Red rice plants that escaped imazamox treatments were sampled in 42 fields across Piedmont and Lombardy regions (Table 1). The largest number of samples (25) was collected in the province of Pavia (PV), in Lombardy. Although it was not possible to obtain complete information about field treatments and practices from all farmers, the data collected show that the rotation of Clearfield^® with conventional rice was implemented in 10 fields. Clearfield^® rice was cultivated for at least 4 consecutive years in 11 fields, with 7 where continuous IMI-tolerant varieties were grown for at least 5 out of 6 years since 2006 and three where Clearfield^® varieties had been cultivated continuously for six years. The latter fields therefore received 12 IMI treatments in 6 years. It seems clear that several farmers did not carefully follow the guidelines for cultivation provided by BASF that are attached to the contract they have to sign with this company in order to be able to cultivate Clearfield^® rice.

In fact, the guidelines recommend that Clearfield^® varieties should not be grown in the same field for more than two consecutive years. Although the situation is rather different, it is worth pointing out that the guidelines used in USA recommend rotating Clearfield^® rice annually with other crops [19]. However, the field histories also prove that virtually all farmers follow the guidelines in relation to the no. of imazamox treatments per year (i.e., two).

In addition, difficulties in controlling Echinochloa spp. and sedges have been reported for seven of the sampled fields (Table 1). The cultivation of Clearfield^® rice is a valuable tool for controlling red rice, but this technology has to be seen in the context of the rice cropping system as a whole, where there is already an intense use of ALS inhibitors. More than 90% of Italian rice fields are treated at least once per year with this group of herbicides and ALS-resistant populations of four other weed species is widespread [17,20]. The selection pressure imposed by the repeated use of imazamox therefore determines an increased risk of resistance evolution in rice weeds, especially in Echinochloa spp. [18].

The first screening was conducted on 17 red rice populations collected in 2010. The three susceptible checks (010-1, 010-2 and 010-3) were totally controlled by imazamox at the recommended field dose (1×, Figure 1). Among the suspected resistant populations, eight were highly resistant to imazamox, with plant survival ranging between 48 and 98% at dose 1× (Figure 1). The biomass of these surviving plants was always high and ranged between 73 and 98 % of the untreated control, indicating that the plants were scarcely affected by the herbicide. In population 010-8 only a few plants were imazamox-resistant but they were badly affected by the treatment, as shown by the VEB value of 48%. The remaining five populations were still susceptible.

In the second screening all resistant populations were retested at two doses of imazamox, 1× and 3×. ANOVA showed that the plant survival did not significantly differ between imazamox doses (F = 0.057; P = 0.82). Moreover, plant survival of all the populations treated at dose 1× was similar to that observed in the first screening, with plant survival varying between 52 and 100% and 74 and 93% at dose 1× and 3×, respectively (data not shown).

The third screening tested the response to imazamox at dose 1× of the 28 red rice populationscollected in 2011 and showed that 18 populations were highly resistant to imazamox at the recommended field dose, with plant survival ranging between 78 and 100% (Figure 2). Only population 011-42 showed a lower value of 35%. Instead, all the survived plants had a high VEB, always above 85% indicating that they were not affected by the herbicide. In population 011-44 only 2 plants survived and their biomass was similar to that of the untreated control.

The confirmation of 26 imazamox-resistant red rice populations indicates that resistance is already present in all major rice producing Italian provinces and that it is significantly more diffused than reported by Busconi et al. [19], where resistant red rice was found in six locations of one province of the Piedmont region.

Genomic DNA was extracted from plants of the 26 resistant populations and from three susceptible populations: 010-1, 010-2 and 011-24. The combination of specific primers ORY-2F/ORY-2R amplified a fragment of the expected size at the 3’end of the ALS gene. The alignment of the nucleotide sequences of the ALS-amplified region of the resistant and susceptible plants showed only one nucleotide substitution at position 653 (position referred to Arabidopsis thaliana (L.) Heynh. ALS sequence). In the susceptible populations, all plants had a AGT encoding a serine, while in all the resistant populations analyzed, plants had an AAT encoding an asparagine (Table 1). The mutation found in the ALS gene is the same as that present in the IMI-R rice cultivar, Libero, commercialized in Italy since 2006 [19].

Most of the resistant plants analyzed were homozygous at codon 653. Only plants of populations 010-5, 010-9 and 010-13 were heterozygous at this locus (Table 1). The presence of homozygous genotypes proves that the hybridization process resulting in transfer of the herbicide resistance trait to red rice occurred at least two years before seed sampling. It is therefore conceivable that the hybridization events took place in the early years of the introduction of the Clearfield^® rice variety Libero. The progenies were then able to multiply following the selection pressure exerted by the continuous use of imazamox. Instead, the presence of heterozygous plants at the ALS locus may be the result of a hybridization event during the sampling year or due to the segregation of the F2 generation [19].

The fact that homozygous plants were found in population 011-25 after one year of Clearfield^® cultivation (see Table 1) suggests that there may have been a contamination of the commercial seed batch used for sowing. This hypothesis is plausible because five seeds of red rice in 500 g of rice is permissible in Italy. An analogous case has been hypothesized in Brazil where imidazolinone resistant weedy rice was reported after the first year of IMI-R rice planting [10].

A morphological observation of the red rice seeds revealed that some of the populations tested had both red as well as a few white and intermediate colored seeds. This may imply that some of the resistant plants, i.e., those originating from white seeds, belong to off-types originating during the selection process of the Clearfield^® variety. Therefore, although gene flow is likely to be the main origin of the IMI-resistant red rice [21,22], further studies are needed to fully clarify the origin of the red rice mutants.

In the next years new IMI-tolerant rice cultivars will be ready for commercialization (10 will be available for the next cropping season) and so it is likely that the area cultivated with Clearfield^® varieties will increase further. To maintain the Clearfield^® technology alive, farmers should strictly follow the guidelines provided by BASF at time of seed purchase. Firstly, only certified seed batches should be used and imazamox should be sprayed at the rate and timings reported on the label; and secondly the cultivation of IMI tolerant varieties for more than two consecutive years must be avoided. Actually, considering the rapid diffusion of IMI-resistant red rice demonstrated in this study, it would appear reasonable to carefully re-evaluate the guidelines and recommend rotating Clearfield^® varieties annually with traditional rice varieties or other crops. In the case of escaped red rice plants, farmers must remove them at any cost before seeds shattering and report to BASF and/or the advisory service of the Ente Nazionale Risi.

Clearfield^® technology should be integrated into a complex weed management strategy that includes the use of all available tools and techniques. Stale seed-bed preparation allows the germination of red rice seeds and the subsequent control of seedlings with a burn-down herbicide before sowing. Soil tillage should be avoided after the harvest, leaving red rice seeds to decompose over the winter or be predated by birds [23]. If the field is tilled, the red rice seeds will be buried and become deeply dormant. Rotation of rice with other crops, such as soybean or maize, and winter flooding can also be used to reduce the red rice seed-bank [24].

Although the best strategy to preserve the Clearfield^® technology is prevention, in the presence of IMI-resistant red rice more specific measures must be implemented. In addition to manual weeding, it is necessary to stop the cultivation of Clearfield^® rice varieties and go for traditional varieties with a short growing cycle that allows to delay the sowing after the second half of May as well as preserving a good yield. In case of heavy infestation of resistant red rice, the only solution is to rotate rice with other crops where herbicides having a different mode of action can be used. This would also be useful for controlling others ALS-resistant weeds which might be present in the same field.

The herbicide resistance issue in rice crops needs a wider and more integrated management approach because the sustainability of the entire Italian rice cropping system is threatened by the very high selection pressure exerted by the intense use of ALS-inhibiting herbicides in time and space and consequent fast selection of resistant populations belonging to various weed species [17,18,20]. The high biological activity and good ecotoxicological and environmental profile of these herbicides, as well as the lack of new modes of action in the near future, impel stakeholders to preserve their efficacy for as long as possible.

Seeds of 42 populations of red rice that had escaped the treatments with imazamox in Clearfield^® rice fields of the north-western Italy were collected in 2010 and 2011. Seeds from three susceptible populations (010-1, 010-2 and 010-3) that had never been treated with IMI-inhibitors were also collected in the same region. Each seed sample was collected from at least 20 plants of red rice, based on morphological traits. Seed samples were then cleaned and stored at room temperature. Every effort was made to collect from farmers as much information as possible on agronomic practices, and especially herbicide use, of sampled fields starting from the introduction of the Clearfield^® varieties in 2006. The complete list of red rice populations used in the experiments is reported in Table 1.

Seeds of red rice were directly sown in plastic boxes (24 cm × 30 cm × 9 cm) filled with a substrate containing 60% silty loam soil, 15% sand, 15% perlite and 10% peat by volume and placed in a greenhouse at Legnaro, north-eastern Italy (45°21' N, 11°58' E). Experiments were carried out during February-May period and plants were light supplemented using 400 W metal-halide lamps, which provided a Photosynthetic Photon Flux Density (PPFD) of about 400 µmol m⁻² s⁻¹ and a 16-hour photoperiod. Minimum temperatures ranged between 14 and 19 °C while maximum temperatures varied between 22 and 34 °C. About twenty days after sowing, seedlings were thinned to 20 plants per box and then plants were treated at the four-leaf stage. The experimental layout was a complete randomised design with two replicates. An untreated control per each population was always included. This protocol was followed in the three experiments performed.

In the first, 14 suspected resistant and 3 susceptible populations collected in 2010 were treated with the recommended field dose (1×) of imazamox (Altorex, BASF, 40 g a.i. L⁻¹, solution) at 35 g a.i. ha⁻¹. The second experiment repeated the first one with the addition of the treatment at three times the recommended field dose (3×). The third experiment was conducted on the 28 red rice populations collected in 2011 and the imazamox treatment was done at the recommended field dose (1×).

The herbicide was distributed by using a precision bench sprayer delivering 300 L ha⁻¹ at a pressure of 215 kPa and speed of about 0.75 m s⁻¹, with a boom equipped with three flat-fan (extended range) hydraulic nozzles (Teejet^®, 11002). Plants were watered daily as required.

Four weeks after herbicide treatment, the number of surviving plants and a visual estimation biomass, giving a score of 10 to the untreated check and 0 to replicates where all plants were dead, were recorded. Plants were assessed as being dead if, regardless of color, they showed no active growth. Standard error (SE) was calculated for each mean. Populations were ascribed to be resistant when more than 20% of plants survived the 1× dose.

An ANOVA (significance level P < 0.05) was performed to analyze the effect of the imazamox dose on plant survival. The statistical analysis was done by using the software Statistica 6.1 (StatSoft, Tulsa, USA; http://www.statsoft.com [25]).

Leaf samples for DNA extraction were done on individual plants of red rice before imazamox treatment. In this way it was possible to analyze the ALS sequence for both resistant and susceptible plants. Genomic DNA was extracted from plants of all the populations reported in Table 1 using the CTAB method [26]. All extractions were performed using 0.1 g of fresh leaves. Specific primers were designed on regions of high homology among the ALS gene sequence available in Genebank: Oryza sativa (Japonica cultivar-group) AY885674.1. The primer combination ORY-2F (5′-CAGGAGTTGGCATTGATCCGC-3′)/ORY-2R (5′-ACACAGTCCTGCCATCACCATC-3′) amplified a genomic fragment of 372 bp at the 3′ of the ALS gene. PCR reactions was conducted using the Advantage^® 2 PCR Kit (Clontech) in a 50 µL mixture of 1× Advantage^® 2 SA buffer, 0.2 µM of each primer, 0.2 mM of each dNTP, 1 µL of proofreading Advantage 2 Polymerase mix and 100 ng of genomic DNA template. The following PCR program was done: 1 min at 95 °C, then 30 cycles of 30 s at 95 °C, 30 s at 60 °C and 40 s at 68 °C. Final extension time was 3 min at 68 °C. PCR products were extracted from 1% agarose gel stained with SYBR^® Safe DNA gel stain (Invitrogen) and purified using MinElute Gel Extraction Kit (Qiagen) following the manufacturer’s instructions. Sequencing was carried out by an ABI 3730XL sequencer (Applied Biosystems). Sequence analysis and alignments were performed using the BioEdit software [27].