The MUR1 gene of Arabidopsis thaliana encodes an isoform of GDP-D-mannose-4,6-dehydratase, catalyzing the first step in the de novo synthesis of GDP-L-fucose

Abstract

GDP-L-fucose is the activated nucleotide sugar form of L-fucose, which is a constituent of many structural polysaccharides and glycoproteins in various organisms. The de novo synthesis of GDP-L-fucose from GDP-D-mannose encompasses three catalytic steps, a 4,6-dehydration, a 3,5-epimerization, and a 4-reduction. The mur1 mutant of Arabidopsis is deficient in L-fucose in the shoot and is rescued by growth in the presence of exogenously supplied L-fucose. Biochemical assays of the de novo pathway for the synthesis of GDP-L-fucose indicated that mur1 was blocked in the first nucleotide sugar interconversion step, a GDP-D-mannose-4,6-dehydratase. An expressed sequence tag was identified that showed significant sequence similarity to proposed bacterial GDP-D-mannose-4,6-dehydratases and was tightly linked to the mur1 locus. A full-length clone was isolated from a cDNA library, and its coding region was expressed in Escherichia coli. The recombinant protein exhibited GDP-D-mannose-4,6-dehydratase activity in vitro and was able to complement mur1 extracts in vitro to complete the pathway for the synthesis of GDP-L-fucose. All seven mur1 alleles investigated showed single point mutations in the coding region for the 4,6-dehydratase, confirming that it represents the MUR1 gene.

Figure 1

Reaction scheme for the de novo synthesis of GDP-l-fucose from GDP-d-mannose, involving 4,6-dehydratase (step 1), which produces a 4-keto-6-deoxy intermediate, 3,5-epimerization (step 2), and the final NADPH-dependent reduction of the 4-keto group (step 3). Expected products from the chemical reduction and hydrolysis of the 4-keto-6-deoxy intermediate, are also included.

Figure 2

Assay of enzyme activities in the de novo synthesis of GDP-l-fucose. (A) Thin layer chromatogram of nucleotide diphospho-sugar products from the in vitro assay, following incubation of GDP-d-[¹⁴C]mannose with wild-type protein extract (lane 1) or mur1 protein extract (lane 2). (B) Thin layer chromatogram of the monosaccharides from the acid hydrolysis of in vitro assay products using GDP-d-[¹⁴C]mannose (lanes 1–4) or a mixture of GDP-d-[¹⁴C]mannose and GDP-4-keto-6-deoxy-d-[¹⁴C]mannose (lanes 5–7) as substrates. Lane 1, wild-type protein extract with exogenous NADPH; lane 2, mur1 protein extract with exogenous NADPH; lane 3, wild-type protein extract without exogenous NADPH; lane 4, mur1 protein extract without exogenous NADPH; lane 5, mur1 protein extract in the presence of NADPH; lane 6, same as lane 5 but using heat-inactivated mur1 protein; and lane 7, same as lane 5 but using wild-type extract. Samples in lanes 3 and 4 were reduced with NaBH₄ before hydrolysis. Positions of authentic standards are shown. Man, mannose; Man-1-P, mannose-1-phosphate; Fuc, fucose; Rha, rhamnose; and dTal, 6-deoxy-talose.

Figure 3

Southern blot analysis of total Arabidopsis DNA at high and low stringency using ³²P-labeled GMD2 as a probe. (A) Digested Columbia DNA probed at high stringency (68°C hybridization, 65°C washes). (B) Digested Columbia DNA probed at low stringency (58°C hybridization, 55°C washes).

Figure 4

Nucleotide sequence of the MUR1 cDNA and derived amino acid sequence of the MUR1 protein. Nucleotide changes and predicted amino acid substitutions in each of seven mur1 alleles are indicated. The consensus sequence for recognition of the initiation codon and a putative polyadenylylation signal are underlined.

Figure 5

Alignment between the derived MUR1 amino acid sequence and putative GDP-d-mannose-4,6-dehydratases. Identical residues are shaded in gray and conserved residues are boxed. Positions of the amino acid substitutions from seven mur1 alleles are indicated above the Arabidopsis sequence. A.t., A. thaliana; M.t., Mycobacterium tuberculosis; P.a., Pseudomonas aeruginosa; PBCV, Paramecium bursaria Chlorella virus; Y.e., Yersinia enterocolitica; V.c., Vibrio cholerae; and E.c., E. coli. Sequence alignments were done with clustal w and viewed with seqvu.

Figure 6

SDS/PAGE analysis of crude extracts from E. coli harboring the pET11d vector with or without the MUR1 coding region. Lane 1, protein extracted from E. coli containing a pET11d vector without an insert; and lane 2, protein extracted from E. coli containing a pET11d vector with the MUR1 coding region. STD, prestained protein molecular weight standards.

Figure 7

Assay of the MUR1 protein for GDP-d-mannose-4,6-dehydratase activity and reconstitution of the l-fucose biosynthetic pathway in mur1-derived protein extracts. Lane 1, protein from E. coli expressing the MUR1 gene, incubated with GDP-d-[¹⁴C]mannose; lane 2, protein from E. coli transformed with pET11d vector minus the MUR1 insert; lane 3, same as lane 1 but with mur1-derived protein extract added; and lane 4, same as lane 2 but with mur1-derived protein extract added. All final products were hydrolyzed and reduced as described earlier.