SEARCH

SEARCH BY CITATION

Keywords:

  • metalaxyl resistance;
  • migration;
  • Phytophthora infestans ;
  • simple sequence repeat markers;
  • Sub-Saharan Africa

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Limited knowledge is available on Phytophthora infestans populations in Sub-Saharan Africa (SSA). Therefore, and in response to recent severe late blight epidemics, P. infestans isolates from potato, tomato and Petunia × hybrida from eight SSA countries were characterized. Isolates were characterized with ‘old’ markers, including mating type (176 isolates), mitochondrial DNA haplotype (mtDNA) (281 isolates), glucose-6-phosphate isomerase (Gpi) (70 isolates), restriction fragment length polymorphism analysis with probe RG-57 (49 isolates), and by metalaxyl sensitivity (64 isolates). Most isolates belonged to the US-1 genotype or its variants (US-1.10 and US-1.11). The exceptions were genotype KE-1 isolates (A1 mating type, mtDNA haplotype Ia, Gpi 90/100 and unique RG-57 genotype), identified in two fields in Kenya, which are related to genotypes previously identified in Rwanda (RW-1 and RW-2), Ecuador and Europe. Metalaxyl-resistant P. infestans isolates from potato were present in all the countries except Malawi, whereas all the isolates from tomato were sensitive. Genotyping of 176 isolates with seven simple sequence repeat (SSR) markers, including locus D13 that was difficult to score, revealed 79 multilocus genotypes (MLGs) in SSA. When this locus was excluded, 35 MLGs were identified. Genetic differentiation estimates between regional populations from SAA were significant when locus D13 was either excluded (= 0·05) or included (= 0·007), but population differentiation was only low to moderate (FST = 0·044 and 0·053, respectively).


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Phytophthora infestans is well known for causing the devastating Irish potato famine in the late 1840s. This famine was caused when the pathogen first migrated from its centre of origin, most likely central Mexico or the Andean region, to the rest of the world (Grünwald & Flier, 2005; Fry et al., 2009). Since then, late blight has continued to be devastating worldwide, causing major losses on potato (Solanum tuberosum) and tomato (Solanum lycopersicum) (Legard et al., 1995; Fry et al., 2009). Subsequent to the Irish potato famine, migrations have continued to play an important role in the history and devastation caused by P. infestans (Fry et al., 2009).

Phytophthora infestans migrations have been monitored since the early 1980s using a standard set of markers. These markers, also referred to as ‘old’ markers, consist of (i) mating type, (ii) mitochondrial DNA (mtDNA) haplotype, (iii) glucose-6-phosphate isomerase (Gpi), (iv) restriction fragment length polymorphism (RFLP) analysis with probe RG-57 and (v) peptidase (Pep) analyses (Cooke & Lees, 2004). More recently, several simple sequence repeat (SSR) markers have been used to investigate migrations (Knapova et al., 2001; Lees et al., 2006). SSR markers have several advantages over the ‘old’ markers because they are (i) co-dominant, (ii) more suitable for high throughput genotyping, (iii) less technically challenging and (iv) highly reproducible between different laboratories, allowing for global population genetic analyses (Cooke & Lees, 2004; Fry et al., 2009).

The historic migration pathways of P. infestans and the genotypes that were involved are controversial. It is generally accepted that at least two major migrations have occurred, one in the 1840s and the other in the late 1970s (Fry et al., 2009). Late blight epidemics that contributed to the 1840s Irish famine were most likely caused by P. infestans populations with a mtDNA haplotype Ia genotype. Subsequently, these populations were probably replaced by US-1 populations (mtDNA haplotype Ib) that were pan-globally distributed prior to the second migrations in the 1970s (Griffith & Shaw, 1998; Ristaino et al., 2001; May & Ristaino, 2004; Fry et al., 2009). The second P. infestans migrations in the late 1970s, which had devastating effects in Europe and North America, involved populations that were comprised of many different A1 and A2 mating type genotypes. These populations almost displaced the US-1 lineage worldwide (Fry et al., 2009). A few countries in SSA (Sub-Saharan Africa: South Africa, Kenya, Tanzania and Uganda) and elsewhere (Vietnam, Peru and Ecuador) were unaffected by these new migrations or the migrations did not result in complete displacement of the US-1 lineage (Griffith & Shaw, 1998; Oyarzun et al., 1998; Vega-Sánchez et al., 2000; McLeod et al., 2001).

Although it is assumed that most P. infestans populations in SSA have not been changed as a result of the late 1970s migrations, this assumption is based on limited genotypic and phenotypic data from predominantly small population sizes in a few SSA countries. Genotyping of P. infestans populations had only been conducted in a few SSA countries up until 2001, including South Africa, Kenya, Rwanda, Tanzania, Uganda and Ethiopia. These studies showed, with the exception of a few isolates in Rwanda and Ethiopia, that all isolates belonged to the US-1 lineage and its variants (Goodwin et al., 1994; Griffith & Shaw, 1998; Vega-Sánchez et al., 2000; McLeod et al., 2001; Ochwo et al., 2002; Schiessendopper & Molnar, 2002). In Rwanda, between 1984 and 1987, isolates of the US-1 lineage were identified along with mtDNA haplotype Ia genotypes with unique RG-57 fingerprints that were designated as RW-1 and RW-2 (Goodwin et al., 1994; Forbes et al., 1998; Gavino & Fry, 2002). Mitochondrial DNA haplotype Ia isolates were also identified in Ethiopia (Schiessendopper & Molnar, 2002). To date, no P. infestans isolates from SSA have been genotyped with SSR markers.

The main aims of the study were to (i) determine whether migrations (genotype flow) and/or metalaxyl resistance have contributed to recent severe epidemics in some Sub-Saharan countries and (ii) expand the genotypic database (‘old’ markers and SSRs) of P. infestans populations in eight Sub-Saharan countries (Burundi, Kenya, Rwanda, Tanzania, Uganda, Malawi, Mozambique and South Africa) by characterizing the isolates that were collected between 2005 and 2009 with ‘old’ markers (mating type, mtDNA haplotyping, Gpi, DNA fingerprinting with probe RG-57 and metalaxyl sensitivity) and with seven recently published SSR markers.

Materials and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Field sampling

The countries sampled were in central-east Africa (Tanzania, Kenya, Uganda, Burundi and Rwanda), southeast Africa (Malawi and Mozambique) and southern Africa (South Africa) (Fig. 1). The choice of countries sampled depended on the country: (i) being a Southern African Development Community (SADC) country, (ii) being in close proximity to a laboratory, (iii) having reliable contacts and (iv) being politically stable. Suitable laboratories for primary isolations were located in Kenya (Kenya Agricultural Research Institute (KARI), Nairobi), Rwanda (Institut des Sciences Agronomiques du Rwanda (ISAR), Ruhengeri) and Malawi (Department of Agricultural Research Services (DARS), Lilongwe). In central-east Africa, only potatoes were sampled; in southern Africa mainly potatoes, but also a few tomato and petunia plants were sampled; and in southeast Africa tomatoes and potatoes were sampled.

image

Figure 1.  Countries and regions in South Africa (KwaZulu-Natal and Western Cape), southeast Africa (Malawi and Mozambique) and central-east Africa (Uganda, Kenya, Rwanda, Burundi and Tanzania) where Phytophthora infestans isolates were collected.

Download figure to PowerPoint

Fields sampled were no less than 40 km apart within a given country. The fields were either intensively sampled (IS) or less intensively sampled (LIS), depending on infection levels and field size. Intensive sampling consisted of taking samples at 20 equally spaced sampling points along a W-shaped sampling path and was used in large fields (0·25 ha or larger) that had severe infections. Less intensive sampling was done in smaller fields (<0·25 ha) and consisted of either taking five equally spaced sampling points along a diagonal path in fields with low infection levels or randomly collecting as many infected leaves/stems as possible in fields with very low infection levels. At each sampling point, three or four infected leaflets showing separate, actively sporulating lesions were taken. Where leaf lesions were sparse, single petiole/stem lesions were also taken. Samples were taken in duplicate, and placed in separate paper bags.

Central-east Africa

In total, 23 potato fields were sampled in central-east African countries in 2007 (Table 1). In Tanzania, sampling was only conducted in northeastern Tanzania (Arusha and Moshi districts), on the slopes of Mt Kilimanjaro and the Ngorogoro Crater regions. In Kenya, the regions sampled were central Kenya (Meru, Nyahururu and Nakuru triangle) and western Kenya (Eldoret and Kitale districts). In Rwanda, the regions sampled included northwest Rwanda (Gisenyi and Musanze districts) and southwest Rwanda (Nyamagabe district). The regions sampled in Burundi consisted of central Burundi (Matana and Jenda districts) and northern Burundi (Muramvya district). In Uganda, only fields in southwestern Uganda (Lyantonde, Mbarara and Kabale districts) were sampled.

Table 1. Characteristics of Phytophthora infestans isolates collected from selected Sub-Saharan African countries as revealed by several genotypic and phenotypic markers
Region/countryHost (no. fields)Mating typeamtDNAb Gpic RG-57 genotypedMetalaxyl sensitivitye
  1. aThe number of isolates that were tested for mating type is shown in brackets.

  2. bMitochondrial DNA haplotype (mtDNA) of isolates, followed by the number of isolates that were tested in brackets.

  3. cGlucose-6-phosphate isomerase (Gpi) genotype of isolates, followed by the number of isolates that were tested in brackets.

  4. dFingerprint pattern of isolates as revealed through Southern analyses using probe RG-57, followed by the number of isolates tested in brackets.

  5. eMetalaxyl sensitivity of isolates was evaluated using an agar plate assay. Isolates were designated as sensitive (S) when radial growth was <40% of the control at 5 and 100 μg mL−1, intermediately resistant (IR) when radial growth was >40% of control at 5 μg mL−1 but <40% at 100 μg mL−1, and as resistant (R) when radial growth was >40% of control at 5 and 100 μg mL−1. The number of isolates tested is indicated in brackets.

Central-east Africa
 BurundiPotato (3)A1 (10)Ib (14)86/100 (3)US-1 (4)IR (2)
 KenyaPotato mtDNA Ib (7)A1 (41)Ib (50)86/100 (17)US-1(11);US-1.11(1)IR (2), S (5)
Potato mtDNA Ia (2)A1 (4)Ia (21)90/100 (6)KE-1 (5)S (2)
 RwandaPotato (3)A1 (14)Ib (20)86/100 (6)US-1 (2)S (1), R (5)
 TanzaniaPotato (4)A1 (27)Ib (31)86/100 (8)US-1 (3)S(1), R(3)
 UgandaPotato (4)A1 (19)Ib (30)86/100 (9)US-1 (4)S (4), R (1)
 Total analysed1 (23)115166493026
Southeast Africa
 MalawiPotato (3)A1 (8)Ib (9)86/100 (1)US-1 (2)S (8)
Tomato (3)A1 (7)Ib (7)86/100 (4)US-1 (3)S (4)
 MozambiqueTomato (1)A1 (1)Ib (3)86/100 (2)US-1 (1)S (2)
 Total analysed2 (7)16197614
South Africa
 KwaZulu-NatalPotato (9)A1 (28)Ib (38)86/100 (10)US-1 (2)R (5)
Tomato (1)A1 (3)Ib (5)86/100 (1)US-1 (2)S (4)
 Western CapePotato (10)A1 (13)Ib (51)86/100 (2)US-1.10 (6)R (14)
Petunia (2)A1 (1)Ib (2)86/100 (1)US-1.10 (3)R (1)
 Total analysed3 (23)4596141324
Total analysed3(53)176281704964
Southeast Africa

In southeast Africa, only seven tomato and potato fields were sampled in 2009 (Table 1), all using the LIS strategy, because of low levels of infection in all fields. The regions sampled in Malawi consisted of northern Malawi (around Mzimba and Mzuzu), central Malawi (Dedza and Ntcheu districts) and southern Malawi (around Mwanza). In Mozambique samples were taken in northern Mozambique in the Angonia district.

South Africa

Potato fields were sampled in only two provinces of South Africa, i.e. KwaZulu-Natal and the Western Cape Province, from 2005 to 2007. Because of the large size of South Africa compared to the other sampled Sub-Saharan countries, and because of the large distance between the two provinces (>15 000 km), the KwaZulu-Natal and Western Cape populations were treated as two separate populations. In KwaZulu-Natal the regions sampled were in the north east around Howick, Hilton, Bulwer and Mooi River. The Western Cape areas sampled were in the south and included the Sandveld, Cape flats and Kouebokkeveld. Three petunia (Petunia × hybrida) samples were collected from greenhouses in the Cape flats in 2005 and 2007.

Isolations

Leaflets and stems were surface-sterilized for 2·5 min in 0·5% sodium hypochlorite solution to which one or two drops of liquid soap were added to break the surface tension. Subsequently, the leaflets and stems were rinsed three times in sterile distilled water (SDW) and dried separately on paper towels. Eight leaf squares (2 × 2 mm) were taken from the margins of active single lesions, with four plated onto wheat agar (McLeod et al., 2001) and four plated onto pea agar (PA) (120 g frozen peas autoclaved in 1 L distilled water, strained and 15 g agar and 2 g CaCO3 added). The media were amended with nystatin or pimaricin (20 mg L−1), benomyl (200 mg L−1, Benlate 50% WP, Dow Agrosciences), ampicillin (10 mg L−1), rifampicin (60 mg L−1) and vancomycin (100 mg L−1). Plates were incubated at 20–25°C in the dark for 10–14 days.

If the above method of isolation failed, isolations were made by first incubating leaflets in a moist chamber at 20°C in the dark for 24 h to induce sporulation. To obtain an axenic culture of P. infestans, a small wheat agar cube (1 × 1 mm) was gently brushed over freshly produced sporangia and placed onto amended PA plates. Isolates were then transferred onto PA slants, incubated for 10–14 days at 20°C in the dark, and stored under medicinal paraffin oil at 20°C.

DNA extractions

Isolates were inoculated into pea broth (120 g frozen peas boiled in 1 L deionized water, strained and 2 g CaCO3 added) and the mycelia harvested after 7–14 days. The mycelia were lyophilized overnight and ground to a fine powder that was used for DNA extraction. DNA was extracted according to Goodwin et al. (1992).

Characterization with ‘old’ markers

Only selected isolates were genotyped for each of the specific markers (Table 1). For Southern analyses and isozyme analyses, at least one isolate per field was characterized. Two to five isolates from IS fields were characterized for mating type and mtDNA haplotype, and no more than two isolates for LIS fields, with the exception of a field in Kenya where the KE-1 genotype was found, and where more isolates were genotyped (see below).

Mating type

The mating type of 176 SSA isolates was determined as previously described (McLeod et al., 2001). Isolates were paired with known A1 (Tester A isolate) and A2 mating (isolate 03119) type isolates kindly provided by D. E. L Cooke (The James Hutton Institute (JHI), UK). The best growth medium for South African isolates was 10% clarified V8 juice agar, whereas Rye-B agar was the most suitable for the other SSA isolates.

mtDNA haplotype

In total, 281 isolates were genotyped for mtDNA haplotype using PCR-RFLP analysis, as described by Griffith & Shaw (1998), with three primer pairs (F1-R1, F2-R2 and F4-R4). The only modifications that were made were for primer pair F2-R2, where 2 mm MgCl2 were used, along with an annealing temperature of 60°C. Because one field in Kenya contained an mtDNA haplotype (Ia) different from the other sampled fields (see Results), more intensive mtDNA genotyping (71 isolates) was conducted for Kenya (Table 1).

Glucose-6-phosphate isomerase

Gpi genotyping was conducted on a subset of 70 isolates, which represented all the regions sampled, hosts and mtDNA haplotypes (Table 1). Reference isolates representing the US-1 (86/100) genotype (SA960001; W. E. Fry, Cornell University, Ithaca, USA) and EC-1 (90/100) genotype (SCRI C7; D. E. L Cooke, JHI, UK and G. Forbes, International Potato Center, Quito, Ecuador) were included in all the analyses. Gpi analysis was done using polyacrylamide gel electrophoresis and isoelectric focusing as previously described (McLeod et al., 2001), with the exception that the gels were run at 10°C for 1 h at 100 V, followed by 1·5 h at 200 V and ending with 1·5 h at 400 V.

RFLP analysis with probe RG-57

Fingerprinting with probe RG-57 was conducted on a subset of 49 isolates that represented the different hosts and genotypes identified with mtDNA haplotyping and Gpi genotyping (Table 1). Southern blot analysis was conducted using the Amersham gene images AlkPhos direct labelling and detection system (GE Healthcare) according to the manufacturer’s instructions. The same reference isolates used for Gpi analyses were also used in RG-57 analyses. Presence or absence of all fingerprint fragments was scored visually.

Unweighted pair group method of averages (UPGMA) analyses of ‘old’ marker data

Data obtained from mating type studies, Gpi, and RG-57 fingerprint genotyping was used in UPGMA analysis. Mitochondrial DNA haplotype data were not included, because multilocus genotypes (MLGs) published from other countries did not report results for this marker (Forbes et al., 1998). The RG-57 fingerprint data used in the UPGMA analysis consisted of 25 fragments. Twenty-three of the fragments were those described by Forbes et al. (1998) while the two additional fragments were 9a for US-1.10 and 24a for JP-1. Fragment 4 was excluded because it has not been consistently observed in all RG-57 analyses conducted by different research groups (Forbes et al., 1998). As fragment 25 was present in all the analyses, it was also excluded from the UPGMA analyses as previously described (Forbes et al., 1998). The UPGMA analysis was conducted using the four genotypes identified in the Sub-Saharan populations (see Results), along with several genotypes reported from other countries, mostly those by Forbes et al. (1998) and Deahl et al. (2009). Specific attention was paid to include genotypes that were related to the EC-1 and RW-1 genotypes, which were found in preliminary analyses to be related to a new Sub-Saharan genotype KE-1 (see Results). Analyses were conducted in mega v.4.0 (Tamura et al., 2007). The Jukes–Cantor model with 1000 bootstraps was selected within mega v.4.0.

Metalaxyl-M sensitivity

Metalaxyl-M sensitivity was determined as previously described (Therrien et al., 1993), using metalaxyl-amended pea agar plates (5 and 100 μg a.i. mL−1) and unamended plates. Only a subset of 64 isolates was tested, because several of the isolates died in storage before tests could be conducted. The subset selection was based on the genotypes identified, host and geographical origin. Isolates were classified as sensitive (radial growth <40% of control at 5 and 100 μg mL−1), intermediately resistant (radial growth >40% of control at 5 μg mL−1, but <40% of control at 100 μg mL−1), and resistant (radial growth >40% of control at both concentrations) based on Therrien et al. (1993) and Daggett et al. (1993).

Simple sequence repeat markers (SSR)

Initially, 19 SSR markers were screened on a subset of 24 isolates that represented the diversity identified in the African populations using the ‘old’ markers. These analyses showed that only markers D13, Pi02, Pi89, Pi70 and Pi56 (Lees et al., 2006) and markers Pi4B and PiG11 (Knapova et al., 2001) were polymorphic. The loci that were excluded from subsequent analyses included Pi04, Pi16, Pi2H, Pi1D, Pi2D, Pi33, Pi63, Pi65, Pi66 (Lees et al., 2006), Pi4G (Knapova et al., 2001), 1180 and 6433 (Garnica et al., 2006). Locus Pi63 was not included because it consistently produced three alleles for all African isolates, while loci Pi04, 1180, 6433 and Pi16 were excluded because they were difficult to score. The remaining markers were not useful because of their monomorphic nature in the African populations.

Amplification of SSR loci

All primers were synthesized by Applied Biosystems. PCR amplification of loci PiG11, Pi02 and Pi4B were multiplexed, whereas amplification of loci Pi70, Pi89, Pi56 and D13 were conducted in single reactions. The multiplex PCR reaction consisted of 1 ng μL−1 template DNA, 1× PCR buffer (Supertherm Gold; JMR Holdings), 1·5 mm MgCl2 (included in the 10× PCR buffer), 0·16 mm dNTPs, 0·0625 U Taq DNA polymerase (Supertherm Gold), 0·2 μm FAM-PiG11F and PiG11R, 0·12 μm PET-Pi02F and Pi02R, 0·168 μm NED-Pi4BF and Pi4BR in a total volume of 25 μL. The PCR reagents for the single reaction loci (Pi70, D13, Pi56 and Pi89) were all similar, except that the concentrations of the primers differed, and consisted of 1–2 ng μL−1 template DNA, 1× PCR buffer (Supertherm Gold), 1·5 mm MgCl2 (included in the 10× PCR buffer), 0·16 mm dNTPs and 0·0625 U Taq DNA polymerase (Supertherm Gold). The primer concentrations for each reaction were 0·125 μm PET-Pi70F and Pi70R, 0·526 μm NED-D13F and D13R, 0·16 μm FAM-Pi56F and Pi56R, and 0·14 μm FAM-Pi89F and Pi89R. PCR amplification conditions for all reactions, except loci Pi70 and D13, consisted of 20 cycles denaturation at 94°C for 40 s followed by annealing at 60°C for 40 s with a decrease of 0·5°C for each cycle, and extension at 72°C for 20 s after each annealing period. This was followed by a second set of 20 cycles that only differed from the first set in that the annealing temperature was 50°C. The amplification cycles ended with an extension cycle of 10 min at 72°C. PCR amplification of loci Pi70 and D13 differed from the other loci in that the annealing temperatures for the first and second sets of 20 cycles were 55 and 45°C, respectively.

The amplified PCR products were diluted with SDW to approximately 10 ng μL−1. Sample dilution depended on band intensities which were visually estimated on 2% agarose gels. The diluted samples were mixed in a 1:1 ratio with Hi-Di formamide in a total volume of 10 μL, to which 0·45 μL LIZ600™ size standard (Applied Biosystems) was added. The denatured samples were separated on an ABI 3130xl Genetic Analyzer by the Central Analytical Sequencing Facility at Stellenbosch University, and the data were analysed using Applied Biosystems GeneMapper v.3.7.

Simple sequence repeat allele sizes of the SSA populations were calibrated to the allele sizes of P. infestans isolates in the Eucablight database (http://www.eucablight.org) (Lees et al., 2006; Fry et al., 2009). This was conducted by analysing PCR fragments from P. infestans isolates SCRI C7, 4224E and 3992H (kindly provided by D. E. L. Cooke, JHI) on the sequencer used in this study, and correcting for length variations between the sequencer of this study and the sequencer at JHI. This calibration of allele sizes allows for the submission of data of the current study to the Eucablight database in the near future.

SSR data analyses

All analyses were conducted with clone-corrected data to prevent over-representation of alleles in frequently occurring clones. Gene diversity and number of alleles in populations were determined in popgene v.1.32 (Yeh & Boyle, 1997). Population differentiations (FST) (Weir, 1997) between pairwise comparisons of regional populations (central-east and southeast Africa and South Africa) were conducted in an amova analysis using GenAlEx 6 (Peakall & Smouse, 2006). The significance of genetic differentiation estimates among populations and groups was evaluated using a permutation test with 999 permutations in GenAlEx 6 (Peakall & Smouse, 2006). Population differentiation analyses were not conducted on genotypes (US-1, US-1 variants and KE-1) or between hosts (tomato and potato) because the clone-corrected population sizes were too small.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Characterization with ‘old’ markers

Mating type

All 176 P. infestans isolates tested were of the A1 mating type (Table 1).

mtDNA haplotype

Isolates from all the countries had mtDNA haplotype Ib, except the Kenyan populations, which were either mtDNA haplotype Ia or Ib (Table 1). The mtDNA haplotype Ia isolates were only identified from two fields in Kenya. The one IS field near Kitale only contained mtDNA haplotype Ia (= 20) isolates, whereas the LIS field near Ruiru contained mtDNA haplotype Ia (= 1) and Ib (= 2) isolates.

Gpi analyses

Similar to the mtDNA analyses, polymorphisms in Gpi were only detected in Kenya, where the mtDNA haplotype Ia isolates all had Gpi genotype 90/100 and the mtDNA haplotype Ib isolates had Gpi genotype 86/100. Isolates in all the other countries had Gpi genotype 86/100 (Table 1).

RFLP analysis with probe RG-57

Most isolates had fingerprints identical to the US-1 lineage reference isolate. The exceptions were (i) isolates from the Western Cape Province (South Africa) that were US-1 variant isolates, hereafter referred to as US-1.10, (ii) one isolate from Kenya that was a US-1 variant isolate, hereafter referred to as US-1.11, and (iii) isolates from Kenya with a mtDNA haplotype Ia and Gpi 90/100 genotype, hereafter referred to as genotype KE-1 (Table 1). The US-1 variants are named here for the first time. In the Western Cape Province in South Africa, the US-1.10 isolates contained band 9a in addition to all the other US-1 fingerprint bands (Fig. 2). All the isolates that were fingerprinted from the Western Cape Province had the US-1.10 genotype (Table 1), and included isolates that were obtained from different hosts (potato and petunia), regions (Kouebokkeveld, Sandveld and Phillipi) and years (2005, 2007 and 2008). The US-1.11 isolate contained fingerprint band 23 in addition to all the other US-1 fingerprint bands (Fig. 2). The isolate co-occurred with two US-1 isolates in a field in Limuru, which was less intensively sampled. The KE-1 genotype isolates had a unique fingerprint (Table 1; Fig. 2), which when coded as described by Forbes et al. (1998) was as follows: 1110111111001101001110011.

image

Figure 2.  RG-57 DNA fingerprint patterns of Phytophthora infestans genotypes (a) US-1 and US-1.10 and (b) US-1.11, KE-1 and US-1 that were identified among Sub-Saharan African populations. RG-57 bands differing from the US-1 fingerprint are indicated by arrows.

Download figure to PowerPoint

Unweighted pair group method of averages (UPGMA) analyses of ‘old’ marker data

Unweighted pair group method of averages analyses showed that genotype KE-1 clustered with genotypes that were previously reported from Rwanda (RW-1 and RW-2), and with genotypes from Ecuador (EC-1), Poland (PO-6) and the Netherlands (NL-3 and NL-5). As expected, US-1 variants from South Africa (US-1.10) and Kenya (US-1.11) clustered with other known US-1 variants (Fig. 3) that were identified by Forbes et al. (1998) as old pre-1970s populations.

image

Figure 3.  Unweighted paired group method with arithmetic averages (UPGMA) dendrogram of Phytophthora infestans isolates from Sub-Saharan African countries (US-1, US-1.10, US-1.11 and KE-1) and other countries, based on RG-57 fingerprint, mating type and glucose-6-phosphate isomerase (Gpi ) identity. Nei and Li distances are shown below. Numbers in the tree represents bootstrap values. Bootstrap support values <60% are not indicated. The isolates from other countries were from the USA (US), Japan (JP), Russia (RU), France (FR), Jersey (Channel Islands) (JE), the Netherlands (NL), Poland (PO) and Canada (CA). The genotypes of these isolates were mainly obtained from the studies of Sujkowski et al. (1994), Forbes et al. (1998) and Deahl et al. (2009). (Correction added after online publication 15 March 2012: Missing bootstrap value of 69 was inserted for branch comprising isolates US-1.2 and US-1.4).

Download figure to PowerPoint

Metalaxyl-M sensitivity

In central-east Africa, isolates were sensitive, intermediately resistant or resistant to metalaxyl (= 26). In Kenya the intermediate resistant isolates were from two different fields, and were US-1 genotypes. In Rwanda, Uganda and Tanzania, highly sensitive isolates co-occurred with resistant isolates within the same field. In southeast Africa, all 14 isolates examined were metalaxyl-sensitive. However, the isolates from potato were less sensitive than those from tomato, having a higher percentage growth rate relative to their unamended control than the tomato isolates at 5 and 100 μg mL−1 (data not shown). In South Africa, isolates were either resistant if sampled from potato or sensitive if sampled from tomato (= 24) (Table 1).

Simple sequence repeat (SSR) marker analyses

On a regional basis (central-east Africa, southeast Africa, South Africa), in general, the largest number of alleles was observed for locus D13 followed by loci PiG11 and/or Pi70 (Table 2). Although locus D13 is highly polymorphic in US-1 and US-1 variant populations in SSA, as well as in European populations (Lees et al., 2006), this locus was unfortunately difficult to score for some SSA populations. For example, the amplification of D13 from some isolates from Kenya, Uganda and South Africa yielded sigmoid peak patterns. Therefore, the possibility of incorrect scoring of alleles at locus D13 cannot be excluded. Consequently, in the following section the SSR analyses results are presented including and excluding this locus.

Table 2. The number of multilocus genotypes, allele number and gene diversity of seven microsatellite loci in clone-corrected Phytophthora infestans populations from Sub-Saharan Africa
Region/countryNo. isolatesNo. MLGs including D13aNo. MLGs excluding D13bNo. allelesGene diversityc
D13Pi70PiG11Pi02Pi4BPi56Pi89D13Pi70PiG11Pi02Pi4BPi56Pi89
  1. aNumber of multilocus genotypes (MLGs) when locus D13 was included in the analysis.

  2. bNumber of MLGs when locus D13 was excluded from the analysis.

  3. cGene diversity of the seven loci was calculated including data from locus D13.

Central-east Africa
 Burundi108652612110·730·50·7500·500
 Kenya (US-1 and US-1.11)3015853332110·620·530·630·290·500
 Kenya (KE-1)158632332220·360·50·590·620·350·50·5
 Rwanda1710562222120·780·50·460·10·500·1
 Tanzania2214652342120·670·50·560·20·500·29
 Uganda1916582312120·730·50·4400·500·54
 Total analysed113602683642220·650·510·570·20·480·080·24
Southeast Africa
 Malawi (Potato)83332332110·670·50·610·610·4400
 Malawi (Tomato)74352422110·690·50·690·50·500
 Mozambique31122222110·50·50·50·50·500
 Total analysed187652422110·620·50·60·540·4800
South Africa
 KwaZulu-Natal (Potato)154252112110·750·47000·500
 KwaZulu-Natal (Tomato)44342222110·660·50·470·470·500
 Western Cape (Potato)237462412110·720·50·3700·500
 Western Cape (Petunia)32222332110·50·50·610·50·500
 Total analysed4516962432110·650·560·380·240·500
Total analysed176793583642220·640·520·520·330·490·030·08

There were a few private alleles within the SSA populations, as well as some isolates that contained three alleles at some loci. For locus Pi89, all the KE-1 genotype isolates had three alleles (of which allele 177 was a private allele). Some KE-1 isolates also had three alleles at loci D13 and PiG11. For data analyses of these isolates, the two most common alleles in a field or population were included. The KE-1 genotype also contained a private allele at locus Pi56 (allele 172). The only other two private alleles identified in the SSA populations were one allele at locus Pi70 (allele 195) that only occurred in Kenyan US-1 and KE-1 genotypes, and allele 156 in one isolate from Tanzania at locus Pi02.

Simple sequence repeat MLG analyses that included and excluded locus D13 revealed substantial differences in the number of MLGs in the total SSA population (Table 2). Analyses that included this locus yielded a total of 79 MLGs, as opposed to 35 MLGs identified when locus D13 was excluded. This higher number of MLGs (∼ 50%) identified when locus D13 was included in the analyses was evident in most of the countries, host and genotype groups. The exceptions were (i) isolates from Burundi, (ii) the KE-1 genotypes in Kenya and (iii) groups where less than eight isolates were genotyped (Table 2).

The ‘old’ markers only identified four MLGs (KE-1, US-1, US-1.10 and US-1.11) in the SSA population, as opposed to the many SSR MLGs (either including or excluding locus D13) that were identified for several host groups, regional populations and ‘old’ marker genotype groups. Of the 162 isolates examined with the old markers (designated using mainly mtDNA genotype, with Western Cape isolates designated as US-1.10), 76 and 34 MLGs were identified with the SSRs when D13 was included and excluded, respectively, and provided a better marker to distinguish among individuals. For example, among the three US-1 petunia isolates there were two unique SSR MLGs, and among the 26 potato US-1.10 isolates (all isolates from the Western Cape Province were assumed to be US-1.10 isolates, see Discussion) four MLGs were detected (D13 excluded). Among the KE-1 genotype isolates from Kenya (15 isolates), six SSR MLGs were detected (Table 2). For all the US-1 genotype isolates from all the countries (141 isolates), a total of 65 (including D13) and 26 (excluding D13) SSR MLGs were detected (data not shown).

This high number of SSR MLGs detected in SSA populations was also reflected in the within-field diversity, as between two and six different SSR MLGs were detected within some fields. In one Kenyan field where 14 isolates were genotyped, eight SSR MLGs, the highest for all the fields sampled, were identified when locus D13 was included and six SSR MLGs when D13 was excluded. Another example was the tomato field in KwaZulu-Natal where only four isolates were genotyped, which represented three SSR MLGs.

Several of the MLGs were restricted to a particular country or region, explaining the weak regional structure shown by the amova. However, some MLGs were shared between distantly located countries, probably as a result of the independent emergence of the same variant in more than one location. When locus D13 was excluded from the analyses, several similar MLGs were detected in central-east Africa and South Africa, e.g. the Western Cape shared a MLG with Uganda, and a shared MLG was found in KwaZulu-Natal and Rwanda. There was a tendency for the same MLGs to be detected in neighbouring countries in central-east Africa, but there were exceptions, e.g. MLGs were shared between Rwanda and Kenya. Several MLGs were also shared between central-east Africa and southeast African countries. When D13 was included in the analyses, similar observations were made, except that there was no sharing of MLGs between central-east Africa and southeast Africa, possibly because of the small population size from southeast Africa.

amova analyses for the regional populations (central-east Africa, southeast Africa and South Africa) were conducted by pooling the host data within regions. When D13 was excluded from the analysis, there was an indication of genetic substructuring between the three regional populations (< 0·05) (Table 3), but genetic substructuring became more pronounced when D13 was included in the data set (< 0·007) (Table 4). The population differentiation was, however, moderate to low for all three region combinations (FST = 0·085, 0·034; = 0·005–0·028) (Table 5). amova results were similar when only the potato isolates were included in the analysis (data not shown).

Table 3. Analysis of molecular variance (amova) for Phytophthora infestans populations in three regions (South Africa, central-east Africa and southeast Africa) in Sub-Saharan Africa using six simple sequence repeat loci (D13 was excluded)
Sourced.f.SS% variance FST P
Among populations24·57540·0440·050
Within populations3836·7910−0·2631·000
Among individuals4168·00096−0·2071·000
Total81109·366100  
Table 4. Analysis of molecular variance (amova) for Phytophthora infestans populations in three regions (South Africa, central-east Africa and southeast Africa) in Sub-Saharan Africa using six simple sequence repeat loci (D13 included)
Sourced.f.SSVariance (%) FST P
Among populations28·28340·0530·007
Within populations8087·3440−0·2781·000
Among individuals83160·50096−0·2101·000
Total165256·127100  
Table 5. Estimates of pairwise FST values (below the diagonal) averaged over six simple sequence repeat loci (D13 included) of Phytophthora infestans populations in three regions (South Africa, central-east Africa and southeast Africa) in Sub-Saharan Africa. Significance values indicated above the diagonal
PopulationSouth AfricaCentral-east AfricaSoutheast Africa
South Africa0·0170·028
Central-east Africa0·0340·005
Southeast Africa0·0850·085

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

This study used several ‘old’ genetic and phenotypic markers to show that the structure of P. infestans populations in central-east Africa, southeast Africa and South Africa is very similar, and that populations have not changed much since 2001 (Goodwin et al., 1994; Griffith & Shaw, 1998; Vega-Sánchez et al., 2000; McLeod et al., 2001; Ochwo et al., 2002). In the SSA regions, the US-1 clonal lineage, as defined by the ‘old’ markers, is still widespread and dominant, along with its variants (US-1.10 and US-1.11) that only have a localized distribution. An additional lineage detected was KE-1, as defined by the ‘old’ markers, which was present only in two fields in Kenya in central-east Africa. The SSR markers, although revealing at least 35 different MLGs in contrast to the four MLGs identified by ‘old’ markers, further confirmed that populations in SSA are similar as there was very limited substructuring. If locus D13 was included, only low to moderate substructuring was present, possibly because of the independent emergence of the same variant in more than one location via rare SSR mutations in local populations.

Two main factors have contributed to the lack of population substructuring in SSA: (i) only the A1 mating type is present, which precludes sexual reproduction, and (ii) introductions in this region are probably restricted to migrations among founder populations in SSA countries. The latter is probably the result of the absence of host trade with countries that have genetically diverse P. infestans populations. The importation of potato tubers is not well documented in central-east Africa and southeast Africa, but it is thought to be limited or absent. For South Africa, since the late 1980s, only the importation of in vitro material or G0 seed has been allowed (Fry et al., 2009). The illegal trade of potatoes across borders probably occurs and there is also some formal trade between some central-east African countries (Okoboi, 2001). In north Africa (Morocco, Egypt and Tunisia) trade differs markedly from SSA, with countries such as Morocco importing up to 40 000 t of seed potatoes annually from Europe (Sedegui et al., 2000; Hammi et al., 2002), resulting in P. infestans populations being characterized by the presence of both mating types and several different genotypes (Baka, 1997; Sedegui et al., 2000; Hammi et al., 2002; Jmour & Hamada, 2006).

The only evidence of a possible migration into SSA is the restricted occurrence of the KE-1 lineage (mtDNA haplotype Ia and Gpi 90/100), which might be a recent migrant; alternatively, it could have been present, but undetected prior to 2001 (Goodwin et al., 1994; Griffith & Shaw, 1998; Vega-Sánchez et al., 2000; Ochwo et al., 2002). The KE-1 lineage might be an asexual descendent of the RW-1 and RW-2 genotypes identified in Rwanda from 1984 to 1987 (Goodwin et al., 1994; Forbes et al., 1998). However, this seems unlikely, because even though these genotypes have the same Gpi genotype and mtDNA type, they differ by five RG-57 fingerprint bands, which is quite a large number of genetic differences considering the known diversity for this marker in P. infestans. Although Forbes et al. (1998) designated the RW-genotypes as belonging to the second migration, in the late 1970s, this hypothesis was made prior to the discovery that the Irish potato famine was probably caused by an isolate of mtDNA haplotype Ia (Ristaino et al., 2001; May & Ristaino, 2004; Fry et al., 2009). It is thus possible that the KE-1 lineage and the RW-2 genotype, both being mtDNA haplotype Ia, were introduced along with the US-1 lineage into central-east Africa during the very first migrations of P. infestans to this region in 1941 (Nattrass & Ryan, 1951). The relatively high number of SSR genotypes among the KE-1 genotypes, at least seven among the 15 genotyped isolates, further supports the fact that this lineage may have been present in central-east Africa for some time. This hypothesis could be tested using SSR genotype data of Irish potato famine herbarium material, or analyses of a worldwide collection of P. infestans SSR genotypes. The presence of three alleles at several loci in the KE-1 isolates, as well as locus Pi63 in all SSA isolates, is notable and could suggest a change in ploidy levels, as previously noted in other P. infestans isolates (Lees et al., 2006).

The ‘old’ markers suggested that mutations and/or mitotic recombination are taking place in the SSA populations. This is not unexpected, because the US-1 clonal lineage has been present in this region for decades (Vega-Sánchez et al., 2000; McLeod et al., 2001), allowing for mutations, mitotic recombination and movement of transposons to take place, as exemplified by the widespread occurrence of the US-1.10 variant in the Western Cape province of South Africa. This genotype could have been introduced with petunia, because trade of ornamental plants is unregulated in South Africa, but this seems unlikely considering the limited occurrence of US-1 and US-1 variants worldwide. Metalaxyl resistance could have provided a selective advantage for the US-1.10 lineage, but this seems improbable as high levels of metalaxyl resistance were already prevalent in the US-1 lineage in this region since 1996 (McLeod et al., 2001), when metalaxyl use was also banned. The selective advantage of lineage US-1.10 might be a result of its ability to infect potato and petunia, better adaptation to climatic conditions in this region, or selection by fungicides other than metalaxyl against oomycetes.

In the current study, only a few tomato isolates were obtained, which precluded population differentiation studies. However, it was, interesting to note that although sample sizes were small in regions such as KwaZulu-Natal, all tomato isolates were metalaxyl-sensitive and all potato isolates were resistant. A similar finding was previously reported for several production regions in South Africa (McLeod et al., 2001). In South Africa, it is unknown whether tomato isolates are adapted to tomato, as reported from central-east Africa (Vega-Sánchez et al., 2000; Mukalazi et al., 2001). In central-east Africa, unlike in South Africa, the tomato isolates all had a US-1 variant genotype (US-1.7), and all the potato isolates were the US-1 lineage (Vega-Sánchez et al., 2000).

Metalaxyl resistance was found in most of the SSA countries with the exception of Malawi (potato and tomato isolates) and Mozambique (only tomato isolates sampled). However, for most countries only a few isolates were tested. Metalaxyl resistance was previously reported in South Africa, Uganda and Kenya (McLeod et al., 2001; Olanya et al., 2001); and thus seems to persist in these regions, even in the Western Cape province where metalaxyl was withdrawn in December 1996. This is in contrast to findings in Mexico (Toluca valley), Ireland and the Netherlands, where withdrawal of metalaxyl resulted in a decline in the proportion of resistant strains, or populations reverting back to sensitive populations (Gisi & Cohen, 1996; Grünwald & Flier, 2005). Except for Rwanda, the frequency of metalaxyl use and its efficacy in central-east Africa is mostly unknown (Muhinyuza et al., 2007). Therefore, it is unknown whether these resistant isolates originated as a result of selection pressure being applied, or whether resistant US-1 lineage isolates pre-exist at a baseline level without exposure to metalaxyl (Gisi & Cohen, 1996; Grünwald & Flier, 2005). It is, however, important to note that, in general, metalaxyl resistance does not occur in the US-1 lineage elsewhere (Goodwin et al., 1996). The co-occurrence of metalaxyl-resistant and -sensitive isolates within the same fields in Kenya, Rwanda, Uganda and Tanzania may suggest that either metalaxyl is not used intensively by resource-poor farmers, or that the resistant and sensitive isolates are of equal fitness.

In conclusion, the characterization of SSA P. infestans populations has shown that migrations have not been the cause of more severe epidemics. Even though the KE-1 genotype was detected, it co-occurred with US-1 isolates within one of the fields and had a limited distribution, suggesting that it is not more aggressive than the US-1 lineage. The more severe epidemics could be the result of: (i) metalaxyl resistance, resulting in less effective spray programmes, but this would require investigation of more isolates and knowledge of the actual usage of metalaxyl in SSA countries; (ii) metalaxyl-resistant isolates being more aggressive than metalaxyl sensitive isolates, as reported elsewhere in the world (Kadish & Cohen, 1992); (iii) the occurrence of mutations or mitotic recombination, resulting in more aggressive and/or better adapted genotypes, for example the US-1.10 lineage; and/or (iv) abnormal weather patterns. Continued monitoring of populations will be important, because the future importation of new strains that exacerbate epidemics cannot be excluded. Future studies should investigate the aggressiveness of the different clonal lineages, and metalaxyl-resistant and -sensitive isolates within lineages. From an evolutionary standpoint, it will be important to compare the fitness and aggressiveness of the SSA US-1 populations to those of new-lineage isolates to determine whether Muller’s Ratchet, a mechanism which describes how certain asexual populations may undergo progressive decline in fitness, might be operating as previously hypothesized for the US-1 clonal lineage (Fry et al., 2009).

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

We are grateful for the financial support received from the National Research Foundation (NRF) of South Africa (Thuthuka program), Stellenbosch University, the Department of Science and Technology South Africa, the South African Pesticide Initiative Program (SAPIP) and Potatoes South Africa (PSA). We would like to thank the International Potato Centre (CIP) in Kenya (Dr P. Gildemacher), Uganda (Dr B. Lemaga) and Malawi (Dr P. Demo) for assistance with arranging the sampling trips. We are also grateful to M. Wakahiu (Kenya), N. Senkesha (Rwanda), Dr S. Kuoko (Tanzania), D. Harahagazwe (Burundi), R. Kakuhenzire (Uganda) and M. Soko (Malawi/Mozambique) for assistance with disease sample collections.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  • Baka ZAM, 1997. Mating type, nuclear DNA content and isozyme analysis of Egyptian isolates of Phytophthora infestans. Folia Microbiologica42, 61320.
  • Cooke DEL, Lees AK, 2004. Markers, old and new, for examining Phytophthora infestans diversity. Plant Pathology53, 692704.
  • Daggett SS, Götz E, Therrien CD, 1993. Phenotypic changes in populations of Phytophthora infestans from eastern Germany. Phytopathology83, 31923.
  • Deahl KL, Perez FM, Thompson JM et al. , 2009. Characterization of Phytophthora infestans isolates from Jersey, Channel Islands. Potato Research52, 34154.
  • Forbes GA, Goodwin SB, Drenth A, Oyarzun P, Ordoñez ME, Fry WE, 1998. A global marker database for Phytophthora infestans. Plant Disease82, 8118.
  • Fry WE, Grünwald NJ, Cooke DEL, McLeod A, Forbes GA, Keqiang C, 2009. Population genetics and population diversity of Phytophthora infestans. In: Lamour K, Kamoun S, eds. Oomycete Genetics and Genomics. Diversity, Interactions and Research Tools. Hoboken, NJ, USA: John Wiley & Sons Inc, 13964.
  • Garnica DP, Pinzón AM, Quesada-Ocampo LM et al. , 2006. Survey and analysis of microsatellites from transcript sequences in Phytophthora species: frequency, distribution, and potential as markers for the genus. BMC Genomics7, 245.
  • Gavino PD, Fry WE, 2002. Diversity in and evidence for selection on the mitochondrial genome of Phytophthora infestans. Mycologia94, 78193.
  • Gisi U, Cohen Y, 1996. Resistance to phenylamide fungicides: a case study with Phytophthora infestans involving mating type and race structure. Annual Review of Phytopathology34, 54972.
  • Goodwin SB, Drenth A, Fry WE, 1992. Cloning and genetic analyses of two polymorphic, moderately repetitive nuclear DNAs from Phytophthora infestans. Current Genetics22, 10715.
  • Goodwin SB, Cohen BA, Fry WE, 1994. Panglobal distribution of a single clonal lineage of the Irish potato famine fungus. Proceedings of the National Academy of Sciences, USA91, 115915.
  • Goodwin SB, Sujkowski LS, Fry WE, 1996. Widespread distribution and probable origin of resistance to metalaxyl in clonal genotypes of Phytophthora infestans in the United States and western Canada. Phytopathology86, 793800.
  • Griffith GW, Shaw DS, 1998. Polymorphisms in Phytophthora infestans: four mitochondrial haplotypes are detected after PCR amplification of DNA from pure cultures or from host lesions. Applied and Environmental Microbiology64, 400714.
  • Grünwald NJ, Flier WG, 2005. The biology of Phytophthora infestans at its center of origin. Annual Review of Phytopathology43, 17190.
  • Hammi A, Msatef Y, Bennani A, Ismaili AEL, Serrhini MN, 2002. Mating type, metalaxyl resistance and aggressiveness of Phytophthora infestans (Mont.) de Bary in Morocco. Journal of Phytopathology150, 28991.
  • Jmour W, Hamada W, 2006. First report of A2 mating type of Phytophthora infestans in Tunisia using molecular markers and some observations on its metalaxyl resistance. Tunisian Journal of Plant Protection1, 8591.
  • Kadish D, Cohen Y, 1992. Overseasoning of metalaxyl-sensitive and metalaxyl-resistant isolates of Phytophthora infestans in potato tubers. Phytopathology82, 8879.
  • Knapova G, Tenzer I, Gessler C, Gisi U, 2001. Characterisation of Phytophthora infestans from potato and tomato with molecular markers. In: Biodiversity in Plant Pathology. Proceedings of the 5th Congress of the European Foundation for Plant Pathology, Taormina, Italy . Pisa, Italy: SIPV, 69.
  • Lees AK, Wattier R, Shaw DS, Sullivan L, Williams NA, Cooke DEL, 2006. Novel microsatellite markers for the analysis of Phytophthora infestans populations. Plant Pathology55, 3119.
  • Legard DE, Lee TY, Fry WE, 1995. Pathogenic specialization in Phytophthora infestans: aggressiveness on tomato. Phytopathology85, 135661.
  • May KJ, Ristaino JB, 2004. Identity of the mtDNA haplotype(s) of Phytophthora infestans in historical specimens from the Irish Potato Famine. Mycological Research108, 4719.
  • McLeod A, Denman S, Sadie A, Denner FDN, 2001. Characterization of South African isolates of Phytophthora infestans. Plant Disease85, 28791.
  • Muhinyuza J-B, Nshimiyimana J-C, Kirk WW, 2007. Susceptibility of commonly grown potato cultivars to potato late blight in Rwanda and control with fungicides. African Crop Science Conference Proceedings8, 83540.
  • Mukalazi J, Adipala EM, Sengooba T, Hakiza JJ, Olanya M, Kidanemariam HM, 2001. Metalaxyl resistance, mating type and pathogenicity of Phytophthora infestans in Uganda. Crop Protection20, 37988.
  • Nattrass RM, Ryan M, 1951. New hosts of Phytophthora infestans in Kenya. Nature168, 856.
  • Ochwo MKN, Kamoun S, Adipala E, Rubaihayo PR, Lamour K, Olanya M, 2002. Genetic diversity of Phytophthora infestans (Mont.) de Bary in the eastern and western highlands of Uganda. Journal of Phytopathology150, 5412.
  • Okoboi G, 2001. The Marketing Potential of Potatoes in Uganda and Market Opportunities for Rwanda. Draft Report. Ibadan, Nigeria: International Institute of Tropical Agriculture.
  • Olanya OM, Adipala E, Hakiza JJ et al. , 2001. Epidemiology and population dynamics of Phytophthora infestans in Sub-Saharan Africa: progress and constraints. African Crop Science Journal9, 18593.
  • Oyarzun PJ, Pozo A, Ordoñez ME, Doucett K, Forbes GA, 1998. Host specificity of Phytophthora infestans on tomato and potato in Ecuador. Phytopathology88, 26571.
  • Peakall R, Smouse PE, 2006. GENALEX 6: genetic analysis in Excel. Population genetic software for teaching and research. Molecular Ecology Notes6, 28895.
  • Ristaino JB, Groves CT, Parra G, 2001. PCR amplification of Irish potato famine pathogen from historic specimens. Nature41, 6957.
  • Schiessendopper E, Molnar O, 2002. Characterization of Phytophthora infestans populations in Sub-Saharan Africa as a basis for simulation modeling and integrated disease management. In: Proceedings of the Global Initiative on Late Blight Conference, Germany, 11–13 July, 2002. 140.
  • Sedegui M, Carroll RB, Morehart A et al. , 2000. Genetic structure of the Phytophthora infestans population in Morocco. Plant Disease84, 1736.
  • Sujkowski LS, Goodwin SB, Dyer AT, Fry WE, 1994. Increased genotypic diversity via migration and possible occurrence of sexual reproduction of Phytophthora infestans in Poland. Phytopathology84, 2017.
  • Tamura K, Dudley J, Nei M, Kumar S, 2007. MEGA4: molecular evolutionary genetics analysis (MEGA) software version 4.0. Molecular Biology and Evolution24, 15969.
  • Therrien CD, Tooley PW, Spielman LJ, Fry WE, Ritch DL, Shelly SE, 1993. Nuclear DNA content, allozyme phenotypes and metalaxyl sensitivity of Phytophthora infestans from Japan. Mycological Research97, 94550.
  • Vega-Sánchez ME, Erselius LJ, Rodriguez AM et al. , 2000. Host adaptation to potato and tomato within the US-1 clonal lineage of Phytophthora infestans in Uganda and Kenya. Plant Pathology49, 5319.
  • Weir BS, 1997. Genetic Data Analysis II. Sunderland, MA, USA: Sinauer Associates Inc.
  • Yeh FC, Boyle T, 1997. POPGENE Version 1.32: Microsoft Window-Based Freeware for Population Genetic Analysis. Edmonton, Canada: University of Alberta.