Molecular characterisation of the gene encoding [Delta 1]-Pyrroline-5- Carboxylate Reductase isolated from Arabidopsis thaliana (L.) Heynh.

Thesis (M.Sc.)-University of Natal, Pietermaritzburg, 1995.

In Arabidopsis thaliana (L.) Heyhn, the size of the pool of free proline increases up to 27-fold in response to osmotic stress. The magnitude of this accumulation is dependent upon the rate of imposition of the stress. Numerous reports have suggested a role for proline accumulation as a general adaptation to environmental stress. However, controversy surrounds the beneficial effect of proline accumulation in plants under adverse environmental conditions. Stress-induced proline accumulation in plants occurs mainly by de novo synthesis from glutamate. The final and only committed step of proline biosynthesis in plants is catalysed by Δ¹-pyrroline-5-carboxylate reductase (P5CR). The sequence of an incomplete 999 bp cDNA encoding P5CR from A. thaliana was determined. This enabled a preliminary molecular study of the structure and function of both the gene and the corresponding enzyme. The 999 bp cDNA insert in the clone Y AP057 was sequenced on the sense and antisense strands following subcloning of four sub-fragments in appropriate orientations. Comparison with known plant P5CR sequences revealed that Y AP057 does not encode the first 23 N-terminal amino acids of P5CR from Arabidopsis. However, it does encode the remaining 253 amino acid residues of Arabidopsis P5CR The cDNA Y AP057 is complete on the 3' end as indicated by the presence of a poly(A) tail. The nucleotide sequence determined shows complete homology to the corresponding exons of the genomic copy of a bona fide gene encoding P5CR in A. thaliana (Verbruggen et al, 1993). The only difference observed between the sequence of Y AP057 and that of a cDNA sequenced by these workers is that polyadenylation was initiated seven nucleotides earlier in Y AP057 than in the sequence of the published cDNA. Genomic Southern analysis suggests the presence of only a single copy of the gene encoding P5CR in Arabidopsis. Restriction mapping and sequencing the ends of another incomplete Arabidopsis P5CR cDNA clone FAFJ25 (664 bp) indicated that the regions sequenced were completely homologous to the corresponding portions of Y AP057. Analysis of codon usage in the Arabidopsis gene encoding P5CR revealed it to closely resemble the consensus pattern of codon usage in A. thaliana. This suggests that the gene is moderately. expressed. Expression of the gene encoding P5CR in Arabidopsis is not likely to be subject to translational control. Although P5CR from A. thaliana has a fairly high composition of hydrophobic amino acid residues, it does not possess any stretches of hydrophobic amino acids of sufficient length to act as membrane-spanning domains or to anchor the enzyme in a membrane. Neither does it contain an N- terminal leader sequence capable of directing it to either the plastid or mitochondrion. The enzyme therefore appears to be cytosolic. The nucleic acid and deduced amino acid sequences of Arabidopsis P5CR were compared with those from·eleven other organisms for which P5CR sequences are currently available. Except among the three different plants examined, P5CR sequences displayed less identity at the amino acid level than at the nucleotide level. The deduced amino acid sequence of Arabidopsis P5CR exhibits high similarity to the corresponding genes and amino acid sequences of P5CR from soybean and pea. Lower but significant similarity was observed to the amino acid sequences of P5CRs from human, Saccharomyces cerevisiae and the bacteria Escherichia coli, Pseudomonas aeruginosa, Thermus thermophilus, Mycobacterium leprae; Treponema pallidum and Methanobrevibacter smithii. Similarity was also observed to the translational product of a gene from Bacillus subtilis with high homology to the E. coli proC gene. However, construction of a phenogram indicating the relatedness of the various P5CR enzymes suggests that sequence analysis of this enzyme is not a good indicator of evolutionary relatedness of organisms from different biological kingdoms. Multiple alignment of the twelve known P5CR sequences indicated homology between the sequences across their entire lengths. Homology was particularly high in the C-terminal portions of the P5CRs studied. It is speculated that this region may be of importance in binding of the substrate Δ¹-pyrroline-S-carboxylate (P5C). Another region displaying high sequence conservation was found in the central portion of all P5CRs. All P5CRs studied, with the exception of PSCR from T. pallidum contained an N-terminal domain capable of binding a nicotinamide dinucleotide cofactor. Comparison of this region with consensus sequences for NADH and NADPH binding sites in proteins suggests that NADPH is the preferred reductant used by P5CRs from plants and human. In contrast, the N-terrninal domains of P5CRs from S. cerevisiae, M smithii, T. thermophilus and M leprae display greater similarity to a consensus NADH-binding site. The definite preference of plant P5CRs for NADPH in comparison with NADH suggests that P5CR may be involved in regulating the redox potential within plant cells and that this step in proline biosynthesis from glutamate may be of importance in overall metabolic regulation. Three amino acid residues are universally conserved in all P5CRs studied. All are found within blocks of high sequence similarity. These residues are likely to be of importance in the structure or catalytic mechanism of P5CR. A number of other residues are common to several of the enzymes examined. These may also be of importance in subsequent manipulation of Arabidopsis P5CR at the molecular level. Prediction of the putative secondary structures of A. thaliana, soybean, pea, human and E. coli indicated a high degree of similarity between the enzymes. This was particularly evident in the region of the putative P5C-binding domain. Considerable similarity exists in hydrophobicity profiles of P5CRs from these five organisms. Proline levels in reproductive organs of unstressed Arahidopsis plants were considerably higher than those in vegetative tissues. This suggests differential expression of enzymes involved in proline metabolism in these organs. In situ hybridisation studies indicated an increase in levels of mRNA transcripts encoding P5CR in stem tissues in response to water deprivation stress. Regulation of levels of mRNA transcript encoding P5CR in Arabidopsis therefore appears to be an osmotically sensitive process. Furthermore, this accumulation of transcript occurred in a tissue-specific manner. In particular, an increase in levels of transcript encoding P5CR was observed in the cortical parenchyma, phloem, vascular cambium and pith parenchyma in the vicinity of the protoxylem. The significance of these findings in contributing to a better understanding of the role of proline in adaptation to environmental stress is discussed.