In this work, we used four different genetic materials which were compared pairwise. The first pair was lpa1-1/lpa1-1 vs. +/+ control in “B73” genetic background kindly provided by Dr. Victor Raboy, USDA ARS, Aberdeen, ID, USA. B73 is an inbred line used in the 80’s for the hybrid production, now widely used as a benchmark. The second pair was lpa1-1/lpa1-1 vs. +/+ control in “B73/Mo17” genetic background. These two genotypes were obtained by crossing lpa1-1/lpa1-1 B73 with “Mo17” inbred line provided by the germplasm bank at DISAA, Department of Agricultural and Environmental Sciences—Production Landscape, Agroenergy, University of Milan. The F1 obtained was selfed, generating an F2 population scored by Chen’s assay and genotyped to isolate lpa1-1/lpa1-1 and +/+ genotypes (as described in the following sections). Following three cycles of sib crossing, we obtained two synthetic populations differing only for the presence of the lpa1-1 in homozygous status (Fig. 1).

Fig. 1.

Pedigree scheme used to obtain the two synthetic populations in B73/Mo17 geneticbackground. The two synthetic populations differ only for the presence of the lpa1-1 in homozygous status.

The Chen assay is a colorimetric method used to distinguish the HIP (high inorganic phosphate) phenotype. The HIP phenotype is diagnostic for the presence of the lpa1 mutation. We used the qualitative Chen assay with some modifications [32]. The seeds were ground in a mortar using a pestle, and 100 mg of the flour obtained was placed in a microtiter plate with 1 mL of 0.4 M HCl solution. After 1 h at room temperature, 100 $\mu$L of the extract was transferred to another microtiter plate, adding 900 $\mu$L of Chen’s reagent (6 NH${}_2$SO${}_4$, 2.5% ammonium molybdate, 10% ascorbic acid, H${}_2$O [1:1:1:2, v/v/v/v]). The blue-colored phosphomolybdate complex was visible after 1 h: the intensity of the blue color is directly proportional to the free phosphate content.

DNA was extracted from wild-type and mutant leaves as described by Dellaporta et al. [33]. PCR was performed with the primers 13L (5′-CTTCATGATCTGCGGTCACG-3′) (forward primer, position +5293) and 51R (5′-AAGCATCAGCTTCGGGTAATGT-3′) (reverse primer, position +6460) and 2 $\mu$L of DNA aliquots were used as a template. The reaction mix underwent an initial denaturation step at 94 °C for 2 min, 37 cycles of denaturation at 94 °C for 15 s, annealing at 65 °C for 30 s, and extension at 72 °C for 30 s. Extension at 72 °C for 5 min was performed to complete the reaction. Amplicon length was verified on 1% (w/v) agarose gels, and the presence of the mutation in lpa1-1 was confirmed by Sanger sequencing.

Experiments were carried out in 2020 and 2021 in the experimental field of the University of Milan located in Landriano (PV), Italy (N45°18${}^{\prime }$, E9°15${}^{\prime }$). The seeds of each genotype were sown at the end of April in both years. The experiment was laid out in randomized blocks, and each variety (B73, B73/Mo17, and the respective lpa1-1 mutants) was cultivated in three plots, for a total of 12 plots. The size of each plot was about 10 m${}^2$ (5 m $\times$ 2.1 m), sown in three rows, with a density of 60 plants per plot. The experimental fields were in maize–maize rotation with standard soil fertilization (about 220 kg/ha of nitrogen). The maize plants were grown by conventional farming methods (pre-emergence herbicide was applied). Water was applied by sprinkler irrigation as needed.

The following agronomic data were collected in the two backgrounds in both years:

Plant height: the distance between the base of the male inflorescence (panicle) and the ground level was measured considering more than 30 plants chosen randomly in the three plots. Plant height was measured in mid-July during full flowering using a measuring rod.

Ear height: insertion height of the first ear measured considering the distance between the ground level and the insertion point of the first ear in the stem. These data were collected from the same plants used for plant height.

Moreover, five representative ears for each plot (15 ears for each genotype) were randomly selected for grain yield estimation in the two backgrounds in both years. The ears were harvested on September 12th, and the seeds were immediately dried to 12–13% of relative humidity. Later, the following parameters were collected:

Ear length: measured with a ruler on 15 ears for each genotype.

Average single seed weight: single representative seeds were selected from each ear and were measured using an analytical weight scale (Precisa XB 220A, 0.1 mg). The test was performed in triplicate.

The number of seeds per ear: evaluated by counting all the seeds of each sampled ear.

Seeds weight/ear: measured by weighing all the seeds that make up an ear.

In parallel, 200 seeds of the inbred line B73 and 200 seeds of the relative lpa1-1 mutant were sown in a non-irrigated part of the field to study drought stress in field conditions.

In particular, root sampling was carried out in the season 2021, and the root system architecture of the wild-type was compared with the relative lpa1-1. Only two days prior to sampling, the fields were irrigated with 12 mm of water to facilitate the excavation of roots. At flowering, ten roots per genotype were excavated in the first 30 cm using standard shovels. The excavated roots were shaken to remove the main fraction of soil adhering to the roots, and the remaining soil particles were removed by vigorous rinsing at low pressure, according to Trachsel et al. [34]. On the cleaned roots, different parameters were collected: number of above-ground whorls occupied with brace roots (BW); brace root number (BO); angle of the first arm of brace roots originating from the first whorl in relation to horizontal (BA1); culm diameter at the base of the sampled plant (CD).

Later, a crown root and a brace root were detached from each root system to estimate the root density. The roots of each sample were scanned (using a high-resolution digital scanner). The images were processed with Adobe Photoshop software: the shadows of the roots and the background of the images were removed, and the color of the roots was changed to green and made uniform. The processed images were analyzed using ImageJ 1.52 [35] and Easy Leaf Area software (University of California, Davis, CA, USA) [36] to calculate the root area/root length ratio.

Different parameters in the epigeal part of the plant were collected: net photosynthesis (P${}_n$), transpiration rate (E) and stomatal conductance (GS) were measured on B73 and its relative lpa1-1 with a portable CIRAS-2 photosynthesis system (PP-Systems, PP-Systems, Amesbury, MA) (CO${}_2$ 350–360 $\mu$mol$\cdot$mol${}^{-1}$; PAR 1700 $\mu$mol$\cdot$m${}^{-2}$$\cdot$s${}^{-1}$; 75% humidity). Measurements were carried out at two different moments: the 6th of July under optimal water conditions (no stress) and the 16th of July under moderate drought stress. All the measurements were taken between 11 and 12 AM on the third last leaf, the temperature was 30 $\pm$ 1 °C and the humidity 75%

Stable carbon isotope analysis was carried out on flag leaves of both B73 and lpa1-1/lpa1-1 plants grown in the field and sampling was done on the 14th of July. Harvested leaves were preventively dried at 80 °C and then ground to a fine powder by using a grinder. Samples were prepared by adding 1 mg of dry powdered plant tissues into 5 $\times$ 9 mm tin capsules. Capsules were carefully closed by folding them with cleaned tweezers and were then transferred to an auto-sampler.

The $\delta$${}^{13}$C values of samples were measured using a Flash 2000 HT elemental analyzer (hermo Fisher Scientific, Cambridge, United Kingdom) coupled, via a ConFLo IV Interface, with a Delta V Advantage isotope ratio mass spectrometer (IRMS) (Thermo Fisher Scientific, Bremen, Germany), interconnected to the software Isodat 3.0 (Isodat 3.0, Thermo Fisher Scientific, Bremen, Germany), according to Bononi et al. [37]. Briefly, the combustion/reducing reactors, combined in a single quartz tube, were heated at 1020 °C. The He gas flow was 120 mL min${}^{-1}$ and 100 mL min${}^{-1}$ for carrier and reference, respectively. The O${}_2$ purge for flash combustion was 3 s at a flow rate of 175 mL min${}^{-1}$ per sample. The GC separation column was maintained at 45 °C. The CO${}_2$ reference gas pulse was introduced three times (20 s each) at the beginning of each run. The run time of the analysis was 600 s for a single run. The analysis of each sample (6 flag leaves for each genotype) was performed five times.

Calibration was performed using three secondary reference materials provided by IAEA: NBS18 ($\delta$${}^{13}$C = –5.014 $\pm$ 0.035‰); IAEA-600 ($\delta$${}^{13}$C = –27.771 $\pm$ 0.043‰); IAEA-612 ($\delta$${}^{13}$C = –36.722 $\pm$ 0.006‰). Two in-house solid standards, sulfanilamide ($\delta$${}^{13}$C = –27.23 $\pm$ 0.06‰) and methionine ($\delta$${}^{13}$C = –30.01 $\pm$ 0.05‰), were used for normalization and quality assurance.

The isotope ratio ${}^{13}$C/${}^{12}$C was expressed using the standard $\delta$${}^{13}$C notation:

(1)$${\delta }^{13}\mathrm{C}=\left[{\left({}^{13}\mathrm{C}/{}^{12}\mathrm{C}\right)}_{\text{sample}}/{\left({}^{13}\mathrm{C}/{}^{12}\mathrm{C}\right)}_{\mathrm{VPDB}}-1\right]\times 1000$$

which expresses the part per thousand deviation of the isotope ratio ${}^{13}$C/${}^{12}$C of a sample relative to an international standard, the Vienna Pee Dee Belemnite [38].

For the quantification of abscisic acid (ABA), 100 seeds of the two genotypes (B73 and the relative lpa1-1 mutant) were germinated under controlled conditions. After 2 weeks, the plant leaves were sampled, and the following method was applied for the quantification of ABA.

Sample preparation and extraction method: samples were ground by an electric blender and an amount of 5.0 $\pm$ 0.1 g of sample was weighed in a 40 mL glass tube and 5.0 $\pm$ 0.1 g of anhydrous sodium sulfate and 10 mL of methanol were added. The sample was vortexed for 10 min and after 5 min was vortexed for more 5 min; then, the mixture was centrifuged for 2 min at 3500 rpm. The supernatant passed through a 0.45 $\mu$m membrane and was directly analyzed by HPLC-DAD (20 $\mu$L).

Reagents and standard: the reagents used for the determination and quantification of ABA were analytical-grade water, methanol and acetic acid (Carlo Erba, Italy). Standard abscisic acid (purity $\ge$98%, lot number BCCG6453) was from Supelco (Milan, Italy).

Apparatus: high performance liquid chromatography analysis of ABA was performed on a Shimadzu instrument (Shimadzu, Milan, Italy) equipped with two pumps (10AD vp), a diode array detector (SPD-M10A vp), a system controller (SCL-10A vp) and a Rheodyne 20 mL injection loop (Thomas Scientific, Cotati, Swedesboro, NJ, USA). The HPLC pumps and diode array system were controlled by computer using a LC Solution version 1.25 SP5 workstation program (Shimadzu, Milan, Italy). The analytical column employed was a Kinetex XB-C18 (150 mm $\times$ 3.0 mm, particle size 5 $\mu$m) by Phenomenex (Phenomenex, Bologna, Italy).

Chromatographic conditions: the mobile phase consisted of (A) water/acetic acid (99:1 v/v) and (B) methanol. The operating conditions were as follows: 0–30 min linear gradient from 10% B to 100% B, 30–50 min isocratic elution 100% B, 50–70 min linear gradient from 100% B to 10% B, and 70–80 min isocratic elution 10% B. The flow rate was 0.4 mL$\cdot$min${}^{-1}$. The injection volume was 20 $\mu$L and the wavelength used for the detection was 265 nm.

Analytical quality assurance: the abscisic acid content was quantified using the calibration curve. A stock standard solution of ABA was prepared dissolving the compound in methanol (concentration 1.0 mg$\cdot$mL${}^{-1}$). Working standard solution of 0.25, 0.50, 1.00, 2.00, 2.50, 4.00, 5.00, 8.00, 10.00, 12.50, and 20.00 $\mu$g$\cdot$mL${}^{-1}$ were prepared by diluting the stock solution with methanol from 0.25 to 20.00 g mL${}^{-1}$. The calibration curve was linear (R${}^2$ = 0.9987) in the range 0.25–20.00 $\mu$g$\cdot$mL${}^{-1}$. The standard deviation was 0.05 (derived from 10 replicated analyses). ABA content was expressed as mg$\cdot$kg${}^{-1}$ of fresh vegetable matrix. The limit of detection (LOD) was 0.01 mg$\cdot$kg${}^{-1}$ and the limit of quantification (LOQ) was 0.1 mg$\cdot$kg${}^{-1}$. The recovery was 98 $\pm$ 3% for concentration between 0.50 and 5.00 mg$\cdot$kg${}^{-1}$, and 102 $\pm$ 3% for concentration between 5.00 and 20.00 mg$\cdot$kg${}^{-1}$.

In this work, we used four different genetic materials compared pairwise. The first pair was lpa1-1/lpa1-1 vs. +/+ control in B73 genetic background. The second pair was obtained by crossing lpa1-1 B73 with Mo17 inbred line provided by our germplasm bank. The F1 obtained was selfed, generating an F2 population scored by Chen’s assay and genotyped to isolate lpa1-1/lpa1-1 and +/+ genotypes (as described in the following sections). Following three cycles of sib crossing, we obtained two synthetic populations differing only for the presence of the lpa1-1 in homozygous status (Fig. 1).

The presence of the lpa trait in the material used in the following experiments was confirmed through chemical and molecular methods. The qualitative Chen assay for free phosphate allowed us to rapidly distinguish the lpa1-1 mutants with respect to the wild phenotype: the former were colored blue in contact with the Chen reagent, while the wild-types remained white due to the low content of free phosphate (Fig. 2A). Later, Sanger sequencing was used to molecularly confirm the presence of the mutation (Fig. 2B): in comparison with the wild-type allele, the coding sequence of the lpa1-1 allele is characterized by the presence of a Single Nucleotide Polymorphism (SNP), C to T, in position 5759 with respect to the starting codon on the genomic sequence (Fig. 2C).

Fig. 2.

Chen assay for free phosphate. (A) microtiter row 1 showed the results of the assay performed on four wild-type seeds; in row 2, the blue-colored phosphomolybdate complex was visible and represented four mutant seeds. (B) The presence of the mutation (C to T) was molecularly confirmed by Sanger sequencing. (C) Structure of the ZmMRP4 gene and position of the Single Nucleotide Polymorphism (SNP), C to T.

The field experiment was organized in randomized blocks; plants were grown under high-input conditions, and water was not a limiting factor during the season. Field emergence percentage was calculated by counting the number of plants present between the V2 and V4 stages of development and dividing by the total number of seeds. This parameter was measured in both the backgrounds and years and appeared to be reduced in lpa1-1 compared to the wild phenotype, as reported in Table 1. This value was much lower than the germination of the same seeds under controlled conditions at the optimal temperature and humidity (data not shown).

Furthermore, the plants of lpa1-1 grown under field conditions were characterized by a lower plant and ear height compared to the relative wild-types in both the years analyzed and in both the backgrounds (Fig. 3A,B). In the control line B73, the plant height reached an average of 189.6 cm in 2020 and 198.3 cm in 2021, while the mutant lpa1-1 introgressed in B73 was 184.3 and 181.1 cm in height, respectively (Fig. 3A). On the same plants used for plant height, ear height was also measured. Ear height was statistically higher in B73 than lpa1-1 in both the years (127.1 vs. 93.4 cm in 2020 and 134.8 vs. 94.3 cm in 2021) (Fig. 3B).

Fig. 3.

Agronomic parameters collected in the field experiment on the first genetic background (B73). Plant height (A) and ear height (B) measured at flowering in B73 and in its relative lpa1-1. The data represent the means of n $\ge$30 biological replicates. Ear length (C), average single seed weight (D), number of seeds/ear (E) and seeds weight/ear (F). The data were collected in 2020 and 2021 and represent the means of 15 biological replicates. Significant differences between the wild-type and lpa1-1 were assessed by Student’s t-test (*p 0.1, **p 0.05 and ***p 0.01).

The same trend was recorded in the synthetic population B73/Mo17, although, in general, these plants were taller than in the first background due to the hybrid vigor. The height of B73/Mo17 plants reached 234.6 cm in 2020 and 243.2 cm in 2021, while the mutant lpa1-1 introgressed in B73/Mo17 was 189.6 and 190.1 cm high, respectively (Fig. 4A). Also, ear height was statistically higher in the wild-type than lpa1-1 in both the years (142.9 vs. 103.7 cm in 2020 and 147.8 vs. 105.3 cm in 2021) (Fig. 4B).

Fig. 4.

Agronomic parameters collected in the field experiment on the second genetic background (B73/Mo17). Plant height (A) and ear height (B) measured at flowering in B73/Mo17 and in its relative lpa1-1. The data represent the means of n $\ge$30 biological replicates. Ear length (C), average single seed weight (D), number of seeds/ear (E) and seeds weight/ear (F). The data were collected in 2020 and 2021 and represent the means of 15 biological replicates. Significant differences between the wild-type and lpa1-1 were assessed by Student’s t-test (*p 0.1, **p 0.05 and ***p 0.01).

After collecting these agronomic data at flowering, other parameters (ear length, average single seed weight, number of seeds/ear, and seeds weight/ear) were measured at harvesting for grain yield estimation, collecting n $\ge$5 ears for each plot (in total, 15 ears per genotype).

Under optimal growth and water conditions, lpa1-1 showed equal or superior performance compared to the inbred line B73. In particular, the ear length and the average seed weight were statistically higher in the mutant in both the years (Fig. 3C,D). In contrast, the number of seeds per ear was statistically the same in 2020 and slightly higher in B73 in the following year (Fig. 3E). An important parameter in grain yield estimation is the seed weight per ear, which was statistically superior in lpa1-1 in 2020 and the same between the two genotypes in 2021 (Fig. 3F).

These results were confirmed in the other genetic background, the synthetic population B73/Mo17. Most of the parameters measured were statistically the same in the wild-type and in the relative mutant, and even the seed weight/ear of lpa1-1 in 2020 was statistically higher than in the control (110.24 vs. 93.38 g).

Another experiment on lpa1-1 was conducted in the field in order to study its response to drought stress conditions. In this case, the plants of B73 and lpa1-1 were grown in a non-irrigated part of the field. Ten roots per genotype were sampled at flowering and then cleaned by vigorous rinsing (Fig. 5A). Photos of the two genotypes were taken (Fig. 5B), and among the parameters collected, no significant differences were found regarding the brace roots, as indicated in Table 2. Also, the diameter of the culm, an epigeal parameter measured at the base of the sampled plant, was statistically the same between lpa1-1 and the wild-type (Table 2).

Fig. 5.

Schematic representation of the root sampling and collected data. Ten roots per genotype were sampled in the field through a shovel and the soil was removed with a vigorous rinsing (A). An example of B73 and lpa1-1 root system architecture (B). A representative crown root and a brace root were detached from each root system (C) and the area occupied per unit of length was calculated with Easy Leaf Area software (D). Significant differences between the wild-type and lpa1-1 were assessed by Student’s t-test (*p 0.1, **p 0.05 and ***p 0.01).

Furthermore, a crown root and a brace root were detached from each root system to estimate root density (Fig. 5C). Images were processed, and the area occupied per unit of length was calculated, as indicated in Fig. 5D. This ratio was statistically higher in the mutant than in the wild-type in both types of post-embryonic roots: in the case of crown roots, the area reached 1.70 cm${}^2$/cm in lpa1-1, while it was 0.89 cm${}^2$/cm in the inbred line B73. The same trend occurred with brace roots: 0.90 vs. 0.60 cm${}^2$/cm, respectively. Considering these data, the root system architecture of lpa1-1 did not appear to be a limiting factor under water limitation.

The results obtained so far seem to exclude the root system as the leading cause of the drought stress in lpa1-1. Therefore, our research focused on the epigeal part of the plant, and several measurements were made at two different moments: under optimal water conditions (16% of volumetric soil moisture) and under moderate drought stress (6% of volumetric soil moisture). In Fig. 6A it is possible to observe how the leaves of the two genotypes appeared when the measurements under moderate drought stress were carried out: the leaves of the mutant appeared altered, stressed, and curled upwards. The three parameters collected with the CIRAS-2 Portable Photosynthesis System highlighted that under optimal water conditions the transpiration rate and stomatal conductance were significantly higher (by 31% in both cases) in lpa1-1 compared to the control line, revealing a greater stomatal opening in the mutant (Fig. 6C,D). Pn did not significantly differ in the two genotypes under non-stressful conditions, although the mean was higher in lpa1-1 than in the wild-type (29.70 vs. 24.70 $\mu$mol m${}^{-2}$ s${}^{-1}$) (Fig. 6B). When water stress occurred, these three parameters did not significantly change in the wild-type, while they dropped in the mutant: the net photosynthesis decreased by 58% (Fig. 6B), the transpiration rate by 63% (Fig. 6C) and the stomatal conductance of the 67% (Fig. 6D), revealing lpa1-1 as alterated in stomatal regulation under more stressful water conditions.

Fig. 6.

Measurements carried out in the field at two different moments, under optimal water conditions (no stress) and under moderate drought stress. (A) Under moderate drought stress, the leaves of lpa1-1 were stressed and curled upwards compared to the wild phenotype. The photos were taken the same day. The CIRAS-2 Portable Photosynthesis System was used in the field to measure the net photosynthetic CO${}_2$ rate (Pn) (B) the transpiration rate (E) (C) and the stomatal conductance (GS) (D). Values represent the means of eight biological replicates. Data were analyzed with one-way ANOVA and post-hoc Tukey test considering four values (B73/lpa1-1, no stress/stress).

To further support this hypothesis, we measured the carbon stable isotope composition of the maize flag leaves of the two genotypes grown under optimal water conditions — i.e., when both transpiration rate and stomatal conductance were higher in lpa1-1 than in the wild-type — since stomatal conductance may significantly alter ${}^{13}$C/${}^{12}$C discrimination during CO${}_2$ fixation [39]. $\delta$${}^{13}$C analysis revealed that B73 and lpa1-1 differently discriminate the carbon stable isotopes under optimal water conditions (–13,41 $\pm$ 0.02‰ vs. –12.58 $\pm$ 0.09‰, respectively) (p 0.01). Hence, the difference between the two values ​​of $\delta$${}^{13}$C was $\mathrm{\Delta }$${}_{(\mathit{\text{lpa1-1}}-\text{B73})}$ = +0.83 $\pm$ 0.11‰, suggesting that a different regulation of stomatal conductance occurred during the growth. Indeed, the $\delta$${}^{13}$C values of the organic matter accumulated during the growth mirrors the carbon stable isotope separations occurring under different stomatal conductance, thus maintaining the memory of all the activities that determined them.