Identification of the UMP synthase gene by establishment of uracil auxotrophic mutants and the phenotypic complementation system in the marine diatom, Phaeodactylum tricornutum 1 .

Uridine-5´-monophosphate synthase (UMPS), the critical step of de novo pyrimidine biosynthesis pathway, which is a housekeeping plastid process in higher plants, was investigated in a marine diatom, the most crucial primary producer in the marine environment. A mutagenesis using an alkylation agent, N-ethyl-N-nitrosourea (ENU) was carried out to the marine diatom Phaeodactylum tricornutum . Cells were treated with 1.0 mg mL -1 of ENU and were screened on agar plates containing 100-300 mg L -1 5-fluoroorotidic acid (5-FOA). Two clones were survived the selection and were designated as Requiring Uracil and Resistant to FOA (RURF) 1 and 2. The fifty-percent-effective concentration (EC 50 ) of 5-FOA on growth of RURF1 was about 5 mM, whereas that in wild-type cells was 30 μ M. The ability to grow in the absence of uracil was restored by a P. tricornutum gene which potentially encoded UMPS or the human umps gene, HUMPS . Because the P. tricornutum gene was able to restore growth in the absence of uracil it was designated as ptumps encoded a major functional UMPS in P. tricornutum . RNA interference (RNAi) to the ptumps targeting the 5´ region of the ptumps resulted in an occurrence of a clear RURF phenotype in P. tricornutum . This RNAi phenotype was reverted to wild type by insertion of the HUMPS , confirmed that the ptumps encodes UMPS. The results showed the direct evidence of occurrence of novel-type UMPS in a marine diatom and also revealed the potential usage of this gene silencing and complementation system for molecular tools for this organism.


INTRODUCTION
Diatoms are an abundant and diverse group of microalgae that are ubiquitous in marine and freshwater environments (Werner, 1977). It has been shown that marine diatoms are responsible for about one fifth of global primary production (Trégure et al., 1995;Falkowski et al., 2000), indicating that they play critical roles in global circulations of carbon and other elements. This relatively recent discovery of the importance of marine diatoms was followed by determinations of whole genome sequences of marine diatoms Thalassiosira pseudonana and Phaeodactylum tricornutum (Armbrust et al., 2004;Bowler et al., 2008).
Due to the comparative analysis of these sequences, it was shown that diatom genomes are mixtures composed of wide range of gene sequences related to those in animals, green algae, red algae, alfa-proteobacteria, and Chlamydia as well as many other genes unique to diatoms (Montsant et al., 2005;Moustafa et al., 2009). This indicated that diatom genomes have been evolved through redundant symbiotic events and established a unique combination of genome structures and metabolisms, which presumably occurred generally in evolutionary histories of the Chromalveolates supergroup (Keeling, 2004).
De novo pyrimidine biosynthesis pathway consists of six emzymatic reactions that react from relatively simple substrates, HCO 3 1 1 C-terminal side (Fig. 4, A and B). Such inverted alignments of these domains were also observed in umps genes in Kinetoplastae, Parabodo caudatus and Trypanosoma brucei (Fig. 4,B) and Oomycetes, Phytophthora sp. (Supplemental Fig. S2). All UMPS shown in Figure   4B have the conserved amino acid D, T, G for the domain of purine/pyrimidine phosphorybosyl transferase (PPRT, which includes OPRT), and D, K, D, I, T for ODC domain (Fig. 4, B, highlighted black). The amino acid sequences E/D and adjacent D in PPRT domain and D, K, D, T in ODC domain (Fig. 4, B, asterisks) were reported to be necessary for activities (Hershey and Taylor, 1986;Kimsey and Kaiser, 1992).
Phylogenetic analysis for these two domains indicated that such inverted order of OPRT and ODC domains in UMPS sequence are specifically observed in classes of Oomycetes and Kinetoplastea which respectively belong to the supergroups, Chromalveolates and Excavates (Supplemental Figs. S2,S3). The OPRT domain of PtUMPS likely belongs to a subclade closely related to Oomycetes but distant from that of mammals and Kinetoplastea (Supplemental Fig. S2). In contrast, the ODC domain of PtUMPS seemed to constitute a unique clade with those from both Oomycetes and Kinetoplastides (Supplemental Fig. S3).
1 3 possess any of these 16 SNPs (Fig. 6, B, asterisks). Besides the full-length amplification of ORF (1-1957), inner sequences of the ptumps ORF in RURF1 genome was also amplified with two sets of primers targeting nucleotides 1008-1593 and 312-1603. But in any case, SNP was not detected (data not shown), strongly suggesting a partial or a total lack of one of ptumps alleles.

Establishment of RURF phenotype by RNA interference
Occurrences of chemically induced RURF phenotype and successful complementation of the phenotype by single gene insertion of the ptumps or the HUMPS, strongly suggested that the ptumps encoded a functional UMPS and umps genes from distant origin can also work in P. tricornutum. These facts prompted us to obtain the RURF phenotype by silencing the ptumps gene by RNAi technique. A vector pFcpApptumpsRNAi1, which had RNAi fragment for the 5 -terminus region of the ptumps cDNA (Fig. 7, A), was introduced into wild-type cells of P. tricornutum. Transformed cells were screened on plates containing 300 mg L -1 of 5-FOA and 50 mg L -1 of uracil. Sixteen clones out of 10 8 potential transformants were survived on 5-FOA in the presence of uracil. The growth characteristics of one clone from these 16 apparent RNAi cells were determined in liquid F/2ASW medium without or with added 200 mg L -1 5-FOA and 50 mg L -1 uracil. As a result, this clone revealed a clear resistance to 5-FOA and growth dependency to uracil (Fig. 7, B). Figure 8A indicates transcript levels of the ptumps in wild-type cells, RURF1, and the transformant with pFcpApptumpsRNAi1 (ptumps-i). The transcript level of the ptumps was undetectable in the ptumps-i cells, in sharp contrast to that in wild-type cells (Fig. 8, A).

Complementation of ptumps-i cells by insertion of the HUMPS gene
The HUMPS gene was introduced into ptumps-i cells using pFcpAphumps vector, which is equipped with Zeocin resistant cassette. Transformed cells were initially screened on 100 mg L -1 Zeocin and 50 mg L -1 uracil. Sixteen clones out of 10 8 cells were survived on the initial screening. After inoculating to a minimal F/2ASW agar plate with Zeocin, 4 clones revealed a clear wild-type phenotype, which did not require uracil and sensitive to 5-FOA.
These revertant cells were transferred to liquid medium of minimal F/2ASW without or with added 200 mg L -1 5-FOA and 50 mg L -1 uracil. Revertants (ptumpsi-HUMPS) grew in the ordinary F/2ASW with an equivalent rate to that of wild-type cells, whereas RURF1 cells and ptumps-i cells failed to grow (Fig. 8, B). In sharp contrast to this, growth of wild-type cells and the ptumpsi-HUMPS cells were severely suppressed in the presence of 5-FOA, whereas RURF1 and ptumps-i cells showed clear tolerances to 5-FOA in the presence of uracil ( Fig. 8, C).

DISCUSSION
Chemical mutagenesis induced by ENU in this study successfully allowed us to obtain two RURF mutants but the frequency of mutation was extremely low. We tentatively ascribe this low mutation ratio to the ploidy of diatom cells in the experimental culture. Although sexual reproduction processes have been observed microscopically in some diatoms, the experimental condition to control the sexual cycle and ploidy is not established in diatoms (Armbrust, 2009). In P. tricornutum, it was suggested that the cells did not represent an www.plantphysiol.org on August 9, 2017 -Published by Downloaded from Copyright © 2011 American Society of Plant Biologists. All rights reserved. alteration of ploidy (Darley 1968) and the cells seem to spend most of their life cycle as a diploid (Geitler 1935). In fact, sequence of the ptumps ORF of wild-type cells in this study strongly suggested that cells at exponential growth phase possess two different alleles of the ptumps (thus heterozygous) (Fig 6B). In this study, we always used cells at the exponential growth phase, which most probably at a diploidal stage. It is thus probable that this low mutation efficiency (2.0 x 10 -6 ) in this study could be due to the diploidal state of diatom chromosome in experimental culture.
The results in the present study indicated that UMPS activity in RURF1 cells is largely defective, but cells still have a sensitivity to 5-FOA, which is about 150 times lower than that of wild-type cells. This implies that RURF1 cells maintain the activity to produce trace amounts of 5´-UMP from orotate. Indeed, transcript levels of the ptumps in RURF1 cells were significantly lower than that of wild-type cells but still existed (Fig. 6, A). The comparison of ptumps ORF sequences between genomic DNA from wild-type and RURF1 cells strongly suggested an occurrence of deletion of the total or a partial sequence of one of alleles of the ptumps in RURF1 mutant (Fig. 6, B). ENU is known to be a strong alkylation agent and thus causes primarily single nucleotide mutations, but it was also pointed out that 1 6 functional. Six amino acids in the deduced polypeptide sequence were changed by the SNPs (data not shown). Although these 6 amino acid replacements occurred at locations distant from either ODC or OPRT domain (data not shown), the remarkable difference in 5-FOA tolerance between wild-type and RURF1 cells may suggest that PtUMPS encoded by the second allele might be functionally defective. There are also possibilities of occurrences of mutations in the promoter region of the ptumps and also in genes encode trans-acting factors for the ptumps promoter, however, these aspects remain untested.
The entire sequence of PtUMPS did not resemble those identified in higher eukaryotes and yeast, but showed about 40% similarities with those in the Bodonid, P. caudatus and the Kinetoplastid, T. brusei; about 50% with those in the Oomycetes, Phytophthora sojae and the results of the present study suggest a significant difference in regulations of supply of uridine nucleotide in diatoms compared to that of higher plants, which may also suggest differences in modes of biosynthesis of polysaccharide, lipid, and secondary metabolites in diatoms. It was recently predicted from genome database analysis that P. tricornutum cells most probably lack in complete oxidative pentose phosphate pathway in the plastid and µmol m -2 s -1 and were allowed to form colonies for 2-3 weeks. The number of colonies was counted and relative survival ratios of ENU-treated cells to that of non-treated cells were calculated. The concentration of ENU, which gave 20 to 30% lethality was employed for the mutagenesis.
The minimal effective concentration of 5-FOA for growth inhibition of wild-type cells of P.
tricornutum was determined. Wild-type cells of initial density of 5, 2.5, 1.25 and 0.625 × 10 5 cells mL -1 were plated as plaques on the F/2ASW agar plates containing 5-FOA at final concentrations of 100, 200 or 300 mg L -1 . Cells were cultured for 2-3 weeks and growth inhibition by 5-FOA to wild-type cells of P. tricornutum were assessed by visual rating.
Wild-type cells of P. tricornutum were mutagenized by 1 mg mL -1 ENU for 30 min and plated on F/2ASW agar plate containing 100 mg L -1 of 5-FOA and 50 mg L -1 of uracil. After 2 to 3 weeks of culture, 5-FOA resistant colonies were inoculated to new F/2ASW plates containing 300 mg L -1 of 5-FOA and 50 mg L -1 uracil and were allowed to grow for 2 to 3 weeks at 20˚C under continuous illumination of PPDF of 50 µmol m -1 s -1 .

RNAi constructs for silencing the ptumps
Two fragments of the ptumps gene were amplified by PCR from the pFcpApptumps using the following primers: ptumps sense fw (5´-GGAATTCATGGCCACCCCCTCTTTT-3´) which contained a EcoRI site, ptumps anti fw (5'-        533-537, 953-957, 1357-1368, 1414-1418, 1437-1441, 1448-1452, 1480-1501, 1712-1721, and 1829-1833 relative to the initial nucleotide of the ORF. Upper law and lower law indicate DNA sequences from RURF1 and wild-type cells, respectively. The SNPs in ptumps alleles are indicated by asterisks. Values are mean ± SD of three separate experiments.          Upper law and lower law indicate sequences from RURF1 and wild-type cells, respectively.
SNIPs of the ptumps allele are indicated by asterisks.