Skip to main content

Molecular design Carbohydrate-binding scaffold



To advance the knowledge and utility of carbohydrates in plant bioscience and biomedicine by development and exploitation of carbohydrate-specific proteins.



The realization that microorganisms could live and thrive at temperatures ranging from the freezing point to above the boiling point of water have greatly expanded the range of possible conditions for enzyme catalysis and the identification of modules capable of binding under a variety of extreme conditions. Enzymes from thermophiles have the interesting property that they function optimally at high temperatures. The overall aim of the project is to evolve modules of thermostable glycosyl hydrolases (in particular xylanases) in order to understand and perfect their target specificity and thermostability properties and to employ these modules in biotechnology and bioanalysis. Our initial studies havebeen focused to a carbohydrate binding module of the xylanase 10A derived from Rhodothermus marinus.

We target genes encoding protein domains with genetic variability, clone the genetic libraries in vectors suitable for subsequent selection of domain variants with superior properties using phage display or array technology. This has allowed us to understand the basic properties of the modules and to eventually evolve reagents with improved properties. Such modules have a great potential for applications concerning the use of plant material as food and feed as well as for the utilization and basic analysis of fibres out of these materials.

Key findings and conclusions

We establish and use combinatorial library technology to find new solutions to basic and applied questions in the field of carbohydrate research.

In the past we have for instance reached the following achievements:

  • Thermostable carbohydrate binding module CBM4-2 has been adopted to molecular evolution by phage display technology.
  • CBM4-2 has been shown to have a capacity to evolve into variants with different carbohydrate specificities.
  • CBM4-2 chas been shown to have a capacity to evolve into variants with protein specificity.
  • CBM has been shown to be amendable to affinity maturation by molecular evolution.
  • Products obtained by molecular evolution of CBM4-2 have been shown retain protein structure, co-ordination of calcium ions and thermostability.
  • Carbohydrate-specific modules evolved from CBM4-2 have been shown to be excellent tools in histochemical analysis and in carbohydrate microarray assays.
  • Through evolution of a unique CBM that targets the XXXG building block we have developed a perfected tool to detect unfucosylated xyloglucan.
  • Through evolution of a unique CBM we have developed a perfected tool to detect xylan.
  • Crystal structures of evolved CBM in complex with ligand have been solved at high resolution (1 Å).
  • Structural features like CH-π interactions as well as structural adaptations of the binding site have been demonstrated to contribute to cross-reactivity of CBM.
  • Neutron crystallography adds substantial value to interpretation of hydrogen bonding networks (incl. those involving water) in CBM-carbohydrate interactions.

Scientific results



The scaffold and carbohydrate specific binders
We have constructed a combinatorial library (Cicortas Gunnarsson et al., 2004; Cicortas Gunnarsson et al., 2006a) based on the thermostable scaffold of the carbohydrate-binding module CBM4-2 derived from the xylanase produced by Rhodothermus marinus. Twelve residues within or in close proximity to the proposed xylan binding site were diversified (Figure 1) and the resulting library was displayed on filamentous phage.


Residues diversified in library of the thermostable carbohydrate-binding module CBM4-2

Figure 1. Residues diversified in combinatorial library using the scaffold of CBM4-2



Molecular variants binding xylan cellulose, mannan and a human glycoprotein (a monoclonal IgG4 antibody) were selected from the library. Selected clones were

  1. well folded as defined by CD
  2. had modified binding patterns against carbohydrates or had lost their xylan-binding capacity altogether while gaining another binding property (IgG4-specific molecular variants)
  3. retained a high level of productivity in E. coli (40-100 mg/l shake flask culture without optimization of production parameters)
  4. were largely thermostable mostly having a melting temperature of 75±5°C.

We conclude that this scaffold has properties very suitable for construction of combinatorial libraries and offers a good starting material to retrieve specific binders against a variety of targets.

Importantly, we have also been able to develop specific recognition units against non-fucosylated xyloglucan (Cicortas Gunnarsson et al. 2006c), a structure against which no specific recognition motif has been available in the past. By using molecular evolution we have also been able to isolate variants with substantially improved specificity against other targets like xylan (Cicortas Gunnarsson et al., 2007). These findings demonstrate the usefulness of using a molecular design approach of carbohydrate binding modules for the development of reagents with novel properties.

The structure of evolved CBM
Recently we determined the structure of a module (XG-34) derived from the CBM4-2 scaffold with specificity for xyloglucan by use of X-ray crystallography (Gullfot et al., 2010) (PDB: 
3JXS). This module retained the overall structure of CBM4-2 as determined by NMR although the R.M.S. difference between these modules was relatively large.

In particular the two loops that are important for formation of the carbohydrate binding cleft appeared to have moved closer together as compared to their distance in the wildtype module. The role of these modifications for creation of a specific binding site remained to be elucidated and future assessment of evolved CBM in complex with carbohydrate was considered important to provide evidence of such a role.

We were able to determine the structure of yet another module, X-2, derived from the scaffold both in the apo form and in complex with a synthetic oligomers representing the target, xylan (PDB: 2Y6J, 2Y6K, 2Y6L). We were also able to identify a variant of this module carrying a single mutation (leucine → phenylalanine) in the waist of the carbohydrate binding grove. This structure of this module without and with ligands could be determined at high resolution using X-ray crystallography (PDB: 2Y6H, 2Y6G, 2Y64).

This module, in contrast to X-2, binds not only xylan but also other carbohydrates like xyloglucan and β-glucan. As we could determine the structure of this module not only with oligomers of xylan but also with oligomers of glucose we were able to assess how ligand cross-reactivity was achieved (von Schantz et al., 2012). 

The ability of the X-2 L110F module but not the X-2 module to cross-react to these other targets was partly explained by an inability to accommodate the extra hydroxymethyl group next to the leucine in the X-2 module. It was also noted that cross-reactivity was linked to a rearrangement of some side chains of X-2 L110F to

  1. allow accommodation of the additional hydroxymethyl group of one of the glucose units of the target and to
  2. create now polar interactions that compensated for other interactions lost upon binding of this cross-reacting ligand (See figure 2) (von Schantz et al., 2012). Thus, CBM crossreactiviy was in part linked to a plasticity of the protein that allowed accommodation of other ligands in the binding site.



CBM X-2 in complex with xylopentaoseCBM X-2 L100F in complex with xylopentaoseCBM X-2 L100F in complex with cellopentaose
Figure 2. Crystal structure of X-2 (left) and X-2 L110F (middle, right) in complex with xylopentaose (magenta) and cellopentaose (pink; only three of five units are visable) (von Schantz et al., 2012). Three protein residues are highlighted, D29 (yellow), L/F110 (green) and R142 (cyan). Note difference in structure of D29 and R142 upon binding of cellopentaose. A few differences in polar interactions between complexes with xylo- and cellooligomers are indicated as dashed lines. 

Protein-carbohydrate interactions often depend on polar interactions mediated through hydrogen bonds. These can be assumed from structures determined using X-ray crystallography but they are in most cases not directly visible using this technology. Consequently we have attempted to exploit neutron diffraction, well-suited to visualize such interactions, to map these interactions of an engineered CBM (X-2 L110F) in complex with a xyloglucan heptameter.

In collaboration with the support lab of ESS (the European Spallation Source in Lund) we were able to obtain crystals of sufficient size (>1 mm3) to obtain a neutron diffraction data set at BIODIFF. A preliminary analysis demonstrated with confidence hydrogen bonds in the complex (Ohlin et al., 2015). Final analysis of the structure highlighted the water-mediated hydrogen bond network within the ligand binding site (Fisher et al., 2015). We with confidence propose that neutron diffraction will add substantial information in our quest to understand CBM-carbohydrate interactions.

Protein specific binders from the carbohydrate-specific scaffold
Interestingly, protein specific variants could be selected from the library of this carbohydrate binding module (Cicortas Gunnarsson et al., 2004). We have characterized these specificities in more detail (Cicortas Gunnarsson et al., 2006b) and we have showed that they target the protein itself and that they are not strongly dependent on an interaction with carbohydrates found on these proteins. Thus, this module has the capacity to transform itself into a protein-binding entity and we therefore conclude that it displays extensive plasticity with respect to the type of target it can bind.

Affinity maturation
Carbohydrate-specific binders regularly tend to be of low affinity. We have however recently (von Schantz et al., 2009) demonstrated that the affinity of an isolated binder retrieved from a library based on the CBM4-2 scaffold can be affinity-matured, much like antibodies, by an in vitro random mutagenesis approach to achieve 100-fold improvement in affinity with retained specificity. Modules derived from this scaffold can thus be perfected to fit applications requiring higher affinity that those initially found.

Applications based on these modules are being developed. In the initial phase we demonstrated their usefulness in chromatographic applications for the separation of oligosaccharides (Johansson et al., 2006). Further investigations have proven the utility of such engineered modules for the assessment of carbohydrate composition not only of plant tissue sections (Filonova et al., 2007; Sandquist et al., 2010; von Schantz et al., 2009) but also for the assessment of biotechnological processes designed to modify pulp fibers (Filonova et al., 2007) and of bioanalytical tools like oligosaccharide microarrays (Pedersen et al., 2012).

We envisage that modules derived from libraries based on the CBM4-2 scaffold will be highly useful in the assessment of the biology of plants as well as of biotechnological procedures. They will thereby have an important impact on carbohydrate research in the future.

Using bioinspired assemblies we have taken the analysis of CBM-carbohydrates one step further to assess the bindning behavior of CBM in complex matrixes (Paës et al., 2015). We have thereby initiated studies of the ways CBM are likely to perform in substrates they are likely to encounter when they are used in real applications beyond model systems commonly used to characterize them. We envisage that these systems will aid in future development of perfected CBM well-suited to fit a specific application.


  1. Cicortas Gunnarsson L, Nordberg Karlsson E, Albrekt A-S, Andersson M, Holst O and Ohlin M (2004) A carbohydrate binding module as a diversity-carrying scaffold. Protein Eng Des Sel 17, 213-221.     
  2. Cicortas Gunnarsson L, Nordberg Karlsson E, Andersson M, Holst O and Ohlin M (2006a) Molecular engineering of a thermostable carbohydrate-binding module. Biocatal Biotransfor 24, 31-37.     
  3. Cicortas Gunnarsson L, Dexlin L, Nordberg Karlsson E, Holst O and Ohlin M (2006b) Evolution of a carbohydrate binding module into a protein-specific binder. Biomol Eng 23, 111-117.
  4. Johansson R, Cicortas Gunnarsson L, Ohlin M and Ohlson S (2006) Thermostable carbohydrate-binding modules in affinity chromatography. J Mol Recognit 19, 275-281.     
  5. Cicortas Gunnarsson L, Zhou Q, Montanier C, Nordberg Karlsson E, Brumer H III and Ohlin M (2006c) Engineered xyloglucan specificity in a carbohydrate-binding module. Glycobiology 16, 1171-1180.     
  6. Cicortas Gunnarsson L (2007) Novel binding specificities engineered into the scaffold of a carbohydrate binding module. Doctoral thesis. (Abstract)
  7. Cicortas Gunnarsson L, Montanier C, Tunnicliffe RB, Williamson MP, Gilbert HJ, Nordberg Karlsson E and Ohlin M (2007) Novel xylan-binding properties of an engineered family 4 carbohydrate-binding module. Biochem J 406, 209-214.
  8. Filonova L, Cicortas Gunnarsson L, Daniel G and Ohlin M (2007) Synthetic xylan-binding modules for mapping of pulp fibres and wood sections. BMC Plant Biol 7, 54 (Abstract)
  9. von Schantz L, Gullfot F, Scheer S, Filonova L, Cicortas Gunnarsson L, Flint JE, Daniel G, Nordberg-Karlsson E, Brumer H and Ohlin M (2009) Affinity maturation generates greatly improved xyloglucan-specific carbohydrate binding modules. BMC Biotechnol 9, 92. (Abstract)
  10. Gullfot F, Tan T-C, von Schantz L, Nordberg Karlsson E, Ohlin M, Brumer H and Divne C (2010) The crystal structure of XG-34, an evolved xyloglucan-specific carbohydrate binding module. Proteins 78, 785-789. (Abstract) (protein structure)
  11. Sandquist D, Filonova L, von Schantz L, Ohlin M and Daniel G (2010) Microdistribution of xyloglucan in differentiating poplar cells. BioResources 5, 796-807. (pdf)
  12. von Schantz L, Håkansson M, Logan DT, Walse B, Österlin J, Nordberg-Karlsson E, Ohlin M (2012) Structural basis for carbohydrate binding specificity - a comparative assessment of two engineered carbohydrate binding modules. Glycobiology 22, 948-961. (Abstract)
  13. Pedersen HL, Fangel JU, McCleary B, Ruzanski C, Rydahl MG, Ralet M-C, Farkas V, von Schantz L, Marcos SE, Andersen MCF, Field R, Ohlin M, Knox JP, Clausen MH, Willats WGT (2012) Versatile high-resolution oligosaccharide microarrays for plant glycobiology and cell wall research. J Biol Chem 287, 39429-39438. (Abstract)
  14. von Schantz L (2012) Engineering protein-carbohydrate interactions - lessons from natural and evolved carbohydrate binding modules. Doctoral thesis. (Abstract)
  15. von Schantz L, Håkansson M, Logan DT, Nordberg Karlsson E, Ohlin M (2014) Polar interactions with branching xyloses and CH-π interactions define carbohydrate binding module recognition of xyloglucan. Proteins 82, 3466-3475. (Abstract)
  16. von Schantz L, Schagerlöf H, Nordberg-Karlsson E, Ohlin M (2014) Characterization of the substitution pattern of cellulose derivatives using carbohydrate-binding modules. BMC Biotechnol 14, 113. (Abstract)
  17. Ohlin M, von Schantz L, Schrader TE, Ostermann A, Logan D, Fisher SZ (2015) Crystallization, neutron data collection, initial structure refinement and analysis of a xyloglucan heptamer bound to an engineered carbohydrate binding module from xylanase. Acta Cryst F 71, 1072-1077. (Abstract)
  18. Paës G, von Schantz L, Ohlin M (2015) Bioinspired assemblies of plant cell wall polymers unravel affinity properties of carbohydrate-binding modules. Soft Matter 11, 6586-6594. (Abstract)
  19. Fisher Z, von Schantz L, Håkansson M, Logan DT, Ohlin M (2015) Neutron crystallographic studies reveal hydrogen bond and water mediated interactions between a carbohydrate-binding module and its bound carbohydrate ligand. Biochemistry 54, 6435–6438. (Abstract)

Contact information

Dr. Mats Ohlin

Dept. of Immunotechnology
Lund University
Medicon Village (Building 406)
S-223 81 LUND
telephone: +46-46-2224322


Mats Ohlin.

Professor Mats Ohlin

Head of the Department
+46 46 222 43 22
skype: matsohlin

Department of Immunotechnology
Lund University
Medicon Village
Building 406
223 81 LUND

Page Manager: | 2021-02-23