Study of Variability in Root System Architecture of Spanish Triticum turgidum L. Subspecies and Analysis of the Presence of a MITE Element Inserted in the TtDro1B Gene: Evolutionary Implications

We analysed nine traits of the root system of 223 genotypes of Triticum turgidum (2n = 4x = AABB) subspecies dicoccoides, dicoccum, turgidum, durum and polonicum, finding a large intra and interspecific variability in both the number and size of roots, as well as in their spatial distribution. We studied the presence of an incomplete MITE (Miniature Inverted-repeat Transposable Element) inserted in the TtDro1B gene, which is present in some genotypes of dicoccoides, dicoccum, and turgidum, but not in polonicum and the 97.9% of the durum accessions. Comparison between genotypes shows that genotypes with the MITE element have smaller and shallower roots. Since Aegilops is considered to be the donor of the wheat B genome, the presence of the same MITE element was analysed in 55 accessions of the species Aegilops speltoides, searsii, bicornis and longissima, and in no case was it detected. We propose that after the emergence of T. turgidum subsp. dicoccoides, the insertion of the MITE element probably occurred in a single plant. Subsequent domestication resulted in genotypes of dicoccum with and without the MITE element, which after selection gave rise to the subspecies turgidum, and durum and polonicum, respectively. The MITE element can be used to differentiate turgidum from the durum and polonicum with high reliability.


RSA Analysis
The RSA study was carried out using the rhizoslide technique according with Ruiz et al. [25] and Boudiar et al. [29]. Briefly, 12 seeds from each of the genotypes analysed were disinfected with sodium hypochlorite solution (1.25%) during 15 min and rinsed 4 times with sterile distilled water. Seeds were placed in Petri dishes with two sheets of filter paper moistened with 4 mL of distilled water were kept at 4 • C for 2 days, and then put in the rhizoslide system and grown in a chamber at 22 • C-18 • C with a photoperiod of 12 h of light for 1 week. The seedlings were then removed from the rhizoslide and the roots were scanned with a Canon "LiDE210" scanner at 300 ppi. Next, the roots of each seedling were manually separated, and a second scan was performed. The first image was used to measure the angles of each root with respect to the vertical, and the second image was used to measure the length, diameter, surface area and volume of each root. All measurements were carried out with the SmartRoot software v.3.32 that is a plugin for ImageJ1.46R (http://imagej.nih.gov/ij/download.html) (Accessed on 1 July 2020). For each seedling, the following variables were obtained or calculated from the different measurements: total root length in cm (TRL), total root surface area in cm 2 (S), total root volume in cm 3 (V), mean root diameter in cm (D), primary root length in cm (PRL), number of roots (NR), mean vertical angle of all the roots in sexagesimal degrees (MRA), the maximum vertical angle in sexagesimal degrees (MxAV) and the most least vertical angle in sexagesimal degrees (MAV).

DNA Extraction
The DNA was extracted from young leaves with the "NZY Plant/Fungi gDNA Isolation kit" (NZYTech, Lisbon, Portugal) following the instructions specified by the manufacturer.

MITE Detection in Triticum turgidum Subspecies
We used 25 ng of DNA for PCR amplification of the specific region of the TtDro1B gene containing the MITE sequence with these primers: TtB1F (5'TGCTCCTCCGAAAAGGGAAT3'), and TtB1R (5'GCTTAGTTGTTGACAGCCTGACTTAT3') designed from T. turgidum TtDro1B sequences (Genbank accession: MZ151532 and MZ151533). Reactions were carried out in a final volume of 25 µL with NZYTaq II 2x Green Master Mix (Nzytech TM ) according to the manufacturer's specifications. The PCR reaction consisted of 1 cycle of 5 min at 94 • C Agronomy 2021, 11, 2294 4 of 14 followed by 35 cycles at 94 • C 30 s; 55 • C 30 s; 72 • C 1 min 30 s followed by 1 cycle at 72 • C 7 min. We used 15 µL of the PCR reaction in a restriction reaction with HaeIII restriction enzyme, and the digestion products were separated on a 1.5% SB-agarose gel.

MITE Detection in Aegilops Species
We used 25 ng of DNA in two independent PCR reactions with TtB1R primer and two different forward primers B1MITEinF (5 CATGTATAAGCTACTCCCTC3 ) with the 3 region inside the MITE sequence and B1MITEoutR (5 ATGCCAGATGAAGCATGT3 ) with the whole sequence outside the MITE element. The PCR reaction consisted of 1 cycle of 5 min at 94 • C followed by 35 cycles at 94 • C 30 s; 55 • C 30 s; 72 • C 1 min 30 s followed by 1 cycle at 72 • C 7 min. The PCR products were separated by electrophoresis on 0.8% TAE-agarose gels.

Statistical Analysis
The means of the RSA variables were compared between subspecies, and between the set of genotypes presenting or not the MITE element. Variables showing equality of variances were compared by ANOVA, and the least significant difference (LSD) test was used to detect differences between pairs of means. The Kruskal-Wallis non-parametric test was used for variables that did not show equality of variances, and the Tukey test was used for comparison between pairs of means. Statistical calculations were performed with StatGraphics plus v.5.1 software.

Results and Discussion
The evolutionary origin of durum wheat is complex, involving one or more hybridisation and polyploidisation events, which have resulted in different lines of the oldest wild relative of durum wheat, T. turgidum subsp. dicoccoides [1,12,42,43]. Domestication of T. turgidum subsp. dicoccoides took place between 10 and 12,000 years ago, giving rise to cultivated emmer wheat T. turgidum subsp. dicoccum [10][11][12]44], that is considered a valuable gene source to improve the elite durum wheat cultivars [45][46][47]. Artificial selection of dicoccum gave rise to other subspecies and a large number of landraces, many of which are maintained in plant germplasm banks. The core collection (CC) obtained from these large collections allows genetic studies and breeding programmes to be carried out [48].

Study of the RSA in T. turgidum Subspecies
The CC of Spanish durum wheat includes 94 accessions of the subspecies durum, turgidum and dicoccum [17]. The analysis of the root system architecture (RSA) of this collection showed great variability in the length, number and diameter of the roots, as well as in the angle of inclination in relation to the vertical of the soil [25]. In the present work, we have extended the RSA study to a total of 223 genotypes belonging to 5 of the 8 subspecies of T. turgidum [49]. Table 1 shows a statistics summary of the nine variables related to RSA in the five subspecies of T. turgidum. Figure 1 shows the means and the confidence intervals for each variable in every subspecies. Variables related to root size characteristics (TRL, S, V, D, PRL and NR) show similar coefficients of variation in the different subspecies, ranging from 4.77 for D in subsp. polonicum to 25.09 for V in subsp. dicoccoides. Nevertheless, the three variables related to root inclination angles (MRA, MAV and MxAV) have larger and more diverse Coefficients of Variation (CVs), with a maximum value of 65.53 for MAV in subsp. polonicum. The nine RSA variables were compared taking subspecies as an independent factor. The variables V, D, PRL, NR, MRA and MAV showed equal variances, and TRL, S and MxAV did not. ANOVA was used for the first group of variables and the Kruskal-Wallis test for the second, and all results showed significant differences (p < 0.05). Comparison between pairs of means was done using the LSD or Tukey test, depending on whether the variables had equal variances or not. Figure 1 shows the means of each genotype for each of the variables and 95% confidence intervals of the LSD or Tukey tests. Table 1. Summary statistics of the RSA variables analysed in the five subspecies of T. turgidum: Total root length in mm (TRL), total root surface area in mm 2 (S) total root volume (V) in mm 3 , mean root diameter in mm (D), primary root length in mm (PRL), total number of roots (NR), mean vertical angle of all the roots in (MRA), the maximum vertical angle in (MxAV) and the most least vertical angle in (MAV). SD (Standard deviation) CV (Coefficient of variation). Min (Minimum value). Max (Maximum value). n = number of accessions analysed.  . In X axis subspecies abbreviations: did, dicoccoides; dic, dicoccum; tur, turgidum; pol, polonicum and dur, durum. Subspecies with the same letter have no statistically significant differences (p > 0.05).

Subspecies
The subspecies dicoccoides has the lowest values for roots' total length, surface area and root volume, which can be explained by the lower number of roots (3.91) compared to the other four subspecies (4.83-5.84). In contrast, the subspecies durum and polonicum have a more developed root system. However, the primary root length is similar in the 5 subspecies analysed, ranging from 20.07 to 22.26 cm. Regarding to the angles formed by the roots, the subspecies polonicum, dicoccoides and durum have the most vertical roots, while dicoccum and turgidum have the shallowest roots. (Table 1 and Figure 1). These results could indicate a greater similarity between the latter two subspecies, as reported by Pascual et al. [50]. According to our data, the domestication of dicoccoides to give rise to dicoccum involved an increase in the number of seminal roots, total length, surface and volume of the root system, and in angles of the roots making them more horizontal. This agrees with Gioia et al. [11], who found an increase in stem and root development when moving from wild to domesticated emmer wheat and then to durum wheat.
The selection by early farmers of dicoccum is likely to have resulted in the subspecies turgidum, polonicum and durum, whose root systems would have been selected according . In X axis subspecies abbreviations: did, dicoccoides; dic, dicoccum; tur, turgidum; pol, polonicum and dur, durum. Subspecies with the same letter have no statistically significant differences (p > 0.05).
The subspecies dicoccoides has the lowest values for roots' total length, surface area and root volume, which can be explained by the lower number of roots (3.91) compared to the other four subspecies (4.83-5.84). In contrast, the subspecies durum and polonicum have a more developed root system. However, the primary root length is similar in the 5 subspecies analysed, ranging from 20.07 to 22.26 cm. Regarding to the angles formed by the roots, the subspecies polonicum, dicoccoides and durum have the most vertical roots, while dicoccum and turgidum have the shallowest roots. (Table 1 and Figure 1). These results could indicate a greater similarity between the latter two subspecies, as reported by Pascual et al. [50]. According to our data, the domestication of dicoccoides to give rise to dicoccum involved an increase in the number of seminal roots, total length, surface and volume of the root system, and in angles of the roots making them more horizontal. This agrees with Gioia et al. [11], who found an increase in stem and root development when moving from wild to domesticated emmer wheat and then to durum wheat.
The selection by early farmers of dicoccum is likely to have resulted in the subspecies turgidum, polonicum and durum, whose root systems would have been selected according to the different cultivation areas. Thus, the subspecies turgidum has long and shallow roots, and is mainly grown in more temperate and humid areas [25], while the subspecies durum and polonicum have long and deep roots. The latter phenotype allows the subspecies durum to be cultivated in extensive hot and dry regions, and the subspecies polonicum, although it has interesting nutritional characteristics, is currently cultivated only in marginal areas of southern Spain and Italy, Algeria and Ethiopia [51,52].

Analysis of a MITE Element
One of the characteristics of the cereal genome, and in particular of durum and common wheat, is the presence of many transposable elements, accounting for 85% of the total nuclear genome [43]. Among the identified transposons are the MITE elements, which are typically located less than 2kb upstream and downstream of the genes [53]. Loarce et al. [35] isolated and determined the sequence of the TtDro1A and TtDro1B genes from accessions BGE045630 and BGE048497 conserved in the CRF, belonging to the durum and turgidum subspecies of T. turgidum, respectively, which showed very different RSAs. Comparison of the sequences of these two genes between the two subspecies showed some differences, the most obvious being the insertion of a fragment of a MITE element in the 5'UTR region of the TtDro1B gene in turgidum subspecies.
In to the different cultivation areas. Thus, the subspecies turgidum has long and shallow roots, and is mainly grown in more temperate and humid areas [25], while the subspecies durum and polonicum have long and deep roots. The latter phenotype allows the subspecies durum to be cultivated in extensive hot and dry regions, and the subspecies polonicum, although it has interesting nutritional characteristics, is currently cultivated only in marginal areas of southern Spain and Italy, Algeria and Ethiopia [51,52].

Analysis of a MITE Element
One of the characteristics of the cereal genome, and in particular of durum and common wheat, is the presence of many transposable elements, accounting for 85% of the total nuclear genome [43]. Among the identified transposons are the MITE elements, which are typically located less than 2kb upstream and downstream of the genes [53]. Loarce et al. [35] isolated and determined the sequence of the TtDro1A and TtDro1B genes from accessions BGE045630 and BGE048497 conserved in the CRF, belonging to the durum and turgidum subspecies of T. turgidum, respectively, which showed very different RSAs. Comparison of the sequences of these two genes between the two subspecies showed some differences, the most obvious being the insertion of a fragment of a MITE element in the 5'UTR region of the TtDro1B gene in turgidum subspecies.
Genomic modifications following polyploidy processes have been studied by several authors [41,43,[54][55][56]. For instance, Hao et al. [57] identified the 4AL-5AL-7BS translocation in eight subspecies of T. turgidum. The 4AL-5AL translocation is present in the diploid species T. urartu and T. monococcum (2n = 14, AA) [58], indicating that the translocation with chromosome 7BS must have arisen when T. turgidum subsp. dicoccoides originated. In this scenario, several hybridisation events could have occurred, giving rise to different lines of T. turgidum subsp. dicoccoides. Some might have the translocation and some might not, but as there are currently no T. turgidum without the translocation, this would imply that the translocation conferred a major evolutionary advantage that would have resulted in the disappearance of the non-translocated cytotypes [57].
Transposons have also been shown to be activated by stresses, including the emergence of new species through hybridisation and chromosome duplication [36,37,41,59]. The presence of the fragment MITE insertion in the TtDro1B gene in some genotypes of the dicoccoides and dicoccum subspecies could be due to hybridisations with different genotypes of Aegilops speltoides (or a related species of the section Sitopsis (S-genome species) that are considered to be the donor of the wheat B genome, which would have the MITE insertion or be absent of it. In an attempt to explain this hypothesis, we analysed the presence of the MITE element in the Dro1B gene in 55 genotypes belonging to the species Ae. speltoides (41), Ae. searsii (5), Ae. bicornis (4) and Ae. longissima (5), respectively.
The detection of the MITE element in the Aegilops species required a new strategy because the HaeIII pattern could not distinguish between the presence and absence of the element in these genotypes. The primers B1MITEinF, with part of its sequence outside the MITE element and the 3' region inside, and the primer B1MITEoutF, with the whole sequence outside the MITE element, were designed. The absence of the MITE element resulted in no amplification in reactions with the B1MITEin primer and a smaller band size than the control MITE-DNA in reactions with B1MITEout ( Figure 3). In none of the accessions of the Aegilops species studied was the presence of the MITE element detected.
Genomic modifications following polyploidy processes have been studied by several authors [41,43,[54][55][56]. For instance, Hao et al. [57] identified the 4AL-5AL-7BS translocation in eight subspecies of T. turgidum. The 4AL-5AL translocation is present in the diploid species T. urartu and T. monococcum (2n = 14, AA) [58], indicating that the translocation with chromosome 7BS must have arisen when T. turgidum subsp. dicoccoides originated. In this scenario, several hybridisation events could have occurred, giving rise to different lines of T. turgidum subsp. dicoccoides. Some might have the translocation and some might not, but as there are currently no T. turgidum without the translocation, this would imply that the translocation conferred a major evolutionary advantage that would have resulted in the disappearance of the non-translocated cytotypes [57].
Transposons have also been shown to be activated by stresses, including the emergence of new species through hybridisation and chromosome duplication [36,37,41,59]. The presence of the fragment MITE insertion in the TtDro1B gene in some genotypes of the dicoccoides and dicoccum subspecies could be due to hybridisations with different genotypes of Aegilops speltoides (or a related species of the section Sitopsis (S-genome species) that are considered to be the donor of the wheat B genome, which would have the MITE insertion or be absent of it. In an attempt to explain this hypothesis, we analysed the presence of the MITE element in the Dro1B gene in 55 genotypes belonging to the species Ae. speltoides (41), Ae. searsii (5), Ae. bicornis (4) and Ae. longissima (5), respectively.
The detection of the MITE element in the Aegilops species required a new strategy because the HaeIII pattern could not distinguish between the presence and absence of the element in these genotypes. The primers B1MITEinF, with part of its sequence outside the MITE element and the 3' region inside, and the primer B1MITEoutF, with the whole sequence outside the MITE element, were designed. The absence of the MITE element resulted in no amplification in reactions with the B1MITEin primer and a smaller band size than the control MITE-DNA in reactions with B1MITEout ( Figure 3). In none of the accessions of the Aegilops species studied was the presence of the MITE element detected. Previous studies have tried to find out whether Triticum turgidum originated after one or several hybridisation and chromosomal duplication events between T. urartu and Ae. speltoides or a related species, which would result in the synthesis of T. turgidum subsp. dicoccoides once or several times [1,2,6,7,13,43]. Our study does not allow us to differentiate Previous studies have tried to find out whether Triticum turgidum originated after one or several hybridisation and chromosomal duplication events between T. urartu and Ae. speltoides or a related species, which would result in the synthesis of T. turgidum subsp. dicoccoides once or several times [1,2,6,7,13,43]. Our study does not allow us to differentiate between the two alternatives; however, the results we obtained with the Spanish landraces that were analysed indicate that a single plant carrying the insertion of the MITE element in the TtDro1B gene probably appeared because of transposition phenomena (Figure 4). This hypothesis is based on the observation that in all the accessions analysed, we detected the same insertion of the truncated MITE element. Subsequent evolution allowed the propagation and expansion of the plants, giving rise to dicoccoides genotypes with or without the MITE insertion. The domestication process between 10,000 and 12,000 years ago [13,44,45], resulted in the emergence of different lines of T. turgidum subsp. dicoccum, some with and some without the MITE element, depending on the type of dicoccoides plant from which they originated. However, in order to have more evidence for this hypothesis, it would be interesting to extend the study to materials from other countries, especially from the Middle East. between the two alternatives; however, the results we obtained with the Spanish landraces that were analysed indicate that a single plant carrying the insertion of the MITE element in the TtDro1B gene probably appeared because of transposition phenomena (Figure 4). This hypothesis is based on the observation that in all the accessions analysed, we detected the same insertion of the truncated MITE element. Subsequent evolution allowed the propagation and expansion of the plants, giving rise to dicoccoides genotypes with or without the MITE insertion. The domestication process between 10,000 and 12,000 years ago [13,44,45], resulted in the emergence of different lines of T. turgidum subsp. dicoccum, some with and some without the MITE element, depending on the type of dicoccoides plant from which they originated. However, in order to have more evidence for this hypothesis, it would be interesting to extend the study to materials from other countries, especially from the Middle East. The selection undertaken by ancient farmers gave rise to the subspecies turgidum, durum and polonicum. Therefore, selection of dicoccum plants with the MITE element resulted in the different landraces of T. turgidum subsp. turgidum, suggesting that there is The selection undertaken by ancient farmers gave rise to the subspecies turgidum, durum and polonicum. Therefore, selection of dicoccum plants with the MITE element resulted in the different landraces of T. turgidum subsp. turgidum, suggesting that there is probably a link between the insertion of the MITE element and the adaptation of this subspecies to a warmer and more humid environment. Similar results were obtained by Muterko and Salina [41] when they analysed, in a collection of hexaploid and tetraploid wheat, the insertion of a transposon of a new family called M882 in the promoter region of the VRN-B3 gene. The selection from dicoccum plants without the MITE element resulted in different landraces of T. turgidum subsp. durum and T. turgidum subsp. polonicum. In this case, the absence of the MITE element would be related to growth in warmer and drier environments. However, we have detected some accessions of subsp. turgidum and subsp. durum that do or do not have the MITE element, respectively. These discrepancies could be explained because of spontaneous crosses between genotypes of both subspecies and subsequent selection by farmers over many generations, resulting in the different landraces preserved in the Genebanks [14,15].

Relationship between the Presence of the MITE Element Insertion and the RSA
In rice, Uga et al. [13] showed that the Dro1 gene is involved in the angle at which roots develop, and Loarce et al. [35] studied the expression of the TtDro1A and TtDro1B genes in eight genotypes of the subsp. turgidum and durum and found differences in expression of the two genes and in both subspecies. Thus, the TtDro1A gene is more highly expressed than the TtDro1B gene. Moreover, the TtDro1A/TtDro1B ratio is higher in the subspecies turgidum than in durum, so that the higher the ratio, the shallower the roots are. The latter authors proposed that the insertion of the MITE element in the TtDro1B gene of the subspecies turgidum leads to a decrease in the expression of this gene and to shallower roots. According to our results, surface roots could have been selected in cultivation areas with higher water availability, facilitating at the same time the acquisition of nutrients such as phosphorus, which accumulate in the surface layers of the soil [60].
We pooled the genotypes of the five subspecies of T. turgidum according to whether or not they had the MITE element inserted in the TtDro1B gene. A statistical summary is show in Table 2. Table 2. Summary statistics of the root system architecture variables in T. turgidum genotypes without and with the MITE element: total root length in mm (TRL), total root surface area in mm 2 (S) total root volume (V) in mm 3   The RSA variable means differed significantly between the sets of genotypes without and with the MITE element (p < 0.05). Figure 5 shows the mean values and 95% confidence intervals of the LSD or Tukey tests. The RSA variable means differed significantly between the sets of genotypes without and with the MITE element (p < 0.05). Figure 5 shows the mean values and 95% confidence intervals of the LSD or Tukey tests. Genotypes with the MITE element inserted have less root system development, and the root angles are larger and therefore the roots grow more horizontal. This phenotype is mainly observed in the subspecies dicoccum and turgidum and confirms the similarity between them found by Pascual et al. [50] and suggests a closer resemblance of the subspecies turgidum to the more ancestral domesticated forms of durum wheat represented by dicoccum, while polonicum and durum would have appeared more recently. Our results confirm those obtained by Tang et al. [7], who studied the DMC1 gene and found that subsp. turgidum clusters with subsp. dicoccum are part of the same clade, while subsp. durum and subsp. polonicum are part of a different clade.

MITE as a Subspecies Marker
Molecular markers have been used in the genus Triticum to differentiate species with high phenotypic similarity. Thus, Czajkowska et al. [61] designed a test based on Ppd-1 gene variation that allows discriminating between the tetraploid species T. turgidum from T. timophevii, which are morphologically very similar and lead to misclassification errors. In our work, the subspecies turgidum and durum are morphologically very similar with naked and nonbrittle spikes. From a practical point of view, the identification of the MITE element insert has a useful application as most subsp. turgidum landraces have the insert (94.6%) while subsp. durum landraces lack it (97.9%), allowing the differentiation of the two subspecies with a high degree of accuracy. Genotypes with the MITE element inserted have less root system development, and the root angles are larger and therefore the roots grow more horizontal. This phenotype is mainly observed in the subspecies dicoccum and turgidum and confirms the similarity between them found by Pascual et al. [50] and suggests a closer resemblance of the subspecies turgidum to the more ancestral domesticated forms of durum wheat represented by dicoccum, while polonicum and durum would have appeared more recently. Our results confirm those obtained by Tang et al. [7], who studied the DMC1 gene and found that subsp. turgidum clusters with subsp. dicoccum are part of the same clade, while subsp. durum and subsp. polonicum are part of a different clade.

MITE as a Subspecies Marker
Molecular markers have been used in the genus Triticum to differentiate species with high phenotypic similarity. Thus, Czajkowska et al. [61] designed a test based on Ppd-1 gene variation that allows discriminating between the tetraploid species T. turgidum from T. timophevii, which are morphologically very similar and lead to misclassification errors. In our work, the subspecies turgidum and durum are morphologically very similar with naked and nonbrittle spikes. From a practical point of view, the identification of the MITE element insert has a useful application as most subsp. turgidum landraces have the insert (94.6%) while subsp. durum landraces lack it (97.9%), allowing the differentiation of the two subspecies with a high degree of accuracy.

Conclusions
There is great variability in the RSA of the Spanish subspecies dicoccoides, dicoccum, turgidum, durum and polonicum, the last two having the longest and deepest roots. The insertion of an incomplete MITE element in the 5' UTR region of the TtDro1B gene has been identified in genotypes of dicoccoides, dicoccum and turgidum subspecies, but not in polonicum and only in the 2.09% of the durum accessions, and in none of the 55 genotypes of Ae. speltoides, searsii, bicornis and longissima studied. The results of this study seem to suggest that it is likely that the insertion of the MITE element occurred in a single plant from which all genotypes with the MITE element are derived. However, in order to have more evidence for this hypothesis, it would be interesting to extend the study to materials from other countries, especially from the Middle East. Genotypes with the MITE element have shallower and less developed roots. The MITE element inserted in the TtDro1B gene serves to differentiate, in most genotypes, the subspecies turgidum and durum.