Differential interactions between strains of Rhizorhapis, Sphingobium, Sphingopyxis or Rhizorhabdus and accessions of Lactuca spp. with respect to severity of corky root disease



Over 100 isolates of Rhizorhapis suberifaciens, Sphingobium (Sb.) sp., Sb. mellinum, Sb. xanthum, Rhizorhabdus sp., and Sphingopyxis sp. (Sphingomonodaceae) were tested for pathogenicity on lettuce (Lactuca sativa) cultivars Salinas and Green Lakes, susceptible and resistant, respectively, or their resistant descendent breeding line (B.L.) 440-8, to R. suberifaciens type strain CA1T. Rhizorhabdus sp. CA15 and NL2, R. suberifaciens CA3, and Sphingopyxis CA32 were equally virulent to Green Lakes or B.L. 440-8 and Salinas. Over 40 accessions from four Lactuca species were tested for resistance to R. suberifaciens CA1T/CA3 or Rhizorhabdus sp. CA15/NL2. All lettuce accessions with resistance to CA1T were susceptible to isolates CA15, NL2 and/or CA3. None of the Lactuca lines were highly resistant to all four isolates. There was a significant differential interaction between eight Lactuca lines and ten isolates of Rhizorhapis and related genera with respect to corky root severity. Three strains of isolates were distinguished: (i) isolates with a similar virulence pattern as R. suberifaciens CA1T, (ii) isolates with a virulence pattern similar to that of R. suberifaciens CA3 and Sphingopyxis sp. CA32, and (iii) isolates of Rhizorhabdus being moderately aggressive to all Lactuca lines. Thus, strains belonging to several genera can cause similar symptoms (a rare phenomenon) but have different virulence patterns on Lactuca species and cultivars.


Corky root disease constitutes an important problem for lettuce (Lactuca sativa) production in the coastal valleys of California (van Bruggen et al., 1988; O'Brien & van Bruggen, 1992b). It has not been reported for the San Joaquin (Central) Valley of California. However, severe corky root was observed on lettuce plants that were continuously sprinkler irrigated (for cooling purposes) near Modesto in the Central Valley, and bacteria causing this disease were isolated and identified (Francis et al., 2014). The disease has also been reported for other parts of North America, in particular Florida, Wisconsin and New York (van Bruggen et al., 1989; Datnoff & Nagata, 1990), Europe (van Bruggen & Jochimsen, 1992), and Australia (van Bruggen & Jochimsen, 1993). Losses in marketable yield of a susceptible cultivar ranged from 30 to 80% in field experiments in Florida and California, respectively (Datnoff & Nagata, 1992; O'Brien & van Bruggen, 1992a,b).

Corky root is caused by Gram-negative bacteria belonging to the genera Rhizorhapis (formerly known as Rhizomonas), Rhizorhabdus (Rr.) and Sphingopyxis (van Bruggen et al., 1990c, 1993; Francis et al., 2014). So far, only one species of Rhizorhapis has been described, R. suberifaciens. Most strains that cause corky root in California, Florida and Australia belong to this species (Francis et al., 2014). Recently, however, corky root-inducing strains have also been found among isolates of Sphingobium (Sb.) mellinum, Sbxanthum, Sphingobium sp., Rhizorhabdus sp., and Sphingopyxis sp. (Francis et al., 2014). It has been suggested that R. suberifaciens be reclassified as Sphingomonas suberifaciens (Yabuuchi et al., 1999) or Sphingobium suberifaciens (Chen et al., 2013), based on a very limited number of strains. However, a recent polyphasic study on a large number of R. suberifaciens strains and of strains from related genera showed that R. suberifaciens causing corky root of lettuce is clearly distinct from Sphingomonas sensu stricto, and that none of the true Sphingomonas strains are pathogenic to lettuce (Francis et al., 2014). All genera mentioned above are related, belonging to the Sphingomonadaceae (Takeuchi et al., 2001; Francis et al., 2014).

The host range has been determined for several R. suberifaciens isolates, but not for many other isolates of Rhizorhapis sp. or related genera. The host range of R. suberifaciens is rather narrow, including lettuce, endive (Cichorium endivia), common sowthistle (Sonchus oleraceus) and prickly lettuce (Lactuca serriola), all belonging to the tribe Cichoreae of the Asteraceae family (van Bruggen et al., 1990a). However, R. suberifaciens has been isolated from various non-host plant roots (van Bruggen et al., 1990b), indicating that it may be a common rhizosphere organism. Rhizorhapis suberifaciens was indeed able to survive on barley (Hordeum vulgare) roots, although the population on barley was not as high as on lettuce or prickly lettuce (O'Brien & van Bruggen, 1991).

Corky root can be partially controlled by soil fumigation with methyl bromide/chloropicrin or methyl isothiocyanate fumigants (O'Brien & van Bruggen, 1990) but none of these fumigants have been registered for lettuce and some provide only variable control. Crop rotation can reduce losses from corky root (Alvarez et al., 1992), and cover cropping with rye reduced the disease to a limited extent (van Bruggen et al., 1990b). However, these practices would need to be accompanied by other cultural practices such as reducing the amount of N fertilizers to control the disease (van Bruggen et al., 1990b). Significant control of corky root was also obtained by using transplants instead of direct seeding (Nagata & Datnoff, 1991; van Bruggen & Rubatzky, 1992).

The most effective method of control is plant resistance to the pathogen. The resistant cultivars Green Lakes, Montello and Marquette were originally developed in Wisconsin (Sequeira, 1970) by screening for resistance in naturally infested soil. The resistance in these cultivars was derived from a romaine lettuce type (P.I. 171669) that originated from Turkey (Ryder & Waycott, 1994). This resistance was initially thought to be controlled by several recessive genes, but a detailed genetic analysis was not performed. Green Lakes and Montello were later used as sources of resistance in breeding programmes in Florida (Guzman, 1984; Guzman et al., 1992; Nagata et al., 1992) and Green Lakes in California (Brown & Michelmore, 1988; Ryder & Waycott, 1994). Breeding lines (B.Ls) resistant to corky root in naturally infested soil were resistant to the most common strain of R. suberifaciens (CA1T), used for screening of Lactuca accessions and genetic analysis (Brown & Michelmore, 1988). Resistance in Green Lakes, Montello, Marquette and various plant introductions (P.Is) of L. serriola, Lactuca saligna and Lactuca dentata was shown to be partial, monogenic and recessive; this single, recessive resistance gene was named cor (Brown & Michelmore, 1988). The cor gene has been mapped and flanking markers are available for marker-assisted selection (McHale et al., 2009). Breeding and selection in California resulted in two other crisphead cultivars (Glacier and Misty Day) with resistance to corky root conferred by cor (Ryder & Waycott, 1994). These cultivars were crossed with other lettuce types to create additional resistant cultivars with cor (Mou et al., 2007; Mou, 2011). Resistant cultivars, of several lettuce types, are now available from various seed companies. However, resistance used by most companies relies on the same cor gene and the potential for emergence of virulent strains that can overcome this single-gene resistance is therefore a realistic possibility (van Bruggen, 1997; Mou & Bull, 2004), particularly because this gene is deployed on a large scale in California and elsewhere. One exception is B.L. 41-52-RZ which derived its resistance from the cultivar King Henry and was patented in 2007 (Schut et al., 2007); the gene conferring corky root resistance is not publicly known in this case.

So far, the resistance based on the cor gene has proved to be durable and effective, with the exception of lettuce fields that are sprinkler irrigated daily to cool down the crop, as in the Modesto area in California. However, there is a potential risk that the resistance will be overcome as a result of the enormous genetic variability in pathogen populations that span several genera in the Sphingomonadaceae (van Bruggen, 1997; Francis et al., 2014). In a previous study, no differential interaction was detected between 12 isolates of R. suberifaciens or Sb. mellinum and cultivars Green Lakes (resistant) and Salinas (susceptible) (van Bruggen et al., 1989). However, there have been preliminary reports of the occurrence of bacterial isolates that were equally virulent on the susceptible cultivar Salinas and the resistant B.L. 440-8 (with the cor gene originating from Green Lakes) (van Bruggen, 1994, 1997). Therefore, an attempt was made to find new sources of resistance by planting cultivars and accessions of L. serriola and Lactuca virosa, preselected after testing with R. suberifaciens CA1T, in a field that presumably contained other pathogenic members of the Sphingomonadaceae (Mou & Bull, 2004; Bull, 2009). With the expansion of the collection of identified corky root-inducing isolates (Francis et al., 2014), it is important to check whether there are strains among this collection that overcome the resistance developed against R. suberifaciens CA1T. It is also important for the lettuce industry to find P.Is with resistance to all known bacterial strains causing corky root before virulence to the currently deployed cor resistance gene becomes widespread.

The objectives of this study were: (i) to identify strains of bacteria belonging to, or related to, Rhizorhapis, Sphingobium, Sphingopyxis and Rhizorhabdus that cause equally severe corky root on cultivars Salinas and Green Lakes or derivatives of Green Lakes; (ii) to determine if selected isolates belonging to, or related to, Rhizorhapis, Sphingobium, Sphingopyxis and Rhizorhabdus have differential virulence to various lettuce cultivars, B.Ls and accessions of L. sativa, L. serriola, L. saligna and L. virosa; and (iii) to find accessions that are equally resistant to all strains causing corky root of lettuce.

Materials and methods

Bacterial strains and inoculum preparation

One hundred and five isolates were tested for virulence in this study. These comprised 53 isolates of R. suberifaciens from California, Florida and Australia, eight of Sb. mellinum from California, Wisconsin and Europe, four of Sphingobium sp. from Florida and Europe, two of Sb. xanthum from Florida, four of Rhizorhabdus sp. from California and Europe, eight of Sphingopyxis from California and Europe, five of Sphingomonas sp. from California, New York and Europe, and 21 unidentified isolates from California and Europe. Isolates from different genera with various reactions to the lettuce accessions that are specifically mentioned in this paper are listed in Table 1. Isolates that were nonpathogenic to lettuce and those that were less pathogenic to lettuce accessions with resistance to R. suberifaciens CA1T than to lettuce accessions susceptible to this isolate are not listed individually.

Table 1. Bacterial strains with various reactions to the lettuce accessions that are specifically mentioned in this study, their source and origin
SpeciesStrainCollection no.aSourceOrigin
  1. a

    LMG, Collection of the Laboratory of Microbiology, Ghent University, Ghent, Belgium.

  2. b

    Isolated from soil using lettuce seedlings as bait.

  3. c

    No LMG number available; culture AU117 was reported in van Bruggen et al. (1993) and was similar to AU44, but was lost over time.

Rhizorhapis suberifaciens CA1TLMG17323TLettuceSalinas Valley, California
CA3LMG 12532SowthistleSalinas Valley, California
FL16LMG 12537SoilbBelle Glade, Florida
AU44LMG 12543SoilLockyer Valley, Queensland, Australia
AU117cSoilLockyer Valley, Queensland, Australia
Rhizorhabdus sp.CA15LMG 12571LettuceNear Watsonville, Pajaro Valley, California
NL2LMG 12576SoilMaasland, The Netherlands
Sphingobium mellinum WI4TLMG 11032TSoilWisconsin
Sphingobium xanthum FL21LMG 12553SoilBelle Glade, Florida
Sphingobium sp.FL18LMG 12552SoilBelle Glade, Florida
FL24LMG 12555SoilBelle Glade, Florida
NL4LMG 12559SoilMaasland, The Netherlands
Sphingopyxis sp.CA32LMG 12569LettuceModesto, Central Valley, California

All isolates were grown on solid S medium (van Bruggen et al., 1988) for 4–5 days at 30°C for inoculum preparation. Suspensions were prepared from bacterial colonies transferred to sterile distilled water. Bacterial densities were assessed with a spectrophotometer and adjusted to 108–109 colony-forming units (CFU) mL−1 using separate standard curves for two classes of bacterial sizes: 1·4 × 0·4 μm (R. suberifaciens CA1T) and 0·9 × 0·5 μm (Sb. mellinum WI4T).

Plant growth, inoculation and disease assessment

Uncoated lettuce seeds were planted in 5 × 5 × 5 cm pots filled with vermiculite in insect-proof cages as described previously (van Bruggen et al., 1989). One week after planting, seedlings were thinned to one plant per pot, and inoculated with 2 mL bacterial suspension (108–109 CFU mL−1) at the base of each plant. Control plants received 2 mL sterile distilled water. All plants received water, 0·5 × Hoagland solution, or a Ca(NO3)2 + KNO3 (each at 5 mm) solution on alternate days (van Bruggen et al., 1989).

Experimental design

The experimental design to test large numbers of bacterial isolates on many plant genotypes was subject to the following constraints: (i) it was impossible to test all isolates on all genotypes; (ii) to avoid cross-contamination by insects, such as fungus gnats, plants needed to be placed in insect-proof cages; and (iii) the number of plants per cage, pseudoreps, had to be limited in addition to true replications and the number of cages per bacterial isolate. Therefore, the following strategy was developed with three consecutive series of experiments: (i) test the whole isolate collection for pathogenicity and virulence on two cultivars to determine the isolates able to overcome the commonly used source of resistance, (ii) test three representative bacterial isolates on a large number of plant genotypes to identify potential differential interaction, (iii) test 10 selected isolates on eight plant genotypes, selected on the basis of the results of the first two sets of experiments, in a three-fold replicated experiment, to maximize the chance of detecting differential interactions.

First, in five tests, 105 isolates of Rhizorhapis and related genera were tested for pathogenicity to cultivar Salinas and B.L. 440-8, an F5 pedigree B.L. of Salinas and Green Lakes with resistance to R. suberifaciens CA1T. Strain CA1T and sterile water were used as controls. There were four plants per isolate per cultivar combination, placed in individual saucers, grouped by bacterial isolate in insect-proof cages to prevent cross-contamination by watersplash and/or insects. Within each cage, all plants were randomized.

Secondly, in two tests, 23 and 20 lettuce cultivars, B.Ls or P.Is were inoculated with one of three bacterial isolates or left uninoculated. In the first test, Rhizorhabdus sp. CA15 and NL2 were compared with R. suberifaciens CA1T; in the second test, Rhizorhabdus sp. CA15 was compared with R. suberifaciens strains CA1T and CA3. Five cultivars or B.Ls were the same in both experiments for comparison of the tests. There were four plants per treatment in both tests. To prevent cross-contamination, plants inoculated with different isolates were placed in separate insect-proof cages. Within each cage, all plants were randomized.

Thirdly, in a replicated experiment, eight lettuce cultivars, B.Ls or P.Is were inoculated with 10 bacterial isolates of R. suberifaciens, Sb. mellinum, Sb. xanthum, Sphingobium sp., Rhizorhabdus sp. and Sphingopyxis sp. The selected isolates originated from a wide range of geographical locations, and exhibited differential virulence to various cultivars and B.Ls in previous tests. Plants inoculated with different isolates were placed in separate insect-proof cages. There were five plants per treatment, and the experiment was replicated three times (considered as blocks).

Disease observations

Three to four weeks after inoculation, all plants were uprooted, and the roots were inspected for corky root symptoms. Symptoms consist of yellow to brown discolourations of the taproot and longitudinal cracks in the cortex in the case of a severely diseased taproot. Lateral roots rot. When the disease is severe, shoots are stunted and may wilt or die. A 0–9 scoring scale (Brown & Michelmore, 1988) was used. This is an ordinal scale, largely based on the percentage of the root area with symptoms, so that the data could be subjected to analysis of variance.

Statistical analysis

Because the experimental units could not be completely randomized in the tests in which large numbers of isolates or plant lines were screened, mean disease severity ratings and their standard deviations were calculated for the first two sets of tests; analysis of variance could not be performed for those tests, because only pseudoreps within the cages were available. Thus, mean corky root scores (0–9 scale) and their standard deviations 3–4 weeks after inoculation of lettuce cultivar Salinas or B.L. 440-8 with various isolates of Rhizorhapis, Sphingobium, Rhizorhabdus and Sphingopyxis were first calculated for individual tests. Secondly, means and standard deviations of corky root scores on cultivars Salinas and Green Lakes or B.L. 440-8 were calculated for various independent tests over several years. Thirdly, means and standard deviations of corky root scores on many cultivars and B.Ls inoculated with three corky root-causing isolates were calculated. Finally, disease severity data from the last experiment were subjected to anova with a split-plot design with three replications (isolates in the main plots, replications over time in blocks, and cultivars/B.Ls in the subplots). The statistical analysis was carried out with proc glm in sas v. 6.3 (SAS Institute Inc.). Residual values were checked for normality with Wilk's lambda test.


Roots of all uninoculated control plants remained healthy in all tests. The average corky root severity caused by R. suberifaciens CA1T on lettuce cultivar Salinas was 6·3 ± 0·7 in the five experiments. Experiments 1 and 3 resulted in lower corky root scores than the other experiments (Table 2). However, the isolates that caused equally severe symptoms on Salinas and B.L. 440-8 were consistent in the various experiments.

Table 2. Mean corky root scores (0–9 scale) on lettuce cultivar Salinas and breeding line (B.L.) 440-8 inoculated with all isolates included in this study. Disease severity was scored 3–4 weeks after inoculation with various isolates of Rhizorhapis suberifaciens or related genera
ExperimentSpeciesIsolateSalinasB.L. 440-8
  1. a

    Mean ( ± standard deviation) of four plants.

  2. b

    Symptoms on lateral roots only.

  3. c

    Various isolates of Rhizorhapis or related genera that were nonpathogenic on either cultivar or weakly pathogenic on B.L. 440-8.

1 R. suberifaciens CA1T3·0 ± 0·0a1·5 ± 0·7
Sphingobium mellinum WI4T5·0 ± 1·41·3 ± 0·5
Sphingobium sp.FL241·0 ± 0·61·0b± 0·5
 NL44·8 ± 0·91·0 ± 0·0
Rhizorhabdus sp.CA153·8 ± 0·54·0 ± 0·0
11 other isolatesc 0·0 ± 0·00·0 ± 0·0
2 R. suberifaciens CA1T7·3 ± 0·62·5 ± 0·6
 FL164·0 ± 0·81·0 ± 0·0
Sphingobium mellinum WI4T7·0 ± 0·01·0 ± 0·8
Sphingobium sp.FL241·8 ± 1·50·5b± 1·2
Rhizorhabdus sp.CA154·5 ± 0·65·3 ± 0·9
 NL23·6 ± 1·13·6 ± 0·5
8 other isolates 0·0 ± 0·00·0 ± 0·0
3 R. suberifaciens CA1T4·8 ± 2·01·3 ± 0·5
 FL162·3 ± 1·51·3 ± 0·5
 CA33·3 ± 0·93·5 ± 2·4
Sphingobium xanthum FL210·0 ± 0·01·3 ± 0·9
Sphingobium sp.FL180·0 ± 0·00·8 ± 0·9
17 other isolates 3·3 ± 2·51·2 ± 1·1
1 other isolate 0·0 ± 0·00·0 ± 0·0
4R. suberifaciens CA1T8·0 ± 0·81·5 ± 1·3
Rhizorhabdus sp.CA153·8 ± 1·02·8 ± 0·5
35 other isolates 3·7 ± 2·31·7 ± 1·3
4 other isolates 0·0 ± 0·00·0 ± 0·0
5 R. suberifaciens CA1T8·3 ± 0·53·8 ± 0·5
Sphingopyxis sp.CA323·3 ± 0·93·0 ± 1·1
4 other isolates 1·8 ± 1·50·9 ± 0·8
13 other isolates 0·0 ± 0·00·0 ± 0·0

Of the 105 isolates of Rhizorhapis and related genera tested for pathogenicity on Salinas and B.L. 440-8, four (R. suberifaciens CA3, Rhizorhabdus sp. CA15 and NL2, and Sphingopyxis sp. CA32) were more virulent on B.L. 440-8 than on Salinas, or equally virulent to both cultivars. However, they were less virulent on Salinas than CA1T (Table 2). CA3, isolated from prickly sowthistle in the Salinas Valley in California, was the only isolate out of the 53 R. suberifaciens isolates tested that was equally virulent on Salinas and B.L. 440-8 (Table 2). Rhizorhabdus isolates CA15 and NL2, isolated from Watsonville, California, and the Netherlands, respectively, are the only pathogenic isolates of Rhizorhabdus sp. found so far, and both are equally virulent on Salinas and B.L. 440-8 (Table 2). CA32 originates from Modesto, California, and belongs to an as yet unnamed species of Sphingopyxis; it was equally virulent on Salinas and B.L. 440-8 (Table 2). One isolate (FL24) of an unnamed species of Sphingobium caused slight symptoms on lateral roots of both lettuce cultivars but not on taproots, whereas two other Sphingobium isolates (Sphingobium sp. FL18 and Sb. xanthum FL21) caused slight corky root on B.L. 440-8 but no symptoms on Salinas (Table 2). A similar virulence pattern as described for the five separate tests (Table 2) was detected when the data of several tests were combined (Table 3). Rhizorhapis suberifaciens CA3 and Rhizorhabdus sp. CA15 and NL2 were equally virulent on Salinas and Green Lakes or B.L. 440-8. In addition, R. suberifaciens FL16 and Sphingopyxis sp. CA32 were moderately and almost equally virulent on all three lettuce lines (Table 3).

Table 3. Mean corky root scores (0–9 scale) on lettuce cultivars Salinas and Green Lakes or breeding line (B.L.) 440-8. Disease severity was scored 3–4 weeks after inoculation with various isolates of Rhizorhapis, Sphingobium, Rhizorhabdus and Sphingopyxis in multiple tests over several years
SpeciesIsolateNumber of plantsSalinasGreen Lakes or B.L. 440-8
  1. a

    Mean ( ± standard deviation) of the number of plants indicated.

Rhizorhapis suberifaciens CA1T356·5 ± 1·7a2·4 ± 0·9
CA3184·6 ± 1·24·1 ± 1·2
FL1683·2 ± 0·93·2 ± 0·9
AU44143·6 ± 0·92·4 ± 2·1
Sphingobium mellinum WI4T185·9 ± 0·81·8 ± 0·5
Sphingobium xanthum FL21140·7 ± 1·00·7 ± 0·5
Sphingobium sp.FL1841·2 ± 0·70·8 ± 0·0
FL2481·2 ± 0·70·9 ± 0·2
NL4144·0 ± 0·71·7 ± 0·6
Rhizorhabdus sp.CA15304·1 ± 0·63·9 ± 0·8
NL2184·4 ± 0·44·0 ± 0·4
Sphingopyxis sp.CA32143·4 ± 1·42·5 ± 0·4

Of the 38 cultivars, B.Ls and P.Is included in the second series of tests, none were completely resistant to all bacterial isolates tested. However, three cultivars, Green Lakes from Wisconsin, and Raleigh and South Bay from Florida (Guzman, 1984), showed moderate resistance to R. suberifaciens CA1T, Rhizorhabdus sp. CA15 and NL2 (Table 4), while B.L. UC91-0003 and L. serriola P.I. 289064-1 from Hungary were moderately resistant to R. suberifaciens CA1T and CA3, as well as Rhizorhabdus sp. CA15 (Table 5). Breeding lines and plant introductions with high levels of resistance to R. suberifaciens CA1T (B.L. 88-67M, B.L. 88-68M, B.L. UC91-0005, and L. saligna P.I. 490999 from Turkey) were quite susceptible to the Rhizorhabdus sp. CA15 and NL2, or R. suberifaciens CA3. However, average disease severities caused by CA3, CA15 and NL2 were at most 5·8–6·5, while the maximum average disease severity caused by R. suberifaciens CA1T reached 7·6–8·8 on the 0–9 scale (Tables 4, 5).

Table 4. Corky root severity (0–9 scale) on various lettuce cultivars, breeding lines (B.L.) and plant introductions (P.I.) inoculated with Rhizorhapis suberifaciens type strain CA1T, Rhizorhabdus sp. CA15 or NL2
Cultivar or plant lineSource R. suberifaciens Rhizorhabdus sp.
  1. a

    Petoseed company, Woodland, California, USA.

  2. b

    Mean ( ± standard deviation) of four plants.

  3. c

    Ed Ryder, USDA, Salinas, California; all breeding lines of Ryder originate from a cross between the susceptible cultivar Salinas and the resistant cultivar Green Lakes (Ryder & Waycott, 1994).

  4. d

    Royal Sluis Company, Salinas, California, USA.

  5. e

    Victor Guzman, University of Florida, Belle Glade, Florida, USA.

  6. f

    Centre for Genetic Resources, Wageningen, The Netherlands.

  7. g

    Luis Sequeira, Madison, Wisconsin, USA.

  8. h

    Calgene Inc., Davis, California, USA.

RomulusPetoseeda7·6 ± 0·5b4·4 ± 0·54·4 ± 0·5
CeltuceRyderc7·0 ± 0·04·0 ± 0·93·6 ± 0·5
PlatoPetoseed7·6 ± 0·54·2 ± 0·44·0 ± 0·0
SalinasRoyal Sluisd6·8 ± 0·45·4 ± 0·54·4 ± 0·5
Tall GuzmaneGuzmane4·8 ± 0·75·6 ± 0·55·0 ± 0·9
PAN-VITCGNf5·8 ± 0·45·8 ± 0·45·6 ± 0·5
AugustusPetoseed3·8 ± 0·75·2 ± 0·45·4 ± 0·8
P.I. 251245Ryder5·6 ± 1·43·6 ± 0·54·0 ± 0·0
GlacierRyder3·6 ± 0·84·6 ± 0·54·0 ± 0·0
B.L. 88-52MRyder3·4 ± 1·04·4 ± 0·55·0 ± 0·6
P.I. 491114 (Lactuca serriola)Ryder5·2 ± 1·04·0 ± 1·54·6 ± 0·5
Misty DayRyder3·2 ± 0·44·8 ± 0·44·6 ± 0·5
P.I. 261653 (Lactuca saligna)Ryder3·0 ± 2·14·8 ± 0·44·6 ± 0·5
RaleighGuzman3·0 ± 0·63·8 ± 0·43·8 ± 0·4
Green LakesSequeirag2·2 ± 1·03·6 ± 0·53·8 ± 0·4
B.L. 440-8Ryder2·6 ± 1·44·0 ± 0·64·2 ± 0·4
B.L. 88-99MRyder2·8 ± 1·84·4 ± 0·54·8 ± 0·4
CRRRISPCalgeneh2·6 ± 1·24·4 ± 0·54·8 ± 0·4
B.L. 88-97MRyder2·4 ± 1·44·8 ± 0·45·8 ± 0·4
South BayGuzman2·2 ± 0·73·4 ± 0·53·4 ± 0·5
CRRRISP-XLCalgene2·0 ± 1·44·0 ± 0·04·8 ± 0·4
B.L. 88-67MRyder0·8 ± 0·74·0 ± 0·65·0 ± 0·0
B.L. 88-68Ryder1·0 ± 0·94·0 ± 0·05·0 ± 0·6
Table 5. Corky root severity (0–9 scale) on various cultivars and breeding lines (B.L.) of lettuce, and plant introductions (P.I.) of Lactuca sativa, L. serriola and L. saligna 4 weeks after inoculation with Rhizorhapis suberifaciens CA1T or CA3, or Rhizorhabdus sp. CA15
Cultivar or plant lineSourceR. suberifaciensRhizorhabdus sp.
  1. a

    Harris Moran seed company, Davis, California, USA.

  2. b

    Mean ( ± standard deviation) of four plants.

  3. c

    Ed Ryder, USDA, Salinas, California, USA.

  4. d

    Michelmore laboratory, Department of Plant Sciences, University of California, Davis, California, USA.

SalinasHarris Morana8·8 ± 0·4b4·8 ± 1·54·3 ± 0·4
CeltuceRyderc8·3 ± 0·44·3 ± 0·83·0 ± 1·4
P.I. 251245UC Davisd8·0 ± 0·75·8 ± 0·82·3 ± 0·8
P.I. 491011UC Davis6·5 ± 1·55·8 ± 0·86·5 ± 0·5
P.I. 491010UC Davis5·5 ± 1·16·0 ± 0·75·8 ± 0·4
B.L. UC91-0001UC Davis5·3 ± 1·35·5 ± 0·93·8 ± 0·8
P.I. 491250UC Davis4·5 ± 2·15·5 ± 0·93·8 ± 1·6
P.I. 491206UC Davis4·0 ± 2·02·8 ± 1·15·8 ± 1·1
B.L. 440-8Ryder3·5 ± 1·55·0 ± 0·74·5 ± 0·5
P.I. 289064-2UC Davis3·3 ± 1·15·5 ± 0·93·3 ± 0·8
B.L. UC91-0007UC Davis3·0 ± 2·34·3 ± 1·15·0 ± 0·7
P.I. 491249UC Davis3·0 ± 1·63·5 ± 0·94·3 ± 0·8
P.I. UC83-UK1UC Davis2·8 ± 1·61·0 ± 0·04·5 ± 0·5
B.L. 255665UC Davis2·0 ± 1·64·3 ± 1·13·8 ± 0·4
P.I. UC83-US1UC Davis2·0 ± 1·21·3 ± 1·16·0 ± 0·0
P.I. 289064-1UC Davis1·8 ± 1·83·5 ± 0·93·3 ± 0·8
B.L. 88-67MRyder1·8 ± 0·84·5 ± 0·54·5 ± 0·5
B.L. UC91-0003UC Davis1·8 ± 1·33·3 ± 0·43·0 ± 1·2
B.L. UC91-0005UC Davis1·5 ± 0·95·0 ± 1·03·5 ± 0·5
P.I. 490999UC Davis1·0 ± 0·71·0 ± 0·76·3 ± 0·4

The third experiment demonstrated that there was a significant interaction (= 0·0001) between plant line and bacterial isolate with respect to corky root severity (Table 6). The bacterial isolates could be grouped into three strains (Fig. 1): (i) isolates more virulent to L. virosa P.I. UC83-UK1, L. sativa ‘Salinas’, and L. serriola P.I. 251245 than to L. saligna P.I. UC83-US1, L. sativa ‘Green Lakes’, B.L. UC91-0003, L. saligna P.I. 490999 and L. sativa B.L. 88-67M, comprising isolates R. suberifaciens type strain CA1T, AU117 and AU44, and various Sphingobium sp., WI4T, FL21 and NL4; (ii) those isolates that were most virulent to L. virosa P.I. UC83-UK1 and L. serriola P.I. 251245, and B.L. UC91-0003, and slightly less virulent to L. sativa ‘Salinas’ and ‘Green Lakes’, L. saligna P.I. UC83-US1, L. saligna P.I. 490999 and L. sativa B.L. 88-67M, comprising R. suberifaciens CA3 and Sphingopyxis sp. CA32; and (iii) those isolates that were moderately aggressive to all lines but most virulent to L. virosa P.I. UC83-UK1, and L. saligna P.I. UC83-US1 and P.I. 490999, comprising Rhizorhabdus sp. NL2 and CA15.

Table 6. Analysis of variance table for the effect of test, isolate, lettuce cultivar or breeding line, and cultivar × isolate interaction using a split-plot model in proc glm in sas v. 6.3 (SAS Institute Inc.)
Sourced.f.Type I SSMean squareF-valuePr > F
  1. a

    Error B to test for isolate effect.

Test × Isolatea18317·45717·63714·480·0001
Cultivar or breeding line71481·968211·710164·140·0001
Isolate × Cultivar63766·09912·1609·430·0001
Figure 1.

Differential interaction between cultivars and breeding lines (B.L.) or plant introductions (P.I.) of Lactuca species, and isolates of Rhizorhapis and related genera with respect to corky root severity (0–9 scale). The bacterial isolates could be grouped into three strains: (i) more virulent to a subset of plant types, (ii) slightly more virulent to a subset of plant types, and (iii) moderately aggressive to all plant types. CA1, R. suberifaciens CA1T; AU117, R. suberifaciens AU117; AU44, R. suberifaciens AU44; WI4, Sphingobium (Sb.) mellinum WI4T; FL21, Sb. xanthum FL21; NL4, Sphingobium sp. NL4; CA3, R. suberifaciens CA3; CA32, Sphingopyxis sp. CA32; CA15, Rhizorhabdus sp. CA15; NL2, Rhizorhabdus sp. NL2; UK1, L. virosa P.I. UC83-UK1; Sal, L. sativa ‘Salinas’; 2512, L. serriola P.I. 251245; 83US1, L. saligna P.I. UC83-US1; GL, L. sativa ‘Green Lakes’; 9103, L. sativa B.L. UC91-0003; 4909, L. saligna P.I. 490999; and 8867M, L. sativa B.L. 88-67M. Error bars indicate standard errors around the means.

P.I. UC83-UK1 was most susceptible to all isolates. There was no plant accession that was least susceptible to all isolates (Fig. 1).


For the first time, this study has demonstrated that there was a significant differential interaction between cultivars and B.Ls or P.Is of Lactuca species, and strains of Rhizorhapis and related genera with respect to corky root severity. One isolate of R. suberifaciens (CA3) from the Salinas Valley in California, one isolate of Sphingopyxis (CA32) from the San Joaquin Valley in California, and two isolates of Rhizorhabdus (CA15 and NL2) from the Pajaro Valley in California and from South Holland in the Netherlands, respectively, were equally virulent to the resistant cultivar Green Lakes and B.L. 440-8, both with the cor resistance gene (Brown & Michelmore, 1988) and to the susceptible cultivar Salinas. Out of all different lettuce accessions included in two tests (38 in total), there were only three cultivars and two B.Ls with moderate resistance to all corky root-causing isolates included in those tests (R. suberifaciens CA1T and CA3, and Rhizorhabdus sp. CA15 and NL2). None of the accessions were highly resistant to all four isolates. All accessions with resistance to R. suberifaciens CA1T were moderately susceptible to CA3, CA15, NL2 or CA32. Three strains of pathogenic isolates could be distinguished: (i) isolates with a similar virulence pattern to R. suberifaciens CA1T, (ii) isolates with a similar virulence pattern to that of R. suberifaciens CA3 and Sphingopyxis sp. CA32, and (iii) isolates of Rhizorhabdus being moderately aggressive to all Lactuca lines. In addition, a large number of closely related isolates were nonpathogenic or weakly pathogenic to any of the lettuce cultivars and B.Ls tested (Francis et al., 2014; this study).

Differential interaction has not been published previously, although it was mentioned in preliminary reports (van Bruggen, 1994, 1997). Bull (2009) isolated several bacterial isolates belonging to the Sphingomonadaceae that were virulent on the cultivars Green Lakes and Glacier with cor resistance to R. suberifaciens CA1T, but these isolates have not been characterized in detail.

The only gene providing resistance to lettuce corky root identified thus far is the cor gene. This resistance is conferred by a recessive allele at a single locus (Brown & Michelmore, 1988). The resistance is not absolute but quantitative, and the level of resistance varies depending on other genes in the host plants and the bacterial isolates used in the resistance tests. Some cultivars (e.g. the butterhead cultivar Diana) are significantly more susceptible to corky root than Salinas; therefore additional, as yet uncharacterized, loci conditioning susceptibility are likely (O. Ochoa and R. W. Michelmore, unpublished data). There are many isolates of the pathogen, belonging to several genera in the Sphingomonadaceae, that can induce corky root symptoms. Pathogenicity factors are still unknown and may be the same or different in the various pathogenic strains. If they are different, resistance to the common strains of R. suberifaciens, such as CA1T, may not hold up in fields that contain primarily pathogenic strains of other species of Rhizorhapis or related genera. This was demonstrated for cultivar Glacier, which has the cor gene but was only partially resistant in a field experiment (Mou & Bull, 2004). This could be as a result of modifying genes or different corky root-inducing bacterial strains. In the same field experiment, four highly resistant P.I. lines were identified that had molecular marker alleles indicative of the susceptible allele Cor (Moreno-Vázquez et al., 2003), indicating that these lines may contain different resistance gene(s) (Mou & Bull, 2004).

Although the potential for the emergence of pathogenic strains that overcome the resistance conferred by the cor gene exists, this gene has been deployed successfully in many commercial cultivars for several decades and there have been no reports from commercial fields that such strains have become problematic in any of the large lettuce production areas. Nevertheless, it is important that two accessions, B.L. UC91-0003 and P.I. 289064-1, were moderately resistant to R. suberifaciens CA3 and Rhizorhabdus sp. CA15 which induced quite severe disease in other cultivars and B.Ls. Whether the resistance in these selected B.Ls that were tested against several pathogenic isolates in the greenhouse will hold up in the field, still remains to be seen.

Selected Lactuca accessions with moderate levels of resistance to all different corky root strains should be planted in replicated field experiments in areas where different strains of the pathogen were identified. Accessions with resistance in all field experiments would need to be checked for the presence of the cor gene. It would be good to identify additional resistance genes for preemptive lettuce breeding activities so as to diversify the selection pressure on the pathogen populations and increase the durability of resistance (Michelmore et al., 2013). For the time being, it would be prudent and informative to use several different bacterial strains to test lettuce B.Ls for resistance to corky root.


The authors thank Ken Jochimsen, Philip Brown and Joe Wakeman for technical assistance and Ed Ryder for providing some of the breeding lines and plant introductions. Funding was provided by the California Iceberg Lettuce Advisory Board.