM, pH 5.5) at 40 and 50 °C; d effect of pH on enzyme stability. The purified enzyme was incubated in 50 mM citrate–Na₂HPO₄ buffer of pH 5.0–6.0. The remaining activities were determined at appropriate intervals. All the results were expressed as the average of triplicate experiments

SeGA exhibited a high specific activity of 412.94 U/mg toward soluble starch, and an additional specific activity of 143.51 U/mg toward raw corn starch (Table 1). SeGA displayed hydrolysis capacity toward various raw starch, with specific activities of 179.90 U/mg, 53.86 U/mg, and 37.71 U/mg for raw rice, wheat and bean starch, respectively (Table 1). SeGA exhibited higher specific activity toward amylose than toward amylopectin, indicating its preference for linear substrates (Table 1). The divalent ion Ca²⁺ had a minimal effect on enzyme activity, resulting in only approximately a 5% increase in enzyme activity. Other metal ions tested, including Fe²⁺, Na⁺, Li⁺ and K⁺, exhibited inhibitory effects on enzyme activity (Table S3). Using soluble starch and raw corn starch as substrates, the K_m values of SeGA were determined as 34.78 and 86.31 mg/mL, respectively, with corresponding catalytic efficiencies (k_cat/K_m) of 13.51 and 5.11 (Table 2).

Table 1.

Specific activity of SeGA toward raw starch and soluble starch

Substrates	Specific activity(U/mg)
	40 ℃	60 ℃
Raw corn flour	61.78±2.87	143.51±3.33
Raw rice flour	59.57±3.99	179.90±3.50
Raw wheat flour	26.14±1.52	53.86±2.80
Raw bean flour	15.12±0.22	37.71±3.14
Soluble starch	165.21±6.12	412.94±6.43
Amylose	60.33±3.23	118.17±5.80
Amylopectin	7.17±0.38	18.18±0.89

Table 2.

Kinetic parameters of SeGA and SeGA-21

Substrate	Enzyme	Km (mg/mL)	kcat (s-1)	kcat/Km (mL mg-1 s-1)
Soluble starch	SeGA	34.78±1.21	469.76±19.71	13.51±1.21
	SeGA-21	33.49±1.22	599.32±16.92	17.90±1.25
Raw corn flour	SeGA	86.31±2.31	441.13±12.32	5.11±1.22
	SeGA-21	81.55±2.42	561.25±14.12	6.88±1.31

PcGA possessed the optimum temperature at a moderate temperature and maintained more than 50% residual activity within the temperature range of 30–50 °C. However, PcGA did not exhibit catalytic activity toward raw starch. In contrast to PcGA, SeGA did not exhibit an optimal catalytic temperature within a moderate temperature range, despite both enzymes being selected using cold-active fungal glucoamylases as templates in the bioinformatics approach. Fortunately, SeGA exhibited a higher specific activity compared to several known fungal glucoamylases, with a specific activity of 412.94 U/mg, exceeding the activity of certain enzymes from Aspergillus spp. and Penicillium spp. (Table 3). Among fungal glucoamylases, several enzymes can directly hydrolyze raw starch granules into glucose, such as those from Aspergillus fumigatus, Aspergillus clavatus, and Penicillium oxalicum (Table 3). However, only a few of these enzymes exhibited high specific activity toward raw starch at 40 °C. The mutant of pRSGA from Aspergillus fumigatus exhibited the specific activity of approximately 100 U/mg toward raw corn starch, whereas others showed activities of around 50 U/mg (Table 3). By contrast, SeGA exhibited higher specific activity toward raw corn starch at 40 and 60 °C, with a specific activity of 61.78 and 143.51 U/mg, respectively, placing it among the enzymes with high activities reported in the literature (Table 3). In addition to raw corn starch, SeGA also exhibited hydrolysis activity toward a variety of raw starch, including rice, wheat, and bean starch (Table 1).

Table 3.

Catalytic characteristics of some known glucoamylases and SeGA

Name	Source	Optimum temperature (oC)	Thermostability	Specific activity (U/mg)		References
				Soluble starch	Raw corn starch
Nm	Fusarium solani	40 ℃	Half-life of 182 min at 50 °C	603	Nm	Bhatti et al., 2007
SeGA	Septoria linicola	60 ℃	Half-life of 192 h at 40 °C Half-life of 24 h at 50 °C	412.94	143.51 at 60 °C 61.78 at 40 °C	This study
SeGA-21	Septoria linicola	60 ℃	Half-life of 240 h at 40 °C Half-life of 28 h at 50 °C	466.22	173.98 at 60 °C 78.29 at 40 °C	This study
TSgam-2	Talaromyces stipitatus	60 ℃	Half-life of more than 300 min at 55 °C and 40 min at 60 °C	360	Nm	[37]
Nm	Aspergillus japonicus	65 ℃	About 75% residual activity after 1 h at 50 °C	301.7	Nm	[18]
GA	Aspergillus niger	60 ℃	Nm	216	Nm	[20]
Nm	Aspergillus clavatus	50 ℃	Remained almost fully active for 28–32 h at 40 °C	205.4	14.3 at 50 °C	[16]
Nm	Aspergillus niger	70 ℃	Half-life time of 20–25 min at 60 ℃	95	Nm	[3]
PoGA15A	Penicillium oxalicum	65 ℃	Half-life of 1 h at 55 °C	81.2	23.7 at 40 °C	[38]
GlaAb	Aspergillus brasiliensis	60 ℃	About 67% residual activity after incubation at 50 °C for 7 h	72	Nm	[39]
pRSGA	Aspergillus fumigatus	70 ℃	About 78% residual activity at 55 °C for 2 h	42.56	48.38 at 40 °C	[28]
pRSGA: N422Q	Aspergillus fumigatus	60 ℃	Nm	Nm	51.19 at 40 °C	[28]
afRSGA: Q113M	Aspergillus fumigatus	70 ℃	Half-life of about 80 min at 60 °C	Nm	≈100 at 40 °C	[28]
Nm	Aspergillus tritici WZ99	45 ℃	About 65% residual activity after 1-h incubation at 45 °C	39.64	Nm	[36]
TmGLA	Tricholoma matsutake	60 ℃	Half-life of 1 h at 60 °C	13.2	Nm	[17]
Nm	Curvularia lunata	50 ℃	About 70% residual activity after 6 h incubation at 50 °C	12.34	Nm	[9]
Nm	Tetracladium sp.	30 ℃	About 10% residual activity after incubation for 1 h at 50 °C	Nm	Nm	[7]
GA2	Aspergillus flavus	70 ℃	About 58% residual activity after incubation at 70 °C for 1 h	Nm	Nm	[12]
RSDG	Penicillium sp. X-1	65 ℃	About 50% residual activity after incubation at 70 °C for 4 h	Nm	58.7 at 60 °C	Sun et al., 2006

Nm Not mentioned.

In general, the saccharification of raw starch typically involves an extended reaction period [28, 38], thereby requiring the amylolytic enzymes to exhibit excellent thermostability. Among the glucoamylases employed for raw starch hydrolysis, the thermostability of the enzymes exhibited relatively low thermostability. For instance, the enzymes from Penicillium oxalicum and Aspergillus clavatus had a half-life of no more than 5 h at 50–55 °C [16, 38] (Table 3). In comparison, SeGA exhibited excellent thermal stability at the application temperature for raw starch hydrolysis, featuring a half-life of 192 h at 40 °C and approximately 24 h at 50 °C. Based on the comprehensive catalytic performance of the candidate enzymes, SeGA demonstrated considerable application potential in raw starch hydrolysis under moderate temperature conditions, owing to its favorable catalytic properties, including optimal reaction temperature, enhanced thermostability, and high specific activity.

Action pattern of SeGA on raw corn starch

The action pattern of SeGA on raw corn starch was further explored using Scanning Electron Microscope (Fig. 4). The untreated starch granules remained intact and smooth (Fig. 4a). Following hydrolysis for 3–6 h, small holes became detectable on the surface of the starch granules, and the hydrolysis rate reached approximately 20%. (Fig.4b, c). When hydrolysis was extended to 9 h, the hydrolysis rate reached up to 30% and the holes become increasingly deeper and more numerous (Fig. 4d). SEM images confirmed that SeGA hydrolyzed raw corn starch granules through punching holes in a random attack mode. Similarly, it has been reported that many pinholes were observed on raw corn starch granules when a raw starch-hydrolyzing glucoamylase was used for hydrolysis [11, 19]. Glucoamylase, either alone or in combination with alpha-amylase, has been utilized for starch modification, such as the preparation of porous starch [4, 10, 11, 35]. In general, α-amylase is capable of creating wide pores on starch granules, whereas glucoamylase can form sharp and deep channels within starch granules [10]. The combination of raw starch hydrolyzing α-amylase and glucoamylase may contribute to the development of tailor-made porous starch with a distinct internal structure [10].

Fig. 4.

Scanning electron microscopy of raw corn starch granules hydrolyzed by SeGA. a Raw corn starch before hydrolysis; b raw corn starch hydrolyzed with SeGA for 3 h; c raw corn starch hydrolyzed with SeGA for 6 h; d raw corn starch hydrolyzed with SeGA for 12 h

FuncLib design of SeGA for improved enzymatic activity

A total of 306 sequences of fungal glucoamylases were retrieved from NCBI with EC number of 3.2.1.3. According to the sequence logo analysis, two catalytic sites were identified to be Glu206 and Glu427 (Fig. 5, SeGA numbering). FuncLib design involved seven less-conserved sites within 6 Å around the active sites, including Gln339, Tyr345, Ala426, Tyr429, Leu437, Ser438, and Ala440 (Figs. 5 and 6). The first 50 designs in FuncLib with the lowest free energy were selected for further analysis (Fig. S6). Five designs with high affinity for the pentasaccharide molecule and superior RMSD core values in docking analysis were selected for experimental validation. The genes of five FuncLib designs—SeGA-4, SeGA-5, SeGA-17, SeGA-21 and SeGA-35—were commercially synthesized and expressed in P. pastoris (Table 4 and Fig. 2b).

Fig. 5.

Less-conserved sites within 6 Å around the active sites of glucoamylases. Sequences were aligned with ClustalX and sequence logos were generated using WebLogo3. Two conserved catalytic residues, E206 and E427, were marked with a hollow triangle symbol. FuncLib design involved seven less-conserved sites within 6 Å around the active sites. These sites were marked with a solid triangle symbol

Fig. 6.

Amino acids selected for mutagenesis following FuncLib design. Two conserved catalytic residues, E206 and E427 were colored in red. The mutation sites in the FuncLib design included Gln339, Tyr345, Ala426, Tyr429, Leu437, Ser438, and Ala440. These sites were marked in color

Table 4.

Mutation sites and specific activity of designs in FuncLib

Designs	Mutation sites						Specific activity (U/mg)
	Q339	Y345Y345Y345	A426	L437	S438	A440	Soluble starch (60 °C)	Raw corn starch (40 °C)
SeGA-4	Q339N	Y345F			S438G	A440L	255.46±3.73	48.92±3.13
SeGA-5		Y345F	A426H	L437T	S438G	A440E	189.55±2.43	21.31±2.13
SeGA-17		Y345F	A426H	L437T	S438G	A440E	212.53±9.32	31.81±1.54
SeGA-21	Q339N		A426H		S438G	A440E	466.22±13.63	78.29±6.15
SeGA-35	Q339H	Y345F			S438T	A440M	223.43±19.42	47.33±2.69

Y429 was not included as a target site for mutation in Funclib design

Compared with the starting enzyme, SeGA-21 exhibited specific activity of 466.22 and 173.98 U/mg toward soluble starch and raw corn starch, respectively, corresponding to 1.13-fold and 1.21-fold of SeGA for each substrate. The optimum temperature and pH of SeGA-21 were comparable to those of SeGA (Fig. 7a, b). The thermostability of SeGA-21 was further enhanced, as evidenced by a half-life of 240 h at 40 °C and 28 h at 50 °C (Fig. 7c, d). Using soluble starch and raw corn starch as substrates, the K_m values of SeGA-21 were determined to be 33.49 and 81.55 mg/mL, respectively, both of which were lower than those of SeGA (Table 2). In addition, SeGA-21 exhibited increased k_cat values of 599.32 and 561.25 s^-1 toward two substrates, representing a 1.28-fold and 1.27-fold of SeGA, respectively (Table 2). Consequently, the catalytic efficiency (k_cat/K_m) of SeGA-21 was 1.32-fold and 1.35-fold of SeGA toward two substrates, respectively.

Fig. 7.

Effects of temperature and pH on activity and stability of SeGA-21. a Optimal temperature for SeGA-21 activity using raw corn flour as substrate. The assays were measured in citrate–Na₂HPO₄ buffer (50 mM, pH 5.5); b optimal pH for SeGA-21 activity using raw corn flour as substrate. The optimum pH was determined in 50 mM citrate–Na₂HPO₄ buffer; c effect of temperature on enzyme stability; d effect of pH on enzyme stability. The residual activities was determined at appropriate intervals. All the results were expressed as the average of triplicate experiments

FuncLib is an efficient strategy for designing a small set of stable, efficient, and functionally diverse multipoint active-site mutants to increase the catalytic rate or specificity of enzymes [13, 32]. In the design of FuncLib, the mutant SeGA-21 exhibited enhanced specific activity and thermostability. The specific activity of SeGA-21 toward soluble starch was slightly lower than that of the enzyme from Fusarium solani but higher than those of several enzymes from Aspergillus spp. and Penicillium spp. (Table 3). Notably, the specific activity of SeGA-21 toward raw corn starch exceeded certain values reported in the literature (Table 3). In the design of FuncLib, SeGA-21 was predicted to demonstrate a higher affinity for the substrate. Consistent with this observation, SeGA-21 exhibited a lower K_m value than the parental enzyme, further supporting the conclusion that its substrate affinity was enhanced. SeGA-21 exhibited a higher k_cat value, suggesting its elevated catalytic efficiency and thereby contributing to the observed increase in specific activity.

In the docking analysis of SeGA-21, the distances between the catalytic residues (E206 and E427) and the substrate were shorter compared to those of SeGA (Fig. 8a, b). Thus, the substrate was more likely to interact with the catalytic residues to form the transition state required for the reaction, which may account for the observed enhancement in specific activity. In SeGA, the amino acids surrounding A426 and A440 were predominantly hydrophilic (Fig. 8c). When A426 and A440 were mutated to the hydrophilic amino acids in SeGA-21, the hydrophilic interactions among these residues were enhanced, likely contributing to the stabilization of the protein’s three-dimensional structure (Fig. 8d).

Fig. 8.

Docking analysis and modeled structures of SeGA and SeGA-21. a Docking analysis of SeGA and pentaose; b docking analysis of SeGA-21 and pentaose. H-bonds were shown in yellow dotted lines. Two conserved catalytic residues, E206 and E427 were colored in purple in (a) and (b); c amino acids surrounding A426 and A440 in SeGA. The sites before mutation were highlighted in gold. d Amino acids surrounding A426 and A440 in SeGA-21. The sites after mutation were highlighted in green. Two conserved catalytic residues, E206 and E427 were colored in red in (c) and (d)

Efficient hydrolysis of raw corn starch under moderate temperature conditions by amylolytic enzymes

The synergistic effects of raw starch hydrolyzing α-amylase (AmyZ1:ΔTG), glucoamylase (SeGA/SeGA-21), and commercial pullulanase was investigated at a moderate temperature of 40 °C under non-gelatinized conditions. The dosage of SeGA or SeGA-21 was optimized to 2.0 mg/g RCF when using 30% raw corn flour as the substrate (Fig. 9a). The hydrolysis rates reached approximately 20% when using AmyZ1:ΔTG or SeGA alone after a 6-h reaction. Even when the reaction time was extended to 15 h, the hydrolysis rates did not exceed 30% (Fig. 9b). The hydrolysis rate reached approximately 31% when using SeGA-21 alone after a 9-h reaction. However, the rate did not increase significantly even when the reaction time was extended. Following a 15-h incubation period, the combined action of AmyZ1:ΔTG, SeGA, and commercial pullulanase achieved a hydrolysis rate of 62.9%. When SeGA was replaced with SeGA-21, the hydrolysis rate increased to 71.5%, which was 1.2-fold of that achieved using SeGA.

Fig. 9.

Hydrolysis of raw corn flour at moderate temperatures by amylolytic enzymes. a Effect of SeGA and SeGA-21 dosage on the yield of reducing sugars. AmyZ1:ΔTG and the commercial pullulanase were added at a final concentration of 4 U/mg raw corn flour (U/mg RCF), and 0.5 U/mg RCF, respectively. b Saccharification of raw corn flour at moderate temperatures. The amylolytic enzymes included 4 U/mg RCF of AmyZ1:ΔTG, 2.0 mg/g RCF of SeGA or SeGA-21, and 0.5 U/mg RCF of commercial pullulanase. Different letters in the figure indicated significant differences between groups (p < 0.05)

The synergistic hydrolysis of concentrated raw starch using α-amylases and glucoamylases in cold starch process has been investigated in only a limited number of studies [25, 26]. This can be primarily attributed to the limited number of known amylolytic enzymes that are capable of effectively hydrolyzing highly concentrated raw starch suspensions [25, 26]. As indicated in Table 5, the raw starch-digesting glucoamylases from Penicillium oxalicum, Aspergillus fumigatus, and Penicillium sp. X-1 were employed to hydrolyze raw flour in combination with some commercial amylolytic enzymes. The reaction temperature of 40 °C was commonly employed in enzymatic reactions, and the substrate concentration typically ranged from 15% to 25% (Table 5). A reaction time of 36 h was required to achieve a 77.2% hydrolysis rate for 25% raw corn starch using pRSGA from Aspergillus fumigatus at 40 °C in the presence of α-amylase and pullulanase [28]. A higher hydrolysis rate toward raw corn starch was achieved using the raw starch-digesting glucoamylase PoGA15A in combination with α-amylase after a 24-h reaction at 40 °C. However, the concentration of raw corn flour used was restricted to only 15% [38]. An alternative strategy to achieve a superior hydrolysis effect is to increase the reaction temperature. For instance, a reaction temperature of 65  C was necessary for the raw-starch-digesting glucoamylase from Penicillium sp. X-1 to achieve approximately 92.4% hydrolysis rate of 15% raw corn flour as the substrate (Table 5) (Sun et al., 2006). Neither increasing the temperature nor prolonging the culture time contributes to achieving the efficient hydrolysis of high concentration raw starch with low energy consumption and low cost. As reported by Bing et al., AmyZ1:ΔTG, which was derived from AmyZ1, was an efficient α-amylase for raw starch digestion and featured an improved catalytic property [6, 8]. As demonstrated in this study, SeGA-21 was an efficient glucoamylase suitable for raw starch hydrolysis at moderate temperatures. When AmyZ1:ΔTG and SeGA-21 were used in combination to synergistically hydrolyze raw starch, a 71.5% hydrolysis rate of 30% raw corn starch (as the substrate) was achieved after a 24-h reaction at 40 °C. In comparison, the advantages of their combined action lay in achieving a high hydrolysis rate within a short reaction time using high-concentration raw corn flour as the substrate. Thus, SeGA and its mutant may serve as promising candidates for integration into current enzyme cocktails and as a beneficial supplement to existing glucoamylases in the starch processing industry. Furthermore, the use of raw-starch amylolytic enzymes enables the processing of raw starch granules at moderate temperatures without prior gelatinization treatment, which may represent a promising approach for energy-efficient starch processing.

Table 5.

Hydrolysis rate of raw corn flour catalyzed by amylolytic enzymes

Origin of RSDG	Combination of amylolytic enzymes	Reaction condition	Hydrolysis rate	Refs
Septoria linicola	2.0 mg/g SeGA, 4 U/mg AmyZ1:ΔTG, 0.5 U/mg commercial pullulanase	30% raw corn flour, 40 ℃ and 15 h	62.9%	This study
	2.0 mg/g SeGA-21, 4 U/mg AmyZ1:ΔTG, 0.5 U/mg commercial pullulanase	30% raw corn flour, 40 ℃ and 15 h	71.5%	This study
Penicillium oxalicum	0.05 U/mg PoGA15A, 0.05 U/mg commercial α-amylase	15% raw corn flour, 40 ℃ and 24 h	Approximately 100%	[38]
Aspergillus fumigatus	0.2 U/mg pRSGA, 0.5 U/mg commercial α-amylase, 0.1 U/mg commercial pullulanase	20% raw corn flour, 40 ℃ and 36 h	92.2%	[28]
		25% raw corn flour, 40 ℃ and 36 h	77.2%
Penicillium sp. X-1	0.07 U/mg RSDG, 0.03 U/mg commercial α-amylase*	15% raw corn flour, 65 ℃ and 2 h	92.4%	[29]

^*: The enzyme dosage was calculated based on the literature, which specified a low dose of 1000 U RSDG and 500 U α-amylase for 15 g of raw corn starch.

Conclusion

Through the implementation of an integrated bioinformatics-assisted strategy, a novel glucoamylase, SeGA, capable of hydrolyzing raw starch at moderate temperatures, was successfully identified. SeGA demonstrated hydrolysis capacity toward various types of raw starch, including corn, rice, wheat, and bean starch. After modification using FuncLib, the mutant SeGA-21 demonstrated enhanced specific activity and thermostability. SeGA-21 can work synergistically with amylolytic enzymes, effectively hydrolyzing high-concentration raw corn flour within a short reaction time. Different from conventional starch processing methods, this process was carried out without requiring gelatinization or high-temperature liquefaction. Due to its ability to hydrolyze raw starch, moderate reaction temperature requirements, and excellent thermostability, SeGA-21 can be considered as an efficient enzyme for the cold hydrolysis of raw starch and starch modification.

Author contributions

SG: investigation, formal analysis and data curation. YS: formal analysis and data curation. TY and XZ: formal analysis and data curation. XZ: review and editing. YZ: software and methodology. YX: funding acquisition, and review and editing. WF: conceptualization, writing, review and editing, and supervision.

Funding

This work was financed by the National Key Research and Development Program [grant number: 2022YFC2805101] and the Key Science and Technology Project of Anhui Province [Grant number: 202423l10050019].

Availability of data and materials

Data are provided within the manuscript or supplementary information files.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interest.