Alleles of HLA-A, B, C, DRB1, DQB1, and DPB1 loci were fully determined in 117 healthy Japanese. A*2402, A*3303, A*1101, A*0201, B*4403, B*5201, Cw*0102, Cw*1403, Cw*0304, Cw*0702, Cw*0801, and Cw*1202 showed frequencies of over 10%. Multi-locus haplotype frequencies were estimated by the maximum likelihood method. Strength of association between C and B loci was comparable with that between DRB1 and DQB1 loci. Alleles unidentified by a serological method and having very similar nucleotide sequences (A2: A*0201, A*0206, A*0207, B61: B*4002, B*4006) were carried by different haplotypes. Several frequent five-locus haplotypes were identified including A*3303-Cw*1403-B*4403-DRB1(*)1302-DQB1(*)0604, and A*2402-Cw*1202-B*5201-DRB1(*)1502-DQB1(*)0601. These sequence-based haplotypes corresponded to serology-based common haplotypes which have already been described in Japanese. These findings indicate that common HLA haplotypes consist of particular sets of HLA alleles and that these haplotypes have been conserved through recent human evolution.

A polymorphic gene, MIC-A, is one of the MIC family of genes which is composed of a group of homologous genes interspersed in the class III and class I regions of the major histocompatibility complex. MIC-A is located 46 kilobases (kb) centromeric of HLA-B, and is preferentially expressed in the epithelial cells and intestinal mucosa. Recently, MIC-A and the closely related MIC-B were reported as the molecules that conferred specificity in the recognition by the Vdelta1gammadeltaT cells. In the present study, polymorphic exons 2, 3, and 4 of the MIC-A gene were analyzed using the polymerase chain reaction-single-strand conformation polymorphism method. The number of patterns found in exons 2, 3, and 4 were 5, 6, and 4, respectively, in 114 healthy Japanese subjects. Eight MIC-A alleles were observed in Japanese individuals, among which one, tentatively named MIC-AMW, has not previously been reported. There was a strong linkage disequilibrium between MIC-A and HLA-B loci: each MIC-A allele showed strong association with a particular HLA-B group. In contrast, B*3901 showed association with multiple MIC-A alleles. Furthermore, the existence of a MIC-A-MIC-B null haplotype, which is associated with HLA-B*4801, was identified. In this haplotype, a large-scale deletion (of approximately 100 kb) including the entire MIC-A gene was indicated and the MIC-B gene possessed a stop codon.

Stress-inducible MICA (MHC class I chain-related A) is known to bind to NKG2D, which is one of the natural killer (NK) cell receptors, and plays a role in immune surveillance. We have reported that a MICA-MICB null haplotype is in linkage disequilibrium with HLA-B*4801 in the Japanese population. In the haplotype, an approximately 100-kb deletion, including the entire MICA gene, was observed and MICB possessed a premature stop codon. In this study, a multiplex polymerase chain reaction (PCR) method was developed for detecting the MICA deletion. MICB alleles were typed by PCR-single-strand conformation polymorphism (SSCP) method and direct sequencing. The frequency of the MICA-MICB null haplotype was 3.7% on the average, and was strongly associated with HLA-B48 in seven East Asian populations. It was presumed that the stop codon of MICB gene generated after the large-scale deletion. The wide distribution of the null haplotype at polymorphic frequencies suggests that the haplotype has been conservatively maintained because of some selective advantage.

Purification of docosahexaenoic acid (DHA) was attempted by a two-step enzymatic method that consisted of hydrolysis of tuna oil and selective esterification of the resulting free fatty acids (FFA). When more than 60% of tuna oil was hydrolyzed with Pseudomonas sp. lipase (Lipase-AK), the DHA content in the FFA fraction coincided with its content in the original tuna oil. This lipase showed stronger activity on the DHA ester than on the eicosapentaenoic acid ester and was suitable for preparation of FFA rich in DHA. When a mixture of 2.5 g tuna oil, 2.5 g water, and 500 units (U) of Lipase-AK per 1 g of the reaction mixture was stirred at 40 o C for 48 h, 83% of DHA in tuna oil was recovered in the FFA fraction at 79% hydrolysis. These fatty acids were named tuna-FFA-Ps. Selective esterification was then conducted at 30 o C for 20 h by stirring a mixture of 4.0 g of tuna-FFA-Ps/lauryl alcohol (1:2, mol/mol), 1.0 g water, and 1,000 U of Rhizopus delemar lipase. As a result, the DHA content in the unesterified FFA fraction could be raised from 24 to 72 wt% in an 83% yield. To elevate the DHA content further, the FFA were extracted from the reaction mixture with n-hexane and esterified again under the same conditions. The DHA content was raised to 91 wt% in 88% yield by the repeated esterification. Because selective esterification of fatty acids with lauryl alcohol proceeded most efficiently in a mixture that contained 20% water, simultaneous reactions during the esterification were analyzed qualitatively. The fatty acid lauryl esters (L-FA) generated by the esterification were not hydrolyzed. In addition, L-FA were acidolyzed with linoleic acid, but not with DHA. These results suggest that lauryl DHA was generated only by esterification.

Steryl esters of long-chain fatty acids have waterholding properties, and polyunsaturated fatty acids (PUFA) have various physiological functions. Because steryl ester of PUFA can be expected to have both features, we attempted to synthesize steryl esters of PUFA by enzymatic methods. Among lipases used, Pseudomonas lipase was the most effective for the synthesis of cholesteryl docosahexaenoate. When a mixture of cholesterol/docosahexaenoic acid (3:1, mol/mol), 30% water, and 3000 units/g of lipase was stirred at 40°C for 24 h, the esterification extent attained 89.5%. Under the same reaction conditions, cholesterol, cholestanol, and sitosterol were also esterified efficiently with docosahexaenoic, eicosapentaenoic, arachidonic, and γ-linolenic acids.Paper no. J9028 in JAOCS 76, 713-716 (June 1999).