Structure, function and subcellular localization of the potato Resistance protein Rx1

E.J. Slootweg

Research output: Thesisinternal PhD, WU

Abstract

Resistance proteins are part of the plant’s immune system and mediate a defence response upon recognizing their cognate pathogens. They are thought to be present in the cell as part of a larger protein complex. The modular architecture of R proteins suggests that they form a scaffold for various interacting proteins, involved in pathogen recognition, downstream signalling or protein stabilization. However, few common interactors have been found for the CC-NB-ARC domains despite extensive screenings for downstream interactors. The objective of thesis was to get new insights in the structure, function and localization of R proteins by using the potato resistance genes Rx1 and Gpa2 as a model system. Initially, a novel T7 phage display method was developed to facilitate high throughput selection of interacting molecules (Chapter 2). However, the use of a T7 cDNA phage library to identify interactors of the CC-NB-ARC domains of Rx1 resulted in the discovery of a large set of highly basic peptides (Chapter 3). In Chapter 4, the functional role of the CC-NB-ARC domain in mediating disease resistance was explored by creating chimeric proteins between Rx1 and Gpa2. This resulted in the observation that the CC-NB-ARC is able to confer both virus and nematode resistance in potato. Furthermore, it was shown that the CC-NB-ARC of Rx1 and the LRR of Gpa2 are incompatible and vice versa. This phenomenon was studied in more detail in Chapter 5, in which a docking model for the interacting surface of these domains was constructed based on the individual structural domains. Finally, the subcellular localisation was investigated to get a better understanding about the R proteins function in the cell (Chapter 6).

The lytic T7 phages form a powerful platform for the display of large cDNA libraries and offer the possibility to screen for strong interactions with a variety of substrates. To visualise these interactions directly by fluorescence microscopy, we constructed fluorescent T7 phages by exploiting the flexibility of phages to incorporate modified versions of its capsid protein (Chapter 2). By applying translational frameshift sequences, helper plasmids were constructed that expressed a fixed ratio of both wild-type capsid protein (gp10) and capsid protein fused to enhanced yellow fluorescent protein (EYFP). The frameshift sequences were inserted between the 3’-end of the capsid gene and the sequence encoding EYFP. Fluorescent fusion proteins are only formed when the ribosome makes a -1 shift in reading frame during translation. As far as we know this is the first report of using a translational frameshift for a biotechnological purpose. The phages formed in this way have capsids composed of three different variants of their capsid protein; EYFP-fused versions derived by frameshift translation, non-fused versions derived by regular translation from the helper plasmid, and versions that display peptides encoded in the library ligated in the phage genome. Using standard fluorescence microscopy, we could sensitively monitor the enrichment of specific binders in a cDNA library displayed on fluorescent T7 phages. Closely monitoring the effect of the selection procedure enables fine tuning, and obviates the need for more laborious ELISA or plaque lift assays. Furthermore, with the fast pace of developments in single molecule detection technologies and sorting systems, these fluorescent phages open the way to high throughput platforms for the direct selection of binding molecules.

In Chapter 3, cDNA phage display was applied as an alternative method to identify additional downstream Rx1 interactors, which could further resolve the Rx1 signalling pathway. In a pilot experiment the value of T7 phage display to identify specific interactors was demonstrated by using an antibody raised against the PVY coat protein. Screening of a PVY-infected N. benthamiana cDNA phage display library resulted in the selection of peptides harbouring the known PRIKAI epitope. Next, phage display was explored as technique to discover proteins interacting with the potato R protein Rx1. The system turned out to be prone to pick up interactors binding to matrices like Ni-NTA or to fusion proteins like thioredoxine. A possible way to circumvent this weakness was to design the selection procedure in such a way that it alternates between different matrices and to limit the number of selection round. This adapted approach resulted in the identification of a series of highly basic protein fragments and random peptides, for which a specific interaction could be shown. Two cDNA sequences encoded the ribosomal proteins L19 and L36a, which showed a stunted growth phenotype upon gene silencing in N. benthamiana using VIGS and a slightly reduced Rx-mediated HR.

The nematode resistance protein Gpa2 and the virus resistance protein Rx1 provide an excellent test system to investigate the exchangeability of recognition and signalling domains and explore the evolutionary flexibility of R proteins, for they confer resistance to completely unrelated pathogens (Globodera pallida and potato virus X, respectively). In Chapter 4, we provide evidence for the hypothesis, that, via intergenic sequence exchanges and various types of mutations, NB-LRR proteins have the potential to alter resistance specificities towards taxonomically unrelated pathogens in relatively short evolutionary time periods. Both the regulatory sequences and CC-NB domains of the paralogs Gpa2 and Rx1 are non-pathogen specific and exchangeable. Remarkably, the genetic fusions of the CC-NB of Rx1 with the LRR of Gpa2 (Rx1CN/Gpa2L) and the reciprocal domain swap (Gpa2CN/Rx1L) were not functional when driven by the endogenous promoters or 35S promoter. Gain of wild type resistance was obtained by re-introducing the first five LRRs of Rx1 in Rx1CN/Gpa2, restoring the compatibility between the N-terminal part of the LRR and the ARC2 domain. Decreasing the expression levels for Gpa2CN/Rx1L resulted in extreme resistance against PVX, indistinguishable from wild type plants. Our results indicate that not only coding sequences, but that also optimizing the expression levels may play a role in generating novel resistances.

The CC, NB-ARC, and LRR domains of the Rx1 and Gpa2 proteins interact with each other and recognition of the elicitor mediated by the LRR is translated in an activation of the NB-ARC. The available functional and evolutionary data make Gpa2 a suitable candidate for structural modelling of the individual domains and their interaction (Chapter 5). A structural model of the NB-ARC / LRR interaction could function as a framework for the interpretation of known empirical data and the design of new experiments to test R protein operational mechanisms (Zhang 2009). Therefore, computer aided modelling of the 3D structure domain models for the NB-ARC and the LRR domains were obtained and used as basis for a domain docking study. The functional interaction between the domains was studied via a detailed analysis of their incompatibility in chimeric Gpa2 and Rx1 proteins. A large set of sequence exchanges between the two proteins was created for that purpose. Both in the LRR and in the ARC2 domain small regions could be identified in which the amino acids differing between Gpa2 and Rx1 led to domain incompatibility. Five of the ARC2 positions required for LRR compatibility and three known autoactivating positions from the RX1 LRR were used as constraints in domain docking computation to limit the potential search space. The resulting docking model indicated an important role in the NB-ARC-LRR interaction for electrostatic and hydrophobic interactions. A loop region rich in acidic residues in the ARC2 domain was found close in space to a patch of basic residues grouped together in the LRR. Hydrophobic residues on both the NB and the ARC2 contacted hydrophobic residues on the surface of the LRR. A correlation analysis of the NB-ARC and LRR subdomains detected coevolution between the interacting surfaces, which supports a direct interaction between these two domains. Site-directed mutagenesis and pull-down experiments were used to test the role of surface features that might play an important role in the interdomain docking interface.

In Chapter 6, we have made use of the characteristic of the Rx1 protein that it remains functional when its domains are co-expressed as separate polypeptides. This allowed us to create fluorescent constructs, not only of the full length protein, but also of the separate subdomains. Most of these tagged constructs still form functional proteins. C- and N-terminal fusions of Green Fluorescent Protein (GFP) variants to Rx, made it possible to study its subcellular localization in Nicotiana benthamiana cells. Contrary to our expectations we observed the presence of Rx1 in both the cytoplasm and the nucleus. Rx1 does not contain known nuclear localization signals and the size of the protein (140 kDa including GFP) exceeds the limit for passive diffusion through the nuclear pore. Fluorescent fusions of a series of deletion constructs, CC-NB-ARC, NB-ARC, NB-ARC-LRR, CC and LRR showed three distinct patterns of subcellular localization. The NB-ARC-LRR and LRR constructs have a cytoplasmic localization and are mostly absent in the nucleus. The NB-ARC and CC-NB-ARC constructs showed equal fluorescence intensities in both the nucleus and the cytoplasm. The CC alone fused to GFP, however, seems to preferentially accumulate in the nucleus resulting in a three to four times higher fluorescence intensity in the nucleus compared to the cytoplasm. The diffusion behaviour inside the nucleus for both the complete CC and a CC fragment containing the two predicted helices downstream of the central turn, showed that their nuclear accumulation coincides with a significantly reduced nuclear diffusion as compared to unfused GFP and the other CC fragments. This difference might point to a potential interaction between the CC and an unknown nuclear component. Furthermore, SGT1 and Rar1 are thought to function as chaperones involved in stabilizing R proteins. Both the silencing experiments with these two proteins and the P-loop mutation show that the nuclear localisation of Rx1 is probably conformation dependent. Two approaches were followed to see if CP recognition or Rx1 signalling pathway were linked to a certain cellular compartment. At one hand the Rx1 protein itself or its subdomains were directed to either the nucleus or the cytoplasm by fusion to exogenous targeting signals (Nuclear Export Signals or Nuclear Localization Signals). On the other hand the elicitor, the PVX coat protein, was directed to the nucleus or cytoplasm. The PVX coat protein is a much smaller protein and can under normal circumstances diffuse freely between the cytoplasm and the nucleus. The surprising result was that no effect was found for retargeting Rx1, but when the elicitor was targeted to the nucleus, it could not activate Rx1 anymore, indicating that recognition might to take place in the cytoplasm.

In the final chapter, the results obtained in this thesis are put into perspective by studying parallels in scientific literature on NB-LRR proteins with similar functions in other organisms.


Original languageEnglish
QualificationDoctor of Philosophy
Awarding Institution
  • Wageningen University
Supervisors/Advisors
  • Bakker, Jaap, Promotor
  • Schots, Arjen, Co-promotor
  • Goverse, Aska, Co-promotor
Award date23 Oct 2009
Place of Publication[S.l.
Publisher
Print ISBNs9789085854678
Publication statusPublished - 2009

Fingerprint

potatoes
proteins
bacteriophages
coat proteins
cytoplasm
green fluorescent protein
peptides
translation (genetics)
nuclear localization signals
capsid
DNA libraries
pathogens
fluorescence microscopy
cDNA libraries
plasmids
promoter regions
Nematoda
fluorescence
recombinant fusion proteins
screening

Keywords

  • solanum tuberosum
  • plant proteins
  • pathogenesis-related proteins
  • binding proteins
  • resistance mechanisms
  • disease resistance
  • plant viruses
  • plant parasitic nematodes
  • bacteriophages
  • gene expression
  • phage display

Cite this

@phdthesis{1911e2f32da8456e90e38fc40e4c3334,
title = "Structure, function and subcellular localization of the potato Resistance protein Rx1",
abstract = "Resistance proteins are part of the plant’s immune system and mediate a defence response upon recognizing their cognate pathogens. They are thought to be present in the cell as part of a larger protein complex. The modular architecture of R proteins suggests that they form a scaffold for various interacting proteins, involved in pathogen recognition, downstream signalling or protein stabilization. However, few common interactors have been found for the CC-NB-ARC domains despite extensive screenings for downstream interactors. The objective of thesis was to get new insights in the structure, function and localization of R proteins by using the potato resistance genes Rx1 and Gpa2 as a model system. Initially, a novel T7 phage display method was developed to facilitate high throughput selection of interacting molecules (Chapter 2). However, the use of a T7 cDNA phage library to identify interactors of the CC-NB-ARC domains of Rx1 resulted in the discovery of a large set of highly basic peptides (Chapter 3). In Chapter 4, the functional role of the CC-NB-ARC domain in mediating disease resistance was explored by creating chimeric proteins between Rx1 and Gpa2. This resulted in the observation that the CC-NB-ARC is able to confer both virus and nematode resistance in potato. Furthermore, it was shown that the CC-NB-ARC of Rx1 and the LRR of Gpa2 are incompatible and vice versa. This phenomenon was studied in more detail in Chapter 5, in which a docking model for the interacting surface of these domains was constructed based on the individual structural domains. Finally, the subcellular localisation was investigated to get a better understanding about the R proteins function in the cell (Chapter 6). The lytic T7 phages form a powerful platform for the display of large cDNA libraries and offer the possibility to screen for strong interactions with a variety of substrates. To visualise these interactions directly by fluorescence microscopy, we constructed fluorescent T7 phages by exploiting the flexibility of phages to incorporate modified versions of its capsid protein (Chapter 2). By applying translational frameshift sequences, helper plasmids were constructed that expressed a fixed ratio of both wild-type capsid protein (gp10) and capsid protein fused to enhanced yellow fluorescent protein (EYFP). The frameshift sequences were inserted between the 3’-end of the capsid gene and the sequence encoding EYFP. Fluorescent fusion proteins are only formed when the ribosome makes a -1 shift in reading frame during translation. As far as we know this is the first report of using a translational frameshift for a biotechnological purpose. The phages formed in this way have capsids composed of three different variants of their capsid protein; EYFP-fused versions derived by frameshift translation, non-fused versions derived by regular translation from the helper plasmid, and versions that display peptides encoded in the library ligated in the phage genome. Using standard fluorescence microscopy, we could sensitively monitor the enrichment of specific binders in a cDNA library displayed on fluorescent T7 phages. Closely monitoring the effect of the selection procedure enables fine tuning, and obviates the need for more laborious ELISA or plaque lift assays. Furthermore, with the fast pace of developments in single molecule detection technologies and sorting systems, these fluorescent phages open the way to high throughput platforms for the direct selection of binding molecules. In Chapter 3, cDNA phage display was applied as an alternative method to identify additional downstream Rx1 interactors, which could further resolve the Rx1 signalling pathway. In a pilot experiment the value of T7 phage display to identify specific interactors was demonstrated by using an antibody raised against the PVY coat protein. Screening of a PVY-infected N. benthamiana cDNA phage display library resulted in the selection of peptides harbouring the known PRIKAI epitope. Next, phage display was explored as technique to discover proteins interacting with the potato R protein Rx1. The system turned out to be prone to pick up interactors binding to matrices like Ni-NTA or to fusion proteins like thioredoxine. A possible way to circumvent this weakness was to design the selection procedure in such a way that it alternates between different matrices and to limit the number of selection round. This adapted approach resulted in the identification of a series of highly basic protein fragments and random peptides, for which a specific interaction could be shown. Two cDNA sequences encoded the ribosomal proteins L19 and L36a, which showed a stunted growth phenotype upon gene silencing in N. benthamiana using VIGS and a slightly reduced Rx-mediated HR. The nematode resistance protein Gpa2 and the virus resistance protein Rx1 provide an excellent test system to investigate the exchangeability of recognition and signalling domains and explore the evolutionary flexibility of R proteins, for they confer resistance to completely unrelated pathogens (Globodera pallida and potato virus X, respectively). In Chapter 4, we provide evidence for the hypothesis, that, via intergenic sequence exchanges and various types of mutations, NB-LRR proteins have the potential to alter resistance specificities towards taxonomically unrelated pathogens in relatively short evolutionary time periods. Both the regulatory sequences and CC-NB domains of the paralogs Gpa2 and Rx1 are non-pathogen specific and exchangeable. Remarkably, the genetic fusions of the CC-NB of Rx1 with the LRR of Gpa2 (Rx1CN/Gpa2L) and the reciprocal domain swap (Gpa2CN/Rx1L) were not functional when driven by the endogenous promoters or 35S promoter. Gain of wild type resistance was obtained by re-introducing the first five LRRs of Rx1 in Rx1CN/Gpa2, restoring the compatibility between the N-terminal part of the LRR and the ARC2 domain. Decreasing the expression levels for Gpa2CN/Rx1L resulted in extreme resistance against PVX, indistinguishable from wild type plants. Our results indicate that not only coding sequences, but that also optimizing the expression levels may play a role in generating novel resistances. The CC, NB-ARC, and LRR domains of the Rx1 and Gpa2 proteins interact with each other and recognition of the elicitor mediated by the LRR is translated in an activation of the NB-ARC. The available functional and evolutionary data make Gpa2 a suitable candidate for structural modelling of the individual domains and their interaction (Chapter 5). A structural model of the NB-ARC / LRR interaction could function as a framework for the interpretation of known empirical data and the design of new experiments to test R protein operational mechanisms (Zhang 2009). Therefore, computer aided modelling of the 3D structure domain models for the NB-ARC and the LRR domains were obtained and used as basis for a domain docking study. The functional interaction between the domains was studied via a detailed analysis of their incompatibility in chimeric Gpa2 and Rx1 proteins. A large set of sequence exchanges between the two proteins was created for that purpose. Both in the LRR and in the ARC2 domain small regions could be identified in which the amino acids differing between Gpa2 and Rx1 led to domain incompatibility. Five of the ARC2 positions required for LRR compatibility and three known autoactivating positions from the RX1 LRR were used as constraints in domain docking computation to limit the potential search space. The resulting docking model indicated an important role in the NB-ARC-LRR interaction for electrostatic and hydrophobic interactions. A loop region rich in acidic residues in the ARC2 domain was found close in space to a patch of basic residues grouped together in the LRR. Hydrophobic residues on both the NB and the ARC2 contacted hydrophobic residues on the surface of the LRR. A correlation analysis of the NB-ARC and LRR subdomains detected coevolution between the interacting surfaces, which supports a direct interaction between these two domains. Site-directed mutagenesis and pull-down experiments were used to test the role of surface features that might play an important role in the interdomain docking interface. In Chapter 6, we have made use of the characteristic of the Rx1 protein that it remains functional when its domains are co-expressed as separate polypeptides. This allowed us to create fluorescent constructs, not only of the full length protein, but also of the separate subdomains. Most of these tagged constructs still form functional proteins. C- and N-terminal fusions of Green Fluorescent Protein (GFP) variants to Rx, made it possible to study its subcellular localization in Nicotiana benthamiana cells. Contrary to our expectations we observed the presence of Rx1 in both the cytoplasm and the nucleus. Rx1 does not contain known nuclear localization signals and the size of the protein (140 kDa including GFP) exceeds the limit for passive diffusion through the nuclear pore. Fluorescent fusions of a series of deletion constructs, CC-NB-ARC, NB-ARC, NB-ARC-LRR, CC and LRR showed three distinct patterns of subcellular localization. The NB-ARC-LRR and LRR constructs have a cytoplasmic localization and are mostly absent in the nucleus. The NB-ARC and CC-NB-ARC constructs showed equal fluorescence intensities in both the nucleus and the cytoplasm. The CC alone fused to GFP, however, seems to preferentially accumulate in the nucleus resulting in a three to four times higher fluorescence intensity in the nucleus compared to the cytoplasm. The diffusion behaviour inside the nucleus for both the complete CC and a CC fragment containing the two predicted helices downstream of the central turn, showed that their nuclear accumulation coincides with a significantly reduced nuclear diffusion as compared to unfused GFP and the other CC fragments. This difference might point to a potential interaction between the CC and an unknown nuclear component. Furthermore, SGT1 and Rar1 are thought to function as chaperones involved in stabilizing R proteins. Both the silencing experiments with these two proteins and the P-loop mutation show that the nuclear localisation of Rx1 is probably conformation dependent. Two approaches were followed to see if CP recognition or Rx1 signalling pathway were linked to a certain cellular compartment. At one hand the Rx1 protein itself or its subdomains were directed to either the nucleus or the cytoplasm by fusion to exogenous targeting signals (Nuclear Export Signals or Nuclear Localization Signals). On the other hand the elicitor, the PVX coat protein, was directed to the nucleus or cytoplasm. The PVX coat protein is a much smaller protein and can under normal circumstances diffuse freely between the cytoplasm and the nucleus. The surprising result was that no effect was found for retargeting Rx1, but when the elicitor was targeted to the nucleus, it could not activate Rx1 anymore, indicating that recognition might to take place in the cytoplasm. In the final chapter, the results obtained in this thesis are put into perspective by studying parallels in scientific literature on NB-LRR proteins with similar functions in other organisms.",
keywords = "solanum tuberosum, planteiwitten, pathogenesis-gerelateerde eiwitten, bindende eiwitten, resistentiemechanismen, ziekteresistentie, plantenvirussen, plantenparasitaire nematoden, bacteriofagen, genexpressie, faag display, solanum tuberosum, plant proteins, pathogenesis-related proteins, binding proteins, resistance mechanisms, disease resistance, plant viruses, plant parasitic nematodes, bacteriophages, gene expression, phage display",
author = "E.J. Slootweg",
note = "WU thesis, no. 4699",
year = "2009",
language = "English",
isbn = "9789085854678",
publisher = "S.n.",
school = "Wageningen University",

}

Slootweg, EJ 2009, 'Structure, function and subcellular localization of the potato Resistance protein Rx1', Doctor of Philosophy, Wageningen University, [S.l..

Structure, function and subcellular localization of the potato Resistance protein Rx1. / Slootweg, E.J.

[S.l. : S.n., 2009. 176 p.

Research output: Thesisinternal PhD, WU

TY - THES

T1 - Structure, function and subcellular localization of the potato Resistance protein Rx1

AU - Slootweg, E.J.

N1 - WU thesis, no. 4699

PY - 2009

Y1 - 2009

N2 - Resistance proteins are part of the plant’s immune system and mediate a defence response upon recognizing their cognate pathogens. They are thought to be present in the cell as part of a larger protein complex. The modular architecture of R proteins suggests that they form a scaffold for various interacting proteins, involved in pathogen recognition, downstream signalling or protein stabilization. However, few common interactors have been found for the CC-NB-ARC domains despite extensive screenings for downstream interactors. The objective of thesis was to get new insights in the structure, function and localization of R proteins by using the potato resistance genes Rx1 and Gpa2 as a model system. Initially, a novel T7 phage display method was developed to facilitate high throughput selection of interacting molecules (Chapter 2). However, the use of a T7 cDNA phage library to identify interactors of the CC-NB-ARC domains of Rx1 resulted in the discovery of a large set of highly basic peptides (Chapter 3). In Chapter 4, the functional role of the CC-NB-ARC domain in mediating disease resistance was explored by creating chimeric proteins between Rx1 and Gpa2. This resulted in the observation that the CC-NB-ARC is able to confer both virus and nematode resistance in potato. Furthermore, it was shown that the CC-NB-ARC of Rx1 and the LRR of Gpa2 are incompatible and vice versa. This phenomenon was studied in more detail in Chapter 5, in which a docking model for the interacting surface of these domains was constructed based on the individual structural domains. Finally, the subcellular localisation was investigated to get a better understanding about the R proteins function in the cell (Chapter 6). The lytic T7 phages form a powerful platform for the display of large cDNA libraries and offer the possibility to screen for strong interactions with a variety of substrates. To visualise these interactions directly by fluorescence microscopy, we constructed fluorescent T7 phages by exploiting the flexibility of phages to incorporate modified versions of its capsid protein (Chapter 2). By applying translational frameshift sequences, helper plasmids were constructed that expressed a fixed ratio of both wild-type capsid protein (gp10) and capsid protein fused to enhanced yellow fluorescent protein (EYFP). The frameshift sequences were inserted between the 3’-end of the capsid gene and the sequence encoding EYFP. Fluorescent fusion proteins are only formed when the ribosome makes a -1 shift in reading frame during translation. As far as we know this is the first report of using a translational frameshift for a biotechnological purpose. The phages formed in this way have capsids composed of three different variants of their capsid protein; EYFP-fused versions derived by frameshift translation, non-fused versions derived by regular translation from the helper plasmid, and versions that display peptides encoded in the library ligated in the phage genome. Using standard fluorescence microscopy, we could sensitively monitor the enrichment of specific binders in a cDNA library displayed on fluorescent T7 phages. Closely monitoring the effect of the selection procedure enables fine tuning, and obviates the need for more laborious ELISA or plaque lift assays. Furthermore, with the fast pace of developments in single molecule detection technologies and sorting systems, these fluorescent phages open the way to high throughput platforms for the direct selection of binding molecules. In Chapter 3, cDNA phage display was applied as an alternative method to identify additional downstream Rx1 interactors, which could further resolve the Rx1 signalling pathway. In a pilot experiment the value of T7 phage display to identify specific interactors was demonstrated by using an antibody raised against the PVY coat protein. Screening of a PVY-infected N. benthamiana cDNA phage display library resulted in the selection of peptides harbouring the known PRIKAI epitope. Next, phage display was explored as technique to discover proteins interacting with the potato R protein Rx1. The system turned out to be prone to pick up interactors binding to matrices like Ni-NTA or to fusion proteins like thioredoxine. A possible way to circumvent this weakness was to design the selection procedure in such a way that it alternates between different matrices and to limit the number of selection round. This adapted approach resulted in the identification of a series of highly basic protein fragments and random peptides, for which a specific interaction could be shown. Two cDNA sequences encoded the ribosomal proteins L19 and L36a, which showed a stunted growth phenotype upon gene silencing in N. benthamiana using VIGS and a slightly reduced Rx-mediated HR. The nematode resistance protein Gpa2 and the virus resistance protein Rx1 provide an excellent test system to investigate the exchangeability of recognition and signalling domains and explore the evolutionary flexibility of R proteins, for they confer resistance to completely unrelated pathogens (Globodera pallida and potato virus X, respectively). In Chapter 4, we provide evidence for the hypothesis, that, via intergenic sequence exchanges and various types of mutations, NB-LRR proteins have the potential to alter resistance specificities towards taxonomically unrelated pathogens in relatively short evolutionary time periods. Both the regulatory sequences and CC-NB domains of the paralogs Gpa2 and Rx1 are non-pathogen specific and exchangeable. Remarkably, the genetic fusions of the CC-NB of Rx1 with the LRR of Gpa2 (Rx1CN/Gpa2L) and the reciprocal domain swap (Gpa2CN/Rx1L) were not functional when driven by the endogenous promoters or 35S promoter. Gain of wild type resistance was obtained by re-introducing the first five LRRs of Rx1 in Rx1CN/Gpa2, restoring the compatibility between the N-terminal part of the LRR and the ARC2 domain. Decreasing the expression levels for Gpa2CN/Rx1L resulted in extreme resistance against PVX, indistinguishable from wild type plants. Our results indicate that not only coding sequences, but that also optimizing the expression levels may play a role in generating novel resistances. The CC, NB-ARC, and LRR domains of the Rx1 and Gpa2 proteins interact with each other and recognition of the elicitor mediated by the LRR is translated in an activation of the NB-ARC. The available functional and evolutionary data make Gpa2 a suitable candidate for structural modelling of the individual domains and their interaction (Chapter 5). A structural model of the NB-ARC / LRR interaction could function as a framework for the interpretation of known empirical data and the design of new experiments to test R protein operational mechanisms (Zhang 2009). Therefore, computer aided modelling of the 3D structure domain models for the NB-ARC and the LRR domains were obtained and used as basis for a domain docking study. The functional interaction between the domains was studied via a detailed analysis of their incompatibility in chimeric Gpa2 and Rx1 proteins. A large set of sequence exchanges between the two proteins was created for that purpose. Both in the LRR and in the ARC2 domain small regions could be identified in which the amino acids differing between Gpa2 and Rx1 led to domain incompatibility. Five of the ARC2 positions required for LRR compatibility and three known autoactivating positions from the RX1 LRR were used as constraints in domain docking computation to limit the potential search space. The resulting docking model indicated an important role in the NB-ARC-LRR interaction for electrostatic and hydrophobic interactions. A loop region rich in acidic residues in the ARC2 domain was found close in space to a patch of basic residues grouped together in the LRR. Hydrophobic residues on both the NB and the ARC2 contacted hydrophobic residues on the surface of the LRR. A correlation analysis of the NB-ARC and LRR subdomains detected coevolution between the interacting surfaces, which supports a direct interaction between these two domains. Site-directed mutagenesis and pull-down experiments were used to test the role of surface features that might play an important role in the interdomain docking interface. In Chapter 6, we have made use of the characteristic of the Rx1 protein that it remains functional when its domains are co-expressed as separate polypeptides. This allowed us to create fluorescent constructs, not only of the full length protein, but also of the separate subdomains. Most of these tagged constructs still form functional proteins. C- and N-terminal fusions of Green Fluorescent Protein (GFP) variants to Rx, made it possible to study its subcellular localization in Nicotiana benthamiana cells. Contrary to our expectations we observed the presence of Rx1 in both the cytoplasm and the nucleus. Rx1 does not contain known nuclear localization signals and the size of the protein (140 kDa including GFP) exceeds the limit for passive diffusion through the nuclear pore. Fluorescent fusions of a series of deletion constructs, CC-NB-ARC, NB-ARC, NB-ARC-LRR, CC and LRR showed three distinct patterns of subcellular localization. The NB-ARC-LRR and LRR constructs have a cytoplasmic localization and are mostly absent in the nucleus. The NB-ARC and CC-NB-ARC constructs showed equal fluorescence intensities in both the nucleus and the cytoplasm. The CC alone fused to GFP, however, seems to preferentially accumulate in the nucleus resulting in a three to four times higher fluorescence intensity in the nucleus compared to the cytoplasm. The diffusion behaviour inside the nucleus for both the complete CC and a CC fragment containing the two predicted helices downstream of the central turn, showed that their nuclear accumulation coincides with a significantly reduced nuclear diffusion as compared to unfused GFP and the other CC fragments. This difference might point to a potential interaction between the CC and an unknown nuclear component. Furthermore, SGT1 and Rar1 are thought to function as chaperones involved in stabilizing R proteins. Both the silencing experiments with these two proteins and the P-loop mutation show that the nuclear localisation of Rx1 is probably conformation dependent. Two approaches were followed to see if CP recognition or Rx1 signalling pathway were linked to a certain cellular compartment. At one hand the Rx1 protein itself or its subdomains were directed to either the nucleus or the cytoplasm by fusion to exogenous targeting signals (Nuclear Export Signals or Nuclear Localization Signals). On the other hand the elicitor, the PVX coat protein, was directed to the nucleus or cytoplasm. The PVX coat protein is a much smaller protein and can under normal circumstances diffuse freely between the cytoplasm and the nucleus. The surprising result was that no effect was found for retargeting Rx1, but when the elicitor was targeted to the nucleus, it could not activate Rx1 anymore, indicating that recognition might to take place in the cytoplasm. In the final chapter, the results obtained in this thesis are put into perspective by studying parallels in scientific literature on NB-LRR proteins with similar functions in other organisms.

AB - Resistance proteins are part of the plant’s immune system and mediate a defence response upon recognizing their cognate pathogens. They are thought to be present in the cell as part of a larger protein complex. The modular architecture of R proteins suggests that they form a scaffold for various interacting proteins, involved in pathogen recognition, downstream signalling or protein stabilization. However, few common interactors have been found for the CC-NB-ARC domains despite extensive screenings for downstream interactors. The objective of thesis was to get new insights in the structure, function and localization of R proteins by using the potato resistance genes Rx1 and Gpa2 as a model system. Initially, a novel T7 phage display method was developed to facilitate high throughput selection of interacting molecules (Chapter 2). However, the use of a T7 cDNA phage library to identify interactors of the CC-NB-ARC domains of Rx1 resulted in the discovery of a large set of highly basic peptides (Chapter 3). In Chapter 4, the functional role of the CC-NB-ARC domain in mediating disease resistance was explored by creating chimeric proteins between Rx1 and Gpa2. This resulted in the observation that the CC-NB-ARC is able to confer both virus and nematode resistance in potato. Furthermore, it was shown that the CC-NB-ARC of Rx1 and the LRR of Gpa2 are incompatible and vice versa. This phenomenon was studied in more detail in Chapter 5, in which a docking model for the interacting surface of these domains was constructed based on the individual structural domains. Finally, the subcellular localisation was investigated to get a better understanding about the R proteins function in the cell (Chapter 6). The lytic T7 phages form a powerful platform for the display of large cDNA libraries and offer the possibility to screen for strong interactions with a variety of substrates. To visualise these interactions directly by fluorescence microscopy, we constructed fluorescent T7 phages by exploiting the flexibility of phages to incorporate modified versions of its capsid protein (Chapter 2). By applying translational frameshift sequences, helper plasmids were constructed that expressed a fixed ratio of both wild-type capsid protein (gp10) and capsid protein fused to enhanced yellow fluorescent protein (EYFP). The frameshift sequences were inserted between the 3’-end of the capsid gene and the sequence encoding EYFP. Fluorescent fusion proteins are only formed when the ribosome makes a -1 shift in reading frame during translation. As far as we know this is the first report of using a translational frameshift for a biotechnological purpose. The phages formed in this way have capsids composed of three different variants of their capsid protein; EYFP-fused versions derived by frameshift translation, non-fused versions derived by regular translation from the helper plasmid, and versions that display peptides encoded in the library ligated in the phage genome. Using standard fluorescence microscopy, we could sensitively monitor the enrichment of specific binders in a cDNA library displayed on fluorescent T7 phages. Closely monitoring the effect of the selection procedure enables fine tuning, and obviates the need for more laborious ELISA or plaque lift assays. Furthermore, with the fast pace of developments in single molecule detection technologies and sorting systems, these fluorescent phages open the way to high throughput platforms for the direct selection of binding molecules. In Chapter 3, cDNA phage display was applied as an alternative method to identify additional downstream Rx1 interactors, which could further resolve the Rx1 signalling pathway. In a pilot experiment the value of T7 phage display to identify specific interactors was demonstrated by using an antibody raised against the PVY coat protein. Screening of a PVY-infected N. benthamiana cDNA phage display library resulted in the selection of peptides harbouring the known PRIKAI epitope. Next, phage display was explored as technique to discover proteins interacting with the potato R protein Rx1. The system turned out to be prone to pick up interactors binding to matrices like Ni-NTA or to fusion proteins like thioredoxine. A possible way to circumvent this weakness was to design the selection procedure in such a way that it alternates between different matrices and to limit the number of selection round. This adapted approach resulted in the identification of a series of highly basic protein fragments and random peptides, for which a specific interaction could be shown. Two cDNA sequences encoded the ribosomal proteins L19 and L36a, which showed a stunted growth phenotype upon gene silencing in N. benthamiana using VIGS and a slightly reduced Rx-mediated HR. The nematode resistance protein Gpa2 and the virus resistance protein Rx1 provide an excellent test system to investigate the exchangeability of recognition and signalling domains and explore the evolutionary flexibility of R proteins, for they confer resistance to completely unrelated pathogens (Globodera pallida and potato virus X, respectively). In Chapter 4, we provide evidence for the hypothesis, that, via intergenic sequence exchanges and various types of mutations, NB-LRR proteins have the potential to alter resistance specificities towards taxonomically unrelated pathogens in relatively short evolutionary time periods. Both the regulatory sequences and CC-NB domains of the paralogs Gpa2 and Rx1 are non-pathogen specific and exchangeable. Remarkably, the genetic fusions of the CC-NB of Rx1 with the LRR of Gpa2 (Rx1CN/Gpa2L) and the reciprocal domain swap (Gpa2CN/Rx1L) were not functional when driven by the endogenous promoters or 35S promoter. Gain of wild type resistance was obtained by re-introducing the first five LRRs of Rx1 in Rx1CN/Gpa2, restoring the compatibility between the N-terminal part of the LRR and the ARC2 domain. Decreasing the expression levels for Gpa2CN/Rx1L resulted in extreme resistance against PVX, indistinguishable from wild type plants. Our results indicate that not only coding sequences, but that also optimizing the expression levels may play a role in generating novel resistances. The CC, NB-ARC, and LRR domains of the Rx1 and Gpa2 proteins interact with each other and recognition of the elicitor mediated by the LRR is translated in an activation of the NB-ARC. The available functional and evolutionary data make Gpa2 a suitable candidate for structural modelling of the individual domains and their interaction (Chapter 5). A structural model of the NB-ARC / LRR interaction could function as a framework for the interpretation of known empirical data and the design of new experiments to test R protein operational mechanisms (Zhang 2009). Therefore, computer aided modelling of the 3D structure domain models for the NB-ARC and the LRR domains were obtained and used as basis for a domain docking study. The functional interaction between the domains was studied via a detailed analysis of their incompatibility in chimeric Gpa2 and Rx1 proteins. A large set of sequence exchanges between the two proteins was created for that purpose. Both in the LRR and in the ARC2 domain small regions could be identified in which the amino acids differing between Gpa2 and Rx1 led to domain incompatibility. Five of the ARC2 positions required for LRR compatibility and three known autoactivating positions from the RX1 LRR were used as constraints in domain docking computation to limit the potential search space. The resulting docking model indicated an important role in the NB-ARC-LRR interaction for electrostatic and hydrophobic interactions. A loop region rich in acidic residues in the ARC2 domain was found close in space to a patch of basic residues grouped together in the LRR. Hydrophobic residues on both the NB and the ARC2 contacted hydrophobic residues on the surface of the LRR. A correlation analysis of the NB-ARC and LRR subdomains detected coevolution between the interacting surfaces, which supports a direct interaction between these two domains. Site-directed mutagenesis and pull-down experiments were used to test the role of surface features that might play an important role in the interdomain docking interface. In Chapter 6, we have made use of the characteristic of the Rx1 protein that it remains functional when its domains are co-expressed as separate polypeptides. This allowed us to create fluorescent constructs, not only of the full length protein, but also of the separate subdomains. Most of these tagged constructs still form functional proteins. C- and N-terminal fusions of Green Fluorescent Protein (GFP) variants to Rx, made it possible to study its subcellular localization in Nicotiana benthamiana cells. Contrary to our expectations we observed the presence of Rx1 in both the cytoplasm and the nucleus. Rx1 does not contain known nuclear localization signals and the size of the protein (140 kDa including GFP) exceeds the limit for passive diffusion through the nuclear pore. Fluorescent fusions of a series of deletion constructs, CC-NB-ARC, NB-ARC, NB-ARC-LRR, CC and LRR showed three distinct patterns of subcellular localization. The NB-ARC-LRR and LRR constructs have a cytoplasmic localization and are mostly absent in the nucleus. The NB-ARC and CC-NB-ARC constructs showed equal fluorescence intensities in both the nucleus and the cytoplasm. The CC alone fused to GFP, however, seems to preferentially accumulate in the nucleus resulting in a three to four times higher fluorescence intensity in the nucleus compared to the cytoplasm. The diffusion behaviour inside the nucleus for both the complete CC and a CC fragment containing the two predicted helices downstream of the central turn, showed that their nuclear accumulation coincides with a significantly reduced nuclear diffusion as compared to unfused GFP and the other CC fragments. This difference might point to a potential interaction between the CC and an unknown nuclear component. Furthermore, SGT1 and Rar1 are thought to function as chaperones involved in stabilizing R proteins. Both the silencing experiments with these two proteins and the P-loop mutation show that the nuclear localisation of Rx1 is probably conformation dependent. Two approaches were followed to see if CP recognition or Rx1 signalling pathway were linked to a certain cellular compartment. At one hand the Rx1 protein itself or its subdomains were directed to either the nucleus or the cytoplasm by fusion to exogenous targeting signals (Nuclear Export Signals or Nuclear Localization Signals). On the other hand the elicitor, the PVX coat protein, was directed to the nucleus or cytoplasm. The PVX coat protein is a much smaller protein and can under normal circumstances diffuse freely between the cytoplasm and the nucleus. The surprising result was that no effect was found for retargeting Rx1, but when the elicitor was targeted to the nucleus, it could not activate Rx1 anymore, indicating that recognition might to take place in the cytoplasm. In the final chapter, the results obtained in this thesis are put into perspective by studying parallels in scientific literature on NB-LRR proteins with similar functions in other organisms.

KW - solanum tuberosum

KW - planteiwitten

KW - pathogenesis-gerelateerde eiwitten

KW - bindende eiwitten

KW - resistentiemechanismen

KW - ziekteresistentie

KW - plantenvirussen

KW - plantenparasitaire nematoden

KW - bacteriofagen

KW - genexpressie

KW - faag display

KW - solanum tuberosum

KW - plant proteins

KW - pathogenesis-related proteins

KW - binding proteins

KW - resistance mechanisms

KW - disease resistance

KW - plant viruses

KW - plant parasitic nematodes

KW - bacteriophages

KW - gene expression

KW - phage display

M3 - internal PhD, WU

SN - 9789085854678

PB - S.n.

CY - [S.l.

ER -