Polyamines metabolism genes have been identified and studied in a few plant species, including Arabidopsis, rice and tomato20,49−55. Here, we have presented a systematic investigation of PMGs at a whole-genome scale in barley. Genes involved in polyamine metabolic pathway in barley were identified and characterized together with their comparative assessment in relation to those known in Arabidopsis, rice and maize. Polyamine metabolism enzymes involved in the synthesis and catabolism of free PAs pool in plant cells are shown in Fig. 7.
We have identified three SAMDC, two ODC, one ADC, one SPDS, two SPMS, five CuAO and seven PAO members (in addition to the previously reported two PAOs56) in barley. The PA metabolic genes ADC, ODC, SPDS and SPMS/tSPMS in plants are highly conserved among different species, and they typically have only two copies each, except one copy each of SPMS and tSPMS49. In animals, the ODC pathway is considered essential in PA biosynthesis as almost all eukaryotes synthesize Put directly from ornithine (Orn) in a reaction catalyzed by ODC. The ODC pathway is present in most of the plant species, except for some species of the Brassicaceae, with ODC being absent in Arabidopsis57. During the early evolution of Brassicaceae, two paralogs (ADC1 and ADC2) of ancestral ADC gene were generated by duplication event and retained even after the genome shrinkage in species such as Arabidopsis58. The duplication of ADC genes might be a compensatory mechanism for the absence of ODC in plants57. Hence, ODC pathway is likely not essential for plants as ADC pathway does the same function. In the analysis presented here, both the ODC and ADC genes were identified in barley along with their developmental or stress specific function. Presence of many family members of isoforms indicates major contribution of gene duplication to functional diversity in higher organisms59. The number of PA metabolic pathway genes varies among species, implying that duplication events have occurred during the evolution of different species or due to their different genome size. Similarly, in barley as well it could be due to both, large genome size or an evolutionary process.
The biosynthetic pathways for PAs are conserved among organisms, from bacteria to animals and plants14. To delineate the comparative phylogenetic relationship of PA metabolic pathway genes in barley with other plant species, our phylogenetic analysis demonstrated that PA metabolic pathway proteins are highly conserved in the four species; barley, Arabidopsis, rice and maize, with the SAMDCs, PAOs, CuAOs, ADCs, ODCs, and SPDS/SPMSs being clustered into one group, respectively. The PA metabolic pathway protein sequences had homologues regions and shared the conserved catalytic active sites with other plant species; SAMDCs60, PAOs56, CuAOs33, ODCs and ADCs34. The SPMS/SPDS protein sequences were highly similar and also shared conserved active regions35,36. The SPDS and SPMS were found not distinct from one another. AtSPMS in the Arabidopsis genome is also designated as AtSPDS3. Thus, the HvSPMS genes identified in barley can also act as SPDS. This analysis confirmed that the PA metabolic pathway is highly conserved in plants.
Plant PAOs are classified into two groups, terminal catabolism (TC) reaction-type and back-conversion (BC) reaction-type. The TC-reaction produces 1,3-diaminopropane (DAP), H2O2, and the respective aldehydes, while the BC-reaction produces Spd from tetraamines, Spm and T-Spm and/or Put from Spd, along with 3-aminopropanal and H2O237. Previous studies on phylogenetic relationships showed that plant polyamine oxidases (PAOs) can be classified into four clades20,38,39. Our study revealed that barley genome contains nine genes coding for polyamine oxidase (PAO) with members classifying into clades II, III, and IV but not into clade-I (Supplementary Fig. S13). Plant PAOs belonging to clade-I seem to catabolize PAs in the BC-type reaction and are localized in the cytoplasm. As none of the barley PAOs belongs to clade-I, it can be assumed that barley PAOs do not have this characteristics with respect to reaction type and localization. The clade-III PAOs are also involved in degradation of PAs in the BC reaction and found to be localized in cytoplasm37,40,61. One barley PAO gene (HvPAO5) falls into this group. Furthermore, clade-III genes of Arabidopsis and rice are intron-less genes37,61,62, so is the HvPAO5. The Arabidopsis member of clade-III, AtPAO5, has been extensively characterized61. Phenotypes of AtPAO5 knock-down mutants (Atpao5) have also higher T-Spm contents than wild-type plants. In our analysis of barley, thermospermine synthase-like (tSPMS/ACULIS5-like) gene was not identified. The evolutionary studies of PAO genes in this clade might be of interest since these genes are intron-less and code for T-Spm specific PAOs. Clade-IV members are also of BC-type but are localized in peroxisomes38,63,64. Three barley PAOs (HvPAO4, HvPAO7 and HvPAO8) belong to this clade, suggesting that they play similar roles as clade-IV members of other species. The clade-II members catabolize PAs differently in the TC-type reaction, and localize either in the vacuole41,65 or in the apoplast37,66,67. In barley, five PAOs (HvPAO1, 2, 3, -6 and − 9) belong to clade-II type PAOs. In previous reports, barley HvPAO1 and maize ZmPAO1 (Clade-II PAOs), were found to be induced by wounding68, while rice OsPAO6 was upregulated by JA (jasmonate) treatment67. It seems that these PAOs (clade-II) originated from monocotyledonous plants, since they are not found in dicot plants such as Arabidopsis and tomato39. Thus, barley has both BC-Type and
TC-Type PAOs, whereas Arabidopsis and tomato plants do not contain clade-II TC-type PAO(s).
The gene duplication events (tandem/segmental) in plants are considered to be one of the main driving forces in the evolution and expansion of the gene family, and in the establishment of new protein functions69. Gene duplication analysis in this study revealed the tandemly and segmentally duplicated HvPMG genes. The HvPMG genes without duplicated sequences might have originated from different progenitors. The occurrence of the genes at the same chromosomal location implies a common origin, from which they might have evolved by a series of duplication events70. The gene duplication analysis suggests that tandem and segmental duplications may have played an important role in the expansion and evolution of the HvPMG gene family in plants, resulting in their structural and functional diversification. In genetics, the Ka/Ks ratio is used as an indicator of selective pressure acting on a protein-coding gene. In our study, the Ka/Ks ratios of all duplicated HvPMG gene pairs were less than one, which supports that evolution of genes may have occurred from intensive purifying selection pressure by natural selection during the evolutionary process. These results are consistant with the other evolutionary studies carried out in barley71,72. In addition, the estimated divergence time for HvPMGs in barley was found to be about 87.90 MYA for tandem duplicated genes and 54.85 MYA for segmental duplicated gene pairs. The majority of gene pairs were found to have diverged long before the divergence time of grass species (56–73 MYA)73,74. Barley and Arabidopsis share a common ancestor but they have diverged considerably since their separation around 140 million years ago. This could be attributed to structural difference in barley and Arabidopsis PMG genes75,76. The gene pair HvPAO7/ HvPAO8 diverged at 145.5864 MYA during the emergence of monocot and dicot plant species (140–150 MYA)76. Although the gene pair (HvCuAO2/ HvCuAO3) may represent a newly duplicated gene pair as it was estimated to diverge about 15.26 MYA, yet all the duplicated gene pairs were estimated to have originated before the divergence of the genus Hordeum (12–13 MYA)73,77. This indicates that the expansion of the HvPMG gene family in barley may be associated with gene duplication events.
The potential regulatory mechanisms controlling HvMPG gene expression both by analyzing the cis-regulatory elements (CREs) and the microRNA target sites in the promoter regions and the coding sequences of HvPMG genes were, respectively, explored. The variation of CREs is critical for phenotypic evolution in all organisms. In plants, broadly, regulatory regions are enriched for loci associated with phenotypic variation, for example, in maize78,79 and rice 80,81. Thus, analysis of CREs is critical to understanding the relationship between phenotype and genotype as they often dictate genes’ spatio-temporal expression82. Here, a total of 961 putative CREs were identified in the promoter regions, the 1000 bp upstream sequences, of HvPMG genes. Among all the identified CREs, CAAT-box and TATA-box in the group growth and development appeared to be the most abundant CREs and were commonly shared by most of the HvPMG genes. These elements are believed to determine the efficiency of transcription83, as CAAT-box is a common cis-acting element in promoter and enhancer regions, and TATA-box is a core promoter element around − 30 of transcription start site. Among others, some of the regulatory elements for stress response (MYB, MYC and STRE), light response (G-Box and GT1-motif) and hormone response (ABRE, CGTCA-motif and TGACG-motif) were common in most of the HvPMGs. The presence of a well conserved TATA-box and other putative cis-acting motifs responsive to light or auxin (G-box, AuxRE-box, TGACG-box and CCAAT-box) in promoter regions of two maize PAO genes are known84. Thus taken together, our results on the identification of cis-regulatory elements indicate that HvPMG genes could be transcriptionally regulated by multiple stimuli, and may participate in various plant metabolic processes as well as spatio-temporal expressions of identified isogenes.
A total of 9 H. vulgare miRNAs (hvu-miR) comprising target sites in 9 HvPMG genes were identified. The accessibility of the mRNA target site to small RNA has been identified as one important factor involved in target recognition. The UPE value representing the target accesibility in this study indicated a better miRNA-target binding85. The target mRNAs show almost-perfect or imperfect complimentarity with the miRNA in terms of mRNA cleavage or miRNA-direct translational inhibition, respectively86,87. In plants, miRNAs have been shown to be involved in various biotic (bacterial and viral pathogenesis) and abiotic stress responses such as oxidative, mineral nutrient deficiency, drought, salinity, temperature and cold88–90. Many reports have shown that miRNAs regulate responses of barley to different stress conditions90–93. Previous studies have shown that some of the identified miRNAs in this study were involved in both abiotic and biotic stress responses. For example, miRNA hvu-miR5049b is up-regulated under drought conditions [Hackenberg et al.94, while upregulation of hvu-miR6196 occurs during salt adaptation of the autopolyploid Hordeum bulbosum [Liu et al.95, and differential expression of hvu-miR6180 fungal stress in wheat has also been reported [Inal et al.96. Thus, exploring the role of miRNAs in HvPMG gene functions in response to various stresses would be of great interest.
The regulatory PPI network for the barley PA metabolic pathway genes indicated considerable interactive networks among the proteins involved in aldehyde dehydrogenase family, Aldedh domain-containing protein, ornithine aminotransferase, Cn hydrolase domain-containing protein, arginase family, PNP_UDP_1 domain-containing protein, Aminotran_1_2 domain-containing proteins, and S-adenosylmethionine synthase 4. Many of these proteins were previously shown to be involved in plant development and stress responses. For example, aldehyde dehydrogenases (ALDH), a family of enzymes involved in plant metabolism and aldehyde homeostasis to eliminate toxic aldehydes, are expressed in response to stress conditions such as high temperature, high salinity, dehydration, oxidative stress, or heavy metals in Arabidopsis97, S. tuberosum and N. benthamiana98, and other plant species99,100.
In plants, the expression of gene isoforms varies depending upon the plant tissue, developmental stage and environmental conditions101. The gene expression analysis in this study revealed that many HvPMGs are expressed in a redundant manner in different tissues during developmental stages in barley, supporting the idea that PA metabolism genes are involved in various tissues during all developmental processes in all living organisms17,102,103. The expression of HvODC1 and 2 was found only in roots which suggests that barley plants utilize the ODC pathway to produce Put only in roots, while in other organs the alternate ADC pathway is active. HvSAMDC3, HvCuAO7, HvPAO4 and HvSPMS1 were expressed in all the developmental stages indicating that they are involved in processes related to plant growth and development. Our results with barley are consistent with other studies of PA metabolism genes. The fact that HvADC1 was up-regulated during heat-drought, salt and simulated drought stress condition while HvSAMDC2 was slightly up-regulated in simulated drought and cold stress conditions is indicative of the fact that PA metabolism genes play an important role in responses to various abiotic stresses. This is consistent with the presence of putative cis-acting elements in the promoter region of PA biosynthetic genes including ADC and SAMDC.
It is known that PAs are anti-senescence in nature104. One such validation has come from genetic dissection of leaf senescence models, including dark-induced leaf senescence46,105,106. These studies indicate that PA catabolism can play a central role in metabolic reprogramming, directing a senescing leaf toward programmed organ death. Thus, depending upon which direction PA metabolism takes, synthesis/accumulation or catabolism that generates H2O2, the plant will either grow or senesce, respectively. DILS is a barley crop model for early and late events as well as for the identification of the critical time limit for reversal of the senescence process that prevents leaves from reaching the cell death phase. The efficiency of regulation of the senescence process is a sign of the vitality of senescing cells, which at each stage must maintain their ability to maintain homeostasis46,47,106. A critical moment in the model that determines the point of no return has been identified46 but the mechanism of its control is still unknown. Some suggestions regarding the PA metabolism gene expression pattern in senescing leaf have been provided by microarray data106. The leaf senescence-associated changes in gene transcripts involved in PA metabolism were also a part of the present study. Clearly, expression of SPDS/SPMS gene family transcripts’ in barley leaf during senescence showed significant changes. The data presented in this study suggests that the HvSPDS1, HvSPMS1, HvSPMS2 and HvSAMDC2 gene isoforms involved in the biosynthesis of PA metabolism are the key genes in polyamine metabolism that may condition senescence-dependent metabolic reprogramming. Interestingly, HvPAO8, involved in the catabolism of PAs, responded significantly during senescence and the phylogenic study showed that HvPAO8 is also a back-conversion type gene. Whether HvPAO8 is active in back-conversion during the senescence process remains to be determined. Interestingly, a correlation between the levels of HvPAO8 and HvSPMS2 transcripts and dark-induced leaf senescence was detected in this study and needs to be further tested.
Genetic mechanisms that lead to stress-induced senescence and delineate processes involved in either delaying or accelerating senescence are important to be deciphered. Development of transgenic barley plants that are defective in specific PA metabolic genes need to be generated via the ubi-overexpression, RNAi approaches or CRISPR/Cas9107–109. This should allow gaining “anti-aging” or “pro-aging” phenotypes as an important intervention. Designing a strategy to systematically knockout the genes involved in the biosynthesis, degradation and sequestering of PAs could provide an array of plants with different senescence phenotypes. In order to have plants that can be grown without restrictions under field conditions, one possible alternative is to select promising CRISPR/Cas9 mutations and mimic them by TILLING screening of a classical mutant collection.