Fusarium wilt of cotton may commonly result from the interaction of Fusarium oxysporum f. sp. vasinfectum with Belonolaimus longicaudatus

Soil and plant samples were collected between June and November in 2015 and 2016 from cotton fields showing symptoms of FW. The survey included six counties and 11 fields in 2015 and seven counties and 17 fields in 2016 (Table 1). The Tifton fields 1, 4, and 5 in 2015 are the same as fields 3, 5, and 4 in 2016, respectively. County extension agents provided general information about soil proprieties such as soil texture, previous crop, and cotton varieties planted.

In total, 8 to 15 samples were collected from each field. For each sample, a single plant showing symptoms of FW was arbitrarily selected, carefully uprooted, and the roots were observed for the presence or absence of root-knot nematode galls. The soil associated with the plant’s root system and a piece of the diseased plant’s stem from near the base were sealed in a plastic bag. The samples were stored at 10°C until processing. Nematodes were extracted from 150 cm³ of soil for each sample by the centrifugal flotation method (Jenkins, 1964), and the number of individuals for each genus of plant-parasitic nematode present was recorded.

For 2015 samples, Fov was isolated by disinfesting infected stem pieces in 0.875% NaOCl for 1 min, and then placing them into Komada selective medium (Komada, 1975) for five days. Five stem pieces per plant were cut open longitudinally and placed on a single plate. Individual colonies were then selected and transferred to a new plate containing Komada medium and allowed to grow for seven days. Then, single-spore isolations were made through successive serial dilution of conidial suspensions.

DNA was extracted separately from at least three different single-spore isolates per field. DNA extraction was accomplished by inoculating 125 ml flasks of potato dextrose broth with a mycelial plug (0.7 cm diameter) from the single-spore isolates and allowing them to grow for four days on a rotational shaker at 100 rpm. All the mycelium was then dried overnight on a sterile plate prior to using the modified DNA extraction method of Abd-Elsalam et al. (2003). Mycelium was macerated in liquid nitrogen and then put into 1.5 ml Eppendorf tubes prior to adding extraction buffer (200 mM Tris pH = 8.5, 250 mM NaCl, 25 mM EDTA, and 0.5% SDS). The sample was mixed on a vortex mixer and 4 µl of RNAase was added prior to incubation in a water bath for 65°C for 10 min. Then, 130 µl of 3 M sodium acetate (pH = 5.2) was added and the tubes were held at −20°C for 10 min. After centrifuging the sample for 15 min at 4,000 rpm, a 400 µl aliquot of the supernatant was collected and transferred to a 1.5 ml Epperndorf tube. The sample was then mixed in a 1:1 ratio with chloroform for 3 min and then centrifuged for 10 min at 8,000 rpm. The supernatant was removed and 650 µl of isopropanol was added to the tube to precipitate the DNA. Isopropanol was removed after centrifuging tubes for 10 min at 4,000 rpm. Finally, 300 µl of 70% ethanol was added into the tubes and centrifuged for 1 min at 4,000 rpm. Ethanol was removed from the tube and DNA allowed to dry overnight. Deionized water (100 µl) was added to rehydrate the DNA. The translation elongation factor (EF–1α) and IGS regions were PCR-amplified and sequenced for the Fov isolates as described by Kim et al. (2005) and Cianchetta et al. (2015). PCR reactions consisted of 12.5 μl of 2X Mean Green Master Mix (Syzygy Biotech; Grand Rapids, MI), 1 μl of each 10 μM primer and 2 μl of genomic DNA in a 25 μl reaction. Amplification reactions were performed in a thermocycler (PTC–100; MJ Research, Watertown, MA) under conditions as described by Cianchetta et al. (2015). PCR products were purified with a GeneJET PCR Purification Kit (Thermo Fisher Scientific, Inc., Waltham, MA) following the manufacturer’s instructions. A 24 μl sample of 40–68 ng/µl of PCR product and 12 μl of each 3 μM primer were sent to Quintara Biosciences (2600 Hilltop Dr, Building B, R332, Richmond, CA) for sequencing. Sequences were manually checked and aligned with known Fov sequences from GenBank database and private database to identify race or genotype.

For 2016 samples, Fov was isolated by washing cotton stems in soapy water, then surface sanitizing for 1 min in 95% ethanol, 2 min in sodium hypochlorite solution and 1 min in sterile deionized water. The outer layer of bark was removed, and small slivers of stem tissue were plated on acidified quarter-strength potato dextrose agar. Petri plates were incubated at 23°C with a 12-hr photoperiod for six to seven days. Single spores of cultures morphologically resembling Fusarium were transferred to quarter-strength potato dextrose agar under a dissecting microscope.

Genomic DNA was extracted from four single-spore isolates per field. Isolates were grown on potato dextrose agar overlain with sterile cellophane for six to seven days, after which time mycelia were harvested and lyophilized. Approximately 50 milligrams of lyophilized mycelia were placed in 2-ml microcentrifuge tubes with sterile glass beads, and macerated into a fine powder in a Geno/Grinder® (SPEX SamplePrep, Metuchen, NJ). DNA was extracted using a DNeasy Plant Mini kit (QIAGEN, Valencia, CA) following manufacturer protocol with the following modification: samples were eluted in 25 μl AE buffer (as opposed to 50 μl as stated in protocol) to increase the final concentration of DNA. PCR amplification, sequencing and genotyping were conducted as previously described.

Fov race 1 was found in all sampled counties in 2015 and 2016. In 2015, no previously unreported races or genotypes were found in Georgia. In addition to race 1, genotype LA108 was found in Lowndes County and LA110 was found in Tift County. In 2016, two new genotypes, LA127/140 and MDS–12, were identified in the fields sampled. Fov races 2 and 8 were found in Lowndes County; race 8 and genotypes LA127/140, LA110, and LA108 were found in Tift County; genotypes LA108 and LA110 were found in Cook County; genotypes LA108, LA110 and MDS–12 were found in Colquitt County; race 2 and genotype LA110 were found in Coffee County; genotype LA108 was found in Ware County; and genotype LA110 was found in Worth and Tattnall Counties (Table 2 and Figure 1).

Fig. 1

For the fields included in our survey in 2015, the predominant cotton cultivar planted was DP 1252 B2RF, which is susceptible to M. incognita. The M. incognita-resistant varieties planted in 2015 were DP 1454 NR B2RF and PHY 487 WRF, which were planted in 18% of the fields sampled. In 2016, the resistant variety DP 1558 NR B2RF was planted in 35% of the fields sampled (Table 1). Belonolaimus longicaudatus was present in all counties sampled in 2015 and in 70% of the counties in 2016 (Table 2). Meloidogyne incognita was found in 50% of the counties in 2015 and 70% in 2016. Other plant-parasitic nematodes associated with FW-damaged plants in our survey were Pratylenchus brachyurus (lesion), Helicotylenchus spp. (spiral), Hoplolaimus columbus (lance), Criconemella spp. (ring) and Rotylenchulus reniformis (reniform) (Table 2). Except for R. reniformis, the other genera were typically found at low population densities (data not shown).

The number of samples for each county in our survey (both years combined) is listed in Table 3. Only samples with one or more of the three species of nematodes reported (M. incognita and B. longicaudatus) or with likely potential (P. brachyurus) to interact with Fov to cause FW are listed. Rotylenchulus reniformis is not included because it was not found in 2015 and was found only in one field in 2016. A majority of the total survey samples (all counties for both years) contained B. longicaudatus either alone (34% of samples) or together with P. brachyurus (20%). Pratylenchus brachyurus and M. incognita were present in 7.3 and 6.6% of the samples, respectively, and both were present in 2.4% of samples. Meloidogyne incognita and B. longicaudatus were together in 2.4% of samples, all of which were in Tift and Worth counties. A greater number of samples were found with B. longicaudatus alone in Tattnall County (64% of samples) and in Ware County (65% of samples) than in the other counties (Table 3).

For comparison to the results in our survey, samples from cotton fields in Georgia submitted to the Extension Nematology Laboratory at the University of Georgia from 2013 to 2016 were examined. During those four years, B. longicaudatus was detected in 1.2% of the samples from all counties throughout Georgia.

Belonolaimus longicaudatus can interact with Fov to cause FW in cotton (Cooper and Brodie, 1963), and control of B. longicaudatus can reduce FW in the field (Brodie and Hauser, 1970). Aldicarb was widely used for nematode control in cotton production before the production and sale of aldicarb by Bayer CropScience was phased out in the USA beginning in 2011 (https://archive.epa.gov/pesticides/reregistration/web/html/aldicarb_fs.html). Aldicarb is now produced and sold by another company (Ag Logic) in the USA, but it is no longer widely used. Although aldicarb was applied in cotton production primarily to manage M. incognita and R. reniformis, non-target nematode species including B. longicaudatus also would have been suppressed. Aldicarb is effective at reducing B. longicaudatus levels in the field (Rhoades, 1981; Weingartner and Shumaker, 1990). In our samples from plants with FW, B. longicaudatus was found much more frequently than expected based on the frequency that the nematode is found in samples from cotton fields submitted to the University of Georgia Extension Nematology Laboratory. The greatly decreased used of aldicarb and reliance on other nematode control tactics may have allowed B. longicaudatus to thrive, resulting in increased incidence and greater population levels of B. longicaudatus, and increases in B. longicaudatus levels could result in more interactions with Fov and greater FW incidence. Better control of B. longicaudatus should help reduce the incidence of FW. In fields where B. longicaudatus was not found, M. incognita was present. FW can be more severe in fields where both B. longicaudatus and M. incognita occur together (Yang et al., 1976). A few individual plant samples had neither nematode found associated with its roots, which could mean that neither nematode was involved in causing FW on that plant or that any nematodes involved were below our detection limit.

Race 1 was the predominant race of Fov in South Georgia in previous surveys (Holmes et al., 2009; Cianchetta et al., 2015). Additionally, races 2 and 8 and genotypes LA108 and LA110 also have been identified in Georgia (Holmes et al., 2009; Cianchetta et al., 2015). However, our survey is the first to report the presence of genotypes MDS–12 and LA127/140 in Georgia cotton fields. MDS–12 has been previously found in Alabama, which was the first report of this genotype in the USA (Bennett et al., 2013). Importantly, our survey did not find Fov race 4 or either of the two Australian biotypes, which cause severe FW without M. incognita infection (Kim et al., 2005).

Although the underlying mechanism of how M. incognita infection increases the severity of FW is unknown, it has been shown that M. incognita population levels positively correlate with FW incidence and severity (Garber et al., 1979). However, some fields in our survey in Berrien, Cook, Lowndes and Tattnall counties with high levels of FW did not have high levels of M. incognita, and the nematode was not detected in some fields. Most of our fields were infested with B. longicaudatus, suggesting that the Fov isolates were likely to be interacting with it to cause FW. Although other nematode species may form a disease complex with Fov (Cooper and Brodie, 1963, Seinhorst and Kuniyasu, 1971), our survey showed strong evidence of B. longicaudatus being the one most commonly associated with FW in Georgia cotton fields. Therefore, although a number of cotton cultivars with high levels of resistance to M. incognita are now available to growers, planting these cultivars may not provide adequate protection against FW in Georgia.

Another possible explanation for the recent increase in the incidence of FW in Georgia could be a genetic change in Fov allowing it to become more virulent. Fov race 4 is highly virulent on many cotton genotypes (Doan and Davis, 2014), and phylogenic trees have shown that LA110 and LA108 are genetically extremely similar to race 4, which may suggest that they share the high virulence of race 4 (Holmes et al., 2009). However, race 4 is reported to be a root-rot pathotype that caused disease following root-dip assays but not following stem injection and is therefore not considered a vascular-competent pathogen, whereas LA110 and LA108 are reported to be vascular-competent pathotypes that interact with nematode infection to increase disease severity in the field (Bell et al., 2017). The genetic differences between race 4 and genotypes LA108 and LA110 could be associated with differences in virulence or the need to interact with nematodes to cause severe disease under field conditions.

The interaction of Fov with nematodes other than M. incognita is not well documented, has been the subject of only limited research, and is seldom mentioned as a possible factor in FW incidence. The relatively common association of Fov with B. longicaudatus resulting in FW identified in our study requires additional research to be more fully understood. Better recognition that B. longicaudatus can interact with Fov to cause FW in cotton may lead to better control of FW by allowing growers to identify at-risk fields and minimize FW through better control of the nematode.