EMS mutagenesis of Noccaea caerulescens accession SF
To determine a suitable EMS concentration for mutagenesis of N. caerulescens, 200 seeds of an inbred line of N. caerulescens accession ‘Saint-Félix-de-Pallières’ (SF), a Zn/Cd hyperaccumulating accession, were soaked in different EMS concentrations (0.2, 0.3, 0.4 and 0.5%). A subset of the germinated seedlings from the 0.2% and 0.5% EMS concentrations were grown to evaluate their effectiveness for mutagenesis. The number of siliques and the number of seeds per silique in the M1 plants were determined (Fig.1). At the higher concentrations of EMS, a stronger negative effect on fertility was observed, especially obvious at 0.5% EMS. At that concentration also the number of seeds per silique was negatively affected with only approximately 2–3 seeds per silique being produced. A reduction in seed number was already observed at 0.4% EMS, indicating 0.4% EMS was an effective concentration for mutagenesis in the SF accession of N. caerulescens. The use of 0.3% EMS was also expected to provide some effect, and resulted in a better fertility. Therefore the 0.3% and 0.4% EMS treatments were used to generate an M1 mutant population of N. caerulescens.
Figure 1. Optimization of the EMS concentration for N. caerulescens mutagenesis.
Seeds were imbibed with four different ethyl methane sulfonate (EMS) concentrations and a reference treatment with only distilled water (control) for 16 hrs in the dark. Plants were grown from these seeds, vernalized and allowed to flower upon which different traits were determined. (A) The number of seeds per silique, the number of siliques per inflorescence, and the number of seeds per inflorescence of seeds from different EMS treatments, presented in grey, white and dark bars. Data show the average of 3–5 plants ± SE. (B) Representative examples of the inflorescences of plants treated with the indicated EMS concentration.
Development of an M2 population in SF background
To create an M2 population that can be used for TILLING, seeds of two sister plants of the SF accession were treated with 0.3 or 0.4% EMS to create two batches of M1 plants, grown in 148 trays of 54 pots per tray, 5000 plants for the 0.3% EMS treatment and 3000 plants for the 0.4% EMS treatment. Not all plants germinated or flowered, leaving around 50 plants per tray, from which M2 seeds were bulk-harvested for each of the 148 M1 trays. 120 M2 seeds per M1 tray were sown in individual pots to generate a population of ~ 7000 flowering M2 plants for the 0.4% EMS treatment, which was used for TILLING, and ~ 5500 flowering M2 plants for the 0.3% EMS treatment, which was used as back-up. Figure 2 represents the workflow and procedure followed to generate and sample the TILLING population.
Fig. 2. A schematic overview of the mutagenesis and TILLING procedure.
M0 seeds from two sister plants of a single-seed recurrent inbred descendant of N. caerulescens accession Saint-Félix-de-Pallières were treated with 0.3 or 0.4% EMS to generate M1 seeds. These were sown in trays with 54 plants each. M1 plants were grown and allowed to self-fertilize to generate M2 seed, which was bulk-harvested per tray. A subset of the M2 seeds harvested per tray were sown to grow M2 plants in the greenhouse. In total 5520 and 7000 M2 plants were grown to obtain the 0.3 and 0.4% populations, respectively, the latter of which was used for TILLING. The M2 generation was phenotyped for plant morphological traits and their ionome profile. M3 seeds were harvested from individual M2 plants of the TILLING population. For TILLING by to High Resolution Melting curve analysis (HRM), the genomic DNA of individual plants was isolated using a Kingfisher DNA isolation robot. DNA of 4 plants was pooled in single wells of 96-well microtiter plates. Each pool was used for PCR amplification. PCR products per pool were subjected to HRM. The wells that contained mutations were selected and subjected to a second HRM screen. Positive samples were again PCR amplified and confirmed by DNA-sequencing the PCR product.
Morphological phenotypes in the mutagenized population
To evaluate the effectiveness of mutagenesis, a broad range of visual morphological phenotypes were monitored in both populations, during the vegetative and reproductive stages. Mutants for plant architectural traits, such as leaf shape, leaf colour, inflorescence number, flower development and flowering time were observed in the M2 population. The most commonly observed phenotypes in both populations were related to anthocyanin pigmentation. The second largest number of the mutants are related to the leaf pigmentation ranging from narrow yellow spots, necrotic spots and yellow sectors (variegation), or generally lighter or darker green plants. On average, the frequency of the pigmentation mutations was 1.2% and 2.1% in the 0.3% and 0.4% EMS populations, respectively, further referred to as the 0.3 and the 0.4 populations. Flower morphology mutants were investigated upon flowering of the plants. 1.9% of the plants in the 0.4 population showed altered flower morphology phenotypes. Eceriferum mutants with a bright green appearance due to an aberrant cuticular wax layer, and plants without secondary shoots account for 0.11 and 0.04%, respectively, in the 0.4 population. Mutants with altered wax layer, flowers with smaller petals and stamens and with multiple petals (double flowering) were exclusively observed in the 0.4 population, accounting for 0.18% in total. The most striking phenotypes are listed in Table 1. As expected, the 0.4 population yielded a higher percentage of mutant phenotypes than the 0.3 population. A selection of morphological phenotypes observed in the M2 populations are presented in Fig. 3. Many more other abnormal phenotypes were observed at different development stages, as listed in Supplementary Figure S1. The frequency and the wide range of the phenotypic mutations observed in these two populations implied that especially the 0.4 population, would be a valuable resource for screening desired N. caerulescens mutations in forward and reverse genetic screens.
Figure 3. Examples of mutant phenotypes in the M2 populations.
Examples of morphological mutants illustrating the mutation spectrum observed in the N. caerulescens M2 populations. A-D: abnormal leaf shape and organization; E-H: chlorophyII phenotypes; I-J: apetala-like phenotype; K: agamous-like ”double flower” phenotype; L: cauliflower-like phenotype; M: no side shoots, N: non-vernalisation required early flowering mutant, O: wild-type SF inflorescence, P: wild-type SF rosette (2 months old). Additional mutant phenotypes are described in Supplementary Figure S1.
Table 1
Numbers (#) and frequencies of indicated mutant phenotypes frequently observed in the 0.3% and 0.4% EMS M2 populations. The frequency was calculated with respect to the total number of plants in each population.
|
phenotypes
|
# mutants
|
frequency (%)
|
0.3%
|
0.4%
|
0.3%
|
0.4%
|
leaf colour
|
anthocyanin
|
105
|
158
|
1.91
|
2.26
|
chlorophyll
|
76
|
144
|
1.20
|
2.06
|
flower
morphology
|
late flowering
|
15
|
50
|
0.27
|
0.71
|
fasciation
|
17
|
40
|
0.31
|
0.57
|
‘loose’ flower
|
3
|
10
|
0.05
|
0.14
|
terminal flower-like
|
2
|
8
|
0.04
|
0.11
|
apetala-like
|
2
|
8
|
0.04
|
0.11
|
cauliflower-like
|
1
|
6
|
0.02
|
0.09
|
early flowering
|
2
|
5
|
0.04
|
0.07
|
no petals and stamens
|
0
|
3
|
0
|
0.04
|
fused petals
|
1
|
2
|
0.02
|
0.03
|
agamous-like
|
0
|
2
|
0
|
0.03
|
stem colour
|
eceriferum-like
|
0
|
8
|
0
|
0.11
|
architecture
|
no side shoots
|
1
|
3
|
0.02
|
0.04
|
Optimization of the High Resolution Melting Curve analysis
DNA samples were prepared from 7000 fertile M2 individuals of the 0.4 population. Samples from four M2 individuals were combined (4x pooling). To reduce cost and work load for screening the mutants in an efficient way, the High Resolution Melting Curve (HRM) analysis was optimized based on the analysis of a known EMS-induced point mutation in the Flowering Locus C (FLC) gene of N. caerulescens [45]. For this test, DNA samples from the flc-1 mutant, carrying a G->A mutation (483 bp after the ATG) were spiked into five wells in a 96-well plate, which contained the 4x pooled DNA samples from wild-type plants. The initial FLC HRM assay of this plate yielded seven positive samples that displayed aberrant melting curves indicating PCR fragments carrying points mutations in these wells ( Figure S2). After the second round of screening, four wells, all containing the flc-1 mutant DNA, were confirmed as positive, meaning only one spiked sample got undetected, but no false positives were found.
Mutation of the NcbZIP19 gene
The A. thaliana bZIP19 (AtbZIP19) gene encodes a basic-region leucine-zipper (bZIP) transcription factor, which acts together with transcription factor bZIP23 in the control of Zn deficiency responsive gene expression of Zn transporters, and other genes involved in early response to Zn deficiency [46]. These genes act largely redundantly, with the bzip19bzip23 mutant being unresponsive to Zn deficiency and Zn deficiency hypersensitive. Orthologues of these genes are found in N. caerulescens, with only the NcbZIP19 gene found to be expressed [47]. In order to study if a mutation of NcbZIP19 affects the function of the gene and the expression of Zn uptake transporters, we screened the M2 population for mutants in NcbZIP19.
In the whole TILLING population, 28 samples were identified as positives in the first round of HRM screening. After the second round of screening, 14 of these were selected as putative mutants. Four mutant alleles in NcbZIP19 were identified after subsequence confirmation by sequencing HRM-positive PCR amplicons (Fig.4B). Of these four mutations, all in exons, two encoded for non-synonymous amino acid substitutions, leading to a change in the predicted amino acid sequence, while the other two caused synonymous changes, which are unlikely to change the function of NcbZIP19. Of the two nonsynonymous mutations, plant T41-104 carried a G>A mutation at position 78 of the coding region, leading to a serine(Ser) to asparagine (Asn) amino acid substitution. In plant T30-36, a G ->A mutation at position 115 of the coding region caused a substitution from aspartic acid (Asp) to asparagine (Asn). The two synonymous mutations were located at positions 196 and 211, respectively in the coding region. Figure4A shows a schematic representation of the NcbZIP19 gene, marking the location of the induced mutations.
Figure 4. Sequencing results of four mutants for NcbZIP19
(A). Genome organization of the bZIP19 genes of A. thaliana (At) and N. caerulescens (Nc). Untranslated regions are indicated with red boxes, exons are indicated with yellow boxes, introns are indicated with black vertical lines. The EMS-induced mutations are indicated by red (non-synonymous) and black (synonymous) arrows below the first exon. (B). Point mutations confirmed by sequencing. The two non-synonymous substations are shown in the red circles. Two synonymous substitutions are shown in black circles
Genome-scale mutant frequency
The frequency of visible phenotypes in a mutant population indirectly reflects the overall, genome-wide mutation frequency. To determine the chance of obtaining a target mutation with an obvious phenotype in the TILLING population, we evaluated the mutation frequency based on the observation of visually mutant phenotypes that correspond to mutations in supposedly or confirmed single genes. A total of five early flowering and 2 double flower mutants were identified (Table 1). Of the early flowering mutants, we identified three mutations in the FLC gene [37], causing aberrant splicing leading to truncated proteins. The two other early flowering mutants were due to mutations in the SVP gene, producing respectively a premature stop codon and a non-synonymous substitution. The double flower phenotype is most likely due to mutations in the AGAMOUS (AG) gene [49], which is a single copy gene in A. thaliana and N. caerulescens. Two to three mutants for each of the three monogenic phenotypes suggests that mutant saturation was almost complete in the TILLING population, with a chance of 13.5% to find no mutants for a gene of comparable size (~ 700 bps in ORF) in this population, based on a Poisson distribution. Based on the two to three mutants obtained for these three genes (FLC, SVP and AGAMOUS), we could estimate the number of knockout mutations per plant in the TILLING population. Assuming one mutant/3000 M2 plants, and around 30,000 genes (29,712) to be predicted for N. caerulescens [47], we expect every M2 plant to carry on average knock-out mutations in 10 genes.
Identification of ionome mutants
Considering that N. caerulescens is a Zn/Cd/Ni hyperaccumulator, leaves of all M2 plants were sampled to determine their leaf ionome by Inductively Coupled Plasma Mass Spectrometry (ICP-MS). Plants in which one of the 20 trace elements that were analysed (Li, B, Na, Mg, P, S, K, Ca, Mn, Fe, Co, Ni, Cu, Zn, As, Se, Rb, Sr, Mo and Cd) was found to have a concentration that was three or more standard deviations higher or lower than the concentration in the WT (Z-score analysis) were considered to be putatively high or low element mutants [48] (Fig.5A). Since no Ni or Cd was supplied to the plants, we initially focused on the identification of low Zn mutants in the 0.4 population. In total, 34 putative low Zn mutants were identified in this M2 population. These mutants were found in different seed batches, suggesting they arose from independent mutation events (Fig. 5B).
Figure 5. Number of putative ionome mutants in the N. caerulescens M2 population. (A).Putative ionome mutant numbers identified in the 0.4% EMS M2 population, consisting of ~ 7000 plants, based on Z-score values (Z ≥ +/-3) to have either a higher (blue) of lower concentration (con) (red) than average for at least the indicated element. (B). Distribution of putative low-Zn mutants (putants) in the M2 population. The 7000 M2 plant ionomes were determined in 20 different runs by ICP-MS. Blue diamonds indicate the average Zn concentration for each run. Red squares indicate the Zn concentration of putative, low-Zn mutants in each run.
To further confirm the element profile of the putative mutants, representative putative low Zn mutants T6-38, T19-7 and T53-16 were selected and 10–15 T3 progeny of those mutants were grown in the greenhouse. The ionome profiles of these progeny plants were not identical, which would be expected for progenies of recessive mutants (Fig.6) and therefore may suggest (partial) dominance. Among the progeny of T6-38, plant T6-38-4 accumulated very little Zn but showed high Sr and Ca concentrations. The four sister plants, however, contained much higher Zn and Mn concentrations than the wild type plants. Plant T6-48-4 was much reduced in size compared to the other plants and displayed strong leaf chlorosis. None of the other putative mutants, neither in the same line nor in others, displayed any obvious aberrant morphological phenotypes (data not shown). As for the progeny of T19-7, plant T19-7-6 showed a very low Zn concentration, but a higher concentration of K and Sr. Most of the progeny of T54-16 did not grow, with only one plant to be alive when sampling. It confirmed the low Zn concentration of the M2 parent and showed a high K concentration. The element profiles of the progenies compared to their mother plant (M2) indicated that the element profiles of the M3 plants T6-38-4, T19-7-6 and T53-16-5 are similar to those found in the M2 generation.
Figure 6. Leaf ionome profiles of representative putative low-Zn N. caerulescens mutants
Ionome profiles based on ICP-MS analysis of one mature expanded leaf per M2 or M3 plant, indicating the Z-scores for each element of indicated plants. The Z-scores for M2 plants were normalized to the average of 7000 individuals. The Z-score for M3 plants were normalized to the average of 50 M3 plants tested. The ionome profiles are shown for (A) SF wild-type (WT) plants; (B) low-Zn M2 plants; (C) M3 progeny of M2 T6-38; (D) M3 progeny of M2 T19-7; (E) M3 progeny of M2 T53-16; (F) putative low-Zn mutants selected from each M3 progeny.
The availability of the element profiles for the whole M2 population allowed us to check the ionome profiles of the two non-synonymous ncbzip19 mutants, T41-104 and T30-36. According to the Z-score analysis, the element profiles of these mutants were not significantly different from the wild type (Fig. 7), although the concentrations of K, Ni, Cu, Zn, Mo, Cd and particularly Mn, were higher in T41-104 than in WT. Unfortunately, neither of these two mutants produced viable offspring (no seeds or no germinating seeds), which made further analysis of them impossible.
Figure 7. The ionome profiles of NcbZIP19 mutants and WT plants.
Plants T41-104 and T30-36 contain a mutation leading to a non-synonymous amino acid substitution in NcbZIP19. Plants T35-111 and T27-1 contain two mutations leading to synonymous substitutions. Profiles of single M2 individuals for each NcbZIP19 mutant are shown. Profile data of WT is the average of six plants. Z-score: standard deviations difference from the means of WT.