Doctor of Philosophy / Plant Pathology / Frank F. White / Rice and wheat are two major crops that suffer losses from the diseases of bacterial blight and bacterial leaf streak, which are caused by Xanthomonas oryzae pv. oryzae (Xoo) and Xanthomonas translucens pv. undulosa (Xtu), respectively. Transcriptional-Activator Like (TAL) effectors, a special family of type III effector proteins from Xanthomonas, have been demonstrated as critical virulence factors that act by inducing corresponding susceptibility (S) genes in several disease complexes of plants. In this study, I analyzed the contributions of TAL effectors from Xoo and Xtu to virulence and in modulating host gene expression to enhance susceptibility. Specifically, the TalC effector from the African Xoo strain AXO1947 was identified as a critical virulence factor, which functions by promoting expression of the gene OsSWEET14 in rice. TalC is interchangeable with other major TAL effectors from Asian strains of Xoo on the basis of functional complementation. The TAL effector PthXo2 from the Asian Xoo strain JXO1 is a major virulence factor and contains 21.5 repeats in the central repetitive region that targets OsSWEET13 in indica rice varieties but not in japonica rice varieties. A one repeat deletion in the PthXo2 effector enabled effector specificity to switch from indica rice to japonica rice. TAL effector genes from a genomic analysis of the Xtu strain XT4699 and related strains were characterized with regards to their involvement in virulence and the modulation of host gene expression in the Chinese Spring wheat cultivar. The identification of TAL effectors with virulence contributions and their target S genes is important for understanding the virulence mechanisms of Xanthomonas bacteria and promises to provide new strategies for disease control.

Maize is the staple food for a majority of people in Southern Africa, but plant diseases are responsible for at least 10% of crop production losses. Bacterial leaf streak (BLS) of maize was first reported in South Africa in 1949 and has not been reported elsewhere. Very little is known about the pathogen involved and therefore it is deemed necessary to compile a characteristic profile for the pathogen to prevent the possibility of major crop losses as a result of this disease. This study aimed to use biochemical and phenotypic methods to determine the specific characteristics of the causal agent of BLS. Diseased plant material showing symptoms of BLS were collected during the maize production seasons of 2012 and 2013 within South Africa’s maize production regions namely the North West, Free State, Gauteng and Northern Cape provinces. To prevent contamination, maize leaves were surface sterilised thoroughly before bacterial isolation commenced. Sections of the infected maize leaves were placed on GYC agar plates on which yellow, mucoid bacterial colonies after incubation for 24 to 48 hrs. The isolated bacteria were purified and the molecular identification of the bacteria was conducted in a related study. Although literature indicates that Xanthomonas campestris pv. zeae is the causal agent of BLS, pure cultures obtained from maize leaves showing characteristic symptoms of BLS were identified as species of Xanthomonas, Pantoea, and Enterobacter. To elucidate the pathogenicity of the isolated strains, pathogenicity tests based on Koch’s postulates were performed. Results from the pathogenicity tests confirmed that only the isolate Xanthomonas species was capable of inducing the characteristic BLS symptoms when healthy maize plants were inoculated with the suspected pathogens. It is important to inoculate the maize seedlings at the correct age (four-leaf stage) and the spray method is recommended. Re-isolation was repeated from the same plant material used during the initial isolation process but the isolation method was amended. The optimised isolation method involved the use of a dilution range and spread plate method. Colonies from this isolation technique grew as bright yellow colonies that were identified as Xanthomonas spp. This outcome indicates the importance of surface sterilisation, pulverisation and subsequent dilution of plant materials for isolation of bacterial pathogens from diseases plants. These isolates were used to create protein profiles with SDS-PAGE electrophoresis and carbon utilisation patterns with the Biolog® GN2 system. Protein profiling banding patterns was assessed based on presence/absence criteria. Highly similar protein profiles were observed among the X. campestris pv. zeae isolates but groupings of different protein profiles were determined when minor differences in the protein profiles was taken into account. Xanthomonas campestris pv. zeae was successfully distinguished from the X. axonopodis pv. vasculorum reference strain through unique SDS banding patterns. Banding patterns obtained from cultures grown in a liquid medium (tryptic soy broth) were of a higher quality than the banding patterns obtained from bacteria harvested from solid media (CYG agar plates). Carbon source utilisation data was used to evaluate the average well colour development obtained from each isolate. Statistically significant differences were found among some of the isolates, with some isolates being metabolically more active than other isolates. Substrate utilisation patterns produced by the isolates corresponded to previously published studies on various Xanthomonas species. The cell count of the samples used during carbon utilisation patterns must be standardised in order to obtain reliable results. During this study, the application of Koch’s postulates and two inoculation techniques confirmed that Xanthomonas campestris pv. zeae is the pathogen responsible for bacterial leaf streak of maize. Members of the Pantoea and Enterobacter genera were found on the leaf surface of maize plants infected with BLS but inoculations of healthy maize plants with these bacteria did not result in bacterial leaf streak symptoms on the maize plants. These bacteria were not pathogenic and were considered endophytes. The identified pathogen was characterised through protein and metabolic profiling. The protein profiles of the pathogen obtained through analysis of the major bands of the SDS-PAGE gels were highly similar and distinguishable from the Xanthomonas reference culture. Groupings within the X. campestris pv. zeae group was found when major and minor bands were considered, this may however be altered when the intensities of the bands are used during analysis. Carbon utilisation patterns were assessed using Biolog® GN2 plates. A metabolic fingerprint was created for the pathogen of BLS, it was possible to distinguish between X. campestris pv. zeae and other Xanthomonas strains based on the fingerprint. This fingerprint could be used to identify the pathogen. / MSc (Environmental Sciences), North-West University, Potchefstroom Campus, 2015

Maize is the staple food for a majority of people in Southern Africa, but plant diseases are responsible for at least 10% of crop production losses. Bacterial leaf streak (BLS) of maize was first reported in South Africa in 1949 and has not been reported elsewhere. Very little is known about the pathogen involved and therefore it is deemed necessary to compile a characteristic profile for the pathogen to prevent the possibility of major crop losses as a result of this disease. This study aimed to use biochemical and phenotypic methods to determine the specific characteristics of the causal agent of BLS. Diseased plant material showing symptoms of BLS were collected during the maize production seasons of 2012 and 2013 within South Africa’s maize production regions namely the North West, Free State, Gauteng and Northern Cape provinces. To prevent contamination, maize leaves were surface sterilised thoroughly before bacterial isolation commenced. Sections of the infected maize leaves were placed on GYC agar plates on which yellow, mucoid bacterial colonies after incubation for 24 to 48 hrs. The isolated bacteria were purified and the molecular identification of the bacteria was conducted in a related study. Although literature indicates that Xanthomonas campestris pv. zeae is the causal agent of BLS, pure cultures obtained from maize leaves showing characteristic symptoms of BLS were identified as species of Xanthomonas, Pantoea, and Enterobacter. To elucidate the pathogenicity of the isolated strains, pathogenicity tests based on Koch’s postulates were performed. Results from the pathogenicity tests confirmed that only the isolate Xanthomonas species was capable of inducing the characteristic BLS symptoms when healthy maize plants were inoculated with the suspected pathogens. It is important to inoculate the maize seedlings at the correct age (four-leaf stage) and the spray method is recommended. Re-isolation was repeated from the same plant material used during the initial isolation process but the isolation method was amended. The optimised isolation method involved the use of a dilution range and spread plate method. Colonies from this isolation technique grew as bright yellow colonies that were identified as Xanthomonas spp. This outcome indicates the importance of surface sterilisation, pulverisation and subsequent dilution of plant materials for isolation of bacterial pathogens from diseases plants. These isolates were used to create protein profiles with SDS-PAGE electrophoresis and carbon utilisation patterns with the Biolog® GN2 system. Protein profiling banding patterns was assessed based on presence/absence criteria. Highly similar protein profiles were observed among the X. campestris pv. zeae isolates but groupings of different protein profiles were determined when minor differences in the protein profiles was taken into account. Xanthomonas campestris pv. zeae was successfully distinguished from the X. axonopodis pv. vasculorum reference strain through unique SDS banding patterns. Banding patterns obtained from cultures grown in a liquid medium (tryptic soy broth) were of a higher quality than the banding patterns obtained from bacteria harvested from solid media (CYG agar plates). Carbon source utilisation data was used to evaluate the average well colour development obtained from each isolate. Statistically significant differences were found among some of the isolates, with some isolates being metabolically more active than other isolates. Substrate utilisation patterns produced by the isolates corresponded to previously published studies on various Xanthomonas species. The cell count of the samples used during carbon utilisation patterns must be standardised in order to obtain reliable results. During this study, the application of Koch’s postulates and two inoculation techniques confirmed that Xanthomonas campestris pv. zeae is the pathogen responsible for bacterial leaf streak of maize. Members of the Pantoea and Enterobacter genera were found on the leaf surface of maize plants infected with BLS but inoculations of healthy maize plants with these bacteria did not result in bacterial leaf streak symptoms on the maize plants. These bacteria were not pathogenic and were considered endophytes. The identified pathogen was characterised through protein and metabolic profiling. The protein profiles of the pathogen obtained through analysis of the major bands of the SDS-PAGE gels were highly similar and distinguishable from the Xanthomonas reference culture. Groupings within the X. campestris pv. zeae group was found when major and minor bands were considered, this may however be altered when the intensities of the bands are used during analysis. Carbon utilisation patterns were assessed using Biolog® GN2 plates. A metabolic fingerprint was created for the pathogen of BLS, it was possible to distinguish between X. campestris pv. zeae and other Xanthomonas strains based on the fingerprint. This fingerprint could be used to identify the pathogen. / MSc (Environmental Sciences), North-West University, Potchefstroom Campus, 2015

All members of the genus Xanthomonas are considered to be plant pathogenic, with specific pathovars infecting several high value agricultural crops. One of these pathovars, X. campestris pv. zeae (as this is only a proposed name it will further on be referred to as Xanthomonas BLSD) the causal agent of bacterial leaf steak of maize, has established itself as a widespread significant maize pathogen within South Africa. Insufficient information about the present distribution of the pathogen is available. The main aim of the study was thus to isolate and characterise the pathogen using molecular methods. Results demonstrated that the causal agent of bacterial leaf streak disease (Xanthomonas BLSD: potentially X. campestris pv. zeae) was widely distributed within the major maize cultivation regions of South Africa. Most of the isolates collected originated from the Highveld maize production provinces (North West, Free State, Gauteng and Mpumalanga provinces) as well as from irrigated maize fields in the Northern Cape province. The XgumD gene marker was used to determine if the isolates belonged to the genus Xanthomonas. The gumD gene fragment is located within the gumB-gumM region of the operon and is conserved among Xanthomonas species. This gene fragment is partially responsible for xanthan production. This marker was amplified from all isolates and a selected number were sequenced. The marker was only able to confirm that the causal agent was a member of the genus Xanthomonas. PCR methods were used for the characterisation of the isolates. This included PCR and sequencing of ribosomal RNA- gyraseB and gumD genes. A fingerprinting method BOX-PCR was also employed. Good quality DNA of sufficient quantities was obtained from the various isolates. Amplification produced no non-specific amplification products. This resulted in good quality sequences that could be analysed using bioinformatics tools. Phylogenetic analyses of the ribosomal RNA and gyraseB genes could not detect differences amongst the 47 Xanthomonas BLSD isolates. However, these genes were able to distinguish between the type strain of these isolates and various Xanthomonas species and pathovars. From all three neighbour joining trees the Xanthomonas BLSD isolates had close association with X. axonopodis pv. vasculorum strain ATCC 35938. For the 16S rRNA gene there exists no sequence differences between Xanthomonas BLSD and X. axonopodis pv. vasculorum strain ATCC 35938. A single nucleotide difference was observed between Xanthomonas BLSD and X. axonopodis pv. vasculorum strain ATCC 35938 for the 23S rRNA gene. The gyraseB gene detected a total of six nucleotide variations between these two Xanthomonas species. For all of the phylogenetic trees there was no clustering of Xanthomonas BLSD with X. campestris pathovars. Genetic profiling (via BOX-PCR) based on present/absent analysis revealed no variations amongst the Xanthomonas BLSD isolates. All isolates shared an identical pattern produced by 12 distinct PCR products. This profiling technique did differentiate between the isolates of Xanthomonas BLSD and X. axonopodis pv. vasculorum strain ATCC 35938. Their profiles shared common bands, but differed in the number and overall pattern of the bands. These results suggest two main conclusions: (i) Xanthomonas BLSD has a clonal origin with geographical separation not impacting genetic variation. The fact that all the isolates appear to be clonal may imply that when resistant maize cultivars are developed these should be resistant to all isolates of the pathovar irrespective of their geographical origin. This is a suggestion that will have to be corroborated using more isolates and additional genetic fingerprinting techniques (ii) the Xanthomonas BLSD isolates from this study may not belong to X. campestris. Further studies using other markers should be conducted to determine the real identity of Xanthomonas BLSD. / MSc Environmental Sciences, North-West University, Potchefstroom Campus, 2015

All members of the genus Xanthomonas are considered to be plant pathogenic, with specific pathovars infecting several high value agricultural crops. One of these pathovars, X. campestris pv. zeae (as this is only a proposed name it will further on be referred to as Xanthomonas BLSD) the causal agent of bacterial leaf steak of maize, has established itself as a widespread significant maize pathogen within South Africa. Insufficient information about the present distribution of the pathogen is available. The main aim of the study was thus to isolate and characterise the pathogen using molecular methods. Results demonstrated that the causal agent of bacterial leaf streak disease (Xanthomonas BLSD: potentially X. campestris pv. zeae) was widely distributed within the major maize cultivation regions of South Africa. Most of the isolates collected originated from the Highveld maize production provinces (North West, Free State, Gauteng and Mpumalanga provinces) as well as from irrigated maize fields in the Northern Cape province. The XgumD gene marker was used to determine if the isolates belonged to the genus Xanthomonas. The gumD gene fragment is located within the gumB-gumM region of the operon and is conserved among Xanthomonas species. This gene fragment is partially responsible for xanthan production. This marker was amplified from all isolates and a selected number were sequenced. The marker was only able to confirm that the causal agent was a member of the genus Xanthomonas. PCR methods were used for the characterisation of the isolates. This included PCR and sequencing of ribosomal RNA- gyraseB and gumD genes. A fingerprinting method BOX-PCR was also employed. Good quality DNA of sufficient quantities was obtained from the various isolates. Amplification produced no non-specific amplification products. This resulted in good quality sequences that could be analysed using bioinformatics tools. Phylogenetic analyses of the ribosomal RNA and gyraseB genes could not detect differences amongst the 47 Xanthomonas BLSD isolates. However, these genes were able to distinguish between the type strain of these isolates and various Xanthomonas species and pathovars. From all three neighbour joining trees the Xanthomonas BLSD isolates had close association with X. axonopodis pv. vasculorum strain ATCC 35938. For the 16S rRNA gene there exists no sequence differences between Xanthomonas BLSD and X. axonopodis pv. vasculorum strain ATCC 35938. A single nucleotide difference was observed between Xanthomonas BLSD and X. axonopodis pv. vasculorum strain ATCC 35938 for the 23S rRNA gene. The gyraseB gene detected a total of six nucleotide variations between these two Xanthomonas species. For all of the phylogenetic trees there was no clustering of Xanthomonas BLSD with X. campestris pathovars. Genetic profiling (via BOX-PCR) based on present/absent analysis revealed no variations amongst the Xanthomonas BLSD isolates. All isolates shared an identical pattern produced by 12 distinct PCR products. This profiling technique did differentiate between the isolates of Xanthomonas BLSD and X. axonopodis pv. vasculorum strain ATCC 35938. Their profiles shared common bands, but differed in the number and overall pattern of the bands. These results suggest two main conclusions: (i) Xanthomonas BLSD has a clonal origin with geographical separation not impacting genetic variation. The fact that all the isolates appear to be clonal may imply that when resistant maize cultivars are developed these should be resistant to all isolates of the pathovar irrespective of their geographical origin. This is a suggestion that will have to be corroborated using more isolates and additional genetic fingerprinting techniques (ii) the Xanthomonas BLSD isolates from this study may not belong to X. campestris. Further studies using other markers should be conducted to determine the real identity of Xanthomonas BLSD. / MSc Environmental Sciences, North-West University, Potchefstroom Campus, 2015

La bactériose vasculaire du riz (BLB) et à stries foliaires (BLS) causées respectivement par Xanthomonas oryzae pv. oryzae (Xoo) et X. oryzae pv. oryzicola (Xoc) sont deux maladies émergentes en Afrique de l'Ouest, suite à l'expansion de la culture du riz et à l'introduction de variétés à haut rendement au cours de ces dernières décennies. Trois nouvelles races de Xoo ont été caractérisées en Afrique dont on a montré, sur la base d'une analyse génétique, leur spécificité africaine. En revanche, une étude réalisée sur une dizaine de souches de Xoc isolées au Mali en 2003, démontre qu'elles sont apparentées à des souches de Xoc asiatiques. En Asie, plusieurs gènes de résistance à Xoo ont été identifiés et déployés dans les programmes de lutte contre BLB. Cependant aucun gène de résistance à Xoc n'a été encore identifié chez le riz. Les objectifs de notre étude étaient (i): d'implémenter les collections de souches de Xoo et Xoc Africaines disponibles mais incomplètes, à l'aide de nouvelles campagnes d'échantillonage réalisées de 2009 à 2012 dans différentes zones agroécologiques du Burkina Faso et du Mali, (ii) de déterminer la diversité génétique de ces souches, (iii) d'identifier et caractériser de nouvelles sources de résistance contre BLB et BLS au sein d'accessions de riz cultivées au Burkina Faso. Nos résultats ont montré que les souches africaines de Xoc sont hautement variables tant d'un point de vue génétique que du pouvoir pathogène. L'analyse par PCR de deux effecteurs de types III conservés (xopAJ et xopW) permet de différencier les souches de Xoc en deux groupes, xopAJ étant absent dans la majorité des souches et une insertion de 1050 bp étant détectée dans la séquence codante de xopW de certaines souches. Néanmoins, il apparait que la forte diversité génétique des Xoc n'est pas corrélée à leur origine géographique, ni à la période de collecte, ou à la nature de l'hôte. Les souches de Xoo caractérisées appartiennent toutes à la race A1 qui n'avait pas encore été signalée au Mali. Au regard de la diversité des souches et de leur évolution, il est important d'envisager un plan de surveillance épidémiologique à plus large échelle des populations de Xo dans les régions concernées en Afrique de l'Ouest. Enfin, nous avons montré que certaines variétés de riz cultivées au Burkina Faso présentent un phénotype de résistance spécifique des souches africaines de Xoo et ce, à tous les stades de développement de la plante. Ces données originales contrastent par rapport au phénotype des lignées de riz résistantes de référence (Xa4, xa5 et Xa7 efficaces uniquement au stade de tallage maximum). Eu égard à l'absence de gènes de résistance dans le riz efficaces contre Xoc, ces variétés qui constituent également une source de résistance efficaces contre la diversité des souches de Xoc africaines, offrent potentiellement un nouveau moyen pour assurer le contrôle du BLS au Burkina Faso et éventuellement dans d'autres pays Africains. / Bacterial Leaf Blight (BLB) and Bacterial Leaf Streak (BLS) diseases respectively caused by Xanthomonas oryzae pv. oryzae (Xoo) and X. oryzae. pv oryzicola (Xoc) are two emerging diseases of rice in West Africa, due to the recent expansion of rice cultivation and introduction of improved rice varieties over the last decade. Three news Xoo races were characterized based on genetic analysis, demonstrating their african specificity. In contrast, a study achevied on about ten Xoc strains isolated in Mali in 2003, show that they are related to asian Xoc strains. In Asia, several R genes against Xoo have been identified and deployed in breeding program to control BLB. In contrast, no R gene against Xoc has been identified in rice. The objectives of this PhD thesis are to (i) complete the Xo collections of African isolates upon annual sampling operated from 2009 to 2012 in various agroecological areas of Burkina Faso and Mali, (ii) determine the genetic diversity of these strains , (iii) identify and characterize news sources of resistance genes to BLB and BLS within rice accessions cultivated in Burkina Faso.Our results showed that african Xoc are highly diverse genetically and phenotypically. PCR-based analyse of two conserved type III effector gene (xopAJ and xopW) differentiated two groups of Xoc strains, with xopAJ not detected in a majority of African Xoc strains and 1050 bp insertion detected in xopW gene for few strains. However, the high genetic diversity observed among the Xoc strains is not correlated to geographical origin, sampling data or host plant species. Xoo strains characterized belong all to race A1 previously reported by Gonzalez et al. (2007) in Burkina Faso. Given the diversity of X. oryzae strains and their evolution, it is essential to establish a large scale epidemiological monitoring of Xo populations in concerned regions in west Africa.At last, some accessions cultivated in Burkina Faso showed specific resistance to african Xoo strains at all plant development stages. These original data contrast with rice lines carring Xa4, xa5 and Xa7 resistance genes against BLB, which are only effective at maximun tillering stage.Given no sources of effective resistance genes against BLS is available in rice, these accessions which were also efficient against a set of Xoc strains representative of the diversity in Africa, represent a huge potential source for the control of BLS in Burkina Faso, and eventually in others african countries.