Single Nucleotide Polymorphisms in Homocysteine Metabolism Pathway Genes. With Hyperhomocysteinemia

Transcription

1 Single Nucleotide Polymorphisms in Homocysteine Metabolism Pathway Genes Association of CHDH A119C and MTHFR C677T With Hyperhomocysteinemia Jitender Kumar, PhD; Gaurav Garg, MSc; Arun Kumar, PhD; Elayanambi Sundaramoorthy, MBBS; Krishna Rao Sanapala, MSc; Saurabh Ghosh, PhD; Ganesan Karthikeyan, DM; Lakshmy Ramakrishnan, PhD; Indian Genome Variation Consortium; Shantanu Sengupta, PhD Downloaded from by guest on June 15, 2018 Background An elevated level of homocysteine (hyperhomocysteinemia) has been implicated as an independent risk factor for cardiovascular diseases. Deficiency of dietary factors like vitamin B 12, folate, and genetic variations can cause hyperhomocysteinemia. The prevalence of hyperhomocysteinemia in the Indian population is likely to be high because most Indians adhere to a vegetarian diet, deficient in vitamin B 12. In the background of vitamin B 12 deficiency, variations in genes involved in homocysteine metabolism might have a greater impact on homocysteine levels. Methods and Results We genotyped 44 nonsynonymous single-nucleotide polymorphisms (nssnps) from 11 genes involved in homocysteine metabolism and found only 14 to be polymorphic. These 14 nssnps were genotyped in 546 individuals recruited from a tertiary care center in New Delhi, India, and it was found that choline dehydrogenase (CHDH A119C) and methylenetetrahydrofolate reductase (MTHFR C677T) were significantly associated with plasma total homocysteine levels (P and P 0.001, respectively). These 2 SNPs were further genotyped in 330 individuals recruited from the same center, and the association remained significant even after increasing the sample size. Furthermore, we found the possibility of a significant interaction between vegetarian diet and the 2 polymorphisms that could explain the variation of homocysteine levels. We also genotyped all the polymorphic nssnps in apparently healthy individuals recruited from 24 different subpopulations (based on their linguistic lineage) spread across the country to determine their basal frequencies. The frequencies of these SNPs varied significantly between linguistic groups. Conclusion Vegetarian diet along with CHDH A119C and MTHFR C677T play an important role in modulating the homocysteine levels in Indian population. (Circ Cardiovasc Genet. 2009;2: ) Key Words: genetics homocysteine polymorphisms CHDH, MTHFR, Indian population An elevated level of homocysteine, a thiol amino acid, has been associated with several diseases and/or clinical conditions including neural tube defects, 1 Alzheimer disease, 2 end-stage renal disease, 3 schizophrenia, 4 non insulindependent diabetes, 5 etc. It has also been implicated as an independent and graded risk factor for cardiovascular disease. 6 In patients with coronary artery disease (CAD), homocysteine level is a significant predictor of mortality, independent of traditional risk factors. 7 Intracellular homocysteine metabolism is well regulated, and its concentration in circulation is 12 mol/l. However, deficiencies in the micronutrients like folate, vitamin B 12, and vitamin B 6 impair homocysteine metabolism leading to elevated intracellular and extracellular levels of homocysteine Apart from these micronutrients, polymorphisms in some of the genes involved in homocysteine metabolism also modulate homocysteine levels. 11,12 Clinical Perspective on p 606 Several studies have found the association of certain nonsynonymous single-nucleotide polymorphisms (nssnps) with homocysteine levels. However, a majority of these studies focused on 1 or 2 SNPs at a time. Only a few studies evaluated the association of multiple ( 10) nssnps with homocysteine levels in a comprehensive manner in a given population. In the largest study carried out so far, Fredriksen et al 12 genotyped a large population of Norway for 13 SNPs present in 1-carbon metabolism genes and found that methylenetetrahydrofolate reductase (MTHFR C677T and MTHFR A1298C), 5-methyltetrahydrofolate-homocysteine Received July 16, 2008; accepted August 21, From the Institute of Genomics and Integrative Biology (J.K., G.G., A.K., E.S., S.S.), Delhi, India; Indian Statistical Institute (K.R.S., S.G.), B.T. Road, Kolkata, India; and All India Institute of Medical Sciences (G.K., L.R.), New Delhi, India. The online-only Data Supplement is available at Correspondence to Shantanu Sengupta, PhD, Institute of Genomics and Integrative Biology, Mall Road, Delhi , India. shantanus@igib.res.in 2009 American Heart Association, Inc. Circ Cardiovasc Genet is available at DOI: /CIRCGENETICS

2 600 Circ Cardiovasc Genet December 2009 Downloaded from by guest on June 15, 2018 Figure 1. Homocysteine metabolism pathway showing the name of genes studied. methyltransferase (MTR A2756G), and cystathionine synthase (CBS 844_845ins68 and CBS C699T) are associated with homocysteine levels. Janosikova et al 13 apparently genotyped healthy individuals from Czechoslovakia in an attempt to obtain the frequency of 42 SNPs in genes involved in 1-carbon metabolism, but no attempt was made to evaluate their effect on homocysteine levels. The situation in India is even worse as there has been no attempt made to even analyze the basal frequencies of SNPs in genes involved in 1-carbon metabolism, despite the fact that India accounts for one sixth of the world population. Furthermore, a majority of the Indian population adheres to a vegetarian diet that predisposes them to vitamin B 12 deficiency. Four decades earlier, Satoskar et al 14 had clearly shown that vegetarians in India have low vitamin B 12 levels. This has been substantiated by several other studies. 11,15 Low vitamin B 12 levels are expected to result in hyperhomocysteinemia, and we and others have shown that homocysteine levels are high among the vegetarians in India. 11,15 We therefore believe that in the background of vitamin B 12 deficiency, polymorphisms in the genes that modulate homocysteine levels might have synergistic effects. In this study, we genotyped 44 nssnps present in 11 genes involved in homocysteine metabolism (Figure 1) to ascertain their role in modulating plasma homocysteine levels. We also determined the frequency distribution of the SNPs in at least 24 subpopulations spread across the country as a part of Indian Genome Variation project to establish the basal frequencies of these SNPs across the country. Methods Study Individuals In this study, we have used 2 different study groups: (1) the All India Institute of Medical Sciences (AIIMS) study group and (2) the Indian Genome Variation (IGV) study group. AIIMS Study This is a hospital-based study group. A total of 876 individuals (mainly Indo-European population from northern part of India) were recruited at the Department of Cardiology, AIIMS, New Delhi (between September 2004 and March 2007). Initially, 546 individuals (243 patients with CAD as confirmed by coronary angiography and 303 individuals with negative treadmill test considered as controls) were recruited (September 2004 to November 2005) to determine the effect of diet and polymorphisms on homocysteine levels. An additional 330 individuals from the same ethnicity (185 patients with CAD and 145 controls) were recruited in the second phase (December 2005 to March 2007) to validate those polymorphisms that were found to be associated with homocysteine in the first phase. Written consent was obtained from all the participants, and the study was conducted in accordance with the principles of the Helsinki Declaration, with the approval of the ethics committee. A questionnaire with relevant clinical information was filled out that included information about the individual s blood pressure, smoking habits, height, weight measurements, dietary status, milk consumption, etc. IGV Study For this study, healthy individuals from various subpopulations based on linguistic lineage and geographical locations were recruited to study the basal frequency of the polymorphisms in Indian population in general. The detailed selection criterion is explained elsewhere. 16 Oligonucleotides Oligonucleotide pairs to amplify all the regions covering 44 nssnps chosen as well as for minisequencing was designed using Primer Select module of DNASTAR software (supplemental Table I). DNA Isolation and Polymerase Chain Reaction Blood samples (fasting) were collected from volunteers into plain tubes (for serum) and tubes containing EDTA anticoagulant (for plasma). Serum and plasma was separated within an hour of collection and stored at 80 C until further analysis. Genomic DNA was isolated from blood samples using the modified salting-out method as described earlier 11 and stored at 20 C until further use. Polymerase chain reaction conditions were optimized for individual oligonucleotide pair by varying MgCl 2 concentrations and annealing temperatures. SNaPshot (Minisequencing) Forty-four nssnps were chosen for the study from the genes involved in homocysteine metabolism either from dbsnp or from the study by Janosikova et al. 13 These selected SNPs were scored using the SNaPshot ddntp primer extension kit (Applied Biosystems, Foster City, Calif), as described earlier. 11 Genotyping Strategy Forty-four nssnps chosen for this study were initially genotyped in a random pick of 95 DNA samples to check their polymorphic status. Only those SNPs that were found to be polymorphic were then genotyped in 546 individuals, and the association of these SNPs with homocysteine levels were ascertained. SNPs that were found to be significantly associated with homocysteine levels were further genotyped in an additional 330 individuals. Genotyping Using Sequenom MassARRAY and Illumina Golden Gate Assay Of 14 nssnps found to be polymorphic, 12 were further genotyped in a large number of apparently healthy individuals from the IGV project. Two SNPs (choline dehydrogenase CHDH T233G and 5-methyltetrahydrofolate-homocysteine methyltransferase reductase MTRR C1349G) failed at the assay-designing step and hence were not pursued. These individuals were recruited from various parts of the country (based on the geographical regions and linguistic lineages) as a part of the IGV project. Two SNPs from MTHFR gene (rs and rs ) were genotyped in 1800 individuals from 55 subpopulations using Sequenom genotyping platform (Sequenom, San Diego, Calif) as per manufacturer s instructions. The rest of the 10 SNPs were genotyped using Illumina golden gate assay platform (Illumina Inc, San Diego, Calif) in 550 individuals from 24 subpopulations according to the protocol. 17

3 Kumar et al Genetics of Hyperhomocysteinemia 601 Downloaded from by guest on June 15, 2018 Biochemical Analysis Plasma homocysteine was determined using high-performance liquid chromatography (Agilent 1100 series, Palo Alto, Calif), as described earlier. 11 A quality control sample (whose concentration was previously determined) was included along with standard curve with each batch to ensure the accuracy of homocysteine determination in unknown samples. Chemiluminescence competitive immunoassay (Immulite 1000 Analyzer, Siemens Medical Solutions Diagnostics, Flanders, NJ) was used to determine the vitamin B 12 concentrations following the manufacturer s instructions. Folate levels were estimated by using SimulTRAC-SNB radioassay (Diagnostics Division, Orangeburg, NY) as per manufacturer instructions. Statistical Analysis Nonparametric tests were used to compute statistical significance of difference in quantitative parameters between multiple groups as the parameters did not follow Gaussian distribution. Genotypic associations were performed under codominant as well as dominant and recessive model using the Mann Whitney or Kruskal-Wallis test, as appropriate depending on the number of genotypic groups. Allelic associations were performed under an additive allelic model using linear regression. De-Finetti program ( was used for checking whether genotypes conformed to Hardy Weinberg equilibrium. Because we performed association tests at multiple polymorphisms and under multiple genetic models, nominal P values do not provide strict statistical evidence of association, and we need to use a more stringent P value for our association findings to be more convincing. For this, we used a false discovery rate correction 18 to account for multiple testing. An overall false discovery rate of 0.05 was used to correct for multiple testing in our study. The q values corresponding to the significantly associated SNPs were calculated using the method of Storey 19 as implemented in the software Q value by Dabney and Storey ( The multiple testing for Hardy Weinberg equilibrium was done separately from those of association. We have checked the association of the SNPs with homocysteine levels only. A total of 52 association tests were performed on homocysteine levels, and the correction was used for 52 tests together comprising 4 models at each of 12 SNPs and 2 models at each of 2 SNPs where minor homozygotes were missing. SPSS version 10 for Windows was used as statistical tool for analyses. To test for the interaction effects between the significant SNPs and diet, we used a generalized linear model involving main and interaction effects as follows: yijkl i j k ij ik jk ijk ijkl where y ijkl is the phenotypic value of the ith individual in the ith genotypic group of MTHFR C677T, jth genotypic group of CHDH A119C, kth diet group; is the overall mean effect, i s, j s, and k s are the effects of MTHFR C677T, CHDH A119C, and diet, respectively, ij s are the 2-way interaction effects of MTHFR C677T and CHDH A119C, ik s are the 2-way interaction effects of MTHFR C677T and diet, jk s are the 2-way interaction effects of CHDH A119C and diet, and ijk s are the 3-way interaction effects of MTHFR C677T, CHDH A119C, and diet; i, j, k 1, 2. Because of the small number of individuals in the minor homozygous genotype at either of the 2 SNPs, we combined the minor homozygotes with the heterozygote group. In other words, our tests for interactions (whether ijk s are zero or not) are based on 8 cells. Results AIIMS Study We recruited 546 individuals in the first phase of study, whose baseline characteristics are shown in the Table 1. We initially screened 95 DNA samples for 44 nssnps present in 11 genes involved in 1-carbon metabolism (supplemental Table II) and found only 14 to be polymorphic (minor allele frequency [MAF] 1%; supplemental Table III). These 14 Table 1. General/Clinical Characteristics of the Study Group Participants Characteristics Phase 1 (N 546) Phase 2 (N 330) Age, y* 50 (43 to 58) 50 (42 to 58) BMI, kg/m 2 * 24.8 (22.4 to 27.3) 24.8 (22.2 to 27.4) Sex, n (% male) 467 (86) 253 (79) Smokers, n (%) 174 (32) 81 (25) Hypertension, n (%) 271 (50) 64 (20) Diabetes mellitus, n (%) 126 (23) 29 (9) Patients with CAD, n (%) 243 (45) 185 (56) Diet, n (% vegetarian) 257 (48) 166 (53) Milk drinker, n (%) 252 (47) 211 (67) Homocysteine, mol/l* 13.4 (9.4 to 18.5) 18.3 (13.9 to 26.9) Cysteine, mol/l* (179.7 to 240.5) (177.1 to 241.3) Folate, nmol/l* 8.6 (5.9 to 14.0) 10.2 (7.2 to 16.5) Vitamin B12, pmol/l* (112.2 to 204.1) (113.7 to 188.2) Creatinine, mol/l 97.2 (70.7 to 123.8) 76.9 (47.7 to 93.3) *Values shown are median (interquartile range). nssnps were then genotyped in 546 individuals. The genotypic and allelic frequencies of these polymorphisms are shown in supplemental Table III. Linkage disequilibrium (LD) was checked for SNPs present within the same gene using Haploview ( Only 2 SNPs from MTRR gene (rs10380 and rs162036) showed partial LD r None of the other SNPs present were found to be in LD. All the polymorphisms, except CHDH rs12676 conformed to Hardy Weinberg equilibrium at the nominal level of However, after using false discovery rate corrections, even this SNP did not show evidence of departure from Hardy Weinberg equilibrium. The distribution of only 1 SNP MTHFR rs was found to be significantly different between patients with CAD and controls (P 0.05). We then checked for the association of these 14 nssnps with homocysteine levels in 546 individuals (Table 2) and found that only 2 SNPs, methylene tetrahydrofolate reductase (MTHFR C677T) and choline dehydrogenase gene (CHDH A119C), were significantly associated with homocysteine levels (Table 2). MTHFR C677T was associated with elevated homocysteine levels under the assumption of codominant and recessive model, whereas CHDH A119C was significantly associated with reduced homocysteine levels under the assumption of codominant, dominant, and recessive models. None of the other SNPs studied showed any association with homocysteine levels. We also performed tests of association at the allelic level for each of the SNPs (Table 2). Only CHDH A119C polymorphism was found to be significantly associated with homocysteine levels under allelic association (P 0.029). The 2 SNPs (MTHFR C677T and CHDH A119C) that were significantly associated with homocysteine levels under multiple genetic models were further genotyped in an additional 330 individuals recruited from cardiology unit, AIIMS. The general/clinical characteristics of these 330 individuals are shown in Table 1. The median level of homocysteine in

4 602 Circ Cardiovasc Genet December 2009 Table 2. Genotypic Association of Polymorphisms With Homocysteine ( mol/l) Levels Under Different Models in Phase 1 (N 546) Codominant Model Dominant Model Recessive Model Allelic Association Downloaded from by guest on June 15, 2018 Polymorphism (rsid) BHMT G716A ( ) CHDH A119C (9001) CHDH T233G (12676) CTH G1076T ( ) MTHFR C677T ( ) MTHFR A1298C ( ) MTHFR G1781A ( ) MTR A2756G ( ) MTRR A66G ( ) MTRR A1049G (162036) MTRR C1243T ( ) MTRR C1349G ( ) MTRR C1783T (10380) SHMT1 C1420T ( ) P Value* P Value P Value P Value WW WM MM (q Value) WW WM MM (q Value) WW WM MM (q Value) W M (q Value) (0.48) (0.48) (0.48) (0.48) (0.044) (0.041) (0.15) (0.036) (0.48) (0.48) (0.48) (0.48) (0.48) (0.48) (0.48) (0.48) (0.027) (0.15) (0.024) (0.024) (0.52) (0.48) (0.51) (0.54) (0.48) (0.48) (0.48) (0.48) (0.51) (0.48) (0.48) (0.47) (0.38) (0.55) (0.47) (0.49) (0.47) (0.51) (0.23) (0.48) (0.47) (0.48) (0.48) (0.48) (0.48) (0.55) (0.47) (0.49) (0.48) (0.48) (0.48) (0.48) Median levels are shown. WW indicates major homozygous genotype; WM, heterozygous genotype; MM, minor homozygous genotype; W, major allele; M, minor allele. *Kruskal-Wallis test. Mann Whitney test. Additive allelic model using linear regression. q Values reported for total samples (phases 1 2). Software Q value used to calculate q values. the second phase was higher than that in the first phase. The exact reason behind the differences in levels of homocysteine between 2 phases, 1 and 2, is unknown. To rule out the possibility of shift in the homocysteine measurements over time, we repeated homocysteine measurements in 5% of the samples that were collected at various time points during the study and found that the homocysteine levels were in good agreement with those that were previously obtained (supplemental Table IV). The differences in the values were nonsignificant by paired t test (P 0.84) and within experimental errors. Thus, there was no indication of systematic shift in the homocysteine measurements over time. When the 2 SNPs (MTHFR C677T and CHDH A119C) were checked for the association with levels of homocysteine in these samples, only MTHFR C677T was found to be significantly associated (dominant model P 0.007, allelic association P , Table 3). This might be due to small sample size. Hence, we pooled the samples recruited from both first and second phase. The 2 polymorphisms showed significant association with homocysteine levels, both at the genotypic and allelic levels, after increasing the sample size, ie, 876 (Table 3). The q values (and the corrections) were calculated for the 2 significant SNPs (phases 1 2) and all other SNPs in phase 1. Taking into account the 52 association tests in the 2 stages (on the 12 polymorphisms tested only in the first stage [that did not provide any significant association finding] and the 2 polymorphisms tested in the second stage), both the MTHFR C677T and the CHDH A119C polymorphisms remained significantly associated even after a false discovery rate correction. The q values of all the SNPs are shown in Tables 2 and 3. We also performed linear regression analysis taking homocysteine as a dependent variable and controlling for various other confounding factors like age, sex, vitamin B 12 levels, folate levels, smoking, CAD status, diabetes mellitus status, hypertension, milk drinking, and creatinine levels. Even after adjusting for these potential confounding factors, the 2 polymorphisms were found to be significantly associated with homocysteine levels (P for MTHFR C677T and P for CHDH A119C). The genetic variance explained by these 2 polymorphisms contributes 1.7% of the variance of homocysteine levels. Because we observed significant association of 2 polymorphisms with modulated levels of homocysteine, we com-

5 Kumar et al Genetics of Hyperhomocysteinemia 603 Table 3. Genotypic Association of Polymorphisms With Homocysteine ( mol/l) Levels Under Different Models Codominant Model Dominant Model Recessive Model Allelic Association Downloaded from by guest on June 15, 2018 Polymorphism (rsid) WW WM MM Phase 2 (N 330) CHDH A119C (rs9001) MTHFR C677T (rs ) Phases 1 2 (N 876) CHDH A119C (rs9001) MTHFR C677T (rs ) P Value* (q Value) WW WM MM bined the protective (CHDH 119CC MTHFR 677CC, N 15, median 11.2 mol/l) and risk (CHDH 119AA MTHFR 677TT, N 20, median 22.0 mol/l) genotypes for 2 polymorphisms and checked their association with homocysteine levels. Median levels of homocysteine were found to be significantly different (P ) between the 2 groups. We note that this P value does not take into account the fact that the combination of SNPs was chosen based on our results rather than any prespecified hypothesis. We and others have previously shown that individuals adhering to a vegetarian diet have high homocysteine levels presumably due to low levels of vitamin B ,15 We thus checked if MTHFR C677T, CHDH A119C, and diet interact with each other and modulate the levels of homocysteine and found the possibility of a significant interaction between vegetarian diet and the 2 polymorphisms that could explain the variation of homocysteine levels (data not shown). However, none of the pairwise or 3-way interactions were significant when the combined effects of the 2 polymorphisms was examined with folate or vitamin B 12 levels on homocysteine levels. IGV Study We checked the basal frequencies of 12 SNPs in various subpopulations as a part of the IGV project. Samples collected in this study were divided into 4 groups based on their linguistic lineage, ie, Austro-Asiatic, Dravidian, Indo-European, and Tibeto-Burman. Distribution of these polymorphisms among the 4 linguistic groups along with the frequencies reported in HapMap and dbsnp ( are shown in Table 4. Some of the SNPs had a lower MAF in the IGV samples as compared with dbsnp, whereas MTHFR A1298C had a higher MAF. We also found that the MAF of most of these SNPs significantly varied between 1 or more linguistic groups (Figure 2). Our AIIMS study consisted of individuals from the Indo-European linguistic group and as shown in Table 4. The MAF of SNPs from AIIMS study was P Value (q Value) WW WM MM P Value (q Value) W M 18.3 (234) 18.4 (77) 14.9 (5) (234) 18.4 (82) (311) 14.9 (5) (545) 18.4 (87) (219) 19.9 (92) 20.8 (7) (219) 19.9 (99) (311) 20.8 (7) (530) 20.0 (106) P Value (q Value) 15.8 (604) 14.1 (225) 12.5 (20) (0.044) 15.8 (604) 13.9 (245) (0.041) 15.4 (829) 12.5 (20) (0.15) 15.5 (1433) 13.8 (265) (0.036) 15.0 (588) 15.4 (241) 22.0 (22) (0.027) 15.0 (588) 15.9 (263) (0.15) 15.0 (829) 22.0 (22) (0.024) 15 (1417) 19 (285) (0.024) Median levels are shown. No. of individuals genotyped shown in parentheses. WW indicates major homozygous genotype; WM, heterozygous genotype; MM, minor homozygous genotype; W, major allele; M, minor allele. *Kruskal-Wallis test. Mann Whitney test. Additive allelic model using linear regression. Software Q value used to calculate q values. similar to that of the Indo-European linguistic group of the IGV study. The distribution of MTHFR C677T and CHDH A119CC across the country (the 2 SNPs that were significantly associated with homocysteine levels in the AIIMS study) is shown in Figure 3. As reported earlier, the frequency of MTHFR 677TT genotype varies throughout the country, and TT genotype was absent in most of the populations studied (29 of 55). 17 Interestingly, the TT genotype was present in all the populations of North India (consisting of both Indo- European and Tibeto-Burman linguistic lineage) albeit with varying frequencies. The risk genotype in choline dehydrogenase (CHDH 119AA) was uniformly distributed throughout the country although the MAF was significantly higher in Austro-Asiatic and Tibeto-Burman populations as compared with Indo-European and Dravidian populations. Discussion During the past decade, research in the field of hyperhomocysteinemia has gained momentum because elevated homocysteine levels have been associated with several disorders. Although the intracellular concentration of homocysteine is well regulated, various factors including diet and genetic variations modulate the level of homocysteine. 8,11,12,15 Most Indians are strict vegetarians that lead to vitamin B 12 deficiency and consequently high homocysteine levels. In the background of vitamin B 12 deficiency, it can be perceived that polymorphisms responsible for modulation of homocysteine levels can have a major impact with respect to hyperhomocysteinemia. In this study, we found that polymorphisms MTHFR C677T and CHDH A119C are associated with homocysteine levels. Only 2 polymorphisms from MTRR genes showed LD, although incomplete, whereas the rest did not show any LD. We had already shown that the functional SNPs in MTHFR (rs and rs ) are not in LD. 11 Although several studies have reported the association MTHFR C677T with elevated homocysteine lev-

6 604 Circ Cardiovasc Genet December 2009 Table 4. Minor Allele Frequency of 14 nssnps in Indian Population Downloaded from by guest on June 15, 2018 IGV AIIMS HapMap Gene Symbol (rsid) AA DR IE TB IGV AIIMS CEU HCB JPT YRI dbsnp BHMT ( ) CHDH (9001) CHDH (12676) NA NA NA NA NA 0.16 NA NA NA NA ND CTH ( ) MTHFR ( )* MTHFR ( ) MTHFR ( )* MTR ( ) MTRR ( ) MTRR (162036) MTRR ( ) MTRR ( ) NA NA NA NA NA NP 0.15 MTRR (10380) SHMT1 ( ) AA indicates Austro-Asiatic; DR, Dravidian; IE, Indo-European; TB, Tibeto-Burman; IGV, Indian Genome Variation; NA, data not available; NP, nonpolymorphic; ND, not determined. *More than 1800 individuals genotyped. (Five hundred forty-eight individuals genotyped for rest of the polymorphisms in the IGV study.) Minor allele is flipped in this population. Figure 2. Comparison of different linguistic groups studied for their heterogeneity. Dark filled boxes show significant difference between 2 linguistic groups, whereas empty boxes show nonsignificant differences. Significance was calculated using 2 test and corrected for multiple testing. els, this is the first study showing the association of CHDH A119C with homocysteine. Choline dehydrogenase (CHDH) gene is responsible for the oxidation of choline, an essential dietary nutrient, into betaine. Betaine provides methyl group to homocysteine for its conversion into methionine. This reaction takes place mainly in liver. CHDH A119C, present in exon 3, changes the amino acid glutamine to alanine at position 40. This polymorphism lies in a domain (COG nih.gov/cog/grace/wiew.cgi?cog2303) where glutamine in position 40 is conserved in chimpanzee, monkey, rat, and mouse. Although the effect of this polymorphism on enzyme activity is not known, there are 2 studies till date that have reported this polymorphism in the context of organ dysfunction. 20,21 da Costa et al 20 showed a protective role of the C allele of this SNP toward susceptibility to liver dysfunction in individuals fed with a low choline diet. In another study, Niculescu et al 21 demonstrated using gene expression profiling experiments that individuals fed with low choline diet, having CHDH 119CC genotype grouped with individuals that did not show any organ dysfunction. These 2 studies confirm a protective role of this polymorphism. We have also found that individuals with CC genotype (homozygous for minor allele) have lower homocysteine levels than the AA genotype. Because the liver has a major role in homocysteine metabolism, this polymorphism could potentially play a vital role in maintaining the levels of homocysteine in circulation. The polymorphism MTHFR C677T plays an important role in increasing the levels of homocysteine. 12 Frosst et al 22 in 1995 identified this polymorphism in exon 4 at nucleotide position 677 that changes cytosine to thymine. This polymorphism has been shown to be associated with homocysteine levels in several populations including India. 11,15 In MTHFR C677T polymorphism, the amino acid alanine is changed to valine making this enzyme thermolabile. 23 The homozygous variant (TT) reduces the enzyme activity by 68% 24 thereby decreasing the conversion of methylene tetrahydrofolate to methyl tetrahydrofolate (a substrate that converts homocysteine to methionine), resulting in intracellular accumulation of homocysteine. We also determined the basal frequencies of these polymorphisms in at least 24 sub populations in the IGV project. 17

7 Kumar et al Genetics of Hyperhomocysteinemia 605 Downloaded from by guest on June 15, 2018 Figure 3. Distribution of risk genotype of MTHFR C677T and CHDH A119C in 24 subpopulations studied. The average frequency of MTHFR 677TT and CHDH 119AA genotypes are 2.5% and 65%, respectively. We found that some of the polymorphisms had lower MAF as compared with dbsnp, whereas MTHFR A1298C had a higher MAF than that reported in dbsnp. This could be attributed to the ethnic and genetic differences in Indian population compared with other populations. 11,25 Earlier studies also have shown that certain polymorphisms prevalent in other populations are nonpolymorphic in Indian population and vice versa. 26,27 Interestingly, even among the 4 Indian linguistic groups, the MAF varied significantly between 1 or more groups in most of the cases. It has been recently reported that there are high levels of genetic divergence between populations that cluster largely on the basis of ethnicity and language. 17 This is also reflected from our study wherein we found that the MAF of all the SNPs studied were similar in individuals recruited for the AIIMS study and the Indo-European samples studied in the IGV project. We have recently shown the distribution of MTHFR C677T polymorphisms in 55 subpopulations spread across the country. 17 Interestingly, 677TT genotype was observed only in Northern part of the country along with few pockets in southern parts and found to be completely absent in rest of subpopulations. This is in contrast to the risk genotype of choline dehydrogenase (CHDH 119AA) that was found to be uniformly distributed throughout the country. In conclusion, our study suggests that 2 polymorphisms MTHFR C677T and CHDH A119C are associated with homocysteine levels. This is the first report showing that the SNP in CHDH is associated with homocysteine levels. Further studies are required to mechanistically understand how this polymorphism regulates homocysteine levels. Acknowledgments The University Grant Commission (J.K.) and CSIR (G.G., A.K., E.S.) are acknowledged for their fellowships. We thank all the participants of this study. We are grateful to the Council of Scientific and Industrial Research (CSIR) task force project team for sample collection done for IGV study. We all thank the sequencing facility, IGIB, for helping us in genotyping. We also thank Dr. Anurag Agarwal and Dr. Mitali Mukerji for critically reviewing the manuscript. We thank the anonymous reviewers whose comments improved the article a lot. Sources of Funding The study was supported by grants received from CSIR (CMM0016) and DBT, Government of India.

8 606 Circ Cardiovasc Genet December 2009 Downloaded from by guest on June 15, 2018 None. Disclosures References 1. Mills JL, McPartlin JM, Kirke PN, Lee YJ, Conley MR, Weir DG, Scott JM. Homocysteine metabolism in pregnancies complicated by neural-tube defects. Lancet. 1995;345: McCaddon A, Davies G, Hudson P, Tandy S, Cattell H. Total serum homocysteine in senile dementia of Alzheimer type. Int J Geriat Psychiatry. 1998;13: van Guldener C, Stehouwer CD. Homocysteine metabolism in renal disease. Clin Chem Lab Med. 2003;41: Applebaum J, Shimon H, Sela BA, Belmaker RH, Levine J. Homocysteine levels in newly admitted schizophrenic patients. J Psychiatr Res. 2004;38: de Luis DA, Fernandez N, Arranz ML, Aller R, Izaola O, Romero EJ. Total homocysteine levels relation with chronic complications of diabetes, body composition, and other cardiovascular risk factors in a population of patients with diabetes mellitus type 2. J Diabetes Complications. 2005;19: Robinson K, Mayer EL, Dave P, Miller MS, Green R, van Lente F, Gupta A, Marchant KK, Savon SR, Selhub J, Nissen SE, Kutner M, Topol EJ, Jacobsen DW. Hyperhomocysteinemia and low pyridoxal phosphate common and independent reversible risk factors for coronary artery disease. Circulation. 1995;92: Yoo JH, Park JE, Hong KP, Lee SH, Kim DK, Lee WR, Park SC. Moderate hyperhomocyst (e) inemia is associated with the presence of coronary artery disease and the severity of coronary atherosclerosis in Koreans. Thromb Res. 1999;94: Mann NJ, Li D, Sinclair AJ, Dudman NP, Guo XW, Elsworth GR, Wilson AK, Kelly FD. The effect of diet on plasma homocysteine concentrations in healthy male subjects. Eur J Clin Nutr. 1999;53: McKinley MC, McNulty H, McPartlin J, Strain JJ, Pentieva K, Ward M, Weir DG, Scott JM. Low-dose vitamin B-6 effectively lowers fasting plasma homocysteine in healthy elderly persons who are folate and riboflavin replete. Am J Clin Nutr. 2001;73: Kumar J, Garg G, Sundaramoorthy E, Prasad PV, Karthikeyan G, Ramakrishnan L, Ghosh S, Sengupta S. Vitamin B12 deficiency is associated with coronary artery disease in an Indian population. Clin Chem Lab Med. 2009;47: Kumar J, Das SK, Sharma P, Karthikeyan G, Ramakrishnan L, Sengupta S. Homocysteine levels are associated with MTHFR A1298C polymorphism in Indian population. J Hum Genet. 2005;50: Fredriksen A, Meyer K, Ueland PM, Vollset SE, Grotmol T, Schneede J. Large-scale population-based metabolic phenotyping of thirteen genetic polymorphisms related to one-carbon metabolism. Hum Mutat. 2007;28: Janosikova B, Zavadakova P, Kozich V. Single-nucleotide polymorphisms in genes relating to homocysteine metabolism: how applicable are public SNP databases to a typical European population? Eur J Hum Genet. 2005;13: Satoskar RS, Kulkarni BS, Rege DV. Serum proteins, cholesterol, vitamin B12 and folic acid levels in lactovegetarians and nonvegetarians. Indian J Med Res. 1961;49: Refsum H, Yajnik CS, Gadkari M, Schneede J, Vollset SE, Orning L, Guttormsen AB, Joglekar A, Sayyad MG, Ulvik A, Ueland PM. Hyperhomocysteinemia and elevated methylmalonic acid indicate a high prevalence of cobalamin deficiency in Asian Indians. Am J Clin Nutr. 2001;74: Indian Genome Variation Consortium. The Indian genome variation database (IGVdb): a project overview. Hum Genet. 2005;118: Indian Genome Variation Consortium. Genetic landscape of the people of India: a canvas for disease gene exploration. J Genet. 2008;87: Benjamini Y, Hochberg Y. Controlling the false discovery rate: a practical and powerful approach to multiple testing. J R Stat Soc Series B Stat Methodol. 1995;57: Storey JD. A direct approach to false discovery rates. J R Stat Soc Series B Stat Methodol. 2002;64: da Costa KA, Kozyreva OG, Song J, Galanko JA, Fischer LM, Zeisel SH. Common genetic polymorphisms affect the human requirement for the nutrient choline. FASEB J. 2006;20: Niculescu MD, da Costa KA, Fischer LM, Zeisel SH. Lymphocyte gene expression in subjects fed a low-choline diet differs between those who develop organ dysfunction and those who do not. Am J Clin Nutr. 2007;86: Frosst P, Blom HJ, Milos R, Goyette P, Sheppard CA, Matthews RG, Boers GJ, den HM, Kluijtmans LA, van den Heuvel LP. A candidate genetic risk factor for vascular disease: a common mutation in methylenetetrahydrofolate reductase. Nat Genet. 1995;10: Kang SS, Zhou J, Wong PW, Kowalisyn J, Strokosch G. Intermediate homocysteinemia: a thermolabile variant of methylenetetrahydrofolate reductase. Am J Hum Genet. 1988;43: Weisberg I, Tran P, Christensen B, Sibani S, Rozen R. A second genetic polymorphism in methylenetetrahydrofolate reductase (MTHFR) associated with decreased enzyme activity. Mol Genet Metab. 1999;64: Kumar J, Sunkishala RR, Karthikeyan G, Sengupta S. The common genetic variant upstream of INSIG2 gene is not associated with obesity in Indian population. Clin Genet. 2007;71: Pemberton TJ, Mehta NU, Witonsky D, Di Rienzo A, Allayee H, Conti DV, Patel PI. Prevalence of common disease-associated variants in Asian Indians. BMC Genet. 2008;4: Sharma S, Sharma A, Kumar S, Sharma SK, Ghosh B. Association of TNF haplotypes with asthma, serum IgE levels, and correlation with serum TNF-alpha levels. Am J Respir Cell Mol Biol. 2006;35: CLINICAL PERSPECTIVE Plasma levels of homocysteine are modulated by nutritional (including deficiency of vitamin B 12 and folate) and genetic factors. Nutritional supplementation, frequent in the developed world, may mask genetic effects in such populations. Because the strict vegetarian diet of most Indians predisposes them to vitamin B 12 deficiency, single-nucleotide polymorphisms in genes involved in homocysteine metabolism may have a greater effect on plasma homocysteine levels. In this study, we genotyped 44 nonsynonymous single nucleotide polymorphisms present in 11 genes involved in homocysteine metabolism and related them to plasma homocysteine levels. We found that polymorphisms in the methylenetetrahydrofolate reductase (MTHFR) and choline dehydrogenase (CHDH) genes were associated with plasma homocysteine levels. Although several studies have shown that individuals with 677TT variant in the MTHFR gene have higher homocysteine levels, particularly in the setting of a low folate status, the association of 119CC variant in CHDH gene with plasma homocysteine levels has not been previously reported. In this study, we also determined the frequency of various nonsynonymous single nucleotide polymorphisms in genes involved in homocysteine metabolism. We studied at least 24 different subpopulations spread across the India based on their linguistic lineages and ethnic background.

9 Single Nucleotide Polymorphisms in Homocysteine Metabolism Pathway Genes: Association of CHDH A119C and MTHFR C677T With Hyperhomocysteinemia Jitender Kumar, Gaurav Garg, Arun Kumar, Elayanambi Sundaramoorthy, Krishna Rao Sanapala, Saurabh Ghosh, Ganesan Karthikeyan, Lakshmy Ramakrishnan, Indian Genome Variation Consortium and Shantanu Sengupta Downloaded from by guest on June 15, 2018 Circ Cardiovasc Genet. 2009;2: ; originally published online September 5, 2009; doi: /CIRCGENETICS Circulation: Cardiovascular Genetics is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX Copyright 2009 American Heart Association, Inc. All rights reserved. Print ISSN: X. Online ISSN: The online version of this article, along with updated information and services, is located on the World Wide Web at: Data Supplement (unedited) at: Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally published in Circulation: Cardiovascular Genetics can be obtained via RightsLink, a service of the Copyright Clearance Center, not the Editorial Office. Once the online version of the published article for which permission is being requested is located, click Request Permissions in the middle column of the Web page under Services. Further information about this process is available in the Permissions and Rights Question and Answer document. Reprints: Information about reprints can be found online at: Subscriptions: Information about subscribing to Circulation: Cardiovascular Genetics is online at:

10 SUPPLEMENTAL MATERIAL 1

11 Supplementary Table 1: List of oligonucleotides used for PCR and genotyping. Gene Symbol SNP ID Primer Type Primer Sequence (5-3 ) Amplicon Length (bp) AHCY FP CACCCTGGCACAGTCGTCTT 1037 RP ACCCCCAGATTCAGCAGTTAC SS GGGCCTGATGCGTATGCGGGAG AHCY FP GCTTCATGTCCCTAATCTCAG 1288 RP GAGTTGGGAAGGAGGTAGTTT SS TTGGTGTGGATGAGGTTGGTGA AHCY FP CACCCGCCATTGTTGCTTTTATTC 922 RP TTGCCTGCCCCTATTGTCTCCT SS CCTGCTACCACCGCTACC AMD FP GACCCATCATCACTAGACTG 321 RP CTTGGTGAGAAGGCTTCATG SS GAAGGCTTCATGAAATTCTTAC AMD FP GACCCATCATCACTAGACTG 321 RP CTTGGTGAGAAGGCTTCATG SS AGAGCTTCTTTTATTCTCGT AMD FP TCACATCACTCCAGAACCAG 323 RP GGACTGACACAGGTCTAATG SS TAAGTCAGACCTCCTATGATGACCTG BHMT FP TTGAAACTGGGAGGCAAATG 305 RP TCATCACCTGCTTTCACCAG 2

12 SS CCCCGTGCGGTACGTTTAGAGGAAGA BHMT G199S* FP TTGAAACTGGGAGGCAAATG 305 RP TCATCACCTGCTTTCACCAG SS TTTGCATGGCGTGCCCCCC BHMT FP CATGTTCTTCCCACTCACAG 330 RP GACAAGACATGAGGAGACAC SS ATCATTGGTGTGAACTGCCACT BHMT FP CATGTTCTTCCCACTCACAG 330 RP GACAAGACATGAGGAGACAC SS AAAAATTTGATTACCACCCCAG BHMT FP CATGTTCTTCCCACTCACAG 330 RP GACAAGACATGAGGAGACAC SS TCAGGTGAGCTTTCAGT BHMT FP GCCTAATGGCATAATGCCAC 331 RP GCAGTCTTCTAAATGGCCTC SS ACGCCAGAGAGGCCTAC BHMT FP GTCAGGAATCATGGTGAAGC 332 RP TCTTTCTGCTGCATCAGCTC SS ACAACCCTTCAATGTCAAAGCCAGAT BHMT Q406H * FP AACTGAGCAGCAGCTGAAAG 332 RP TCCTGTGCAGCAGTGATTTC SS TATCGAAGATAGCTCCGAT 3

13 CBS * FP TTACACCTCATGTCCTCCTG 332 RP TCTCCTTGCTTTGCCAGGTC SS GGTGTAGTGGTGTGACGGG CBS FP AGTGTGAGGGTGAGTTACAG 286 RP GTTGTTAACGGCGGTATTGG SS GGATGGACCCTTCGGGATCCACCCCA CBS FP TCACAGCATTAAGAGGCTGG 413 RP TTCGGAGTGTGTCTGACATG SS GTCGGGCAGAATGACCACGCAG CHDH 9001 FP TAGCACCAGTTGTACCTGTC 324 RP AAGCATGTGGTGTCTCCTAC SS CCGCGCCCACCACCACATAGCTGTAC CHDH FP TAGCACCAGTTGTACCTGTC 324 RP AAGCATGTGGTGTCTCCTAC SS GCCGCTTGCTCCCCGCG CTH FP TGCTTTATCTTGGGCATGGG 341 RP GCAAAGGCTCATTGTTGGTC SS TTATAGCACCCTCCAAGTGGAA MAT1A * FP CATCCTCCTCATTTCTGTCC 334 RP CAGTAAAATGCACCACGGAG SS AATCCCCAGATATTGCCCA MTHFR FP TTTTCTGCCGGGTTAAAGGG 457 4

14 RP TCCTGTGTGGAAATACAGCG SS GAGATTGACAGCTCCCTCAGCAGTT MTHFR FP AGAGGAGATCTGGGAAGAAC 329 RP CATCCCTATTGGCAGGTTAC SS AAGCTGCGTGATGATGAAATCG MTHFR FP CTCACTTTGTGACCATTCCG 440 RP CCTCCAGACCAAAGAGTTAC SS TGCTTCACTGGTCAGCTC MTHFR FP CTCACTTTGTGACCATTCCG 440 RP CCTCCAGACCAAAGAGTTAC SS AAGACTTCAAAGACACTT MTHFR FP AATCCCTCCCCACGGTTTTC 231 RP TGGGACTCCAGTTGTTCTTG SS GCTGGATGATCTCTCGC MTHFR FP AGTGGAGTTCCCAAGAGAAG 405 RP TTCCGGGAAATGTCCTGTTG SS CTCCTCATACAGCTTTCCCCAC MTR FP CCTTTTTGGGCAGCGTTTTC 330 RP GTCATTGTTGCCTTTCAGCG SS CTAAACGAAGAACACTTCC MTR * FP CTCTTGGCACATGTCTGTTC 195 RP AGCCATCAGAAAAGGTTCCC 5

15 SS GCGTGTCTCATGGAGAACCACTCT MTR * FP TCCTATGTCTAGGGCATGTG 260 RP TGGAAGGACACAATTCAGCC SS ATGCAGGATTTTGCTATG MTR * FP AAGGAGAGGACGACTTCTTG 315 RP ACACTCTATGACGTTGGCAC SS GAAGTAGTTTTCGGTACTGGTG MTR FP ACAGTTGGTGAAGGGAGAAG 332 RP CAAAGCCTTTTACACTCCTC SS CACTTACCTTGAGAGACTCATAATGG MTR FP TGCCCAGAGAAAAAGGCAAC 325 RP GGCACCCTAGGTTGTATTTC SS CTGAGGTTGAGAAATGGCTTGGACCC MTRR FP GCTCATTTGAGATTAGTGCTG 332 RP TCCACTGTAACGGCTCTAAC SS GGCCATCGCAGAAGAAAT MTRR FP GCCAGCCCTCAGAAAGCA 403 RP AAAAGCCCAAAGCAAACATCCA SS TGGCATCACCTGCATCCT MTRR FP GCCAGCCCTCAGAAAGCA 403 RP AAAAGCCCAAAGCAAACATCCA SS AATATTTACAGGTACATCTGCAGGAG 6

16 MTRR FP GCTCTTCTCTAGAATACAGAC 332 RP TACCAATACCAGCGTATGCC SS GACGCAGTGCTCTCTTTTATCTTCAA MTRR FP GCTCTTCTCTAGAATACAGAC 332 RP TACCAATACCAGCGTATGCC SS GAAAATAAAGGCAGACACAA MTRR FP TAGCTCCTAGTGGCTTTCTC 328 RP TGTGAAAGCGACTACTCACC SS GCAGGCACAGGCATCTCGTACAAAGC MTRR FP CCATTCCTAGAAGCCTTTGC 336 RP ATAGTAGTACCTTGCACACG SS GAACATCTTCCTAAACTTCAAC MTRR FP GGTTTGACCCATATGTGTAG 334 RP ATGTGGGTGAGACAACACTG SS ATGAAGACAGCGGGAAAG MTRR FP ACTCCAAGAACAACACCCAG 324 RP CCATGAAGCTGGATGTTGTC SS AAGATCCCATGCTTAAGGAAAT SHMT FP ATTCCCATCATCTGTGGCTC 323 RP GCCGAAGCAAGCTATGACTCTGG SS AACTGCAGGGCTCTGTCT SHMT FP CTCCTTTAGAAGTCAGGCAG 308 7

17 RP CCCGTAAGCATGGAGACCTTGTG SS GCCAGGCAGAGGGAAGA FP- Forward Primer, RP- Reverse Primer, SS- SNaPshot Primer, *SNPs selected from Janosikova B et. al.,

18 Supplementary table 2: List of all the 44 polymorphisms studied. S. No Gene Symbol SNP (rs ID) Base Change AA Change 1 AHCY C112T Arg38Trp 2 AHCY C403A Leu135Ile 3 AHCY T641A Lys214Met 4 AMD C340T Arg114Cys 5 AMD A343G Lys115Glu 6 AMD A358G Ile120Val 7 BHMT C566A Pro189Gln 8 BHMT G199S* G595S Glu199Ser 9 BHMT T656G Phe219Cys 10 BHMT T657G Phe219Leu 11 BHMT G716A Arg239Gln 12 BHMT A868C Asn290His 13 BHMT G1114T Gly372Cys 14 BHMT Q406H* G1218T Gln406His 15 CBS * C209T Pro70Leu 16 CBS A833C Ile278Ser 17 CBS G1106C Arg369Pro 18 CHDH 9001 A119C Glu40Ala 19 CHDH T233G Leu78Arg 20 CTH G1076T Ser359Ile 21 MAT1A * G357T Gln119His 22 MTHFR G203A Arg68Gln 23 MTHFR C677T Ala222Val 24 MTHFR G1269T Glu423Asp 25 MTHFR A1298C Glu429Ala 26 MTHFR G1697A Gly566Glu 27 MTHFR G1781A Arg594Gln 28 MTR G155A Arg52Gln 9

19 29 MTR * A764G Tyr255Cys 30 MTR * G940A Asp314Asn 31 MTR * G1485A Met495Ile 32 MTR A2756G Asp919Gly 33 MTR A3775G Ile1259Val 34 MTRR A66G Ile22Met 35 MTRR C524T Ser175Leu 36 MTRR T769A Ser257Thr 37 MTRR C997G Leu333Val 38 MTRR A1049G Lys350Arg 39 MTRR C1243T Arg415Cys 40 MTRR C1349G Pro450Arg 41 MTRR C1544T Ala515Val 42 MTRR C1783T His595Tyr 43 SHMT G1018C Glu340Gln 44 SHMT C1420T Leu474Phe * SNPs selected from Janosikova B et. al.,

20 Supplementary Table 3: Distribution of genotype and allele frequency in Phase 1 (N=546). S. No Polymorphism rs ID AA Change WW WM MM Obs MAF Rep MAF 1 BHMT G716A Arg239Gln 275 (52%) 208 (39%) 48 (9%) CHDH A119C 9001 Glu40Ala 372 (69%) 148 (28%) 15 (3%) CHDH T233G Leu78Arg 371 (73%) 120 (23%) 21 (4%) 0.16 ND 4 CTH G1208T Ser403Ile 252 (48%) 236 (44%) 42 (8%) MTHFR C677T Ala222Val 370 (69%) 150 (28%) 15 (3%) MTHFR A1298C Glu429Ala 161 (30%) 273 (51%) 101 (19%) MTHFR G1793A Arg594Gln 400 (75%) 128 (24%) 6 (1%) MTR A2756G Asp919Gly 241 (46%) 230 (43%) 58 (11%) MTRR A66G Ile22Met 155 (29%) 263 (49%) 117 (22%) MTRR A1049G Lys350Arg 371 (70%) 138 (26%) 19 (4%) MTRR C1243T Arg415Cys 485 (93%) 35 (7%) 0 (0%) MTRR C1349G Pro450Arg 495 (94%) 30 (6%) 1 (0%) MTRR C1783T His595Tyr 386 (74%) 121 (23%) 13 (3%) SHMT1 C1420T Leu474Phe 344 (69%) 148 (29%) 10 (2%)