Ghent University Academic Bibliography

Advanced

Screening and engineering of sucrose phosphorylase for the glycosylation of small molecules

Dirk Aerts UGent (2012)
abstract
Glycosylation can significantly improve the physicochemical and biological properties of small molecules like vitamins, antibiotics, flavors, and fragrances. The chemical synthesis of glycosides is, however, far from trivial and involves multistep routes that generate lots of waste. Therefore using enzymes forms a good alternative, because they catalyze their reactions with both stricter specificities and higher yields. Sucrose phosphorylase (SP) is of particular interest, because it can transfer a glycosyl group to a wide variety of alternative acceptors. In addition, SP transfers the glucosyl moiety from both -glucose 1-phosphate and sucrose as donor substrate with sucrose preferably used because higher synthetic yields can be obtained and it is cheap. However, for many industrial relevant acceptor molecules SP’s natural activity is poor or negligible and so far no thermostable representatives are known. Therefore, this PhD thesis’ goal is to develop an efficient glycosylation technology for small molecules using SP as biocatalyst. To that end, the poor activity towards alternative acceptors and the insufficient stability are addressed by the search for new SP variants and enzyme engineering. However, natural enzymes can be exploited for the synthesis of glycosides, engineering the enzyme’s properties is expected to result in new applications of biocatalytic glycosylation reactions in various industrial sectors. Therefore, convenient and efficient high-throughput expression of enzyme libraries in mini-cultures is highly desirable. To that end, a constitutive expression system was developed that comprises four different promoters of varying strength. Applying this system reduces the amount of consumables and number of pipetting steps in high-throughput screening. The developed system was validated by the expression of different sucrose phosphorylase enzymes from Leuconostoc mesenteroides, Lactobacillus acidophilus and Bifidobacterium adolescentis in 96-deep- and low-well plates. The constitutive expression of SPs in low-well plates resulted in a level of activity that is equal or even better than what was achieved by inducible expression. Therefore, our plasmid set with varying constitutive promoters will be an indispensable tool to optimize enzyme expression for high-throughput screening. Next, the transglucosylation potential of these SP enzymes was compared using eighty putative acceptors from different structural classes. To increase the solubility of hydrophobic acceptors, the addition of various co-solvents was first evaluated. All enzymes were found to retain at least 50 % of their activity in 25 % dimethylsulfoxide, with the enzymes from Bifidobacterium adolescentis and Streptococcus mutans being the most stable. Screening of the enzymes’ specificity then revealed that the vast majority of acceptors are transglucosylated very slowly by SP, at a rate that is comparable to the contaminating hydrolytic reaction. However, high activity could only be detected on L-sorbose and L-arabinose, besides the native acceptors D-fructose and phosphate. Improving the affinity for alternative acceptors by means of enzyme engineering will, therefore, be a major challenge for the commercial exploitation of the transglucosylation potential of SP. Among those alternative acceptors, L-ascorbic acid is a very attractive molecule with high potential applications in cosmetics and tissue culturing. However, its low stability toward heat, light and other oxidizing agents hampers its applications. Its glucoside L-ascorbic acid-2-O-glucopyranoside, on the other hand, is very stable and finds applications in skin-care products. However efficient one-step production processes are so far not yet reported. Hence, protein engineering was applied to optimize the active site of SP of Bifidobacterium adolescentis for the transglucosylation of L-ascorbic acid. However, a screening assay was established based on the determination of released phosphate from -Glc-1-P in the transglucosylation reaction, no improved variants were found from saturation libraries from 6 positions that comprise the acceptor subsite of SP. The main raisons for the failure of the applied approach are expected to be the high noise in the screening assay and a too small sampling of the sequence space by our mutagenesis strategy. The former is attributed to the contaminating hydrolytic activity of the wild type SP that prevents the detection of small improvements in transglucosylation activity of variants in the screening assay. The latter can be solved by sampling a larger part of the sequence space, but more sparsely. A larger part of the sequences space is sampled when libraries are constructed involving multiple simultaneous mutations. However, the screening effort for these libraries increases exponentially with every additional position. To limit the screening effort, smaller but higher quality enzyme libraries are required. The use of sequence-function relationships, revealed by correlated mutation analysis, is a promising method, however, this method has rarely been applied by protein engineers to identify ‘hot spots’ to alter enzymes’ specificity. This is particularly caused by the lack of knowledge on how these co-evolution networks are related with different enzyme properties. Sequence-function relationships were deduced with the aim of finding hot spots that determine the substrate specificity of enzymes classified in the -amylase family. It was found that by changing the composition of the multiple sequence alignment, different co-evolution networks are found defining different enzyme properties. Substantial evidence for our predictions, identifying the role of the correlated positions, was found from mutational studies and by introducing a part of the network occurring in amylosucrase into sucrose phosphorylase, confirming that the correlated positions are also structurally correlated. Finally, the co-evolution networks and the predictions of their role presented in this work, provide protein engineers with invaluable information to optimize the acceptor specificity of sucrose phosphorylase or its paralogues for alternative acceptors. As discussed before, SP is a promising biocatalyst for the glycosylation of a wide range of acceptor molecules, but its industrial application has been hampered by the low thermostability of known representatives. In this work, homologous enzymes from thermophilic bacteria are disclosed for the first time. Three of them were recombinantly expressed in Escherichia coli and thoroughly characterized. Surprisingly, only the enzyme from Thermoanaerobacterium thermosaccharolyticum was found to have significant activity and is the most stable SP reported to date with a half-life of 60 hours at the industrially relevant temperature of 60°C. In contrast, the enzymes from Meiothermus silvanus and Spirochaeta thermophila did not catalyze the phosphorolysis of sucrose, although they were expressed in soluble form. All three sequences belong to glycosidase subfamily GH13_18, which is shown here to fall apart in two phylogenetic branches, of which only one contains active SP enzymes. Unfortunately, the true specificity of the other branch could not be identified, despite numerous efforts employing a wide range of substrates. Besides the isolation of thermostable variants from natural sources, protein engineering can also be applied to increase the stability of enzymes. Here, consensus engineering, mutating amino acids to the most frequently occurring amino acid of a multiple sequence alignment (MSA) has been applied to increase the stability of SP. The application of this concept over the entire sequence for enzymes, however, has been limited to a few examples. One of the difficulties applying consensus engineering to enzymes is that the MSAs of enzymes are not composed of independent sequences, but they are biased in terms of phylogenetic distribution, which conflicts with one of the premises of the concept. In this work, three consensus SPs were created based on three different MSAs with varying composition and were shown to be active stable SP enzymes. However, it was found that for improving the stability of enzymes by the consensus concept, the composition of the MSA is very important in the design strategy and should be composed as much as possible of independent sequences that fold to the same structure. In addition, from studies that determine the individual effects of consensus mutations, it has been observed that only 50 % of the consensus mutations are stabilizing and that multi-mutants of stabilizing mutations are not always more stable. Co-evolution has been assigned as one of the possible causes, and thus avoiding highly correlated positions in consensus libraries can increase the probability of finding stabilizing mutations. However, correlation mutation analysis revealed that by taking the consensus over the entire sequence, the co-evolution network of highly correlating positions is not disturbed in our consensus enzymes and thus correlation mutation analysis could be used as a tool to check if the MSA is composed of the right set of sequences. In contrast, at weakly correlating positions many differences are observed between the consensus enzymes and their natural variants. In conclusion, in this PhD thesis a constitutive expression system was developed that can be considered as a robust, convenient and effective system to accelerate high-throughput screening in protein engineering. In addition, a clear picture has been drawn on the opportunities and challenges of exploiting SP for the transglucosylation of alternative acceptors such as vitamins, flavors, fragrances or neutraceuticals. Unfortunately, protein-ligand interactions for the transglucosylation of L-ascorbic acid could not be introduced in SP by classical protein engineering techniques. However, an alternative procedure involving sequence-function relationships, which is new to the field of enzyme engineering, has been studied in detail and provides the enzyme engineer with hot spot positions for future engineering purposes. Moreover, a novel thermostable SP was discovered that is the most stable representative reported to date. Further, new insights were gained in the consensus engineering of enzymes, which allows protein engineers making better experimental set-ups for improving the thermostability of enzymes.
Please use this url to cite or link to this publication:
author
promoter
UGent and UGent
organization
alternative title
Screening en engineering van sucrosefosforylase voor de glycosylatie van kleine moleculen
year
type
dissertation
publication status
published
subject
keyword
Sucrose phosphorylase, Glycosylation, Evolution, Protein engineering, Substrate specificity, Thermostability
pages
238 pages
publisher
Ghent University. Faculty of Bioscience Engineering
place of publication
Ghent, Belgium
defense location
Gent : Het Pand (Bibliotheek)
defense date
2012-12-19 16:30
ISBN
9789059895836
language
English
UGent publication?
yes
classification
D1
additional info
dissertation consists of copyrighted material
copyright statement
I have transferred the copyright for this publication to the publisher
id
3071604
handle
http://hdl.handle.net/1854/LU-3071604
date created
2012-12-14 14:28:43
date last changed
2017-01-16 10:40:50
@phdthesis{3071604,
  abstract     = {Glycosylation can significantly improve the physicochemical and biological properties of small molecules like vitamins, antibiotics, flavors, and fragrances. The chemical synthesis of glycosides is, however, far from trivial and involves multistep routes that generate lots of waste. Therefore using enzymes forms a good alternative, because they catalyze their reactions with both stricter specificities and higher yields. Sucrose phosphorylase (SP) is of particular interest, because it can transfer a glycosyl group to a wide variety of alternative acceptors. In addition, SP transfers the glucosyl moiety from both \unmatched{f061}-glucose 1-phosphate and sucrose as donor substrate with sucrose preferably used because higher synthetic yields can be obtained and it is cheap. However, for many industrial relevant acceptor molecules SP{\textquoteright}s natural activity is poor or negligible and so far no thermostable representatives are known. Therefore, this PhD thesis{\textquoteright} goal is to develop an efficient glycosylation technology for small molecules using SP as biocatalyst. To that end, the poor activity towards alternative acceptors and the insufficient stability are addressed by the search for new SP variants and enzyme engineering. 
However, natural enzymes can be exploited for the synthesis of glycosides, engineering the enzyme{\textquoteright}s properties is expected to result in new applications of biocatalytic glycosylation reactions in various industrial sectors. Therefore, convenient and efficient high-throughput expression of enzyme libraries in mini-cultures is highly desirable. To that end, a constitutive expression system was developed that comprises four different promoters of varying strength. Applying this system reduces the amount of consumables and number of pipetting steps in high-throughput screening. The developed system was validated by the expression of different sucrose phosphorylase enzymes from Leuconostoc mesenteroides, Lactobacillus acidophilus and Bifidobacterium adolescentis in 96-deep- and low-well plates. The constitutive expression of SPs in low-well plates resulted in a level of activity that is equal or even better than what was achieved by inducible expression. Therefore, our plasmid set with varying constitutive promoters will be an indispensable tool to optimize enzyme expression for high-throughput screening.
Next, the transglucosylation potential of these SP enzymes was compared using eighty putative acceptors from different structural classes. To increase the solubility of hydrophobic acceptors, the addition of various co-solvents was first evaluated. All enzymes were found to retain at least 50 \% of their activity in 25 \% dimethylsulfoxide, with the enzymes from Bifidobacterium adolescentis and Streptococcus mutans being the most stable. Screening of the enzymes{\textquoteright} specificity then revealed that the vast majority of acceptors are transglucosylated very slowly by SP, at a rate that is comparable to the contaminating hydrolytic reaction. However, high activity could only be detected on L-sorbose and L-arabinose, besides the native acceptors D-fructose and phosphate. Improving the affinity for alternative acceptors by means of enzyme engineering will, therefore, be a major challenge for the commercial exploitation of the transglucosylation potential of SP.
Among those alternative acceptors, L-ascorbic acid is a very attractive molecule with high potential applications in cosmetics and tissue culturing. However, its low stability toward heat, light and other oxidizing agents hampers its applications. Its glucoside L-ascorbic acid-2-O-glucopyranoside, on the other hand, is very stable and finds applications in skin-care products. However efficient one-step production processes are so far not yet reported. Hence, protein engineering was applied to optimize the active site of SP of Bifidobacterium adolescentis for the transglucosylation of L-ascorbic acid. However, a screening assay was established based on the determination of released phosphate from -Glc-1-P in the transglucosylation reaction, no improved variants were found from saturation libraries from 6 positions that comprise the acceptor subsite of SP. The main raisons for the failure of the applied approach are expected to be the high noise in the screening assay and a too small sampling of the sequence space by our mutagenesis strategy. The former is attributed to the contaminating hydrolytic activity of the wild type SP that prevents the detection of small improvements in transglucosylation activity of variants in the screening assay. The latter can be solved by sampling a larger part of the sequence space, but more sparsely. 
A larger part of the sequences space is sampled when libraries are constructed involving multiple simultaneous mutations. However, the screening effort for these libraries increases exponentially with every additional position. To limit the screening effort, smaller but higher quality enzyme libraries are required. The use of sequence-function relationships, revealed by correlated mutation analysis, is a promising method, however, this method has rarely been applied by protein engineers to identify {\textquoteleft}hot spots{\textquoteright} to alter enzymes{\textquoteright} specificity. This is particularly caused by the lack of knowledge on how these co-evolution networks are related with different enzyme properties. Sequence-function relationships were deduced with the aim of finding hot spots that determine the substrate specificity of enzymes classified in the -amylase family. It was found that by changing the composition of the multiple sequence alignment, different co-evolution networks are found defining different enzyme properties. Substantial evidence for our predictions, identifying the role of the correlated positions, was found from mutational studies and by introducing a part of the network occurring in amylosucrase into sucrose phosphorylase, confirming that the correlated positions are also structurally correlated. Finally, the co-evolution networks and the predictions of their role presented in this work, provide protein engineers with invaluable information to optimize the acceptor specificity of sucrose phosphorylase or its paralogues for alternative acceptors. 
As discussed before, SP is a promising biocatalyst for the glycosylation of a wide range of acceptor molecules, but its industrial application has been hampered by the low thermostability of known representatives. In this work, homologous enzymes from thermophilic bacteria are disclosed for the first time. Three of them were recombinantly expressed in Escherichia coli and thoroughly characterized. Surprisingly, only the enzyme from Thermoanaerobacterium thermosaccharolyticum was found to have significant activity and is the most stable SP reported to date with a half-life of 60 hours at the industrially relevant temperature of 60{\textdegree}C. In contrast, the enzymes from Meiothermus silvanus and Spirochaeta thermophila did not catalyze the phosphorolysis of sucrose, although they were expressed in soluble form. All three sequences belong to glycosidase subfamily GH13\_18, which is shown here to fall apart in two phylogenetic branches, of which only one contains active SP enzymes. Unfortunately, the true specificity of the other branch could not be identified, despite numerous efforts employing a wide range of substrates. 
Besides the isolation of thermostable variants from natural sources, protein engineering can also be applied to increase the stability of enzymes. Here, consensus engineering, mutating amino acids to the most frequently occurring amino acid of a multiple sequence alignment (MSA) has been applied to increase the stability of SP. The application of this concept over the entire sequence for enzymes, however, has been limited to a few examples. One of the difficulties applying consensus engineering to enzymes is that the MSAs of enzymes are not composed of independent sequences, but they are biased in terms of phylogenetic distribution, which conflicts with one of the premises of the concept. In this work, three consensus SPs were created based on three different MSAs with varying composition and were shown to be active stable SP enzymes. However, it was found that for improving the stability of enzymes by the consensus concept, the composition of the MSA is very important in the design strategy and should be composed as much as possible of independent sequences that fold to the same structure. In addition, from studies that determine the individual effects of consensus mutations, it has been observed that only 50 \% of the consensus mutations are stabilizing and that multi-mutants of stabilizing mutations are not always more stable. Co-evolution has been assigned as one of the possible causes, and thus avoiding highly correlated positions in consensus libraries can increase the probability of finding stabilizing mutations. However, correlation mutation analysis revealed that by taking the consensus over the entire sequence, the co-evolution network of highly correlating positions is not disturbed in our consensus enzymes and thus correlation mutation analysis could be used as a tool to check if the MSA is composed of the right set of sequences. In contrast, at weakly correlating positions many differences are observed between the consensus enzymes and their natural variants. 
In conclusion, in this PhD thesis a constitutive expression system was developed that can be considered as a robust, convenient and effective system to accelerate high-throughput screening in protein engineering. In addition, a clear picture has been drawn on the opportunities and challenges of exploiting SP for the transglucosylation of alternative acceptors such as vitamins, flavors, fragrances or neutraceuticals. Unfortunately, protein-ligand interactions for the transglucosylation of L-ascorbic acid could not be introduced in SP by classical protein engineering techniques. However, an alternative procedure involving sequence-function relationships, which is new to the field of enzyme engineering, has been studied in detail and provides the enzyme engineer with hot spot positions for future engineering purposes. Moreover, a novel thermostable SP was discovered that is the most stable representative reported to date. Further, new insights were gained in the consensus engineering of enzymes, which allows protein engineers making better experimental set-ups for improving the thermostability of enzymes.},
  author       = {Aerts, Dirk},
  isbn         = {9789059895836},
  keyword      = {Sucrose phosphorylase,Glycosylation,Evolution,Protein engineering,Substrate specificity,Thermostability},
  language     = {eng},
  pages        = {238},
  publisher    = {Ghent University. Faculty of Bioscience Engineering},
  school       = {Ghent University},
  title        = {Screening and engineering of sucrose phosphorylase for the glycosylation of small molecules},
  year         = {2012},
}

Chicago
Aerts, Dirk. 2012. “Screening and Engineering of Sucrose Phosphorylase for the Glycosylation of Small Molecules”. Ghent, Belgium: Ghent University. Faculty of Bioscience Engineering.
APA
Aerts, Dirk. (2012). Screening and engineering of sucrose phosphorylase for the glycosylation of small molecules. Ghent University. Faculty of Bioscience Engineering, Ghent, Belgium.
Vancouver
1.
Aerts D. Screening and engineering of sucrose phosphorylase for the glycosylation of small molecules. [Ghent, Belgium]: Ghent University. Faculty of Bioscience Engineering; 2012.
MLA
Aerts, Dirk. “Screening and Engineering of Sucrose Phosphorylase for the Glycosylation of Small Molecules.” 2012 : n. pag. Print.