Metabolism and biological functions of the nitrogen-rich compound guanidine have long been neglected. The discovery of four classes of guanidine-sensing riboswitches and two pathways for guanidine degradation in bacteria hint at hitherto overlooked but widespread sources of unconjugated guanidine in nature. So far, only three enzymes from a narrow range of bacteria and fungi have been shown to produce guanidine, with the ethylene-forming enzyme as the most prominent example. Here, we show that a related class of Fe2+- and 2-oxoglutarate-dependent dioxygenases (2-ODD-C23) highly conserved among plants and algae catalyze the hydroxylation of homoarginine at the C6-position. Spontaneous decay of 6-hydroxyhomoarginine yields guanidine and 2-aminoadipate-6-semialdehyde. The latter can be reduced to pipecolate by pyrroline-5-carboxylate reductase but is more likely to be oxidized to aminoadipate by aldehyde dehydrogenase ALDH7B in vivo. Arabidopsis has three 2-ODD-C23 isoforms, among which Din11 is unusual because it also accepted arginine as substrate, which was not the case for the other 2-ODD-C23 isoforms from Arabidopsis or other plants. Guanidine contents were typically between 10 and 20 nmol*(g fresh weight)-1 in Arabidopsis but increased to 100 or 300 nmol*(g fresh weight)-1 after homoarginine feeding or treatment with Din11-inducing methyljasmonate, respectively. In 2-ODD-C23 triple mutants, the guanidine content was strongly reduced, whereas it increased in overexpression plants. We discuss the implications of the finding of widespread guanidine-producing enzymes in photosynthetic eukaryotes as a so far underestimated branch of the bio-geochemical nitrogen cycle and propose possible functions of natural guanidine production.
This fundamental study advances our understanding of nitrogen metabolism by identifying a new type of guanidine forming enzyme in eukaryotes. The evidence supporting this pioneering research is convincing, with rigorous biochemical, cellular, and in vivo experiments.
Nitrogen availability is a limiting factor for productivity in many ecosystems and plants are the major contributors to assimilation of nitrogen into organic matter from inorganic precursors. The largest proportion of nitrogen is used in primary metabolism for the synthesis of proteins and nucleic acids. Additionally, many secondary metabolites that serve for internal signaling, adaptation or interactions with other organisms are nitrogen-containing molecules. Guanidine, which exclusively occurs as the protonated guanidinium cation under physiological conditions, has been discovered as a degradation product of guanine and arginine under chemically harsh conditions (Bénech & Kutscher, 1901; Strecker, 1861). Thirty-two years after its discovery in 1861, guanidine was detected in plants (Schulze, 1893). However, guanidine has no known function in plants, and it is unclear to which extent and how plants can synthesize and metabolize this nitrogen-rich compound (Marsden et al., 2015). In contrast, two enzymatic pathways for guanidine catabolism were recently discovered in bacteria and many bacteria express guanidine importers or exporters in response to externally supplied or internally produced guanidine (Funck et al., 2022; Nelson et al., 2017; Schneider et al., 2020; Sinn et al., 2021; Wang et al., 2021). Bacterial genes involved in guanidine metabolism are often regulated by members of four independent guanidine riboswitch classes, which may indicate an ancient evolutionary origin of guanidine metabolism (Lenkeit et al., 2020; Nelson et al., 2017; Salvail et al., 2020; Sherlock & Breaker, 2017; Sherlock et al., 2017).
Only two enzymes are known to release guanidine from arginine in side reactions. The arginine-4,5-desaturase NapI induces the decay of 5-hydroxyarginine to glutamate-5-semialdehyde (GSA) and guanidine when the initial hydroxylation occurs at the C5 atom of arginine instead of the C4 atom, where it would induce the desired desaturation of the C4-C5 bond (Dunham et al., 2018). The Ethylene Forming Enzyme (EFE) from Pseudomonas savastanoi converts 2-oxoglutarate (2-OG) into ethylene and this activity depends on structural changes induced by binding of arginine to the active site (Fukuda et al., 1992; Martinez et al., 2017). In a second reaction with roughly 50% turnover rate, EFE also catalyzes the hydroxylation of arginine at the C5 atom. Both enzymes belong to the class of Fe2+ and 2-OG-dependent dioxygenases (2-ODD), which use molecular oxygen to catalyze a broad range of reactions involving an FeIV-oxo-intermediate formed after the initial conversion of 2-OG to succinate (Farrow & Facchini, 2014; Hausinger, 2015; Herr & Hausinger, 2018). A third guanidine-producing enzyme, the guanylurea hydrolase GuuH from Pseudomonas mendocina strain GU, is from a different protein family and may have evolved recently to utilize guanylurea generated by metformin degradation (Tassoulas et al., 2021).
In plants, the family of 2-ODDs has diversified immensely, and it has been divided into three subfamilies according to sequence similarity and the types of reactions that the respective enzymes catalyze (Kawai et al., 2014; Prescott & John, 1996). Subfamily A comprises RNA demethylases and related enzymes, subfamily B comprises prolyl-hydroxylases that function in posttranslational protein modification, and the largest subfamily C comprises enzymes that act in the synthesis or modification of plant hormones, pigments and other secondary metabolites. Besides the 2-ODDs with known enzymatic function, subfamily C also contains many proteins of unknown function. Among these, Arabidopsis Din11 (Dark-induced11; At3g49620) has been identified because its expression was induced by prolonged dark treatment, which leads to C-starvation and the induction of autophagy (Fujiki et al., 2001; Magen et al., 2022). Din11 orthologs are encoded in the genomes of most analyzed plants and green algae, and Arabidopsis as well as many other Brassicacea express three paralogs of this protein: Din11, At3g49630 and At3g50210, among which At3g50210 seems to be the most ancestral form (Supplementary Fig. S1). In rice, the single Din11 ortholog has been named 2-ODD33 and was reported to have weak melatonin-2-hydroxylase activity (Byeon & Back, 2015). Because the numbering of the 2-ODDs in rice does not reflect their phylogenetic relationship and the phylogenetic study by Kawai et al (2014) used DOX as rather uncommon acronym for the 2-ODDs, we propose a hybrid nomenclature with 2-ODD followed by a denominator for the phylogenetic clade as defined in Kawai et al (2014). According to this nomenclature, we propose 2-ODD-C23.1 as systematic name for At3g50210, 2-ODD-C23.2 for Din11 and 2-ODD-C23.3 for At3g49630.
An important N-containing class of secondary plant metabolites are non-proteinogenic amino acids, which may have metabolic, storage, defense, or signaling functions (Bell, 2003). Homoarginine, the C6-analog of arginine, is accumulated to high concentrations in seeds of several Lathyrus species, where it is assumed to be a nitrogen storage form that is not easily accessed by bacteria or insects (Bell, 1962; Bell, 1980; Lambein et al., 1992; Rao et al., 1963). Trace amounts of homoarginine were also detected in other plants, and in humans it is thought to interfere with nitric oxide production and signaling (Koch et al., 2022; Lambein et al., 2019; Sulser & Sager, 1976). In man and animals, homoarginine is produced from lysine and arginine in a side reaction of glycine amidinotransferase (Choe et al., 2013). A protein with similar activity has been isolated from Lathyrus sativus, but the molecular identity of this protein remains unknown and we did not detect a homolog of human glycine amidinotransferase in any annotated plant genome (Srivenugopal & Adiga, 1980). No other homoarginine-metabolizing enzymes have so far been reported from plants.
Here we describe the identification of plant 2-ODD-C23 isoforms as homoarginine-6-hydroxylases, which initiate the decay of 6-hydroxyhomoarginine to 2-aminoadipate-6-semialdehyde (AASA) and guanidine. Within the 2-ODD-C23 clade, we found Din11/2-ODD23.2 to be special because it accepted also arginine as substrate to produce 5-hydroxy-arginine. This finding identifies 2-ODD-C23s as the first eukaryotic enzymes that release free guanidine from conjugated precursors, an activity previously only known as side reaction of rare bacterial enzymes. AASA accumulation is toxic because it inactivates the enzyme cofactor pyridoxal-5-phosphate (PLP) by adduct formation, therefore, AASA is detoxified by oxidation to aminoadipate by ALDH7B4 (AASA-DH, At1g54100; (Končitíková et al., 2015; van Karnebeek et al., 2016)). In solution, AASA is in equilibrium with the cyclic 81-piperideine-6-carboxylate (P6C), which can be reduced to pipecolate by pyrroline-5-carboxylate reductase (P5CR) (Fujii et al., 2002; Struys et al., 2014). Our comparison of the activities of P5CR and ALDH7B4 indicates that under physiological conditions, the oxidation of AASA is the prevalent metabolic route in plants.
Expression of plant EFE-like proteins causes guanidine production in E. coli
To determine if plant 2-ODD-C23 enzymes, which share the highest degree of similarity with EFE among the plant ODDs, also catalyze the production of guanidine, selected proteins were overexpressed in E. coli. We focused on the three 2-ODD-C23 isoforms from Arabidopsis (Din11/2-ODD-C23.2, At3g49630/2-ODD-C23.3 and At3g50210/2-ODD-C23.1) and the orthologs from rice and soybean, which are encoded by single-copy genes. We also included Ava5009 from Anabaena variabilis as representative of the most similar group of cyanobacterial 2-ODDs, although it is probably not orthologous to 2-ODD-C23 from plants and green algae. When we compared the annotation of Din11 with RNAseq data, we noticed that only very few RNAseq reads cover the 5′-end of the mRNA, whereas the coverage increases strongly 89 bp downstream of the predicted transcription start site, at a position 13 bp after the annotated start codon (Supplementary Fig. S2). A second ATG codon in the same reading frame is located 62 bp downstream of the major transcription start site. Comparison of the predicted protein sequences showed that the first 24 amino acids of the annotated Din11 (At3g49620.1, here termed Din11L) do not align with 2-ODD-C23 homologs from other plant species, whereas the variant starting with the second ATG codon (termed Din11s) aligns well over its entire length. Nevertheless, we were able to amplify the coding sequence of Din11L from Arabidopsis cDNA.
Of the three Arabidopsis 2-ODD-C23 isoforms, only the short variant of Din11 induced guanidine production when overexpressed in E. coli (Fig. 1). Guanidine production was stimulated by addition of arginine to the culture medium, whereas addition of 2-OG had no effect (Supplementary Fig. S3). We speculated that the other 2-ODD-C23 isoforms might use a compound similar to arginine as substrate and supplemented the cultures with various other guanidine compounds known to occur in nature (methylguanidine, guanidinoacetate, guanidinopropionate, guanidinobutyrate, agmatine, homoarginine, taurocyamine). Only when the culture medium of the cells was supplemented with homoarginine, all 2-ODD-C23 isoforms from Arabidopsis except Din11L induced guanidine production in the E. coli cultures (Fig. 1). The 2-ODD-C23 orthologs from soybean and rice (2-ODD33) mediated homoarginine-dependent guanidine accumulation to a similar degree than Din11s and At3g50210. The guanidine concentration was very similar in the E. coli cells and in the supernatant, indicating that guanidine was exported from the cells until an equilibrium with the medium was reached (Supplementary Fig. S3). In contrast to the 2-ODD-C23 isoforms, the 2-ODD Ava5009 from the cyanobacterium Anabaena variabilis did not mediate guanidine production from arginine or homoarginine.
The detection of free guanidine indicated that 2-ODD-C23 might be homoarginine-6-hydroxylases, as 6-hydroxyhomoarginine is expected to decay spontaneously to yield guanidine and 2-aminoadipate-6-semialdehyde (AASA). The latter is in spontaneous equilibrium with 81-piperideine-6-carboxylate (P6C), which can be converted to pipecolate by P5CR (proC in E. coli, (Fujii et al., 2002)). Indeed, we detected pipecolate in the E. coli cultures in quantities proportional to the amount of guanidine (Fig. 1). Only in the cells expressing Din11s, the proportion of pipecolate was lower, indicating that Din11s used both arginine and homoarginine as substrates. In contrast to guanidine, pipecolate was predominantly retained inside the cells (Supplementary Fig. S3). By GC-MS analysis of the headspace of bacterial cultures expressing Din11s or At3g50210 and incubated with 5 mM arginine or homoarginine for 6 h in air-tight tubes, no ethylene production was detected (Supplementary Fig. S3). The guanidine concentration in the cultures was above 100 µM, indicating that there was enough oxygen in the tubes to allow enzyme activity.
Plant 2-ODD-C23 enzymes are homoarginine-6-hydroxylases
To determine if the Arabidopsis 2-ODD-C23 isoforms directly catalyze the release of guanidine from homoarginine, we purified the recombinant proteins by Ni-affinity chromatography. Measurement of the rate of O2 consumption at variable substrate concentrations allowed the characterization of the three Arabidopsis enzymes (Fig. 2). With a fixed and saturating concentration of 0.5 mM 2-OG, At3g50210 had the lowest KM for homoarginine (0.68 mM) and the highest specific activity (Amax = 17 nmol s-1 mg-1). Din11s had a KM of 1.68 mM for homoarginine and a lower specific activity (Amax = 10 nmol s-1 mg-1). With a KM of 4.8 mM and an Amax of 6 nmol s-1 mg-1, At3g49630 had the lowest activity of the three Arabidopsis 2-ODD-C23 paralogs. LC-MS/MS analysis of the reaction products confirmed the formation of P6C, succinate and guanidine. When O-(2,3,4,5,6-Pentafluorbenzyl)-hydroxylamine was added during the reaction, the derivatization product of AASA was also detected. No products with the masses of dehydro-homoarginine or hydroxy-homoarginine were accumulating in the reaction mixture. We concluded that all three Arabidopsis 2-ODD-C23 isoforms specifically catalyze the C6-hydroxylation of homoarginine, which is followed by the spontaneous decay to guanidine and AASA, which subsequently cyclizes to P6C. As expected from the analysis of the E. coli cultures, only Din11s showed O2 consumption also with arginine, although the KM (3.3 mM) was higher compared to homoarginine and the specific activity was lower (Amax = 3 nmol s-1 mg-1). Neither lysine nor agmatine or any other guanidine-containing compound induced O2 consumption with one of the purified enzymes. The reaction of Din11s with arginine yielded guanidine and GSA with its cyclic counterpart 81-pyrroline-5-carboxylate (P5C), indicating that arginine is hydroxylated at the C5 position as in EFE from Pseudomonas savastanoi (Fukuda et al., 1992).
When we tested the effect of 1 mM arginine on the reaction of the 2-ODD-C23 isoforms with homoarginine, we detected a competitive inhibition of Din11s and At3g50210, whereas At3g49630 was insensitive to arginine (Fig. 2 and Supplementary Fig. S4). We further tested the effect of the arginine antimetabolite canavanine on the reaction of the 2-ODD-C23 isoforms. Canavanine is the 5-oxa-analog of arginine and therefore cannot be hydroxylated at this position. Similar to arginine, which is both substrate and inhibitor for Din11s, canavanine showed competitive inhibition of the reaction of Din11s with homoarginine. In contrast, the inhibition of At3g50210 was of the mixed type and At3g49630 was not inhibited by canavanine.
When the concentration of the co-substrate 2-oxoglutarate was varied at near-saturating concentration of homoarginine (10 mM), all three 2-ODD-C23 isoforms showed KM values between 20 and 100 µM and the reaction was inhibited by high 2-OG concentrations (Supplementary Fig. S4 and Supplementary Table S1). For At3g50210 we determined the pH-dependence of the enzymatic reaction and observed maximal activity around pH 7.2 (Supplementary Fig. S4), whereas the enzyme was virtually inactive below pH 5 and above pH 9.
Arabidopsis 2-ODD-C23 isoforms are cytosolic enzymes and mediate guanidine production in vivo
The protein sequences of the 2-ODD-C23 isoforms from Arabidopsis do not contain any recognizable targeting signals. Consistently, the analysis of plants expressing 2-ODD-C23:GFP fusion proteins under control of the constitutive UBQ10-promoter demonstrated the cytosolic localization of all three isoforms (Fig. 3). Among 41 plants transformed with a construct for expression of Din11L:GFP, no plants with detectable GFP fluorescence were identified. Among 80 lines carrying a construct for expression of Din11s:GFP, several lines showed cytosolic GFP signals but only in one line, the expression of Din11s:GFP was stable over several generations. For both At3g49630:GFP and At3g50210:GFP, several independent lines were identified that stably expressed the GFP fusion protein. GFP signals were typically strongest in At3g50210:GFP-expressing plants.
In order to determine whether the 2-ODD-C23 isoforms catalyze C6-hydroxylation of homoarginine under physiological conditions, we identified and generated Arabidopsis mutants. Several T-DNA insertion lines were obtained from the Nottingham Arabidopsis Stock Centre and were backcrossed at least 3 times to Col-0 wildtype (WT) plants to obtain single insertion lines (Supplementary Fig. S5 and Supplementary Table S2). Because Din11 and At3g49630 are located directly next to each other on chromosome 3, we used mutagenesis with CRISPR/Cas9 with a gRNA targeting the third exon of both Din11 and At3g49630 to generate double mutants. We obtained one line, in which the 5′-end of Din11 is fused to the 3′-part of At3g49630. Insertion of 1 bp at the fusion site causes a frame shift and inactivates both genes. A second line contains a 27 bp deletion including the border between intron 2 and exon 3 of Din11 and a 1 bp deletion in exon 3 of At3g49630. The CRISPR/Cas9-generated mutants also were backcrossed to WT plants to eliminate the CRISPR/Cas9 expression cassette and potential off-target mutations. To generate triple mutants, the double mutants were crossed with the mutant line at3g50210-1, which carries a T-DNA insertion in the third exon of At3g50210. All single, double, and triple mutants were phenotypically not different from WT plants when growing under normal conditions in sterile culture or in the greenhouse.
Basal guanidine contents were around 10-15 nmol [g fresh weight (Fw)]-1 equivalent to 100-200 nmol (g dry weight)-1 in the rosettes of axenically grown Arabidopsis WT plants and were increased to 20-30 nmol (g Fw)-1 in 2-ODD-C23:GFP overexpressing plants (Fig. 4). The guanidine content of soil grown plants was usually slightly higher, and in legume seeds and seedlings we detected up to 2.4 µmol (g dry weight)-1 (equivalent to 145 ppm) guanidine (Supplementary Fig. S6). The guanidine content in rosettes of 2-ODD-C23-triple mutants was consistently lower than in WT plants. Surprisingly, the basal homoarginine content (10-20 nmol (g Fw)-1 in WT rosettes) was reduced to around 5 nmol (g Fw)-1 in rosettes of both the triple mutants as well as in At3g50210:GFP overexpressing plants (Fig. 4). When WT plants and At3g50210:GFP-expressing plants were transferred to culture plates with 0.2 mM homoarginine in the medium, they accumulated both homoarginine and guanidine in a time-dependent manner, reaching 100 nmol (g Fw)-1 guanidine in WT rosettes and 600 nmol (g Fw)-1 in the rosettes of the overexpressor plants. Like WT and overexpressor plants, the 2-ODD-C23 triple mutants accumulated homoarginine to around 800 nmol (g Fw)-1, whereas the guanidine content in the rosettes stayed nearly unchanged. In din11 and at3g49630 single mutants, the guanidine content was not reduced compared to WT plants after external supply of homoarginine. In at3g50210 single mutants, however, the guanidine content was as low as in the triple mutants (Supplementary Fig. S5).
Addition of homoarginine or arginine to the growth medium impaired root growth of Arabidopsis seedlings (Supplementary Fig. S7). On media containing either 1 mM arginine or 0.1 mM homoarginine, the roots of 1-week old plants had only half the length of the roots of control plants grown without amino acid supplement. Above 0.2 mM homoarginine, also the growth of the rosettes was impaired (Supplementary Fig. S7). However, these inhibitory effects were not changed in the 2-ODD-C23 triple mutants and 2-ODD-C23:GFP overexpressing plants, indicating that hydroxylation of homoarginine likely does not serve to detoxify high concentrations of homoarginine. Addition of 1 mM guanidine to the growth medium had no effect on seedling growth, whereas 5 mM guanidine was lethal (Supplementary Fig. S7 and data not shown).
Originally, Din11 was named for its induction by prolonged darkness, a condition now understood to induce both C-starvation and autophagy. However, we did not detect phenotypical differences between din11 or triple mutants and WT plants after prolonged dark treatment (Fig. 5). It had further been reported that expression of Din11 was induced by treatment with methyljasmonate (MeJA) (Caarls et al., 2017) and by perturbations of ROS homeostasis (Gadjev et al., 2006), whereas analysis of publicly available RNAseq data did not reveal a treatment that induced all three Arabidopsis 2-ODD-C23 isoforms (Supplementary Fig. S8) (Zhang et al., 2020). Treatment of WT plants with 10 or 50 µM MeJA did not induce consistent and marked changes of the content of homoarginine and arginine in the rosettes of Arabidospis seedlings. However, MeJA treatment strongly induced the guanidine content in Arabidopsis rosettes and this induction was lost in din11 single mutants and not decreased much further in 2-ODD-C23 triple mutants (Fig. 5).
Metabolic fate of AASA in plants
The different rates of homoarginine metabolism in WT plants, triple mutants and At3g50210:GFP overexpressing plants, as evidenced by different levels of guanidine accumulation, were surprisingly not reflected in different rates of homoarginine accumulation after external supply (Fig. 4). Therefore, we wondered if AASA/P6C production by 2-ODD-C23 might serve a signaling rather than a metabolic function. As in the E. coli cultures, P5Cr might convert AASA/P6C to pipecolate, which is the direct precursor for the mobile defense signal N-hydroxypipecolate (Zeier, 2021). Purified, recombinant plant P5CR converted P6C to pipecolate with a KM of 173 ±18 µM or 257 ±23 µM and an Amax of 7 or 13 nmol s-1 mg-1, when NADH or NADPH were used as electron donors, respectively (Fig. 6). An alternative route for AASA/P6C metabolism would be oxidation of AASA to aminoadipate by an aldehyde dehydrogenase. For maize and pea, it was shown that ALDH7B4 has a high affinity for AASA (Končitíková et al., 2015). Indeed, purified, recombinant AtALDH7B4 (At1g54100) converted AASA to aminoadipate with a KM of 11.9 ±1.0 µM and an Amax of 529 ±6 nmol s-1 mg-1 with NAD+ as electron acceptor. For the cofactor NAD+, ALDH7B4 had a KM of 324 ±17 µM, whereas the enzyme was almost inactive with NADP+ as electron acceptor and neither was NADPH formation linear over time nor did it show a linear correlation with the amount of enzyme (Supplementary Fig. S9). ALDH7B4 also accepted GSA/P5C as substrate, but the conversion rates were 60 times lower compared to AASA/P6C and also in this case, the reaction rate was not linearly correlated to the amount of enzyme (Supplementary Fig. S9). The 10-fold lower KM of AASA oxidation to aminoadipate vs. P6C reduction to pipecolate suggests that the oxidative route is predominantly occurring under physiological conditions. Independent of genotype and homoarginine supply, the pipecolate content of seedlings was below 2 nmol (g Fw)-1 under our conditions and could not be quantified because the peak was not clearly separated from other compounds with the same M/Z ratios. The aminoadipate content in Arabidopsis rosettes was also below the level for reliable quantification [1 nmol (g Fw)-1] but increased to almost 6 nmol (g Fw)-1 in At3g50210:GFP-overexpression plants after supply of external homoarginine (Supplementary Fig. S10). This increase was absent in 2-ODD-C23 triple mutants and barely detectable in WT plants, indicating that aminoadipate metabolism is faster than the production via AASA derived from homoarginine hydroxylation by 2-ODD-C23.
In this study, we discovered a novel enzymatic activity for a previously uncharacterized clade of plant 2-ODDs. All analyzed 2-ODD-C23 proteins mediated the hydroxylation of homoarginine at the C6 position, thereby inducing its decay to AASA and guanidine (Fig. 2). This is in sharp contrast to P. savastanoi EFE, which was virtually inactive with homoarginine (Martinez & Hausinger, 2016). The KM values for homoarginine hydroxylation by the plant enzymes were higher than the KM of P. savastanoi EFE for arginine (50 µM) and the Amax values were intermediate between the values for ethylene formation (54 nmol s-1 mg-1) and arginine hydroxylation (1.2 nmol s-1 mg-1) reported for EFE (Martinez & Hausinger, 2016). Among the biochemically characterized plant 2-ODDs from class C, DMR6/S5H had a lower KM (5.1 ±1.4 µM) and a similar Amax (5.9 ±1.1 nmol s-1 mg-1) for salicylate hydroxylation (Zhang et al., 2017). Din11, which is exclusively present in brassicaceae, additionally accepted arginine as substrate. Although Din11 has a lower specific activity and a higher KM for the reaction with arginine, the low concentration of homoarginine in Arabidopsis may lead to a preferential hydroxylation of arginine by Din11. 2-ODD33, the 2-ODD-C23 ortholog from rice, was previously reported to have melatonin-2-hydroxylase activity (Byeon & Back, 2015). However, the reported rate of 2-hydroxymelatonin accumulation by purified, recombinant 2-ODD33 was 1000-fold lower than the hydroxylation of homoarginine in our experiments, indicating that homoarginine is the preferred substrate of 2-ODD-C23 enzymes.
The evolutionary origin of 2-ODD-C23 function is at present ambiguous. Loss of the gene in Chlamydononas reinhardtii has previously led to the interpretation that these enzymes were specific to plants (Kawai et al., 2014). However, putative orthologs of 2-ODD-C23 are present in other green algae, whereas the protein with the highest degree of similarity in the cyanobacterium Anabaena variabilis (Ava5009) did not have homoarginine-6-hydroxylase activity when expressed in E. coli (Fig. 1). It is thus conceivable that the production of guanidine from homoarginine or arginine evolved in photosynthetic eukaryotes. The single-copy 2-ODD-C23 isoforms from rice and soybean accepted exclusively homoarginine as substrate, indicating that arginine hydroxylation by Arabidopsis Din11 derived from neofunctionalization after at least two duplication events in the Brassicaceae family (Supplementary Fig. S1). Sequence alignments between bacterial and fungal EFEs and plant 2-ODDs did not reveal a close relationship between Din11 and EFE (data not shown), indicating an independent evolution of arginine-5-hydroxylase activity in both enzymes.
The guanidine content was strongly reduced mutant plants devoid of 2-ODD-C23 expression, indicating that 2-ODD-C23s are the major source of free guanidine in Arabidopsis. At3g50210 mediated the major part of the homoarginine-dependent guanidine production (Supplementary Fig. S5) which is consistent with the highest specific activity (Fig. 2) and the highest basal expression level of At3g50210 (Supplementary Fig. S8). In contrast, Din11 expression has the highest dynamic range and induction of Din11 by MeJA induced a strong increase in guanidine production, presumably from arginine as precursor (Fig. 5). This increase was virtually absent in din11 single mutants and the triple mutants, indicating that it was exclusively attributable to Din11 activity. Based on a basal guanidine content of 120 nmol (g dry weight)-1 (equivalent to 7 ppm) in Arabidopsis and our observation that other plants have similar or even higher contents of guanidine (Supplementary Fig. S6), we estimated the global annual production of guanidine by plants. With a total annual biomass production of 5*1010 t by land plants, this estimation yields an annual guanidine production of 3.5*105 t by plants, corresponding roughly to the 2.5-fold of the industrial production in 2016 (Precision Reports, 2021).
Plants were shown to transport but barely metabolize guanidine (Marsden et al., 2015), indicating that nitrogen released as guanidine cannot be efficiently recycled by plants. Considering the often growth-limiting function of nitrogen for plants, this raises the question whether guanidine may serve as a defense or signaling compound. In soil, a greater proportion of bacteria contains genes for guanidine import and assimilation than among bacteria associated with human skin or the gut, which preferentially contain genes for guanidine exporters (Sinn et al., 2021). Since bacteria differ in their responses to guanidine, it is possible that guanidine accumulation or even secretion is a way for plants to modulate their microbiomes (Sharma et al., 2023). A similar function was proposed for hydroxyguanidine that was secreted by Pseudomonas canavaninivorans after degradation of plant-produced canavanine (Hauth et al., 2022). Induction of Din11 expression and guanidine production by wounding and MeJA may also point to a function of guanidine in the defense against or repulsion of herbivores (Caarls et al., 2017; Griffiths, 2020). Expression of Din11 was additionally induced in several mutants or after treatments that cause ROS accumulation, indicating that ROS may act as second messenger in the induction of Din11 expression (Supplementary Fig. S8) (Gadjev et al., 2006). Further studies will be required to determine the biological purpose of guanidine production from homoarginine or arginine by 2-ODD-C23 enzymes.
In addition or alternatively, the second reaction products of 2-ODD-C23s, AASA/P6C or GSA/P5C could represent physiologically functional products. GSA/P5C produced by cytosolic Din11 (Fig. 3) is most likely converted to proline by P5CR, which is a cytosolic enzyme in plants as well (Funck et al., 2012). Thereby, the activity of Din11 may explain the observation that besides glutamate also arginine can serve as precursor for proline synthesis in plants (Adams & Frank, 1980; Winter et al., 2015). So far, no combination of enzymes residing in the same compartment was known that could catalyze the conversion of arginine to proline without glutamate as intermediate (Trovato et al., 2019). However, the exclusive occurrence of Din11 orthologs in brassicaceae indicates that this metabolic pathway is not common among plants (Supplementary Fig. S1). The maximal guanidine concentration of 400 nmol (g Fw)-1 observed in MeJA-treated Arabidopsis plants (Fig. 5) argues against a major contribution to stress-induced proline accumulation, which typically reaches 5-10 µmol (g Fw)-1 (Forlani et al., 2019). Besides the induction by MeJA, expression of Din11 was also induced by prolonged dark treatment, which led to its initial identification and naming (Fujiki et al., 2001). Prolonged dark treatment causes C-starvation in plants and results in the induction of autophagy to mobilize energy by breakdown of cellular components and organelles (Magen et al., 2022). Degradation of homoarginine by 2-ODD-C23s is unlikely to be a relevant contribution to energy production during C-starvation because the observed homoarginine content was very low in Arabidopsis (Fig. 4). In contrast, degradation of arginine via Din11 to GSA/P5C and guanidine may serve to mobilize the C-skeleton of arginine without overloading the nitrogen metabolism. Degradation of arginine via arginase to ornithine and urea would result ultimately in glutamate and ammonia formation by glutamate dehydrogenase and urease (Winter et al., 2015). High ammonia concentrations are toxic for plants, especially when the energy or the enzymes to re-assimilate it into glutamine are lacking (Moreira et al., 2022). However, the extent of chlorophyll degradation was not different between WT plants, 2-ODD-C23 triple mutants and overexpressing plants, indicating that the efficiency of autophagy in remobilizing carbon and nitrogen does not depend on 2-ODD-C23 activity (Fig. 5). We are not aware of any reports describing a function of homoarginine in plants except the storage and defense function in seeds of hyperaccumulating Lathyrus species (Bell, 1980). The unexpected reduction of the homoarginine content in both, 2-ODD-C23 overexpression plants and 2-ODD-C23 triple mutants in Arabidopsis may indicate the existence of an auto-regulatory signaling loop.
AASA originating from the decay of 6-hydroxyhomoarginine can be converted to pipecolate by P5CR after spontaneous cyclisation to P6C (Fujii et al., 2002; Struys et al., 2014). In this way, 2-ODD-C23s might contribute to the production of the mobile defense signal N-hydroxypipecolate (Zeier, 2021). However, the direct comparison of the kinetic properties of P5CR and ALDH7B4 suggested that oxidation of AASA to aminoadipate by ALDH7B4, which is a shared reaction with the lysine degradation pathway, is the main pathway under physiological conditions (Fig. 6) (Arruda & Barreto, 2020). We did not find information about the rate of AASA-cyclisation to P6C or the equilibrium concentrations in plants. In human plasma from patients with ALDH7A1 defects, approximately equal concentrations of both compounds were detected after FMOC-derivatization (Xue et al., 2019). Consistent with the kinetic properties of P5CR and ALDH7B4, no increase of the pipecolate content after feeding with external homoarginine was observed in Arabidopsis plants overexpressing the 2-ODD-C23 isoform At3g50210 (Supplementary Fig. S10). Instead, a small increase in aminoadipate content was observed, indicating that AASA was indeed oxidized by ALDH7B4, but the rate of aminoadipate metabolism apparently exceeded the rate of synthesis by 2-ODD-C23 and ALDH7B4. E. coli lacks a specific AASA-DH and consequently cyclisation to P6C followed by reduction to pipecolate by the constitutively expressed P5CR (proC) is the main route of homoarginine metabolism by 2-ODD-C23 expressing E. coli cells (Fig. 1).
In humans, homoarginine is formed from lysine and arginine by mitochondrial glycine amidinotransferase (Choe et al., 2013). It is discussed that homoarginine interferes with NO production from arginine and thus might be an important regulator of NO signaling (Koch et al., 2022). Plants seem to lack NO synthases that use arginine as substrate and do not have proteins with significant sequence similarity to glycine amidinotransferase, although a protein with similar activity has been isolated from Lathyrus sativus seedlings (León & Costa-Broseta, 2020; Srivenugopal & Adiga, 1980). Thus, the biosynthesis and physiological function of homoarginine in plants will need to be addressed in future studies. In humans, the accumulation of AASA as consequence of mutations in ALDH7A1/Antiquitin, which is the ortholog of Arabidopsis ALDH7B4, leads to developmental retardation and epilepsy (Bouchereau & Schiff, 2020; Brocker et al., 2013). The disease symptoms are caused by formation of adducts of AASA with pyridoxal-5-phosphate, thereby inactivating pyridoxal-5-phosphate-dependent enzymes. In plants, ALDH7B enzymes may similarly serve to detoxify AASA derived from homoarginine hydroxylation by 2-ODD-C23s or from lysine degradation by the bifunctional lysine-ketoglutarate reductase/saccharopine dehydrogenase (LKR/SDH, At4g33150) (Arruda et al., 2000; Goncalves-Butruille et al., 1996). Mutation of Arabidopsis ALDH7B4 was reported to increase drought and salt sensitivity, whereas the sensitivity to herbicides was not significantly altered (Gil-Monreal et al., 2017; Kotchoni et al., 2006). In rice, mutation of ALDH7B6 reduced seed viability and caused the accumulation of pigments in the endosperm (Shen et al., 2012; Shin et al., 2009). Our analyses show that Arabidopsis ALDH7B4 has an even lower KM (11.9 µM, Fig. 6) for AASA than the previously analyzed orthologs from rice and pea (97 and 89 µM respectively; (Končitíková et al., 2015)). It remains to be determined if ALDH7B4 functions specifically to prevent AASA accumulation or if it additionally acts nonspecifically on a broad range of aldehydes and semialdehydes.
Overall, our characterization of plant 2-ODD-C23s as homoarginine-6-hydroxylases is to our knowledge the first description of eukaryotic enzymes that mediate the production of guanidine. Although homoarginine represents a low abundant metabolite of unknown function and origin in most analyzed plants, the large amount of plant biomass produced every year implies that substantial amounts of guanidine enter the biological nitrogen cycle via the action of 2-ODD-C23 enzymes. The second product of 6-hydroxyhomoarginine decay, AASA/P6C, was primarily fed into the lysine degradation pathway by ALDH7B. So far, Din11 from Arabidopsis was the only 2-ODD-C23 isoform that accepted also the abundant arginine as substrate, but its low expression level and low specific activity apparently prevented excessive diversion of arginine towards guanidine and proline production. Consistently, the reduced guanidine and homoarginine contents were so far the only phenotypical difference that we detected in Arabidopsis mutants lacking all three 2-ODD-C23 isoforms. Although neither homoarginine nor guanidine are highly abundant metabolites in Arabidopsis, the observed differences in the triple mutants and overexpressor plants confirm that the 2-ODD-C23 isoforms catalyze homoarginine hydroxylation in vivo and that potentially also production of homoarginine is a regulated process. Induction of Din11 by prolonged dark treatment, wounding, and reactive oxygen species provides hints that guanidine, AASA, or GSA may function in defense responses or other biotic interactions. We speculate that plants use guanidine production to influence guanidine-sensing or guanidine-assimilating microbes and plan to test this hypothesis in future studies.
Materials and Methods
DNA constructs and recombinant protein production
Bacterial expression constructs for Arabidopsis Din11, At3g49620, At3g50210, and ALDH7B4, as well as Os2-ODD33 and Gm2-ODD-C23 were generated by amplifying the coding sequence from cDNA with primers that introduce overlaps with a modified pET vector containing an N-terminal His-tag and a TEV cleavage site (Supplementary Table S3). The resulting PCR products were fused with the linearized vector by Gibson assembly (NEBuilderHiFi). Protein expression in E. coli SoluBL21 cells was induced with 1 mM IPTG at 18 °C overnight. The cells were lysed by sonication in extraction buffer (50 mM Hepes pH 7.5, 50 mM NaCl, 1x protease inhibitor cocktail (EDTA-free cOmplete, Roche) and the His-tagged proteins were purified with HisBind NiNTA-agarose (Qiagen) according to the recommendations of the supplier. Imidazole was removed from the purified proteins by passage through PD MidiTrap G-25 columns (GE lifescience) equilibrated with extraction buffer. Protein concentrations were determined according to Bradford (1976) with BSA as standard. All activity assays were performed with enzymes purified from at least three independent bacterial cultures.
Enzyme activity assays
Activity of ODDs was measured by following the oxygen consumption at 30 °C in a Clark-type electrode (Hansatech Oxygraph+) calibrated according to the instructions of the supplier. The reaction mix contained 50 mM MOPS pH 7.2, 70 mM NaCl, 20 mM KCl, 0.4 mM Na-ascorbate, 40 µM Fe(NH4)2(SO4)2, and variable concentrations of substrate or cosubstrate. The reactions were started by adding the purified enzymes and the specific activities were calculated from the difference between the oxygen consumption rate in the initial linear phase and the rate observed prior to the addition of the enzymes (typically around −20 nmol O2 ml-1 s-1, corresponding to 1 to 2% of the maximal rate in the presence of enzyme).
The activity of P5CR was calculated from the difference in the rates of NAD(P)H oxidation at pH 7.5 and 35 °C in the presence or absence of DL-P5C or L-P6C as described previously (Forlani et al., 2015). DL-P5C was synthesized by the periodate oxidation of δ-allo-hydroxylysine (Sigma H0377) and purified by cation-exchange chromatography, as described previously (Forlani et al., 1997). L-P6C generation from L-allysine ethyleneacetal (Sigma-Aldrich 714208) roughly followed a previous protocol (de La Fuente et al., 1997). A solution of 50 mg allysine ethyleneacetal in 5 ml water was mixed with 300 mg amberlite IR120 (Fluka 06428) and gently stirred for 20 min. The supernatant was discarded and the resin was loaded into a mini-column, washed with 10 mL of water and eluted with 10 mL of 1 M HCl in 1 ml fractions. The concentration of P6C was measured by mixing a proper dilution in 60 µl with 120 µl 1 mg ml-1 o-aminobenzaldehyde in ethanol, and the increase of absorbance at 465 nm was followed until a plateau was reached. The extinction coefficient of χ465 = 2800 M-1 cm-1 was used (Fothergill & Guest, 1977). Both DL-P5C and L-P6C were kept as stocks in 1 M HCl at 4 °C in the dark and neutralized with 1 M Tris-base immediately before use. The standard enzyme assay contained 50 mM Tris-HCl, 0.5 mg ml-1 NAD(P)H and 2 mM DL-P5C or 1 mM L-P6C. The activity of ALDH7B4 was calculated from the difference in NAD(P)+ reduction at pH 7.5 and 35 °C in the presence or absence of variable concentrations of DL-P5C or L-P6C. The standard reaction mixture contained 50 mM Tris-HCl, pH 7.5, 1 mM NAD(P)+ and 2 mM DL-P5C or 1 mM L-P6C. The extinction coefficient of ε340 = 6.22 mM-1 cm-1 for NADH and NADPH was used to calculate specific activities.
Plant material, plant transformation and cultivation conditions
Arabidopsis [Arabidopsis thaliana (L.) Heynh.], ecotype Col-0 and T-DNA insertion lines (Supplementary Table S2) were obtained from the NASC. Surface-sterilized seeds were sown on solid half-strength Murashige and Skoog medium including Gamborg B5 vitamins (pH 5.8, supplemented with 2% [w/v] sucrose and 0.8% agar) in Petri dishes sealed with Leukopor tape (Leukoplast) and cultivated under a long-day light regime (16 h photoperiod; 110 µmol m-2 sec-1 at 21 °C). For seed production, seedlings were transferred to commercial substrate (“Einheitserde”, type P, Gebr. Patzer, Sinntal-Altengronau, Germany) and cultivated in a greenhouse under long-day conditions. All T-DNA insertion mutants were backcrossed at least three times with WT plants, until the segregation pattern was consistent with a single T-DNA insertion. Constructs for the expression of GFP fusion proteins were obtained by inserting the coding sequences into pENTRa1 (Invitrogen) followed by LR-recombination with pUBC-GFP-Dest (Grefen et al., 2010). The construct for simultaneous CRISPR/Cas9-mediated inactivation of Din11 and At3g49630 was generated by ligating two matching oligonucleotides (Supplementary Table S3) into the BsaI-linearized vector pHEE401E (Wang et al., 2015). The resulting gRNA matches a Cas9 cleavage site in both genes with one mismatch in Din11 and no strongly predicted off-targets. All constructs were introduced into Arabidopsis plants by floral dip with Agrobacterium tumefaciens strain GV3101 (Clough & Bent, 1998). Transgenic plants were selected by spraying of seedlings with BASTA solution (50 µg ml-1) or growth on sterile culture plates containing 30 µg ml-1 hygromycin B (Sigma). For the GFP constructs, lines with stable expression and a single-copy T-DNA insertion were selected by segregation analysis. CRISPR/Cas9-induced mutations were detected by high-resolution melting analysis using gene-specific primers (Supplementary Table S3) and confirmed by sequencing of the PCR products. The CRISPR/Cas9 construct was eliminated by backcrossing to WT plants. The absence of functional transcripts was confirmed by RT-PCR and sequencing of the products.
For feeding experiments, homoarginine or arginine were added to the culture medium from filter-sterilized stocks (100 mM, pH 6). For treatment with MeJA, 100 µl of 5 mM MeJA in 50% (v/v) EtOH were distributed on the lid of the Petri dishes, which were consecutively sealed with Parafilm. Control plates were prepared with 50% EtOH. For the analysis of starvation and autophagy, the plants were cultivated on half-strength MS plates without sucrose under short-day conditions at 21 °C. Plates with 8-day-old seedlings were wrapped in aluminum foil and incubated at 21 °C for 8 days.
Other plant material was collected around the University of Konstanz or was purchased in a garden store. Seeds of alfalfa, broccoli and rocket were surface sterilized as above and germinated on filter discs soaked in distilled water.
Protoplasts were isolated from leaves of 2-ODD-C23:GFP expressing and WT plants by overnight incubation in protoplast medium (0.45 M sorbitol, ½ strength MS salt mixture) supplemented with 10 mg ml-1 cellulase and 2.5 mg ml-1 macerozyme. Spectral images were recorded with a Zeiss LSM880 with a 63x water immersion lens. The samples were excited at 488 nm and images were recorded with 20 spectral channels (9 nm bandwidth) between 490 and 668 nm. A chlorophyll spectrum was obtained from WT protoplasts and a GFP spectrum from an epidermis protoplast that did not contain chloroplasts. Linear unmixing with background subtraction was performed with the ZEN software (Zeiss). Channel overlay, false coloring and adjustment to identical gain, offset and contrast settings were performed in ImageJ and Adobe Photoshop.
E. coli cultures were lysed by ultrasonication prior to or after the addition of a 20-fold excess of MeOH spiked with 10 µM 13C15N-guanidine (Sigma-Aldrich 607312). Plant samples were homogenized at −80 °C in a TissueLyser (Qiagen) and resupendend in a 10-fold excess of 80% (w/v) MeOH containing 10 µM 13C15N-guanidine and 20 µM 13C15N-arginine (Cortectnet CCN250P). After ultrasonication for 5 min, the suspension was centrifuged for 10 min at maximum speed in a tabletop centrifuge and the supernatant was used directly for metabolite quantification.
Metabolite concentrations in the samples were detemined by high performance liquid chromatography coupled with mass spectroscopy on a LCMS-2020 Single Quadrupole mass spectrometer (Shimadzu). 2 μL sample volume was injected into a Nucleodur HILIC column (250 mm * 2 mm, 3 μm particle size, Macherey-Nagel). Metabolites were eluted at 20 °C with an acetonitrile gradient (90% (v/v) to 60% during 6 min, to 45% in 1 min, constant for 2.5 min, back to 90% in 1.5 min) in 10 mM ammonium formate, 0.2% (v/v) formic acid at a flow rate of 0.15 ml min−1. The separated compounds were analyzed by single ion monitoring (SIM) for positively (and negatively when appropriate) charged ions. Quantification of guanidine, pipecolate and aminoadipate was performed by calibrating peak areas with pure standards and by normalization to the added 13C15N-guanidine. Homoarginine was quantified by calibrating the peak areas with a pure standard and by normalization to the added 13C15N-arginine.
J.S.H. acknowledges the ERC CoG project 681777 “RiboDisc” for financial support. We thank Astrid Joachimi, Dmitry Galetskiy, and Janine Dietrich for technical assistanceas well as Nicolai Müller and Dieter Spiteller for help with metabolite detection. Additionally, we thank Svetlana Boycheva Woltering for help with autophagy assays as well as Martin Stöckl and the Bioimaging Centre of the University of Konstanz for assistance with confocal microscopy and image processing. We are grateful to Erika Isono and the team of gardeners of the University of Konstanz for access to plant cultivation facilities and plant care.
D.F., M.S. and J.S.H. conceived the project. D.F., M.S., and G.F. performed recombinant enzyme expression, purification, and activity assays. D.F. generated and characterized transgenic plants and mutants. D.F. and J.S.H. wrote the manuscript with input from all authors. D.F. and G.F. prepared figures. The manuscript was reviewed and approved by all coauthors.
Competing Interest Statement
All authors declare the absence of any competing interests.
- Metabolism of proline and the hydroxyprolinesAnnual Review of Biochemistry 49:1005–1061https://doi.org/10.1146/annurev.bi.49.070180.005041
- Lysine catabolism through the saccharopine pathway: Enzymes and intermediates involved in plant responses to abiotic and biotic stressFrontiers in Plant Science 11https://doi.org/10.3389/fpls.2020.00587
- Regulation of lysine catabolism in higher plantsTrends in Plant Science 5:324–330https://doi.org/10.1016/S1360-1385(00)01688-5
- α,γ-diaminobutyric acid in Seeds of twelve species of Lathyrus and identification of a new natural amino-acid, L-homoarginine, in seeds of other species toxic to man and domestic animalsNature 193:1078–1079
- The non-protein amino acids of higher plantsEndeavour 4:102–107https://doi.org/10.1016/0160-9327(80)90056-3
- Nonprotein amino acids of plants: Significance in medicine, nutrition, and agricultureJournal of Agricultural and Food Chemistry 51:2854–2865https://doi.org/10.1021/jf020880w
- Die Oxydationsprodukte des ArgininsHoppe-Seyleŕs Zeitschrift für Physiologische Chemie 32:278–280https://doi.org/10.1515/bchm2.1901.32.3-4.278
- Inherited disorders of lysine metabolism: A reviewJournal of Nutrition 150:2556–2560https://doi.org/10.1093/jn/nxaa112
- Aldehyde dehydrogenase (ALDH) superfamily in plants: gene nomenclature and comparative genomicsPlanta 237:189–210https://doi.org/10.1007/s00425-012-1749-0
- Molecular cloning of melatonin 2-hydroxylase responsible for 2-hydroxymelatonin production in rice (Oryza sativa)Journal of Pineal Research 58:343–351https://doi.org/10.1111/jpi.12220
- Arabidopsis JASMONATE-INDUCED OXYGENASES down-regulate plant immunity by hydroxylation and inactivation of the hormone jasmonic acidProc Natl Acad Sci U S A 114:6388–6393https://doi.org/10.1073/pnas.1701101114
- Homoarginine levels are regulated by L-arginine:glycine amidinotransferase and affect stroke outcome: results from human and murine studiesCirculation 128:1451–1461https://doi.org/10.1161/circulationaha.112.000580
- Floral dip: a simplified method for Agrobacterium-mediated transformation of Arabidopsis thalianaPlant Journal 16:735–743https://doi.org/10.1046/j.1365-313x.1998.00343.x
- Δ-1-piperideine-6-carboxylate dehydrogenase, a new enzyme that forms alpha-aminoadipate in Streptomyces clavuligerus and other cephamycin C-producing actinomycetesBiochemical Journal 327:59–64
- Two distinct mechanisms for C–C desaturation by iron(II)- and 2-(oxo)glutarate-dependent oxygenases: Importance of α-heteroatom assistanceJournal of the American Chemical Society 140:7116–7126https://doi.org/10.1021/jacs.8b01933
- Functional diversity of 2-oxoglutarate/Fe(II)-dependent dioxygenases in plant metabolism [Review]Frontiers in Plant Science 5https://doi.org/10.3389/fpls.2014.00524
- Functional properties and structural characterization of rice delta(1)-pyrroline-5-carboxylate reductaseFrontiers in Plant Science 6https://doi.org/10.3389/fpls.2015.00565
- Δ1-Pyrroline-5-carboxylate dehydrogenase from cultured cells of potato (purification and properties)Plant Physiology 113:1413–1418https://doi.org/10.1104/pp.113.4.1413
- Regulation of proline accumulation and its molecular and physiological functions in stress defenceOsmoprotectant-Mediated Abiotic Stress Tolerance in Plants: Recent Advances and Future Perspectives (pp. 73-97) https://doi.org/10.1007/978-3-030-27423-8_3
- Catabolism of L-lysine by Pseudomonas aeruginosaJ Gen Microbiol 99:139–155https://doi.org/10.1099/00221287-99-1-139
- Biotransformation of L-lysine to L-pipecolic acid catalyzed by L-lysine 6-aminotransferase and pyrroline-5-carboxylate reductaseBiosci Biotechnol Biochem 66:622–627https://doi.org/10.1271/bbb.66.622
- Dark-inducible genes from Arabidopsis thaliana are associated with leaf senescence and repressed by sugarsPhysiologia Plantarum 111:345–352https://doi.org/10.1034/j.1399-3054.2001.1110312.x
- Two reactions are simultaneously catalyzed by a single enzyme: The arginine-dependent simultaneous formation of two products, ethylene and succinate, from 2-oxoglutarate by an enzyme from Pseudomonas syringaeBiochemical and Biophysical Research Communications 188:483–489https://doi.org/10.1016/0006-291x(92)91081-z
- Discovery of a Ni2+-dependent guanidine hydrolase in bacteriaNature https://doi.org/10.1038/s41586-022-04490-x
- Requirement of proline synthesis during Arabidopsis reproductive developmentBMC Plant Biology 12https://doi.org/10.1186/1471-2229-12-191
- Transcriptomic footprints disclose specificity of reactive oxygen species signaling in ArabidopsisPlant Physiology 141:436–445https://doi.org/10.1104/pp.106.078717
- Induction of the PDH bypass and upregulation of the ALDH7B4 in plants treated with herbicides inhibiting amino acid biosynthesisPlant Science 264:16–28https://doi.org/10.1016/j.plantsci.2017.08.003
- Purification and characterization of the bifunctional enzyme lysine-ketoglutarate reductase-saccharopine dehydrogenase from maizePlant Physiology 110:765–771https://doi.org/10.1104/pp.110.3.765
- A ubiquitin-10 promoter-based vector set for fluorescent protein tagging facilitates temporal stability and native protein distribution in transient and stable expression studiesPlant Journal 64:355–365https://doi.org/10.1111/j.1365-313X.2010.04322.x
- Jasmonates: biosynthesis, perception and signal transductionEssays Biochem 64:501–512https://doi.org/10.1042/ebc20190085
- Biochemical diversity of 2-oxoglutarate-dependent oxygenases2-Oxoglutarate-Dependent Oxygenases :1–58https://doi.org/10.1039/9781782621959-00001
- Canavanine utilization via homoserine and hydroxyguanidine by a PLP-dependent gamma-lyase in Pseudomonadaceae and RhizobialesRSC Chem Biol 3:1240–1250https://doi.org/10.1039/d2cb00128d
- Amazing diversity in biochemical roles of Fe(II)/2-oxoglutarate oxygenasesTrends in Biochemical Sciences 43:517–532https://doi.org/10.1016/j.tibs.2018.04.002
- Evolution and diversity of the 2-oxoglutarate-dependent dioxygenase superfamily in plantsPlant Journal 78:328–343https://doi.org/10.1111/tpj.12479
- Homoarginine in the cardiovascular system: Pathophysiology and recent developmentsFundamental & Clinical Pharmacology 36:1–11https://doi.org/10.1111/fcp.12858
- Role and structural characterization of plant aldehyde dehydrogenases from family 2 and family 7Biochemical Journal 468:109–123https://doi.org/10.1042/BJ20150009
- Over-expression of different aldehyde dehydrogenase genes in Arabidopsis thaliana confers tolerance to abiotic stress and protects plants against lipid peroxidation and oxidative stressPlant Cell Environ 29:1033–1048https://doi.org/10.1111/j.1365-3040.2005.01458.x
- Free amino acids and toxins in Lathyrus sativus seedlingsPlanta Med 58:380–381https://doi.org/10.1055/s-2006-961491
- Grass pea (Lathyrus sativus L.): orphan crop, nutraceutical or just plain food?Planta 250:821–838https://doi.org/10.1007/s00425-018-03084-0
- Discovery and characterization of a fourth class of guanidine riboswitchesNucleic Acids Research 48:12889–12899https://doi.org/10.1093/nar/gkaa1102
- Present knowledge and controversies, deficiencies, and misconceptions on nitric oxide synthesis, sensing, and signaling in plantsPlant Cell Environ 43https://doi.org/10.1111/pce.13617
- Metabolism and autophagy in plants - a perfect matchFEBS Letters 596:2133–2151https://doi.org/10.1002/1873-3468.14359
- Plant acquisition and metabolism of the synthetic nitrification inhibitor dicyandiamide and naturally-occurring guanidine from agricultural soilsPlant and Soil 395:201–214https://doi.org/10.1007/s11104-015-2549-7
- Structures and mechanisms of the non-heme Fe(II)- and 2-oxoglutarate-dependent ethylene-forming enzyme: Substrate binding creates a twistJournal of the American Chemical Society 139:11980–11988https://doi.org/10.1021/jacs.7b06186
- Biochemical and Spectroscopic Characterization of the Non-Heme Fe(II)- and 2-Oxoglutarate-Dependent Ethylene-Forming Enzyme from Pseudomonas syringae pv. phaseolicola PK2Biochemistry 55:5989–5999https://doi.org/10.1021/acs.biochem.6b00890
- Glutamine synthetase: an unlikely case of functional redundancy in Arabidopsis thalianaPlant Biol (Stuttg 24:713–720https://doi.org/10.1111/plb.13408
- Metabolism of free guanidine in bacteria is regulated by a widespread riboswitch classMolecular Cell 65:220–230https://doi.org/10.1016/j.molcel.2016.11.019
- Precision Reports. (2021). Global guanidine market research report 2021. Retrieved 2023-02-23 from https://www.precisionreports.co/global-guanidine-market-17253362
- Dioxygenases: Molecular structure and role in plant metabolismAnnual Review of Plant Physiology and Plant Molecular Biology 47:245–271https://doi.org/10.1146/annurev.arplant.47.1.245
- The isolation and characterization of L-homoarginine from seeds of Lathyrus sativusBiochemistry 2:298–300https://doi.org/10.1021/bi00902a019
- Biochemical validation of a fourth guanidine riboswitch class in bacteriaBiochemistry 59:4654–4662https://doi.org/10.1021/acs.biochem.0c00793
- Solving the conundrum: Widespread proteins annotated for urea metabolism in bacteria are carboxyguanidine deiminases mediating nitrogen assimilation from guanidineBiochemistry 59:3258–3270https://doi.org/10.1021/acs.biochem.0c00537
- Ueber einige stickstoffhaltige Bestandtheile der Keimlinge von Vicia sativaZeitschrift für Physiologische Chemie 17:193–216
- Biotic stress-induced changes in root exudation confer plant stress tolerance by altering rhizospheric microbial communityFrontiers in Plant Science 14https://doi.org/10.3389/fpls.2023.1132824
- Mutation of OsALDH7 causes a yellow-colored endosperm associated with accumulation of oryzamutaic acid A in ricePlanta 235:433–441https://doi.org/10.1007/s00425-011-1477-x
- Biochemical Validation of a Third Guanidine Riboswitch Class in BacteriaBiochemistry 56:359–363https://doi.org/10.1021/acs.biochem.6b01271
- Biochemical Validation of a Second Guanidine Riboswitch Class in BacteriaBiochemistry 56:352–358https://doi.org/10.1021/acs.biochem.6b01270
- Rice aldehyde dehydrogenase7 is needed for seed maturation and viabilityPlant Physiology 149:905–915https://doi.org/10.1104/pp.108.130716
- Widespread bacterial utilization of guanidine as nitrogen sourceMolecular Microbiology https://doi.org/10.1111/mmi.14702
- Partial purification and properties of a transamidinase from Lathyrus sativus seedlings. Involvement in homoarginine metabolism and amine interconversionsBiochemical Journal 189:553–560https://doi.org/10.1042/bj1890553
- Untersuchungen über die chemischen Beziehungen zwischen Guanin, Xanthin, Theobromin, Caffeïn und KreatininJustus Liebigs Annalen der Chemie 118:151–177https://doi.org/10.1002/jlac.18611180203
- Human pyrroline-5-carboxylate reductase (PYCR1) acts on Δ1-piperideine-6-carboxylate generating L-pipecolic acidJournal of Inherited Metabolic Disease 37:327–332https://doi.org/10.1007/s10545-013-9673-4
- Identification of uncommon amino acids in the lentil seed (Lens culinaris MedExperientia 32:422–423https://doi.org/10.1007/BF01920772
- Filling in the gaps in metformin biodegradation: a new enzyme and a metabolic pathway for guanylureaApplied and Environmental Microbiology 87https://doi.org/10.1128/AEM.03003-20
- Proline metabolism and its functions in development and stress toleranceOsmoprotectant-Mediated Abiotic Stress Tolerance in Plants: Recent Advances and Future Perspectives :41–72https://doi.org/10.1007/978-3-030-27423-8_2
- Pyridoxine-dependent epilepsy: An expanding clinical spectrum [Review]Pediatric Neurology 59:6–12https://doi.org/10.1016/j.pediatrneurol.2015.12.013
- A guanidine-degrading enzyme controls genomic stability of ethylene-producing cyanobacteria [Article]Nature Communications 12https://doi.org/10.1038/s41467-021-25369-x
- Egg cell-specific promoter-controlled CRISPR/Cas9 efficiently generates homozygous mutants for multiple target genes in Arabidopsis in a single generationGenome Biol 16https://doi.org/10.1186/s13059-015-0715-0
- Physiological implications of arginine metabolism in plantsFrontiers in Plant Science 6https://doi.org/10.3389/fpls.2015.00534
- Simultaneous quantification of alpha-aminoadipic semialdehyde, piperideine-6-carboxylate, pipecolic acid and alpha-aminoadipic acid in pyridoxine-dependent epilepsySci Rep 9https://doi.org/10.1038/s41598-019-47882-2
- Metabolic regulation of systemic acquired resistanceCurrent Opinion in Plant Biology 62https://doi.org/10.1016/j.pbi.2021.102050
- A comprehensive online database for exploring 20,000 public Arabidopsis RNA-Seq librariesMolecular Plant 13:1231–1233https://doi.org/10.1016/j.molp.2020.08.001
- S5H/DMR6 encodes a salicylic acid 5-hydroxylase that fine-tunes salicylic acid homeostasisPlant Physiology 175:1082–1093https://doi.org/10.1104/pp.17.00695
- Genome-wide insertional mutagenesis of Arabidopsis thalianaScience 301:653–657https://doi.org/10.1126/science.1086391
- GABI-Kat SimpleSearch: new features of the Arabidopsis thaliana T-DNA mutant databaseNucleic Acids Research 40:D1211–1215https://doi.org/10.1093/nar/gkr1047
- A high-throughput Arabidopsis reverse genetics systemPlant Cell 14:2985–2994https://doi.org/10.1105/tpc.004630
- A comprehensive online database for exploring 20,000 public Arabidopsis RNA-Seq librariesMolecular Plant 13:1231–1233https://doi.org/10.1016/j.molp.2020.08.001