1. Introduction
Carbohydrate epimerases are a subclass of isomerases capable of inverting the configuration of hydroxyl groups of both substituted (phosphate and nucleotide diphosphate) and unsubstituted sugars [
1]. During recent years, more attention has been drawn towards the latter, evidenced by the growing number of characterized enzymes and industrial applications [
2,
3,
4,
5,
6,
7].
d-tagatose 3-epimerase, for example, plays an important role in the production of
d-psicose (showing potential as no-calorie sweetener and antidiabetic [
8]) from sucrose where it is combined with two other enzymes in a one-pot cascade reaction [
9]. Another example is cellobiose 2-epimerase, which catalyzes the interconversion of glucose to mannose at the reducing end of disaccharides (e.g., cellobiose and lactose) and oligosaccharides (e.g., cellotriose and cellotetraose) [
10]. Common applications are the production of epilactose starting from milk ultrafiltrate containing lactose or the production of lactulose from whey powder. Epilactose is regarded as a new prebiotic and lactulose is a potential low-calorie sweetener that could substitute for lactose in the food industry [
11,
12].
Even though cellobiose 2-epimerase enzymes have a rather relaxed substrate specificity, their activity is typically limited to oligosaccharides [
5]. Until today, only the enzyme from
Caldicellulosiruptor saccharolyticus (
CsCE) has been shown to also accept monosaccharides as substrate, a feature that was exploited for the production of mannose from glucose [
4]. However, similar enzymes might still exist as not all current representatives have been screened for their monosaccharide activity and, therefore, it is possible that interesting epimerase reactions might have gone unnoticed. This paper focuses on the cellobiose 2-epimerase from
Rhodothermus marinus, which has already been characterized regarding activity on oligosaccharides and has displayed the ability to produce epilactose from lactose efficiently. Furthermore, the enzyme was shown to be stable at temperatures up to 80 °C, making it highly attractive for potential production processes [
5]. In this research, the substrate specificity of
RmCE was explored with a particular emphasis on the C2 epimerization of galactose to talose, a rare sugar with potential applications in the pharmaceutical industry [
13,
14,
15,
16]. It can, for example, serve as a precursor molecule for several antibiotics of which the best example is Caminoside A, an antimicrobial compound isolated from the sponge
Caminus sphaeroconia, containing a 6-deoxytalose group [
13]. Moreover, the O2 and O3-methylated forms have been shown to be submillimolar inhibitors of galactose-binding galectin-4 and galectin-8, proteins which are involved in inflammation and cancer [
17].
Currently, talose can only be produced by a rather inefficient chemical process, requiring either high amounts of dibutyltin oxide [
18] or five consecutive chemical steps and a lot of solvents [
19]. Therefore, an efficient biochemical route would provide a clear added value to the production of talose.
3. Discussion
Most cellobiose 2-epimerase enzymes are characterized by a broad substrate specificity meaning that they can be used for the production of several important molecules (epilactose, mannose, lactulose) from their cheaper counterparts [
4,
11,
12,
30,
31,
32]. The specificity of cellobiose 2-epimerase from
Rhodothermus marinus was explored further and revealed side-activity on several monosaccharides comprising the conversion of
d-glucose to
d-mannose (comparable to
CsCE) and
d-galactose to
d-talose (a new reaction). As a consequence, the previous assumption that CE enzymes are only active on β-1,4-substituted sugars might be questioned and more cellobiose 2-epimerases might still exist in nature that exhibit monosaccharide activity. In order to further clarify the activity of CE enzymes on monosaccharides, galactose was docked in the active site of
RmCE and polar contacts were evaluated (
Figure S4). Interestingly, the residues responsible for binding of glucose (reducing end of cellobiose) are the same as for galactose, so no clear determinant for monosaccharide activity could be discovered. While all CE enzymes are probably able to recognize monosaccharides in their subsite, it remains unclear why only some representatives display significant activity. Moreover, the monosaccharide specificity among these representatives differs, as evidenced by the activity on galactose by
RmCE which was not present in the case of
CsCE.
Interestingly, the
kcat of the galactose to talose reaction turned out to be relatively high, even though it is only considered a minor activity. The combination of a medium-high activity, a stable enzyme and a cheap substrate led us to believe that a viable production process was possible. Moreover, the product talose, its derivatives and glycoconjugates are valuable molecules with several important industrial applications, like anti-tumor [
15] and antimicrobial [
33,
34] activities as well as recrystallization-inhibition (RI) activity [
35]. The only problem in the entire process was the by-product tagatose that was being formed through the isomerization side reaction of the enzyme, similar to the formation of fructose in the production of mannose by
CsCE [
4]. Therefore, reaction and process optimization were performed to chart the purity and yield of the product by varying substrate concentration and reaction time. This revealed that product purity remained >99% at the beginning of the reaction, but decreased steadily afterwards. In contrast, the product yield kept increasing during the course of the reaction, reaching a maximum yield of about 20%. Upscaling of the reaction was performed and resulted in 23 g/L of
d-talose with a purity of 86% and a yield of 8.5%. Depending on the application, a trade-off should be considered between purity and product yield, decreasing reaction time if the former is favored and increasing the reaction time if the latter is preferred. Alternatively, either preparative high-performance liquid chromatography (HPLC) [
36,
37] or simulated moving bed chromatography might be used to separate monosaccharides, as established by the production of psicose from sucrose with a three-enzyme cascade reaction [
9]. A more elegant solution might also be provided by adding an additional enzyme to the reaction mixture, creating a coupled process. This has already been successfully demonstrated in the production of
N-acetylneuraminic acid, the starting material for anti-influenza agents such as zanamivir and inavir, using an
N-acyl-
d-glucosamine 2-epimerase [
38].
In an attempt to increase the activity on monosaccharide substrates, two mutant libraries were constructed and in total 2700 mutants were screened for increased activity. Unfortunately, no improved mutants could be detected, indicating that amino acid positions Tyr-124, Tyr-307, Trp-321 and Trp-385 are not ideal hotspots for gaining kcat improvements. Targeting other positions, possibly further away from the active site, could prove to be the key to unlocking higher activities on mannose, galactose and other monosaccharide substrates.
Literature and our own results demonstrated that the
kcat of
RmCE is higher than that of
CsCE. As
CsCE already holds an application in the production of glucose from mannose, it was decided to try and transfer this property from
RmCE to
CsCE [
4]. Variant S99M/Q371F was able to accomplish this by doubling the
kcat while more or less retaining the same
Km value. This shift in activity can probably be attributed to the different orientation of Trp-385 (Trp-372 in
CsCE) effectuated by the presence of Phe-384 (Gln-371 in
CsCE) and Met-109 (Ser-99 in
CsCE).
In conclusion, this paper presents two enzymes that can be applied for industrial production processes: the wild-type RmCE which can efficiently produce d-talose from the cheap substrate d-galactose on the one hand; and a mutant CsCE with a higher kcat for the production of d-mannose on the other hand.