Effect of soybean ureases on seed germination and plant development

Abstract Urease catalyzes the hydrolysis of urea to ammonia and carbon dioxide. The ammonia (nitrogen (N) product of urease activity) is incorporated into organic compounds. Thus, urease is involved in N remobilization, as well as in primary N assimilation. Two urease isoforms have been described for soybean: the embryo-specific, encoded by the Eu1 gene, and the ubiquitous urease, encoded by Eu4. A third urease-encoding gene was recently identified, designated Eu5, which encodes the putative protein product SBU-III. The present study aimed to evaluate the contribution of soybean ureases to seed germination and plant development. Analyses were performed using Eu1/Eu4/Eu5-co-suppressed transgenic plants and mutants of the Eu1 and Eu4 urease structural genes, as well as a urease-null mutant (eu3-a) that activates neither the ubiquitous nor embryo-specific ureases. The co-suppressed plants presented a developmental delay during the first month after germination; shoots and roots were significantly smaller and lighter. Slower development was observed for the double eu1-a/eu4-a mutant and the eu3-a single mutant. The N content in transgenic plants was significantly lower than in non-transgenic plants. Among the mutants, eu3-a presented the lowest and eu1-a the highest N content. Altogether, these results indicate that increased ureolytic activity plays an important role in plant development.


Introduction
Nitrogen (N) is the most limiting plant nutrient, possibly after fixed carbon, for plant growth and development (Marschner, 2012). Therefore, efficient mechanisms both to take up N in its various forms and to reallocate it are necessary for optimal N use efficiency (Witte, 2011). In plant cells, urea is an important internal and external source of N that must be converted to ammonia for N assimilation (Wang et al., 2008). In nature, two major biochemical processes lead to urea production: (1) arginase-catalyzed production of urea (and ornithine) from arginine (a major N storage form); (2) purine degradation to glyoxylate and urea. While arginases are active in plants, purine degrada-tion exclusively through urea, though occurring in many bacteria, fungi and algae, does not occur in soybean and Arabidopsis. Rather, in these dicotyledonous plants, the purine degradation product allantoin, which contains the four ring N atoms of purine, is degraded to four ammonia molecules, bypassing a urea intermediate (summarized in Witte, 2011).
Nonetheless, plant assimilation of arginine-derived urea is important for efficient N use, both in mobilization of seed N reserves during germination, as well as in remobilization of N in senescing tissues. In soybean, arginine is the major amino acid repository of seed N (Micallef and Shelp, 1989), and one of the predominant amino acids in angiosperm seed protein in general (Van Etten et al., 1963). In soybean, arginase action during germination releases much urea, which is hydrolyzed by urease action Polacco, 1999, 2000) The importance of urease for recycling N was highlighted in aged Arabidopsis thaliana seeds that failed to germinate when urease was chemically inhibited, but could be rescued by an external N source (Zonia et al., 1995). And, according to Bohner et al. (2015) 13% of N exported out of senescing leaves of A. thaliana via the petiole is urea. Urease-negative soybean (mutants and nickel-deprived wildtype) accumulate urea in necrotic leaf tips (Stebbins et al., 1991) to levels approaching 2.5% dry weight (Eskew et al., 1983). We note that in Stebbins, (1991) and Eskew et al. (1983) available N was not limiting. Urea can be hydrolyzed by two different enzymes: urease and an ATP (and biotin)-dependent urea carboxylase/allophanate hydrolase. The latter, found in some fungi, algae and at least one bacterium (Kanamori et al., 2004), has never been reported in plants. Rather, all plants appear to have a urease (Hogan et al., 1983 and our own observations). In soybean and Arabidopsis, urea nitrogen is only available after urea hydrolysis by urease (Goldraij et al., 2003;Witte, 2011;Polacco et al., 2013). Urease (EC 3.5.1.5) was the first identified nickel-dependent metalloenzyme (Dixon et al., 1975), and much has been learned of the construction and function of its metallocenter (Carter et al., 2009). Urease catalyzes the hydrolysis of urea to ammonia and carbon dioxide. In addition to plants, ureases are synthesized by bacteria, fungi and algae (Krajewska, 2009). The N product of urease activity -ammonia -is incorporated into organic compounds mainly by glutamine synthase activity (Mobley et al., 1995;Sirko and Brodzik, 2000). Thus, urease is involved in N remobilization, as well as in primary N assimilation (Cao et al., 2010).
In addition to the N assimilatory function of urease, plant ureases appear to have defensive roles against herbivore and fungal attack . In soybean, three urease isoforms have been described. The ubiquitous urease, encoded by the Eu4 gene, is expressed at low levels in all tissues and is responsible for recycling both metabolically-derived and exogenous urea (Polacco et al., 1985;Torisky et al., 1994;Goldraij et al., 2003). The embryo-specific urease, encoded by Eu1, is highly expressed in developing embryos and accumulates in mature seeds (Polacco and Havir, 1979;Polacco and Winkler, 1984;Polacco and Holland, 1993). A third urease-encoding gene was recently identified in the soybean genome (Polacco et al., 2011(Polacco et al., , 2013. This gene was designated Eu5, and its putative protein product was named SBU-III. Eu5 is expressed in the first stages of root development and during seed maturation. Its transcript levels are lower than those of the other two soybean urease isoforms (Wiebke-Strohm et al., 2016).
Urease-negative mutant soybean plants were examined to ascertain the role(s) of the urease isoforms. An embryo-specific urease null mutant (eu1-a) seems not to exhibit an altered physiology (Polacco et al., 2011). Ubiquitous urease missense mutants (eu4-a and eu4-b) produce an inactive protein and display no ureolytic activity in leaves, roots and hypocotyls (Torisky and Polacco, 1990;Stebbins et al., 1991;Witte et al., 2002, Goldraij et al., 2003Polacco et al., 2011). eu4 callus cultures cannot use 5 mM urea as N source, but are resistant to 50 mM urea in the presence of a standard NH 4 + + NO 3 -N source, and show growth responses contrary to those of Eu4 cultures (Goldraij et al., 2003). The eu1-a/eu4-a double mutants were considered virtually devoid of ureolytic activity Goldraij et al., 2003).
A null mutant for the Eu3 gene, which encodes an accessory protein, UreG, necessary for urease activation, has also been characterized (Freyermuth et al., 2000). There is only a single copy of this gene in the soybean genome, and the deletion mutant eu3-a exhibits a complete loss of urease activity Polacco et al., 2011;Tezotto et al., 2016).
A previous study was performed by our team aiming to overexpress Eu4 in soybean plants. Unexpectedly, the transgenic plants exhibited co-suppression of the endogenous and the introduced Eu4 transgene, resulting in decreased ureolytic activity (Wiebke-Strohm et al., 2012). As null mutants for the ubiquitous urease have not been obtained to date, the co-suppressed transgenic plants represent a powerful tool for functional gene studies. Here, we sought to determine the roles of urease in soybean development by elimination of all urease isoforms.

Plant material and growth conditions
Homozygous eu1-a, eu4-a, eu1-a/eu4-a and eu3-a mutants have been described previously. All, except eu1-a, were recovered from EMS (ethyl methane sulfonate) mutagenesis of cv. Williams, and were subsequently outcrossed to Wiliams 82. The original eu1-a mutation was recovered from the "Itachi' landrace and introgressed into Williams by Dr Dick Bernard (University of Illinois-Champaign-Urbana) by five crosses. Thus, the genetic background of these mutants is the Williams (eu1-a) and Williams82 (eu4-a, eu1-a/eu4-a and eu3-a) cultivars. Williams and Williams82 are supposedly isogenic, except for a fungal resistance gene introgressed into Williams 82 (Bernard and Cremeens, 1988).
Two independent transgenic events (A3 and A8) of soybean cultivar IAS5 that presented co-suppression of eu4 were obtained from bombarded embryogenic tissue. The vector used for transformation contained the Eu4 and the gfp-encoding sequences (Wiebke-Strohm et al., 2012). Plants derived from non-transgenic embryogenic tissues submitted to the same culture conditions were recovered and used as a control. Subsequent generations were obtained by self-fertilization of plants.
Transgenic seeds of A3 and A8 events (from T 1 , T 2 and T 3 ) were placed in Petri dishes containing sterile filter paper moistened with sterile distilled water for 24 h. Seeds expressing the gfp reporter were selected under blue light using a fluorescence stereomicroscope (Olympus ® ), equipped with a BP filter set with a 488 nm excitation filter and a 505-530 nm emission filter. GFP-positive and negative plants were also PCR-screened to confirm presence/absence of the transgene using the protocol described by Wiebke-Strohm et al. (2012). Positive transgenic, as well as non-transgenic seeds, were sown in organic soil and maintained in a greenhouse until maturity at FUNDACEP-CCGL (Cruz Alta, RS, Brazil) and supplemented with a nutrient solution containing either NO 3 or NH 4 (as N source).
For seed germination and developmental evaluation, GFP-positive transgenic (T 2 ), mutants, IAS5 non-transgenic and Williams82 seeds were sown in pots containing vermiculite and maintained for one month in a growth chamber at 26 ± 1º C with a 16/8 h light/dark cycle at a light intensity of 22.5 mEm -2 s -1 . Plants were not supplemented with any nutrient solution during the first 30 days of development.

RNA extraction, cDNA synthesis and quantitative real-time PCR (RT-qPCR)
Total RNA was extracted from roots with Trizol reagent (Invitrogen) and treated with DNase I (Invitrogen) according to the manufacturers' instructions. First-strand cDNAs were obtained using 1 mg of DNA-free RNA, M-MLV Reverse Transcriptase SystemTM (Invitrogen) and oligo(dT) primers.
RT-qPCR was performed on a StepOne Real-time Cycler TM (Applied Biosystems). PCR-cycling conditions were implemented as described: 5 min at 94°C, followed by 40 cycles of 10 s at 94°C, 15 s at 60°C and 15 s at 72°C. A melting curve analysis was performed at the end of the PCR run, over the range of 55-99°C, increasing the temperature stepwise by 0.1°C every 1 s. Each 25-mL reaction comprised 12.5 mL cDNA (1:50 dilution), 1x PCR buffer (Invitrogen), 2.4 mM MgCl 2 , 0.024 mM dNTPs, 0.1 mM of each primer, 2.5 mL of SYBR-Green (1:100,000, Molecular Probes) and 0.03 U of Platinum Taq DNA Polymerase (5 U/ml, Invitrogen). All PCR assays were performed in technical quadruplicates and 10 biological samples. Reactions lacking cDNA were used as negative controls.
Transcript levels of the three urease-encoding genes were evaluated. The F-box protein and a metalloprotease were used as references for gene expression normalization (Jian et al,. 2008;Libault et al., 2008). Primer sequences are presented in Table 1. The expression data analyses were performed after comparative quantification of amplified products using the 2 -DDCt method (Livak and Schmittgen, 2001).

Ureolytic activity
Urease activity in transgenic (A3 and A8) and nontransgenic plants was determined with a urease indicator solution: 6 g urea, 10 mL cresol red (1 mg/mL ethanol), 10 mL KH 2 PO 4 /K 2 HPO 4 /EDTA (10 mM KPi/1 mM EDTA pH 7.0) and 1 mL sodium azide 20% (w/v) per liter (Meyer-Bothling and Polacco, 1987). Powdered leaves and roots ( ± 100 mg) of two-week old plants, were incubated in a 1 mL urease indicator solution for 24 h at 60 o C, as well as mature seed slices that were incubated for 20 min at room temperature. The eu3-a mutant was used as negative control. As urea is hydrolyzed by urease, the ammonia released increases the pH and turns to pink the initial yellow coloration of the cresol red pH indicator.

Germination
Germination capacity of seeds (T 2 ) from two transgenic events (A3 and A8) was compared with that of IAS5 non-transgenic seeds. For eu1-a, eu4-a, eu1-a/eu4-a and eu3-a mutant seeds germination capacity was compared with that of Williams82 non-mutant seeds. Seed germination, defined as radicle protrusion, was recorded over one month.

Developmental pattern
Seven, 14, 21 and 30 days after sowing, seedlings from all genotypes were classified according to their developmental stage following the categories proposed by Neumaier et al. (2000): VE = emergence of cotyledons; VC = completely opened cotyledons; V1 = completely developed unifoliate leaf pair; V2 = completely developed first trifo-Role of ureases in plant development 211 In addition, the dry matter and length of roots and shoots were also determined one month after sowing.

Grain yield
The number of seeds produced per plant in three generations of transgenic (T 1 , T 2 and T 3 ) and non-transgenic plants were compared.

Nitrogen content
Shoot nitrogen content (N mobilized from the cotyledons) was measured by the Kjeldahl method according to the methodology described by Tedesco et al. (1995) one month after sowing. This analysis evaluated 18 plants derived from each line A3 and A8 and 10 non-transgenic ones. Eighteen plants from each mutant (eu1-a, eu4-a, eu1-a/eu4-a and eu3-a) and non-mutant plants were also evaluated.

Statistical analysis
A Student's t-test was used to compare the expression levels (RT-qPCR) of urease-encoding genes (Eu4, Eu1, Eu5) in roots of transgenic vs. non-transgenic plants and mutant vs. non-mutant plants.
In order to compare developmental stages, a score was attributed to each developmental category (VE = 1, VC = 2, V1 = 3, V2 = 4, V3 = 5). A generalized linear model for repeated measures was used to compare plant development among genotypes (IAS5 transgenic vs. IAS5 non-transgenic plants; eu1-a or eu4-a or eu3-a or eu1-a/eu4-a mutant vs. Williams82 non-mutant plants). ANOVA followed by Bonferroni's post hoc test were performed on dry matter, length and weight of roots and shoots data. A Student's t-test was carried out in order to compare the number of seeds produced per transgenic and non-transgenic plant in different generations (T 1 , T 2 and T 3 ). Data on shoot nitrogen content was compared among genotypes by ANOVA followed by Tukey's post hoc test. Analyses were performed using SPSS 18.0 software.

Gene expression
Roots were chosen for gene expression analysis by RT-qPCR because Eu5 transcripts are mainly detected in this organ (Wiebke-Strohm et al., 2016). As expected, the progeny of transgenic plants showed lower Eu4 transcript levels than non-transgenic controls, suggesting that the co-suppressed phenotype was maintained. Additionally, it was observed that the other urease-encoding genes, Eu1 and Eu5, were also down-regulated ( Figure 1A).
The transcript levels of all three urease-encoding genes in the mutants plants followed the predicted pattern: eu4-a and eu3-a displayed normal mRNA levels of the three genes; eu1-a and eu1-a/eu4-a presented lower levels of Eu1, but normal Eu4 and Eu5 expression levels (Figure1B). It is worth noting that although not affecting the mRNA expression levels, the eu4-a (and eu4-b) mutant produces a non-functional enzyme with a single amino acid replacement (Goldraij et al., 2003).

Ureolytic activity
Ureolytic activity in transgenic and non-transgenic plants was evaluated with cresol red pH indicator by the seed chip assay of dried samples of leaves, roots and seeds. As expected, samples containing non-transgenic tissues showed pink coloration, indicative of urea hydrolysis. On the other hand, leaf and root samples of transgenic plants showed no observable color change even after 24 h incubation at 60 ºC, indicating absence or drastic reduction of urease activity. Slices of mature seeds exhibited little or no activity. Very low urease activity was confirmed comparing transgenic with eu3-a seeds (used as a negative control) (Supplemental Figure S1).
Taken together, urease expression and activity assays indicate that transgenic plants have had all three urease-212 Rechenmacher et al. encoding genes silenced in all tissues, reinforcing the potential of these plants for functional studies.

Germination
The germination rate of T 2 seeds from the two independent transgenic events was evaluated and compared to that of non-transgenic seeds. No differences were detected, suggesting that the absence of all three ureases did not affect germination. The same result was observed on germination of the eu1-a, eu4-a, eu1-a/eu4-a and eu3-a mutants and non-mutant Williams82. Germination rates were higher than 90% for all genotypes (data not shown).

Developmental pattern
Plant development was evaluated 7, 14, 21 and 30 days after sowing. T 2 transgenic plants and non-transgenic plants, as well as eu1-a, eu4-a, eu1-a/eu4-a and eu3-a mutant and non-mutant plants were classified into developmental categories. Interaction among genotype, developmental categories and time-course was highly significant (p < 0.01). The two independent transgenic events showed a significant delay in development when compared with non-transgenic plants (Figure 2A and Figure 3). Size and dry weight of shoots and roots were significantly lower in transgenic plants when compared with non-transgenic (Table 2).
The developmental pattern of mutant plants was compared to non-mutants. The eu1-a single mutant showed a pattern similar to that of the non-mutant. The eu4-a mutant developed faster than control plants. Slower development was observed for the eu1-a/eu4-a double mutant and the eu3-a single mutant when compared to the non-mutant plants ( Figure 2B). One month after germination, no differences were observed among the shoot and root sizes of eu1-a and eu4-a mutants and non-mutant plants. The shoot sizes of eu1-a/eu4-a double mutants and eu3-a single mutants were significantly smaller than those of the other genotypes. Regarding shoot and root weight, the eu1-a/eu4-a and eu3-a mutants were lighter than the other two mutants, but did not differ significantly from control. Total root length analysis showed that mutant eu3-a had the smallest root system (Table 2).

Nitrogen content
The nitrogen content was measured in shoots of onemonth-old plants. The N content in transgenic plants of the two independent events was significantly lower than that present in non-transgenic plants ( Figure 4A). The comparison among mutants showed that eu3-a presented the lowest and eu1-a the highest N content ( Figure 4B)

Grain yield
The number of seeds produced by three generations of transgenic plants was compared with those produced by non-transgenic plants. A significantly lower number of seeds was obtained for transgenic plants ( Figure 5).

Discussion
The present study aimed to evaluate the contribution of soybean ureases to seed germination and plant development. Analyses were performed using co-suppressed transgenic plants and plants with mutations in urease-related genes. The transgenic plants were the progeny (T 2 ) of two independent events in which the Eu4 gene was downregulated as previously described (Wiebke-Strohm et al., 2012). Molecular analyses showed that Eu4 co-suppression was maintained in the transgenic progeny. In addition it was verified that transgenic plants exhibited very low transcript levels of the other two ureases encoded by the Eu1 and Eu5 genes. The phenomenon of co-suppression by transgenic DNA has been observed in many organisms, Role of ureases in plant development 213  with the introduction of transgenic copies of a gene resulting in reduced expression of the transgene, as well as of the endogenous gene. This effect depends on the sequence identity between transgene and endogenous gene (Ketting and Plasterk, 2000). Soybean plants with mutations in urease genes were also evaluated confirming the expected expression pattern. Normal transcript levels of all three urease-encoding genes were detected for eu4-a and eu3-a mutants. This result is consistent with the eu4-a allele encoding G468E missense-altered ubiquitous urease (Goldraij et al., 2003). Eu3 is the only UreG-encoding gene in the soybean genome. UreG is essential for urease activation, and the eu3-a mutant presents a complete loss of urease activity (Freyermuth et al., 2000), consistent with a > 90% deletion of the UreG ORF (Tezotto et al., 2016). The eu3-a mutation did not alter Eu4, Eu1 and Eu5 expression levels, indicating a lack of feedback control on urease structural gene transcription by apo-urease(s). The mutants eu1-a and eu1-a/eu4-a exhibited similar expression patterns: lower Eu1, but normal Eu4 and Eu5 transcript levels. Since eu1-a is a null mutant (Polacco and Holland, 1993), low levels of Eu1 transcripts are expected, and indeed, Torisky et al. (1994) employing Eu1-specific PCR primers, recovered no detectable product from eu1-a embryo cDNA.
Ureolytic activity in transgenic and non-transgenic plants was evaluated by the seed chip assay. As expected, leaf and root samples of transgenic plants showed no color change even after 24 h incubation. The urease activity level can be inferred based on the time change "yellow to pink" on the seed chip assay (Polacco et al., 2011). According to the authors, 0.2% normal urease specific activity requires 10 hours for changing the solution color. When the activity decreases to 0.15%, the time required for solution color change is 48 hours. Based on these data, we conclude that the urease activity in leaves and roots of transgenic plants was absent or less than 0.2%, since no change in color was observed after a 24 hours incubation. The reaction cata- 214 Rechenmacher et al.   lyzed by urease is essential to allow most organisms (those lacking urea carboxylase) to use external or internally generated urea as a nitrogen source (Mobley and Hausinger, 1989;Mobley et al., 1995). It has been demonstrated that aged A. thaliana seeds fail to germinate when urease was chemically inhibited, but seed viability could be rescued by an external N source (Zonia et al., 1995). In the present study no differences were detected in germination rates of transgenic and mutant seeds. However, it is important to highlight that soybean seeds were not aged and have a much higher protein content than A. thaliana seeds.
An association between urease activity and developmental pattern was observed. Transgenic plants, as well as eu3-a and the double eu1-a/eu4-a mutants, showed a delay over the first month after sowing. The delay in development was maintained even in adult transgenic plants and may be the cause of lower seed production. In A. thaliana, both urease transcripts and ureolytic activity increased after germination, especially in 8/9-day-old wild-type seedlings (Zonia et al., 1995). Embryo-specific urease (Eu1) activity in young soybean plants was also observed by Torisky and Polacco, (1990). Similarly, high transcript level of Eu4 and moderate transcript levels of Eu1 and Eu5 were detected in soybean seeds one day after dormancy break (Wiebke-Strohm et al., 2016). Taken together, these results indicate that increased urease content and/or ureolytic activity play a role in early stages of plant development.
A role in making nitrogen available during plant development has been attributed to soybean ubiquitous urease due to its catalytic activity and tissue distribution (Stebbins et al., 1991). However, our results indicate that Eu1 and Eu5 have a contribution in the developmental process as well. This is supported by the finding that transgenic A3 and A8, mutant eu1-a/eu4-a and eu3-a plants showed a delay in the first developmental stages. In transgenic plants, the impairment in development was confirmed by the significant reduction in size and weight of roots and shoots one month after germination. These data are consistent with significantly lower nitrogen content detected in transgenic plants. Regarding mutant plants, eu3-a tends to be smaller and lighter than the other genotypes, although significant differences were only detected for roots and shoots size. The reduction in dry matter is also reflected in the lowest nitrogen content. An unexpected result was the significantly higher nitrogen content present in eu1-a. The differences in nitrogen content might be due to differences in mutant genetic backgrounds. Williams is the background for eu1-a, while Williams82 is the one for the other mutants. In our experiment, non-mutant Williams82 was used as control.
Based on bioinformatics analyses, Eu5 has been suggested not to be a functional ureolytic enzyme due to a number of mutations, including deletions (Witte, 2011). However, according to our results the product of this gene might be involved in plant development. Further studies are necessary to elucidate whether the ureolytic activity and/or other non-enzymatic property of ureases are involved on plant development.