Rothamsted Repository Download

Key message Degradation of nitrogen-rich purines is tightly and oppositely regulated under drought and low nitrogen supply in bread wheat. Allantoin is a key target metabolite for improving nitrogen homeostasis under stress. Abstract The metabolite allantoin is an intermediate of the catabolism of purines (components of nucleotides) and is known for its housekeeping role in nitrogen (N) recycling and also for its function in N transport and storage in nodulated legumes. Allantoin was also shown to differentially accumulate upon abiotic stress in a range of plant species but little is known about its role in cereals. To address this, purine catabolic pathway genes were identified in hexaploid bread wheat and their chromosomal location was experimentally validated. A comparative study of two Australian bread wheat genotypes revealed a highly significant increase of allantoin (up to 29-fold) under drought. In contrast, allantoin significantly decreased (up to 22-fold) in response to N deficiency. The observed changes were accompanied by transcriptional adjustment of key purine catabolic genes, suggesting that the recycling of purine-derived N is tightly regulated under stress. We propose opposite fates of allantoin in plants under stress: the accumulation of allantoin under drought circumvents its degradation to ammonium (NH 4+ ) thereby preventing N losses. On the other hand, under N deficiency, increasing the NH 4+ liberated via allantoin catabolism contributes towards the maintenance of N homeostasis.


Dear Author,
Here are the proofs of your article.
• You can submit your corrections online, via e-mail or by fax.
• For online submission please insert your corrections in the online correction form. Always indicate the line number to which the correction refers.
• You can also insert your corrections in the proof PDF and email the annotated PDF.
• For fax submission, please ensure that your corrections are clearly legible. Use a fine black pen and write the correction in the margin, not too close to the edge of the page.
• Remember to note the journal title, article number, and your name when sending your response via e-mail or fax.
• Check the metadata sheet to make sure that the header information, especially author names and the corresponding affiliations are correctly shown.
• Check the questions that may have arisen during copy editing and insert your answers/ corrections.
• Check that the text is complete and that all figures, tables and their legends are included. Also check the accuracy of special characters, equations, and electronic supplementary material if applicable. If necessary refer to the Edited manuscript.
• The publication of inaccurate data such as dosages and units can have serious consequences. Please take particular care that all such details are correct.
• Please do not make changes that involve only matters of style. We have generally introduced forms that follow the journal's style. Substantial changes in content, e.g., new results, corrected values, title and authorship are not allowed without the approval of the responsible editor. In such a case, please contact the Editorial Office and return his/her consent together with the proof.
• If we do not receive your corrections within 48 hours, we will send you a reminder.
• Your article will be published Online First approximately one week after receipt of your corrected proofs. This is the official first publication citable with the DOI. Further changes are, therefore, not possible.
• The printed version will follow in a forthcoming issue.

Please note
After online publication, subscribers (personal/institutional) to this journal will have access to the complete article via the DOI using the URL: http://dx.doi.org/ [DOI]. If you would like to know when your article has been published online, take advantage of our free alert service. For registration and further information go to: http://www.link.springer.com.
Due to the electronic nature of the procedure, the manuscript and the original figures will only be returned to you on special request. When you return your corrections, please inform us if you would like to have these documents returned.
Nitrogen (N) is a macronutrient required in large quantities for plant development and growth with N deficiency causing chlorosis in older leaves and significant yield losses. Under N deficiency and natural senescence plants translocate available N from source tissues to sink tissues, such as young leaves (Masclaux-Daubresse et al. 2010) and developing grains, accounting for 60-92% of total grain N ( Barbottin et al. 2005). The glutamine synthetase-glutamate synthase (GS-GOGAT) cycle plays an important role in this process since it recycles N liberated from the catabolism of N-rich macromolecules, such as protein and nucleic acids, into lowmolecular-weight organic compounds for long-distance N transport (Lea and Miflin 2010).
Purines are the most abundant N heterocyclic compounds in nature and are found in nucleic acids (DNA, RNA) and many other cellular components, such as ATP, GTP or NADH (Werner and Witte 2011). Plants undergo the Electronic supplementary material The online version of this article (https ://doi.org/10.1007/s1110 3-019-00831 -z) contains supplementary material, which is available to authorized users. complete breakdown of the purine ring via a catabolic pathway enabling the recycling of both carbon (C) and N (Fig.  S2). Overall, the oxidation of one molecule of xanthine to one molecule of glyoxylate liberates three molecules of CO 2 and four molecules of ammonium (NH 4 + ), which are likely to be reassimilated by the GS-GOGAT cycle into amino acids. The pathway starts with the conversion of xanthine to urate catalysed by xanthine dehydrogenase (XDH) (Triplett et al. 1982;Werner and Witte 2011). Urate is further processed by urate oxidase (UOX) producing 5-hydroxyisourate (5-HIU), 5-HIU is then converted to allantoin via the 2-oxo-4-hydroxy-4-carboxy-5-ureido-imidazoline (OHCU) intermediate by allantoin synthase (AS) (Hanks et al. 1981;Ramazzina et al. 2006;Kim et al. 2007;Lamberto et al. 2010;Pessoa et al. 2010). Allantoin is metabolised to allantoate by allantoinase (ALN) and then to ureidoglycine by allantoate amidohydrolase (AAH) (Yang and Han 2004;Todd and Polacco 2006;Werner et al. 2008). The last two enzymatic steps are catalysed by ureidoglycine aminohydrolase (UGAH), which converts ureidoglycine to ureidoglycolate (Serventi et al. 2010). Ureidoglycolate amidohydrolase (UAH) converts ureidoglycolate to hydroxyglycine and, lastly, hydroxyglycine decays to glyoxylate by a nonenzymatic reaction (Werner et al. 2010).
In addition to its housekeeping role in N recycling, the purine catabolic pathway has an important function in certain dinitrogen (N 2 )-fixing legumes (Schubert 1986;Sinclair and Serraj 1995;Alamillo et al. 2010;Coleto et al. 2014). Allantoin and allantoate (also known as ureides) are the main products of atmospheric N 2 fixation in root nodules, which are then translocated to the shoot (Herridge et al. 1978;Pate et al. 1980). In recent years, the ureide allantoin has gained attention by the scientific community as several metabolomics studies reported this metabolite to accumulate in a broad range of plant species under drought (Bowne et al. 2011;Oliver et al. 2011;Silvente et al. 2012;Degenkolbe et al. 2013;Yobi et al. 2013;Casartelli et al. 2018), high salt (Kanani et al. 2010;Wu et al. 2012;Nam et al. 2015;Wang et al. 2016), cold (Kaplan et al. 2004) and sulfate starvation (Nikiforova et al. 2005). In contrast, allantoin was found to be reduced under prolonged N deficiency in maize and rice (Amiour et al. 2012;Coneva et al. 2014). Further, the work of Watanabe et al. (2014) revealed that Arabidopsis AtALN mutants, which constitutively accumulated allantoin, were more tolerant to desiccation stress. It was demonstrated that allantoin mediates abscisic acid (ABA) signalling by stimulating the activity of genes and enzymes belonging to the ABA-producing pathways. In addition, allantoin was shown to accumulate in Arabidopsis leaves grown under drought and salt stress in coordination with transcriptional changes of a number of purine catabolic genes (Irani and Todd 2016;Lescano et al. 2016). To date, there are few studies on the regulation of the purine catabolic pathway under N deficiency, however, early reports showed that AtALN was strongly up-regulated when Arabidopsis seedlings were grown under N starvation (Yang and Han 2004).
Given the potentially important role of allantoin in N metabolism and stress tolerance but the limited information available in cereals, the aim of the present study was to characterise the purine catabolic pathway in bread wheat. For this, we selected two wheat genotypes (RAC875 and Mace) that are adapted to Australian environments. The gene loci were identified based on the reference genome of the cultivar Chinese Spring and their chromosomal location was experimentally verified. Quantification of allantoin in different tissues throughout plant development revealed accumulation of allantoin under drought and reduced allantoin levels under N limitation and this was associated with differential regulation of the genes in the purine catabolic pathway. Monitoring of N-metabolite pools in seeds revealed that allantoin levels progressively increase in developing grains and that genotypic differences exist for accumulation of allantoin and other important N-containing metabolites.

Plant growth
The experiment was conducted in a controlled environment with day/night cycle of 12 h/12 h at a flux density at canopy level of 300 µmol m −2 s −1 , 20 °C/15 °C day/night temperature and 82% average humidity. Potting mixture was composed of river sand and coco-peat and prepared according to Melino et al. (2015). Granular urea was provided as basal N application with rates of 150 and 75 mg N kg −1 for high and low N treatments, respectively. At stem elongation (39 days after sowing) a third of the basal urea rates for each treatment were applied by soil drenching. A soil water retention curve was constructed by measuring the pre-dawn leaf water potential of 3-week old seedlings under progressive drought stress with a Scholandertype pressure chamber (Soil Moisture Equipment Corp., Santa Barbara, USA) (Fig. S1b). Overall, the experiment was comprised of two cultivars (RAC875 and Mace), three treatments (high nitrogen well-watered (HN-WW), low nitrogen well-watered (LN-WW) and high nitrogen drought (HN-D)), six sampling time points and six biological replicates for a total of 216 experimental units. Individual seeds were germinated in black square plastic pots (11 × 11 cm 2 area, 14 cm height) containing 1.5 kg of soil mix arranged in a randomised complete block design. WW conditions were maintained by daily watering of the pots to 20% soil water content (SWC), calculated as SWC = [(m wet soil − m dry soil ):m dry soil ] × 100. Drought was induced by withholding water until signs of leaf rolling appeared (approximately 6.5% SWC; Fig. S1a). Samples collected during the reproductive stages were harvested according to the anthesis date of individual plants and water was first withheld 2 weeks after anthesis. Whole spikes were collected from the main stem of each plant and to minimise the differences in development of the grains along the spike, only the middle part of the spike was employed for further analyses. Rachis was removed and the vegetative part of the spike (hereafter referred as spikelet) was separated from the developing grains. Details on the sample collection throughout the experiment are given in Fig. S1a. Samples for molecular analysis were snap-frozen in liquid nitrogen and stored at − 80 °C until further use.  Table S1 for details). Each purine catabolic gene was assigned with the name of the Arabidopsis orthologous according to Watanabe et al. (2014). For wheat, the sub-genome localisation (e.g., Chr 1AL) was additionally included in the gene name to allow distinction of the three homeologous sequences (Table S1).

Synteny analysis
Synteny analysis were conducted based on available genome sequences of Sorghum bicolor, Oryza sativa and Brachypodium distachyon using the Phytozome 11 'Ancestry' tool (phytozome.jgi.doe.gov). Generally, five genes upstream and downstream of the respective purine catabolic pathway gene were included in the orthology analysis. These genes in bread wheat were identified by TBLASTN searches of Brachypodium genes against the TGACv1 wheat genome assembly (http://pre.plant s.ensem bl.org/Triti cum_aesti vum/Info/Index ). Genes were considered syntenic if they were present on the same TGACv1 scaffold where the wheat purine catabolic genes were annotated.

Wheat nulli-tetrasomic lines
To verify the chromosomal location of the wheat purine catabolic gene Chinese Spring nulli-tetrasomic (NT) lines were used in which individual wheat chromosomes were replaced (nullisomic) by an extra pair (tetrasomic) of their homeologs (Sears 1954). PCR was performed using 150 ng of genomic DNA of the NT lines as template using standard reaction conditions of ThermoPol (BioLabs, USA) using the primers provided in Table S1.

Metabolite extraction
Metabolites were extracted from 10 mg of homogenised, freeze-dried tissue with 500 µl of 100% (v/v) methanol containing 12.5 µM 13 C 15 N-allantoin (internal standard), except for mature grains samples for which 25 µM 13 C 15 Nallantoin was used. Samples were vortexed and incubated in an Eppendorf Thermomixer at 1400 rpm and 30 °C for 15 min followed by a 15 min centrifugation at 13,000 rpm (4 °C). The supernatant was transferred to a new tube. 500 µl of milli-Q water was added to the remaining pellet, vortexed and centrifuged for 15 min at 15,000 rpm. The supernatant was combined with the previous one, vortexed for 30 s and centrifuged for 15 min at 15,000 rpm. The resulting supernatant was transferred to a new tube and 300 µl of 100% (v/v) chloroform was added, vortexed and centrifuged for 5 min at 15,000 rpm. 800 µl of the top (polar) phase was transferred into a new tube for allantoin analysis. All samples were re-constituted in 10 µl of methoxyamine hydrochloride (30 mg ml −1 pyridine) and derivatised at 45 °C for 60 min at 500 rpm before adding 20 µl of N-methyl-N-(tert-butyldimethylsilyl)trifluoroacetamide (MTBSTFA) with 1% (w/v) trimethyl chlorosilane (TMCS) and incubated at 45 °C for 45 min.

GC-MS instrument conditions
1 µl of derivatised sample was injected into a GC-QqQ-MS system comprised of a Gerstel 2.5.2 Autosampler, a 7890A Agilent gas chromatograph and a 7000 Agilent triple-quadrupole MS (Agilent, Santa Clara, USA) with an electron impact (EI) ion source. The GC was operated in constant flow mode with helium as carrier gas. The MS was adjusted according to the manufacturer's recommendations using tris-(perfluorobutyl)-amine (CF43). A J&W Scientific VF-5MS column (30 m long with 10 m guard column, 0.25 mm inner diameter, 0.25 µm film thickness) was used. The injection temperature was set at 250 °C, the MS transfer line at 290 °C, the ion source was adjusted to 230 °C and the quadrupole at 150 °C. Helium was used as carrier gas at flow rate of 1 ml min −1 . Nitrogen (UHP 5.0) was used as collision cell gas at flow rate of 1.5 ml min −1 . Helium (UHP 5.0) was used as quenching gas at a flow rate of 2.25 ml min −1 . Gain factor for the triple axis detector was set at 2. Derivatised sample was injected into the column at 100 °C followed by 1 min hold, followed by a ramp of 25 °C min −1 to 325 °C.

Method optimization
Allantoin and 13 C 15 N-allantoin standards were purchased from Sigma Aldrich (Australia). Retention time, a corresponding unique precursor ion and product ions were identified on the GC-QqQ-MS instrument for each standard. Collision energy was optimised between 0 and 20 V for the identified precursor to product ion transitions (Multiple Reaction Monitoring; MRM). A product ion was selected as the Target ion (T) and the other subsequent MRM transition was set as the Qualifier ion (Q) (see Table S2 for details).
Linearity of the method was tested using serial dilutions of the calibration standard and showed a linear calibration range between 0.98 µM to 250 µM 13 C 15 N-allantoin. Correlation coefficient (R 2 ) of the calibration curve was 0.99 for the target allantoin MRM (398 → 171). The limit of detection for the allantoin MRM, was 0.294 µM based on a signal to noise ratio of 3. Data was processed using Agilent MassHunter QQQ Quantitative Analysis software (B.07.00). Allantoin was quantified by single point calibration based on the relative response (the response area of allantoin (MRM 398 → 171) divided by the response area of 13 C 15 N-allantoin (MRM 400 → 173)) and 13 C 15 N-allantoin concentration. Sample dry weight and extraction volume were taken into consideration when calculating the final allantoin concentration.

Metabolome-wide analysis
Metabolites were extracted from 60 mg dry weight (DW) of freeze-dried wheat spikelet and developing grain samples using 100% methanol containing 13 C-sorbitol as internal standard. Chloroform followed by water were added to the mixture. Polar phase aliquots were freeze-dried using a speed vacuum. Aliquots for GC-time of flight (TOF)-MS analysis were reconstituted in N-Methyl-N-(trimethylsilyl) trifluoroacetamide (MSTFA) and methoxyamine hydrochloride as derivatizing agents. Samples were analysed as described previously by Lisec et al. (2006) and Erban et al. (2007). Aliquots for ion chromatography (IC) were reconstituted in milli-Q water and analysed according to Moschen et al. (2016). Aliquots for free spermidine analysis were reconstituted in 0.2 N perchloric acid solution, dansylated and quantified by HPLC according to Do et al. (2013).

Amino acid analysis
Amino acids were extracted from 50 mg DW of freeze-dried wheat samples with 1 ml of 10 mM sodium acetate containing 250 nmol ml −1 Norvaline (internal standard). Amino acids were quantified on a Waters Acquity™ UPLC system using the Waters AccQ-Tag Ultra Chemistry Kit following the manufacturer's instructions (Waters Corp., USA). Chromatograms were analysed with Empower® 3 Software (Waters Corp., USA).

Total N analysis
Total N concentration was determined using the Elementar rapid N exceed ® (Elementar Analysensysteme GmbHanalyser, Germany) using 100 mg DW of homogenised wheat grain samples. Aspartic acid (250 mg) was used as standard for calibration: the theoretical aspartic acid N% (10.52) was divided by the N% measured by the instrument generating a N factor. The N factor was then used to correct the N% measured for each sample.

RNA preparation
Preparation of cDNA libraries were prepared from snapfrozen samples of the youngest fully emerged leaf (YFEL), flag leaf, stem, spikelet and developing grain under control (HN-WW) and treatment (HN-D and LN-WW) conditions. Samples were ground to a fine powder using a 2010 Geno/ Grinder ® (SPEX SamplePrep, USA) and total RNA was extracted from frozen samples with a phenol and guanidine thiocyanate buffer according to Chomczynski (1993). To extract RNA from developing grains which have a high polysaccharide content, an extraction buffer (1% (w/v) sarcosyl, 150 mM NaCl, pH 9) and a guanidine hydrochloride-based buffer for purification according to Singh et al. (2003) was employed. Genomic DNA was removed using the TURBO DNA-free™ Kit (Ambion®, Thermo Fisher Scientific, USA) following the manufacturers' instructions. High RNA quality was confirmed in a 2% (w/v) agarose gel visualized under UV light and RNA concentrations were quantified using a ND-1000 spectrophotometer (NanoDrop Technologies, USA). 1 µg of RNA was used for cDNA synthesis using the SuperScript® III kit (Thermo Fisher Scientific, USA) as per manufacturers' instructions. cDNA quality was verified by PCR amplification of the actin gene (Table S3).

Primer design and quantitative reverse transcription PCR (qRT-PCR) analysis
Quantitative real-time PCR was performed with KAPA SYBR ® Fast qPCR kit Master Mix, and amplification was real-time monitored on a QuantStudio™ 6 Flex Real-Time PCR System (Applied Biosystems, USA). Gene-specific primers targeted to amplify all three homeologs simultaneously were designed with AlleleID® software (Premiere Biosoft) (Table S3) and the specificity of each pair was verified by melting curve analysis and sequencing of the products. Change in gene expression were calculated using qBASE + software and reported as calibrated normalised relative quantities (CNRQ) that represents the relative gene expression level between different samples for a given target gene: NRQ is then divided by a calibration factor (CF) (Hellemans et al. 2007). Four reference genes (Table S4) were quantified by qRT-PCR: TaActin, TaGAPdH, TaCyclophilin and TaEFA2. CNRQ values were calculated using the most stable genes within a specific tissue (selected by qBASE + software).

Data analysis
Statistical analyses were performed using GraphPad Prism version 7.00 for Windows (GraphPad Software, La Jolla California USA. http://www.graph pad.com). All data are reported as mean ± SEM. Significant differences between means of two groups of data were tested by Student's t-test. Significant differences between means of more than two groups were tested by two-way ANOVA. For metabolomewide analysis, metabolites levels were log transformed to improve the normality of the data set and then scaled by subtracting the median metabolite value in each metabolite distribution. Hierarchical clustering by Pearson's correlation distance and PCA analyses were performed with the support of ClustVis web tool (Metsalu and Vilo 2015). Transcriptional data was reported as log2 ratio of the foldchange between treatment (drought or low N) and control conditions.
To experimentally validate the chromosomal localization of the wheat orthologues genes predicted by the TAGCv1 and IWGSC css assembly, nulli-tetrasomic (NT) lines of the wheat cultivar Chinese Spring were used (Fig. 1b;Sears 1954; see M&M for details) and homeolog-specific primers (Table S1) designed for PCR analysis of genomic DNA from the NT lines. The absence of an amplicon in the respective NT line confirms that the target gene is physically located on that nullisomic chromosome pair.
To further corroborate the orthology between the identified loci we analysed the synteny of the genomic regions between bread wheat, Brachypodium, rice and sorghum (for gene IDs see Supplementary Tables S1 and S5). The fragmented nature of the bread wheat TGACv1 assembly (Clavijo et al. 2017) may have reduced the resolution of the analysis, however, wheat showed a high degree of synteny with the three diploid genomes included in the analysis and the position of several purine catabolic genes and their neighbouring genes was highly conserved (Fig. 2). The only exception was XDH, for which we could identify only one syntenic gene (Bradi1g15910, depicted in blue) among all analysed genomes. This gene is located upstream of XDH in sorghum, downstream of XDH in rice and Brachypodium, and on chromosome group 1 in wheat. Interestingly, we identified two Brachypodium genes (Bradi1g15820 and Bra-di1g15826, depicted in brown and grey, respectively) with orthologous sequences on wheat chromosome group 6, supporting the evidence that a duplication of the genomic region harbouring XDH may have occurred during wheat genome evolution. Analysis of the predicted open reading frames (ORFs) of the TaXDH genes revealed five premature stop codons in TaXDH2-6AS and several mutations in TaXDH2-6BS, both likely to result in non-functional proteins. In contrast, the TaXDH2-6DS ORF appeared to translate into a functional protein (data not shown). This was supported by in-silico analysis of publicly available Chinese Spring RNAseq transcriptomics data which revealed that TaXDH2-6DS was the only homeolog expressed (Fig. S3a). Synteny analysis further revealed the presence of two copies of UAH in the Brachypodium genome of which only one gene (BdUAH1) showed a syntenic relationship with the other genomes.

The effects of low N and drought on plant performance
RAC875 and Mace plants were initially grown under wellwatered (WW) conditions divided in two subsets supplied with either high N (HN) or low N (LN). During the course of the experiment, a subset of plants grown under HN were subjected to drought (D) at tillering stage and during grain filling, (Fig. S1a). Therefore, the experiment comprised of three treatments: control (high N, well-watered; HN-WW), low N (low N, well-watered; LN-WW) and drought (high N, drought; HN-D). Analysis of plants at maturity revealed that the low N treatment reduced above ground biomass by 20.8% and 23.3% in RAC875 and Mace, respectively, whilst grain yield was reduced by 24.4% and 25.2% in RAC875 and Mace as compared to the controls (HN-WW), respectively (Fig. 3). When plants were grown under drought conditions, above-ground biomass was reduced by approximately 15% in both genotypes as compared to their respective controls. Under drought conditions, Mace plants were also higher yielding (grain yield) than RAC875 plants (p < 0.02 by Student's t-test); this corresponded to a 16.0% and 26.6% yield loss for Mace and RAC875, respectively, relative to control conditions (Fig. 3).

Allantoin accumulation under low N and drought stress
To assess whether N and water stress alter allantoin levels in the two wheat genotypes, allantoin concentration was quantified in the youngest fully emerged leaf (YFEL) during vegetative growth and in flag leaf, stem, spikelet and developing grain samples during reproductive growth in plants grown under HN-WW, LN-WW and HN-D (Fig. 4). Nitrogen deficiency significantly reduced allantoin concentration at different time points across different tissues and in both genotypes when compared to control conditions (Fig. 4a). In both genotypes grown under LN-WW conditions, allantoin levels were generally below 50 nmol g −1 DW in YFEL, flag leaves and stems, but much higher (50 to 100 nmol g −1 DW) in spikelets and developing grains. The highest reduction of allantoin in both genotypes was measured in the stem under LN-WW as compared to the HN-WW control, with a 22-fold and 10-fold reduction in RAC875 and Mace, respectively (Fig. 4a). Interestingly, under control conditions (HN-WW) Mace flag leaves significantly accumulated much higher levels of allantoin at 19 DAA (305 nmol g −1 DW) as compared to RAC875 (34 nmol g −1 DW), whilst a significant but relatively smaller difference was recorded under low N at 17 DAA (36 and 12 nmol g −1 DW for Mace and RAC875, respectively) (Fig. 4a). Similarly, allantoin concentration in Mace developing grains was significantly higher compared to RAC875 at 22DAA under both control (HN-WW) and low N (LN-WW) conditions.
In    allantoin accumulation was not significant (Fig. 4b). During the tillering stage, under severe drought stress (day 8; 6.3% SWC) allantoin significantly accumulated 3.6-and 2.5-fold in the YFEL of RAC875 and Mace, respectively. The higher magnitude of allantoin accumulation under drought in RAC875 were due to significantly higher levels of allantoin in the YFEL of RAC875 under drought compared to Mace (Fig. 4b). Similarly, during grain filling, allantoin significantly accumulated in Mace flag leaves at day 5 and 8 of the drought treatment (8.6% and 6.6% SWC, respectively), whilst it significantly accumulated in RAC875 flag leaves only at day 8 (Fig. 4b). In spikelet and developing grains, allantoin significantly increased only at day 8 in both genotypes, when drought stress was most severe (Fig. 4b). Interestingly, flag leaves of RAC875 progressively accumulated allantoin during the course of the drought treatment, increasing from 136 nmol g −1 DW at day 5 to 1197 nmol g −1 DW at day 8, representing an almost 30-fold increase relative to HN-WW. On the other hand, Mace accumulated 1614 nmol of allantoin per gram DW after only 5 days of drought treatment and levels remained constant until day 8. However, the magnitude of allantoin accumulation under HN-D in Mace was only 2.5fold as compared to HN-WW at day 8 (Fig. 4b). The large difference in the magnitude of allantoin accumulation between RAC875 and Mace flag leaf can be associated with the higher allantoin levels in Mace already present under control conditions at day 5 (19DAA).

Transcription of purine catabolic genes under low N and drought stress
To assess stress responsiveness of wheat purine catabolic genes, transcript abundance was quantified in all of the RAC875 samples collected during the course of the experiment (Fig. 5) and on YFEL (Z22) Mace samples (Fig. S4).
The data for each gene and treatment is presented in Fig. 5 as log2 ratio of the calibrated normalised relative quantities (CNRQ) (Hellemans et al. 2007) between treatment (LN-WW or HN-D) and control conditions (HN-WW). For this analysis qRT-PCR primers were designed to amplify all three homeologs of each purine catabolic gene. The only exception was the design of primers to specifically amplify TaXDH2-6DS, which is the only TaXDH2 homeolog that is expressed (Fig. S3a). This likely explains why the abundance of TaXDH2-6DS transcripts in RAC875 was approximately 20 times lower than in TaXDH1 (Fig. S3b). Analysis of the transcriptional regulation of purine catabolism under LN-WW revealed that purine catabolic genes were up-regulated in RAC875 in most tissues and developmental stages analysed (Fig. 5). In particular, the genes TaALN and TaAAH, putatively coding for the key enzymes involved in allantoin and allantoate degradation, respectively, were highly responsive to the LN-WW treatment. The differential expression of purine catabolic genes in RAC875 flag leaves, stem and spikelets at 17DAA corresponded to a significant reduction of allantoin measured in the same tissues as presented in Fig. 4a (time point highlighted in green in Fig. 5). At 19DAA, all purine catabolic genes were significantly up-regulated in flag leaves and, with a lower magnitude, also in stems and spikelets (Fig. 5).
In contrast to the transcriptional response under low N conditions, RAC875 droughted plants displayed a differential regulation of specific sets of purine catabolic genes (Fig. 5). Generally, TaALN was down-regulated in all analysed tissues of droughted plants (HN-D) relative to plants grown under HN-WW, except in the flag leaves and spikelets where no significant changes in transcription of TaALN were detected. A decrease in transcript abundance of TaALN in the YFEL occurred already after exposure to mild drought conditions (day 3), suggesting that this gene is particularly drought responsive in this tissue (Fig. 5). TaALN down-regulation at day 6 in YFEL corresponded with accumulation of allantoin as presented in Fig. 4b (time point highlighted in green in Fig. 5). The TaXDH1 and TaUOX genes, coding for the enzymes that synthesise allantoin, were up-regulated under HN-D across the analysed tissues, except in developing grain (Fig. 5). Interestingly, TaXDH2 showed the opposite transcriptional regulation to its paralog, TaXDH1. In fact, TaXDH2 was down-regulated under HN-D in the YFEL, the flag leaf and the stem (Fig. 5). TaAAH and TaU-GAH presented different drought responses depending on the tissue analysed (e.g. were down-regulated in stem and up-regulated in the spikelet).
Analysis of the transcriptional profile of purine catabolic genes in the YFEL of Mace are largely in agreement with those in RAC875. N deficiency increased transcription of TaXDH1, TaALN and TaAAH in Mace YFEL; whilst TaALN was down-regulated andTaXDH1 up-regulated under drought (Fig. S4).

Metabolite signature of spikelet and developing grain under drought and low N conditions
To assess the involvement of allantoin in N and C metabolism in wheat under drought and N limitation during grain filling, the metabolic signatures from the spikelet and the developing grain of RAC875 and Mace were assessed and visualised by hierarchical clustering (Figs. 6, 7) and PCA (Supplementary Fig. S5 and S6).
This analysis revealed that allantoin accumulation under drought conditions (HN-D) at day 8 showed the same pattern with additional metabolites forming a distinct cluster in both the spikelet (D.S.ii; Fig. 6a) and the developing grain (D.Dg.ii; Fig. 6b) of RAC875 and Mace. In both tissues, these clusters contained mostly amino acids including the well-known drought responsive amino acids proline and 4-amino-butanoic acid (GABA) and, in the developing grain, the polyamine putrescine. Generally, metabolites belonging to these clusters appeared to accumulate with higher magnitude in RAC875 compared to Mace. A genotype-specific response was also apparent in two additional clusters in developing grain with metabolites specifically accumulated (D.Dg.i) and reduced (D.Dg.iv) in RAC875 under drought, respectively (Fig. 6b).
Clusters containing metabolites with an apparent reduction under drought in both genotypes were also identified, in both the spikelet (D.S.i; Fig. 6a) and the developing grain (D.Dg.iii; Fig. 6b).These clusters contained sucrose, myo-inositol and quinic acid. In addition, in spikelets, this cluster contained other organic acids (e.g., isocitric acid, 3-phosphate-glyceric acid) as well as the tri-saccharides raffinose and 1-kestose. The cluster in developing grain further contained glucose-6-phosphate and fructose-6-phosphate, as well as tryptophan and glutamate.
Analysis of the metabolite signature under low N condition revealed contrasting patterns compared to the drought treatment. Generally, the drought treatment at day 8 clearly separated metabolites between the two treatments (HN-WW and HN-D; Fig. S5). However, the low N treatment appeared to influence the metabolite levels to a lower extent as supported by PCA analysis (Fig. S6). In particular, in spikelet the major dividing component of the data points was PC2 that separated according to the genotype (Fig. S6a). On the other hand, PCA of developing grain revealed no major differences between genotypes and treatment, although a weak separation according to the time point, especially for 22 DAA, was detected along PC1 (Fig. S6b). Hierarchical clustering of the spikelet data under different N conditions placed allantoin into a representative cluster (Ln.S.i) which contained a large number of metabolites, mostly carbon-rich compounds (Fig. 7a). Specifically, this allantoin cluster contained eleven organic acids (including citric acid, malic acid, succinic acid, glyceric acid and shikimic acid) and ten sugars (including fructose, sucrose, myo-inositol, raffinose, 1-kestose). These metabolites appeared to accumulate to higher levels in spikelet of Mace as compared to RAC875, which showed a small or no treatment effect (Fig. 7a).
Similarly, in the developing grain allantoin grouped in a representative cluster (Ln.Dg.iii) which was composed of allantoin and three carbon-rich metabolites, including raffinose, galactinol and 1-kestose. This is in contrast to the metabolite profile in the drought treatment, where allantoin grouped with amino acids. These amino acids, under low N conditions, are scattered between two representative clusters in spikelets (Ln.S.ii and Ln.S.iii; Fig. 7a) or, in developing grain, largely not assigned to any representative cluster (Fig. 7b). Although, in the developing grain, this large area clustered according to the time points (DAA).
However, the accumulation of allantoin, galactinol and raffinose in developing grain under N starvation appears to be developmentally rather than representing a specific response to N starvation as similar metabolic changes occurred also under control conditions at 22DAA (Fig. 8b). In agreement with the quantitative data on allantoin shown in Fig. 4a, Mace appeared to accumulate higher levels of allantoin and the other metabolites in cluster Ln.Dg.iii suggesting genotypic differences within wheat.
In summary, the metabolite profiling showed that, under drought, allantoin clustered predominantly with N-rich compounds (amino acids), whereas under N-starvation allantoin clustered with C-rich compounds (sugars and organic acids) and therefore allantoin represented the only N-rich compound.

Allantoin accumulation in the grain
To assess the contribution of allantoin to the overall N pools in the grain under control (HN-WW) and stress conditions (HN-D and LN-WW), the plants minus the sampled tiller ( Supplementary Fig. S1) were grown until maturity, corresponding overall to 16-18 plants per treatment. Total N, allantoin and free amino acid concentrations were then determined in grain harvested from the tallest remaining tiller (Fig. 8). Analysis of grain N% highlighted significant differences between treatments (two-way ANOVA, p < 0.0001), in particular, it showed a 20% reduction in grain N% of plants grown under LN-WW with no evident genotypic differences between RAC875 and Mace (Fig. 8a). Allantoin concentration was significantly responsive to the applied treatments and genotypes (two-way ANOVA, Genotype × Treatment, p = 0.0028) (Fig. 8b). Particularly, allantoin was reduced under LN-WW by 44% and 49% in RAC875 and Mace plants, respectively. In contrast, under HN-D, allantoin in RAC875 grains increased by 39% relatively to HN-WW. Interestingly, the largest difference in allantoin concentration between genotypes was observed under HN-WW, as previously observed for the flag leaf (Fig. 4). In fact, Mace grains accumulated 66% more allantoin than RAC875, whilst there were no significant differences between the genotypes under LN-WW and HN-D, as mentioned above (Fig. 8b).
Analysis of the free amino acids content of grain from plants grown under HN-WW revealed that allantoin accumulated to levels comparable to other amino acids, such as glutamine, aspartate, asparagine and arginine (Fig. 8c). When considering that each allantoin molecule contains four N atoms (1:1 nitrogen to carbon ratio, N:C), the overall N stored in allantoin in grain was even larger than the N present in glutamate and aspartate, which were the most abundant free amino acids identified but have only one N atom (1:4 N:C) (Fig. 8c). However, the N-rich amino acid arginine (2:3 N:C) retained approximately double the N content of allantoin in both RAC875 and Mace grains. The analysis also showed significant genotypic differences in certain amino acids between RAC875 and Mace. In particular, Mace had 70% and 31% higher concentration of arginine and alanine than RAC875, whilst RAC875 had 179%, 43% and 35% higher concentration of tryptophan, aspartate and serine than Mace, respectively (Fig. 8c).

Discussion
The aim of this study was to assess the role of the purine intermediate allantoin and the purine catabolic genes under water and nutrient stress in bread wheat. For this, we selected two Australian genotypes, specifically RAC875, a breeding line that has been characterised as tolerant to drought (Izanloo et al. 2008;Bennett et al. 2012;Bonneau et al. 2013) and the locally adapted variety Mace, which is suggested to have a higher N-use efficiency based on preliminary studies (Mahjourimajd et al. 2016).

The wheat purine catabolic genes show high synteny but also differences compared with other grasses
The genes of the purine catabolic pathway, identified in several grass genomes including hexaploid bread wheat, showed a high degree of synteny among three grass genomes, suggesting that the identified wheat loci are true gene orthologues (Fig. 2). However, the poor synteny displayed by XDH and its adjacent genes even in diploid genomes (Brachypodium, rice and sorghum) suggests that XDH is located in an unstable genomic region prone to rearrangements. The majority of the grass genomes had only one copy of the purine catabolic genes (Table S5), corresponding to three homeologs in wheat (Table S1), with the only exceptions of the fore-mentioned XDH, and UAH, for which there was no synteny for the Brachypodium orthologs BdUAH2. For XDH we have identified a second copy on chromosome 6 of which, based on in-silico expression analysis (Fig. S3a), only the homeolog on chromosome TaXDH2-6DS is a functional gene explaining its low expression level compared with TaXDH1 ( Fig. S3b; Fig. 5). Interestingly, TaXDH2 shows an opposite transcriptional response to drought (reduced transcript level) compared with TaXDH1 suggesting different roles and/or functional divergence. A differential response to drought and other treatments (salt, cold, ABA) has also been shown in Arabidopsis, which carries a tandem duplication of AtXDH1 and AtXDH2 on a single chromosome (Hesberg et al. 2004). It will therefore be interesting to investigate the role and enzymatic function of TaXDH2 in more detail in relation to drought and other stresses.

Wheat highly accumulates allantoin under drought stress
There was a significant accumulation of allantoin in all analysed tissues and genotypes of wheat plants exposed to drought (Fig. 4b). This was accompanied by a significant reduction in transcript levels of TaALN, i.e., the first step in allantoin degradation (Fig. 5).
Allantoin was previously reported to accumulate in wheat and rice in response to abiotic stresses, including drought (Bowne et al. 2011;Degenkolbe et al. 2013;Casartelli et al. 2018), however, these studies did not investigate the regulation of the purine catabolic genes under those conditions.
Analysis of the expression of purine catabolic genes in Arabidopsis (Irani and Todd 2016) recently showed that leaf accumulation of allantoin upon drought stress was associated with transcriptional up-regulation of the purine catabolic genes leading to allantoin synthesis (AtXDH1, AtXDH2, AtUOX, AtAS) whilst expression of the allantoin-degrading gene AtALN was only marginally increased. Similarly, Yesbergenova et al. (2005) showed that, in tomato, LeXDH1 and LeXDH2 were up-regulated under drought in leaf and root tissues. In Arabidopsis, the allantoin pathway was also implicated with salt stress, showing reduced AtALN and AtAAH expression and increased AtXDH1 expression (Irani and Todd 2016;Hesberg et al. 2004), as well as an accumulation of allantoin accompanied by AtUOX and AtALNS up-regulation and AtALN down-regulatoin (Lescano et al. 2016).
Our data agree with the above studies confirming that allantoin significantly accumulates under drought (Fig. 4b) and that this is paralleled by the up-regulation of genes putatively encoding enzymes for allantoin synthesis (TaXDH1 and TaUOX) and/or down-regulation of TaALN (Fig. 5). Although we have not quantified xanthine, the gene expression data suggest that under drought, an increased amount of xanthine is likely to feed into this pathway which, in combination with a reduced allantoin degradation, results in the accumulation of allantoin under drought which was observed in our study.
In contrast, studies in common bean (Phaseolus vulgaris) showed up-regulation of both allantoin-synthetising genes (PvUOX) and allantoin-degrading genes (PvALN and PvAAH) in leaves under drought and this corresponded to increased levels of allantoin and allantoate (Alamillo et al. 2010;Coleto et al. 2014). Interestingly, Coleto et al. (2014) reported that these ureides were more concentrated in the tissues of drought-sensitive genotypes. This is in contrast to our data and other studies showing that allantoin specifically accumulates in drought tolerant genotypes of wheat, rice and resurrection plants (Bowne et al. 2011;Oliver et al. 2011;Degenkolbe et al. 2013;Yobi et al. 2013;Casartelli et al.  . This suggests that purine catabolism is regulated differently under drought stress in ureidic legumes. Evidence of the functional role of purine catabolism under abiotic stress were revealed using reverse genetics approaches. Arabidopsis XDH mutants (xdh) showed that disrupting the first step of purine catabolism caused hypersensitivity to water stress (Watanabe et al. 2010) and impaired recovery from extended dark exposure (Brychkova et al. 2008), which in both cases led to excessive reactive oxygen species (ROS) accumulation compared to WT plants. In contrast, the constitutive accumulation of the intermediate allantoin in Arabidopsis ALN knock-out mutants (aln) led to enhanced performance under dehydration, drought and salt stress, which also corresponded to reduced ROS levels (Watanabe et al. 2014;Irani and Todd 2016). These studies further showed that increasing allantoin concentration, either constitutively (aln mutants) or by exogenous supplementation, stimulated ABA and jasmonic acid (JA) metabolism, key components of abiotic stress responses. However, a previous study suggested that allantoin does not possess any in-vitro antioxidant activity (Wang et al. 2012), prompting to speculate that allantoin and purine catabolism play a role in stress sensing and regulation rather than a direct ROS scavenging function.
Metabolomic profiling of the spikelet (defined here as vegetative part of the spike) and the developing grain after 8 days of drought stress showed that allantoin clustered with a set of highly drought-responsive metabolites (Fig. 6a, b). These clusters were composed mainly of amino acids, with alanine, valine, leucine, and proline common to the allantoin clusters in spike samples (D.S.ii) and developing grain (D. Dg.ii). Proline is amongst the best characterized droughtrelated amino acids and has been shown to have a range of functions under stress, e.g., osmolyte, regulator of redox potential, molecular chaperone, ROS scavenger and signalling molecule (Yoshiba et al. 1995;Hare and Cress 1997;Verbruggen and Hermans 2008;Szabados and Savoure 2010;Mohanty and Matysik 2001;Khedr et al. 2003). The specific accumulation of allantoin under drought in a wide range of different plant species and the suggested role in stimulating the ABA and JA pathway (see above) justifies further investigations into the importance of allantoin in the mitigation or tolerance to drought.

Preventing N losses under drought and allantoin as a source of N under nutrient deprivation
Under drought, RNA and DNA as well as protein degradation and nutrient remobilisation caused by premature leaf senescence (Munné-Bosch and Alegre 2004) and increased photorespiration (Mattsson et al. 1997;Wingler et al. 1999;Kumagai et al. 2011b) are sources of high tissue NH 4 + and related emission of volatile ammonia (Mattsson and Schjoerring 2002). In principle, free NH 4 + can be recycled by the GS-GOGAT cycle (Fig. S2), however, this pathway is negatively responsive to drought (Nagy et al. 2013) and might therefore not be sufficiently effective in capturing NH 4 + . Increased ammonia emission due to inhibition of GS with MSO has been demonstrated (Mattsson and Schjoerring 1996) and has been directly linked with enhanced photorespiration under high light and high O 2 stress (Kumagai et al. 2011b), as well as with senescence (Parton et al. 1988) and heat stress (Mattson et al. 1997).
Ammonia emission can be considered an efficient, though very wasteful mechanism, to prevent the build-up of high tissue concentrations, which is toxic to plants (for a review see Britto and Kronzucker 2002). The accumulation of allantoin (e.g. 30-fold accumulation in RAC875 flag leaves, Fig. 4b) would therefore be beneficial to plants because it prevents accumulation of NH 4 + to toxic levels and also, to some extent, to retain organic N in the plant that could be lost to the atmosphere in the form of volatile ammonia. Although allantoin represents just a fraction of the total organic N in a wheat plant, allantoin accumulation under drought appears to occur simultaneously with the accumulation of several other small metabolites containing N, notably amino acids such as proline (Fig. 6). This response seems to be shared by other plant species (see Introduction for references), in fact we previously reported high accumulation of allantoin and several amino acids when rice plants were subjected to drought stress . Therefore, this could underlie a global strategy that plants adopt to improve N balance under stress, when GS-GOGAT enzymatic activity is reduced. Differences in the regulation of GS (Singh and Gosh 2013) and ammonia emission (Kumagai et al. 2011) between drought tolerant and intolerant genotypes have been shown in rice and genetic diversity and drought-tolerant specific GS responses have also been shown in tomato (Sanchez-Rodriguez et al. 2011) and wheat (Nagy et al. 2013).We suggest that the selection of high allantoin and maintenance of GS activity (reduced N losses) under drought stress could be a specific target for breeding (Fig. 9).
The importance of allantoin and purine catabolism in N metabolism have long been established in ureidic legumes, which employ allantoin and allantoate as major form for long-distance transport of N (Schubert 1986;Sinclair and Serraj 1995). However, recently its important role in maintaning N homeostasis in non-ureidic plant species is becoming evident. Soltabayeva et al. (2018) reported that Arabidopsis Atxdh1, Ataln and Ataah mutants displayed an early senescence phenotype when grown under low NO 3 − conditions and that the activity of nitrate reductase (NR) was increased in leaves of Atxdh1 mutants suggesting a higher demand for nitrate than wild type plants. Similarly, Nakagawa et al. (2007) reported an early onset of senescence in RNAi-induced xdh mutants which also displayed reduced chlorophyll content and increased cytosolic GS1 protein, which is known to be induced during senescence (Bernard et al. 2008).
In our study, the quantification of allantoin in the wheat plants grown under N-limited conditions revealed  Fig. 9 Schematic model for the dual-role of allantoin under stress in relation to N homeostasis. Under low N conditions, increased expression of ureide-degrading genes ALN and AAH indicate that NH 4 + is liberated from purines providing an internal source of organic N that can be re-assimilated by the GS-GOGAT cycle. As a result, allantoin concentration in plant tissues is reduced. In contrast, during drought stress, allantoin accumulates due to down-regulation of the ALN gene preventing the liberation of NH 4 + that could lead to cellular toxicity and part of it might be lost as volatile ammonia due to droughtimpaired GS-GOGAT activity a significant reduction of allantoin in all tissues analysed (Fig. 4a) which is in contrast to the observed accumulation of allantoin under drought (Fig. 4b) as discussed above. A decrease of allantoin under N deprivation was also reported in other cereal species, specifically in maize and rice showing reduced allantoin in leaves and roots, respectively (Amiour et al. 2012;Coneva et al. 2014). In addition, high induction of OsALN expression under low N conditions in rice was recently shown (Lee et al. 2018), this was previously also observed in Arabidopsis and the leguminous tree Robinia pseudoacacia (Yang and Han 2004). In agreement with this, the reduction of allantoin under low N reported in this study (Fig. 4a) was accompanied by increased transcription of purine catabolic genes, and this was particularly evident in genes encoding for the ureide degrading enzymes TaALN and TaAAH (Fig. 5). Downstream compounds of the purine catabolic pathway were recently analysed in more detail under N deprivation using the same wheat genotypes, RAC875 and Mace  showing, in addition to reduced levels of allantoin, a significant reduction of allantoate in leaves and an accumulation of glyoxylic acid, the end product of purine degradation. The assumption that the N (and C) remobilised from allantoin supports plant growth is supported by studies showing that Arabidopsis and rice seedlings could grow with ureides as a sole N source (Desimone et al. 2002;Brychkova et al. 2008;Lee et al. 2018), although growth was delayed in comparison to plants supplied with inorganic N. This is in contrast to wheat seedlings that, resupplied with allantoin and xanthine as a sole source of N after NO 3 − starvation, grew and photosynthesised as well as those re-supplied with NO 3 − . Taken together, these findings suggest that the enhanced activity of the purine catabolic pathway provides an internal source of organic N used to maintain homeostasis under low N conditions whereas the accumulation of allantoin under drought releases pressure from the GS-GOGAT cycle thereby preventing accumulation of toxic levels of NH 4 + and possibly N losses due to volatilization (Fig. 9).

Allantoin is a relevant component of grain nitrogen in wheat
In wheat, N translocated to the grain during plant senescence is an important determinant of grain quality and there is currently little information on the contribution of allantoin and/or other purine catabolites to grain N in non-ureide plants. The data reported here show progressive accumulation of allantoin during grain filling (Fig. 4) up to more than 1400 nmol g −1 DW at maturity in Mace (Fig. 8b). This was confirmed by metabolomics analyses of developing grains showing that allantoin levels correlated with other metabolites specifically accumulating at 22DAA, in contrast to the majority of metabolites that showed an opposite trend (Fig. 7).
Although we calculated that allantoin accounted for only about 0.3% of the total N measured in wheat grains, this represents a significant portion of the soluble N pool, comparable to the most concentrated and N-rich amino acids, such as asparagine and arginine (Fig. 8c). Previous studies have shown that allantoin stored in wheat grains is quickly utilised, from as early as 1 day after germination (Montalbini 1992), suggesting that it can be used as a readily available N substrate.
The data reported here further show that genotypic differences exist among bread wheat genotypes, in fact, Mace accumulated 66% more grain allantoin than RAC875, despite similar % grain N (Fig. 8a, b). This difference was also particularely evident in flag leaves of Mace, which had up to 10-fold higher allantoin content compared to RAC875 under both, control and low N conditions (Fig. 4a). Mace is widely grown in Australia because of its superior yield under drought and nitrogen-use efficiency (Mahjourimajd et al. 2016). High allantoin concentrations in the leaves and grain might indeed be a positive contributing factor to high NUE in Mace. This is in agreement with findings in rice where differences in allantoin levels in genotypes with contrasting drought tolerance were already apparent under control conditions . In addition to greater accumulation of allantoin, mature grains of Mace also accumulated 70% more arginine than RAC875 (Fig. 8c). Arginine is the proteinogenic amino acid with the highest N:C ratio (3:2), known to represent a major N form for storage and transport.
Our data thus indicate that the two analysed wheat genotypes prioritize different N compounds for transport and grain loading. Previous studies in Mace suggest that this cultivar has high N use efficiency (Mahjourimajd et al. 2016) and this study and Kastury et al. (2018) additionally demonstrated that Mace is higher yielding under drought than RAC875. The possibility that the superior performance of Mace may be related to the preferential use of metabolites with high N:C ratio, which are more energy effective forms for transporting and storing N, is an intriguing hypothesis that requires further validation. In addition, allantoin and arginine may also play a regulatory role in plant growth and development as they participate in ABA, JA, and nitric oxide (NO) signalling, respectively (Watanabe et al. 2014;Winter et al. 2015;Takagi et al. 2016).

Summary
In summary, our data are in support of the converging evidence that allantoin has a relevant role in non-ureide plants under water stress and in N homeostasis. In addition to the proposed role of allantoin in drought-induced ABA responses, accumulation of this N-rich molecule under drought would contribute to optimise N (and C) balance in the plant by preventing toxic build-up of NH 4 + and possible N losses through volatile ammonia due to a reduced activity of the GS-GOGAT cycle (Fig. 9). Under low N conditions, induced expression of allantoin catabolic genes and decreased allantoin levels indicate that allantoin serves as an internal organic N source (Fig. 9). In wheat, allantoin is also a relevant component of grain N and comparable to that of other major amino acids.

Author Query Form
Please ensure you fill out your response to the queries raised below and return this form along with your corrections

Dear Author
During the process of typesetting your article, the following queries have arisen. Please check your typeset proof carefully against the queries listed below and mark the necessary changes either directly on the proof/online grid or in the 'Author's response' area provided below