Restricted glyphosate translocation in Lolium multiflorum is controlled by a single incomplete dominant nuclear gene

ABSTRACT

The mode of inheritance for herbicide resistance was investigated in a population of Lolium multiflorum (Italian ryegrass) with restricted glyphosate translocation mechanism of resistance. The degree of dominance for glyphosate resistance in Italian ryegrass was evaluated in the progenies of F1 families which were created by reciprocal-crossing between resistant (R) and susceptible (S) parental individuals. The results showed an intermediate level of glyphosate resistance for F1 families compared to that of the parental types. The phenotypic resistance segregation was also investigated using backcross families which were created by pair-crossing the F1 individuals to the susceptible population, and the observed segregation values fitted a one-gene model very well. Thus it appears that glyphosate resistance in the Italian ryegrass obtained from a New Zealand vineyard with restricted herbicide translocation is controlled by a single incompletely dominant nuclear gene.

KEYWORDS:

• Glyphosate resistance
• inheritance
• Italian ryegrass
• nuclear genome
• one-gene

Introduction

Glyphosate has been the main herbicide for controlling weedy species within orchards and vineyards in New Zealand for some time (Harrington & Ghanizadeh 2016). Hence, the evolution of resistance to this herbicide was not surprising (Ghanizadeh et al. 2015a). Italian ryegrass (Lolium multiflorum L.) was one of the first weedy species found to be resistant to glyphosate in vineyards in New Zealand (Ghanizadeh et al. 2013). It has also been noted that glyphosate-resistant Italian ryegrass species are also cross-resistant to glufosinate (Ghanizadeh et al. 2015a). Further investigations showed that the restricted herbicide translocation mechanism is associated with resistance to glyphosate in these Italian ryegrass populations from New Zealand (Ghanizadeh et al. 2016). The restricted herbicide translocation has been the common mechanism of resistance to glyphosate in all cases of glyphosate-resistant ryegrass found in New Zealand to date (Ghanizadeh et al. 2015b; Ghanizadeh et al. 2016; Ghanizadeh et al. 2015c).

Mode of inheritance plays a crucial role in transmission of resistance alleles from one generation to another. The transmission of resistant alleles in resistant weedy species can occur through pollen and ovules (Jasieniuk et al. 1996). The inheritance of glyphosate resistance with the restricted herbicide translocation mechanism of resistance has been found to be governed by a single incompletely dominant nuclear gene for Lolium rigidum (Lorraine-Colwill et al. 2001), Conyza canadensis (Zelaya et al. 2004) and Lolium perenne (Ghanizadeh et al. 2016a). Wakelin and Preston (2006) however reported that glyphosate resistance in four populations of L. rigidum with restricted herbicide translocation was inherited as a single dominant gene.

As Italian ryegrass is a major weed in New Zealand vineyards, it was important to understand the basis of the inheritance of glyphosate resistance in Italian ryegrass, to assist with its management. The results of inheritance studies can be used to interpret the dynamics of herbicide-resistant phenotypes and help develop management strategies (Wakelin & Preston 2006). Thus the objective of the present study was to determine the mode of inheritance in a New Zealand glyphosate-resistant population of Italian ryegrass with the restricted glyphosate translocation mechanism of resistance.

Material and methods

Plant materials

In this study, plants from a glyphosate-resistant Italian ryegrass population (R) were grown together in a glasshouse within pollen-proof cloth in order to obtain seeds. The plants were the clones of resistant parents that had been shown to be resistant to glyphosate in two dose–response experiments (Ghanizadeh et al. 2013). One known susceptible population of Italian ryegrass (S) was also included in this study (Ghanizadeh et al. 2015a).

First filial generations (F₁)

The glyphosate-resistant (R) and susceptible (S) plants were vernalised in a cool-room set at 8°C for 4 weeks. The vernalised R and S plants were crossed using the method described by Ghanizadeh et al. (2016a). Briefly, the vernalised plants were shifted to a glasshouse with supplementary lights to encourage seed-head formation. The flowering process of each plant was recorded and one resistant (R) plant was crossed with one susceptible (S) plant based on their flowering synchronisation. The tillers of R and S plants were enclosed using pollen-proof bags and kept in the glasshouse within frames enclosed with a cloth. The crossed plants were kept in a glasshouse with the daily maximum/minimum temperatures of 26.2/16.7°C. Seeds were collected from R and S maternal plants separately at maturity. In our study, four reciprocal crosses between four R and four S plants were successfully achieved, hence, in total, eight F₁ families were obtained. For F₁ families, four families had resistant maternal (pollen receptor) (R♀) and susceptible paternal (pollen donor) (S♂) parents (F₁ R♀×S♂, denoted as RF₁). The other four had F₁ susceptible maternal (pollen receptor) (S♀) and resistant paternal (pollen donor) (R♂) parents (F₁ S♀×R♂, denoted as SF₁). The original R and S plants that were used for generating F₁ families were re-potted and used for dose–response experiments for the F₁ families. Dose–response experiments (four dose–response experiments for the four reciprocal crosses obtained in this study) were conducted on the F₁ families in order to evaluate the dominance level of F₁ families using the method described by Ghanizadeh et al. (2016a). Pots were sprayed with five rates of glyphosate (0, 100, 250, 625 and 1562 g ae/ha as Roundup 360 Pro, an isopropylamine salt) using a laboratory track sprayer calibrated to deliver 254 l/ha of spray solution at 200 kPa. All herbicide treatments contained 0.1% organosilicone surfactant (Pulse Penetrant). The daily maximum/minimum temperatures in the 2 weeks following application averaged 26.1°C and 12.2°C, respectively. The above-ground plant material was harvested and oven-dried 4 weeks after treatment.

Generation and phenotyping of backcross families

Backcross (BC) families were created using the method described by Ghanizadeh et al. (2016a). Briefly, the F₁ resistant maternal family (RF₁) were backcrossed to the S plants. In total four BC families were obtained and at maturity, the seeds were collected from the RF₁ that had been backcrossed with S plants. The response of the BC families was evaluated at two glyphosate rates of 360 and 480 g ae/ha. The susceptible and resistant parental, and also RF₁ families, were also included for comparison. All seeds were pre-germinated using the method described by Ghanizadeh et al. (2015a) and seedlings were potted using a similar method described by Ghanizadeh et al. (2016a). When the seedlings were at the three-leaf stage, they were treated with 360 and 480 g ae/ha of glyphosate. Four weeks after glyphosate treatment, the seedlings were scored as dead (not actively growing) or alive (actively growing).

Statistical analysis

The dose–response experiments were conducted in a randomised complete block design with 15 replicates of each herbicide rate. The plant material dry weight data were fitted to a three-parameter logistic model as described by Streibig (1988) using Sigmaplot (V. 13.0), and a one-way ANOVA was used to statistically compare the estimated parameters of the three-parameter logistic model as suggested by Ghanizadeh et al. (2015b). Goodness of fit chi-square tests (x² -tests) were used to determine the observed survival segregating values for each BC family according to the one-gene model (0.5F₁:0.5S) (Tabashnik 1991). P-values at 5% probability were obtained in order to validate the one-gene model for each BC family tested. The variation among BC families was evaluated using a heterogeneity x² -test as suggested by Sokal and Rolf (1969).

Results

Dose–response bioassay for F₁ families of Italian ryegrass

The results of dose–response experiments on F₁ families (the mean of all RF₁ and SF₁ families), and glyphosate-resistant (R) and susceptible (S) original plants are shown in Table 1. The GR₅₀ (the glyphosate rate required to reduce the plant shoot dry weight by 50%) for the R population was 1095 g ae/ha compared with a significantly lower GR₅₀ of 121 g ae/ha for the S population (Table 1). This result suggests that the R population was nine times more resistant to glyphosate than the S population. The GR₅₀ values from pooled data from all eight F₁ families (four RF₁ and four SF₁) showed that higher rates of glyphosate were required to reduce shoot dry weight by 50% compared to the S population (Table 1). Based on the GR₅₀ values, the RF₁ and SF₁ were estimated to be 5.9 and 4.0 times more resistant to glyphosate, respectively, than the parent S population, and thus showed an intermediate level of resistance to glyphosate compared to parent R population (Table 1).

Table 1. Parameters estimated from the nonlinear regression analysis of glyphosate dose–response experiments of Italian ryegrass families at 4 weeks after application of glyphosate. R was the glyphosate-resistant parental population, S was the glyphosate susceptible parental population, family RF₁ corresponds to the progeny of reciprocal cross R♀×S♂ and SF₁ corresponds to progeny of the reciprocal cross S♀×R♂.

Family	d (%)	b	GR50 (g ae/ha)^a	R/S GR50 ratio	R^2
R	99.6	1.5	1095 a	9.0	0.99
S	100.3	1.7	121 d	–	0.98
RF1	100.3	1.2	712 b	5.9	0.99
SF1	99.3	1.2	484 c	4.0	0.98
P value	.99	.52	<.0001

Notes: d = the upper limit, b = the slope around the GR₅₀, GR₅₀=the rate of herbicide (g ae/ha) required to reduce dry weight by 50%, R/S GR₅₀= resistant/susceptible GR₅₀ ratio. R²= coefficient of determination.

^aMean values within each column followed by the same letters are not different at 5% probability according to Tukey’s tests.

The segregation of BC families

Among BC families treated with 360 g ae/ha of glyphosate, only one family (BC₄) did not fit well with the expected one-gene segregation as the survival value for this family was significantly higher than expected (Table 2). However, the segregation of the other three BC families fitted the expected one-gene segregation well at the glyphosate rate of 360 g ae/ha. The heterogeneity test was not significant when tested against the one-gene model (x² = 6.5, P = .09) and therefore, the data of all four families were pooled. The x²-test of pooled data was not significantly different from the one-gene model (P = .09) at the glyphosate rate of 360 g ae/ha. When the BC families were treated with 480 g ae/ha of glyphosate, the observed segregation pattern of all BC families also fitted well with the expected segregation pattern of the one-gene model (Table 2). The heterogeneity in resistance segregation for tested BC families was not significant (x² = 4.8, P = .19), and as with the families treated with 360 g ae/ha of glyphosate, the pooled data of all four BC families treated with 480 g ae/ha of glyphosate fitted best with the one-gene model as shown by the x² -test of pooled data (P = .34) (Table 2).

Table 2. Chi-square analysis and heterogeneity test for goodness of fit of the observed segregation values for glyphosate resistance in BC families according to the one-gene model following application of a low glyphosate rate (360 g ae/ha) and a high glyphosate rate (480 g ae/ha).

Family^a	Total treated	Survival observed	Survival expected	x^2	P
Glyphosate 360 g ae/ha
R	45	45
S	55	10
F1	36	11
BC1	40	15	15.9	0.04	.78
BC2	50	23	19.8	0.51	.36
BC3	46	19	18.2	0.03	.82
BC4	48	27	19.0	3.33	.02
Pooled	184	84	72.9	1.67	.09
Heterogeneity test				6.5	.09
Glyphosate 480 g ae/ha
R	45	35
S	60	0
F1	55	14
BC1	40	5	8.0	1.13	.24
BC2	58	15	11.6	0.99	.26
BC3	36	8	7.2	0.09	.74
BC4	50	14	10.0	1.60	.16
Pooled	184	42	36.8	0.73	.34
Heterogeneity test				4.8	.19

^aR = resistant parental (R), S = susceptible parental (S), F₁=first generation filial and BC = backcross families.

Discussion

The restricted translocation mechanism is the most common non-target site mechanism of resistance to glyphosate in weeds (Ghanizadeh & Harrington 2017a). The results of our study showed that the restricted translocation mechanism in Italian ryegrass can be transferred via pollen and ovules, hence it is a nuclear-coded gene. However, the genetic element involved in glyphosate-resistant Italian ryegrass is not completely dominant. This is implied by the level of resistance to glyphosate for the progenies of F₁ families of both resistant maternal (RF₁) and susceptible maternal (SF₁) being half of that for the R population. Nuclear-coded genes are a particularly common inheritance mode amongst the majority of weed populations resistant to herbicides (Jasieniuk et al. 1994; Sabba et al. 2003; Wakelin & Preston 2006; Riar et al. 2011). The mode of inheritance found for Italian ryegrass in our study is similar to that for glyphosate-resistant populations of L. rigidum from Australia (Lorraine-Colwill et al. 2001) and Lolium perenne from New Zealand (Ghanizadeh et al. 2016a). In both L. rigidum and Lolium perenne with the restricted glyphosate translocation mechanism, the mode of inheritance was found to be governed by an incompletely dominant nuclear gene.

We also noted that glyphosate resistance in Italian ryegrass was controlled by a single gene (monogenic inheritance). The results of phenotypic segregations of glyphosate resistance in BC families revealed that glyphosate resistance in the New Zealand population of glyphosate-resistant Italian ryegrass fits well with a one-gene model. The non-target site mechanism of glyphosate resistance for L. rigidum from Australia (Lorraine-Colwill et al. 2001; Wakelin & Preston 2006), Lolium perenne from New Zealand (Ghanizadeh et al. 2016a) and Conyza canadensis from USA (Zelaya et al. 2004) have also been reported to involve monogenic inheritance.

The evolution of herbicide resistance occurs in situations where the herbicide exerts a continuous selection pressure on weeds (Neve 2007). In New Zealand, the evolution of resistance to glyphosate in ryegrass species has only been reported in some vineyards where there has been a persistent use of glyphosate over the previous 10–15 years (Ghanizadeh et al. 2015a; Ghanizadeh et al. 2016b). The increased utilization of glyphosate in New Zealand vineyards is because application of other herbicides such as paraquat and residual herbicides had been discouraged by marketing bodies for the wine industry (Ghanizadeh et al. 2015a; Harrington et al. 2016). Given the results found in our study and the possible gene flow via pollen in allogamous species such as ryegrass species (Busi et al. 2008), the frequency of resistant phenotypes could increase in just a few generations following persistent use of a herbicide (Neve & Powles 2005). This could explain the evolution of glyphosate-resistant ryegrass after continuous use of glyphosate in New Zealand vineyards.

The results of our study, however, revealed that the genetic element of the restricted mechanism of resistance to glyphosate in Italian ryegrass is not completely dominant. This suggests that a dilution effect may occur in progenies of the cross between glyphosate-resistant and susceptible plants since the resistance levels in the progenies are significantly lower than that of original resistant plants. This is particularly important as the rate of transmission of alleles that cause a high level of resistance could be slower for phenotypes with incompletely dominant alleles when no selection pressure is imposed by herbicide applications (Preston & Wakelin 2008). In addition, the potential fitness cost caused by the restricted herbicide translocation mechanism that is governed by a single incompletely dominant nuclear gene (Fernández-Moreno et al. 2017; Ghanizadeh & Harrington 2017b) may further contribute to the reduction in the frequency of glyphosate-resistant phenotypes if glyphosate applications should be eliminated.

In summary, an investigation of the mode of inheritance for the restricted herbicide translocation mechanism in Italian ryegrass revealed that this mechanism is governed by a single incompletely dominant nuclear gene that can be transmitted via both pollen and ovules. Understanding the biological and ecological aspects of herbicide resistance in weeds is essential for developing effective weed management tactics (Ghanizadeh & Harrington 2017b; Neve et al. 2009). Further investigations are needed to identify and quantify the principal dispersal vectors for glyphosate-resistant ryegrass once it develops within vineyards. Since pollen may have an important role in the transmission of resistance alleles, it is necessary to investigate factors affecting pollen viability, longevity and dispersal. Such studies would improve our knowledge of potential transmission of glyphosate resistance alleles in ryegrass.

Acknowledgements

The authors wish to thank the donors of plant material for this study, and the staff of the Plant Growth Unit of Massey University for help growing the ryegrass.

Disclosure statement

No potential conflict of interest was reported by the authors.

ORCID

Hossein Ghanizadeh http://orcid.org/0000-0001-5381-7880

Kerry C. Harrington http://orcid.org/0000-0003-2238-516X