A goal of modern human genetics is to use our knowledge of specific mutations to help understand and treat genetic disease. In most cases after a disease gene is isolated, investigators identify DNA alterations in this gene and attempt to correlate these alterations with phenotypic variation in the disease. However, the results of such studies can be confusing. It is difficult to prove that a DNA alteration actually causes disease and is not a rare polymorphism in the patient. What is needed is a combination of traditional mutation detection methods with systems that allow examination of the effect of a DNA alteration on the function of the coded protein. Such functional systems are also needed for getting at the mechanistic questions of how mutations actually disturb protein function. The ability to answer these questions could potentially open new avenues of therapy.
The genetic disorder we have chosen to study is cystathionine [beta]-synthase (CBS) deficiency. CBS deficiency is an autosomal recessive disorder which results in extremely elevated levels of total plasma homocysteine (tHcy) and methionine (1 ). In individuals with this disease, the major clinical complication is an increased incidence of thromboembolic events, although other problems such as lens dislocation, osteoporosis and neurological defects are frequently observed. Several studies suggest high plasma homocysteine levels may be directly responsible for the vascular lesions observed in patients (1 ,2 ).
There is a wide spectrum of phenotypic variation in individuals with CBS deficiency. About 50% of patients respond to high doses of pyridoxine with a marked reduction in their homocysteine level, and have a better clinical prognosis. Studies on CBS activity in cell lines from these patients demonstrate some residual enzyme activity, whereas cells from pyridoxine non-responders have no detectable activity (1 ).
The cause of CBS deficiency is mutations in the human CBS gene. Over 20 different point mutations have been identified in patients with CBS deficiency (3 -9 ). However, studies on the CBS gene also provide a good example of the problems inherent in distinguishing disease-causing mutations from polymorphisms. A 68 bp duplication of the region containing the intron 7-exon 8 junction was initially reported as a mutation in an Italian CBS deficient patient (8 ). However, more recent studies indicate that this duplication is in fact a benign polymorphism found on 3-6% of all CBS containing chromosomes (10 ,11 ).
In this study we have combined conventional mutation screening with our previously developed yeast functional assay for human CBS to analyze 12 CBS deficient patients from 10 unrelated Norwegian families. The major finding of this study is the identification of a novel point mutation, G797A, found in the majority of pyridoxine-responsive patients. When expressed in yeast, this mutation shows a pyridoxine-responsive phenotype. These results demonstrate the usefulness of yeast functional assays in human genetic analysis.
Our patient population consisted of 12 Norwegian individuals from 10 unrelated families. The patients all had homocystinuria due to suspected CBS deficiency. Details of each of the patients are shown in Table 1 . Given that there were 10 families, there were at most 20 different CBS alleles present in this population.
Initially, we screened for two point mutations, G919A (G307S) and T833C (I278T), which have been observed in multiple patients in other studies (9 ,12 ). For the screening we used PCR amplification of DNA isolated from frozen whole blood followed by restriction digest with specific enzymes to detect the mutations (PCR/RFLP, see Materials and Methods and Table 4 ). We found that four of the 20 alleles in the patients contained the G919A mutation and two of the alleles contained the T833C mutation.
The assay used to detect T833C also allowed us to assay for the presence or absence of the 68 bp duplication of the intron 7-exon 8 boundary which was previously described as a benign polymorphism (Fig. 1 ). We discovered that nine of the 20 CBS alleles present (45%) contained the 68 bp duplication allele. This frequency was much higher than the 6% allele frequency reported by Tsai et al. (10 ) in their study population from the upper Midwest. Because the allele frequency of the 68 bp duplication in Norwegians has not been previously determined, we examined the allele frequency in unaffected Norwegian chromosomes and found it to be ~5.5% (2/36).
Table 1
Using PCR/RFLP methodology, we screened for two other mutations originally identified in non-Norwegian CBS deficient patients (WDK, unpublished). These mutations included C1105T (R369C) and T959C (V320A). We found one C1105T allele and two T959C alleles in this CBS-deficient population (Fig. 2 ). No C1105T or T959C alleles were identified in 36 control chromosomes screened.
The high frequency of the 68 bp duplication on the diseased chromosomes suggested that there might be linkage disequilibrium between a disease causing mutation and this polymorphism. Therefore, we sequenced the DNA of exon 7 in one of the probands homozygous for the 68 bp duplication (patient N1) (Fig. 3 ), and found that he was homozygous for a novel mutation, G797A, which results in an arginine to lysine change at position 266 of the CBS protein.
We developed a PCR/RFLP method for detecting the G797A mutation and screened the rest of our probands (Materials and Methods). Notably, seven of the 20 affected chromosomes in our population contained the G797A allele, and all of them were found in association with the 68 bp duplication. Patients N1, N4ab and N9 were homozygous for both G797A and the 68 bp duplication, indicating physical linkage between the two alleles. Patient N8 was heterozygous for both alleles, but we lacked family material to establish linkage. We did not observe any G797A mutations in 100 control chromosomes.
We also tested to see if a significant fraction of 68 bp duplication containing chromosomes also contained G797A. We screened DNA obtained from 242 non-homocystinuric individuals for the 68 bp duplication and found 15 individuals who contained at least one 68 bp duplication allele. None of these individuals contained the G797A mutation. These findings indicate that the G797A mutation is rare in both the general population and in individuals containing the 68 bp duplication.
In parallel with this work we also screened all of the probands by SSCP analysis. Exons 1-10 and 12 were analyzed (exon 11 could not be amplified.) These studies confirmed the presence of the G797A, C785T, G919A and T833C alleles (data not shown). The C1105T and T959C alleles were not identified by SSCP.
By using all the methods described, we identified mutations which alter the predicted CBS amino acid sequence in 18 of the 20 alleles present in our CBS deficient population (Table 2 ).
Table 2 .
To determine the functional consequences of these CBS missense mutations we expressed them in a strain of Saccharomyces cerevisiae in which the endogenous CBS gene is deleted (WK63[Delta]yCBS) (5 ). This strain cannot synthesize cysteine and thus requires media with cysteine for growth (Cys+ media). Expression of a functional human CBS protein allows cysteine production and thus allows growth on Cys- media. All of the identified mutations, G797A, C785T, C1105T, T833C, G919A and T959C, were engineered by site-directed mutagenesis into the human CBS encoding cDNA. Each mutant cDNA was subsequently cloned into an expression vector and transformed into WK63[Delta]yCBS (see Materials and Methods).
Strains containing the T833C, G919A and C785T mutations were unable to form colonies on standard Cys- media (Fig. 4 ). This result indicates that these mutations severely affect CBS enzyme function. Cells containing the T959C mutation were able to form single colonies on standard Cys- media, but these colonies were significantly smaller than those formed by yeast cells expressing wild-type human CBS. Yeast containing the G797A and the C1105T mutation grew just as well as wild-type. These results indicate that the R369C and R266K alterations in the CBS protein have less severe effects on enzyme function than the G307S, I278T, V320A and T262M alterations.
In this study we analyzed 12 CBS-deficient probands from 10 unrelated Norwegian families using a combination of conventional mutation detection methods and a yeast functional assay for the human CBS gene. We were able to identify six different missense mutations present in 18 of the 20 alleles in the patient group: G919A, C785T, T959C, T833C, G797A and C1105T. Two of the mutations, G919A and T833C, have been previously reported, while the other four are novel.
We expressed all of the mutant human CBS proteins above in our previously described yeast system to examine their function (5 ). In this system the mutant CBS proteins are expressed in a yeast strain deleted for the endogenous CBS gene, and growth behavior of the yeast is observed on standard yeast media lacking cysteine. The two known mutations, G919A and T833C, along with one new mutation, C785T, exhibited a very strong no growth phenotype. The T959C allele gave a distinct slow growth phenotype. Our results from the yeast assay strongly suggest that these four DNA alterations are disease causing mutations and not benign polymorphisms. Interestingly, almost all of these mutant alleles (with the exception of one T833C mutation) are found in the pyridoxine non-responsive patients.
Two other mutations detected, G797A and C1105T, did not give distinguishable phenotypes in yeast under the standard assay conditions. These two mutations were found exclusively in pyridoxine-responsive patients. Since standard yeast media is rich in pyridoxine (1.9 µM in yeast media compared with ~0.3 nM in human serum), we examined the phenotype of these two mutations at low pyridoxine levels. When grown under these conditions yeast expressing the wild-type allele are able to form colonies, while strains carrying the G797A allele are not. In liquid cultures with high pyridoxine the doubling time of yeast carrying the G797A allele was similar to wild-type, but growth became noticeably slower with decreasing pyridoxine. Thus, G797A allele expressed in yeast is pyridoxine responsive, similar to what is observed for this allele in humans.
The C1105T mutation did not exhibit a no growth phenotype in yeast, even when grown at low pyridoxine concentrations. In fact, yeast carrying this allele actually seem to grow slightly better than yeast carrying the wild-type allele under these conditions. Thus our functional data would suggest that the C1105T alteration is not pathogenic, but is simply a polymorphic variant of the protein. In support of this, we have found two additional alleles of C1105T in screening of 200 non-homocystinuric individuals (WK, unpublished). Thus, it may in fact be a rare polymorphism. If this were the case, we would assume that there must be another mutation in cis with this alteration that we failed to discover during mutation screening. Certainly our mutation screening was not exhaustive as we failed to discover at least two other mutant alleles in this group of patients. Alternatively, the allele may be pathogenic, but is not properly modeled in yeast. It is possible that some mutations which affect protein stability or folding may not be properly modeled in the yeast system. In addition, mutations which affect the interaction of CBS with co-factors other than pyridoxine (e.g., heme or S-adenosylmethionine) may not be modeled well because these factors are present in different concentrations in yeast versus mammalian cells.
Our functional data in yeast suggest that the bonefida T833C mutation is not pyridoxine-responsive. However, several reports have found an association between the presence of the T833C mutation and pyridoxine responsiveness in patients (9 ,13 ,14 ). Our findings concerning linkage disequilibrium between the 68 bp duplication and the G797A allele could explain this apparent contradiction. The 68 bp duplication of the intron 7-exon 8 border is not a perfect duplication. There is a single nucleotide change in the 5' repeat which is identical to the mutation found in patients with the T833C alteration. This nucleotide change creates a novel BsrI restriction site (8 ). However, individuals with this duplication produce fully functional CBS mRNA because only the second splice acceptor site is utilized and thus the mutation is not introduced into the mRNA (10 ,11 ). We presume that the bonefida T833C mutation arose due to intra-chromosomal recombination between the two repeats. All of the reports showing association of pyridoxine-responsiveness with T833C were done before the existence of the 68 bp duplication was known. Thus, it is possible that some of the T833C alleles identified in these studies were actually 68 bp duplication alleles. We suggest that these patients should be reexamined in light of this information.
The spectrum of mutations discovered in Norwegian homocystinurics is different than that described in other populations. In Ireland, which has a relatively high rate of homocystinuria, the G919A allele accounts for 71% of the mutant alleles (12 ). In our Norwegian group this allele accounts for only 20%. In Italian patients the G919A allele was not observed at all (15 ). In Norwegians the G797A allele appears to be quite frequent (35%), but this allele has not been reported elsewhere. The differential distribution of disease-causing mutations in different populations implies that most of the mutations causing CBS deficiency are of relatively recent origin. Consistent with this hypothesis is the linkage disequilibrium observed between the G797A mutation and the 68 bp duplication. Presumably, the original G797A mutation formed on a chromosome containing the 68 bp duplication, and then expanded in the population due to genetic drift.
The work described in this paper has potential implications in regard to two facets of CBS deficiency. The ethnic differences in the distribution of different mutations should have important implications for genetic testing. As testing for CBS mutations becomes more widespread, knowledge of an individual's ethnic background will allow genetic testing to be done more efficiently and cost-effectively. More important are the implications for the functional modeling of mutant CBS proteins in yeast. We have shown here that it is possible to identify a substance, i.e., pyridoxine, that can rescue a specific allele in yeast, and this substance can also have beneficial effects in humans. By having model systems available for other mutant alleles it may be possible to identify other substances which can rescue mutant alleles and potentially be beneficial to patients suffering from CBS deficiency.
More generally, this study demonstrates the utility of combining a yeast functional assay with conventional mutation detection in the analysis of human genetic disease. As more human genes are functionally modeled in yeast, this type of analysis should be applicable to a wide range of human diseases.
Eleven patients were recruited from different medical practices in the south, west and middle part of Norway. The one remaining patient was diagnosed in the Hordaland Homocysteine Study (16 ). The present study was approved by the Regional Ethics Committee and all of the patients signed informed consent agreements. Details about the patients are given in Table 1 .
Table 4 .
Table 5 .
In eight patients, pretreatment tHcy levels were available. These were categorized as pyridoxine responsive if the tHcy level showed a 50% or greater reduction after initiation of pyridoxine (250-500 mg/day) combined with low dose folic acid (0.4 mg/day) and intramuscular B12 injections. In four patients (diagnosed >10 years ago) the pre-treatment levels of plasma tHcy are not known. In these patients a positive cyanide-nitropusside test in the urine combined with clinical features were considered sufficient for the diagnoses. Two of these patients (N4a and b) have a normal tHcy level with only pyridoxine substitution and are considered pyridoxine responsive. One patient (N8) uses high doses of both pyridoxine and folic acid, but her particularly low tHcy level indicates pyridoxine responsiveness. The fourth patient has had varying tHcy concentrations (30-80 µmol/l) and uses high doses of pyridoxine, folic acid and betaine and is therefore considered non-responsive.
We had two control groups. A Norwegian control group consisting of 50 controls, 25 males and 25 females, was randomly picked from a selection of 329 anonymous subjects participating in the Hordaland Homocysteine Study. A second control group consisting of 242 subjects from the the Portland Oregon area is described in ref. 17 .
DNA was isolated from frozen whole blood using Instagene (Biorad). For PCR-RFLP assays, 10 µl of DNA was used in 50 µl PCR reactions. Table 4 shows the conditions and enzymes used for each individual assay. After PCR, products were precipitated with ethanol, resuspended in 20 µl TE with restriction buffer, and then digested for 3 h. Products were then analyzed on 3% agarose gels.
For the direct sequencing of mutations, the PCR products were purified after gel electrophoresis using QIAquick Gel Extraction KitTM (Qiagen), and sequenced at the Automated Sequencing Facility at the Fox Chase Cancer Center. The primers for PCR amplification of the direct sequencing products are as follows. For exon 7 (G797A and C785T mutations) we amplified with 5'-AAGCTGGACATGCTGGTGGC-3' and 5'-CCACTCACCCTGCATCGAGG-3'. For the T959C mutation (exon 9) the primers used were 5'-ATCATTGGGGTGGATCCCGA-3' and 5'-CGTTGCTCTTGAACCACTTG-3'. For the C1105T mutation, the primers used were the same as in Table 4 .
SSCP was adapted from ref. 18 . Initial PCR reactions were prepared using the primer sequences given in Table 5 and a second PCR was then performed using the same primer pairs and 2 µl of the initial PCR product as template. Amplification was performed in a 20 µl volume containing 20 pM of each primer, 200 mM each of dATP, dGTP, dTTP, 0.2 mM of dCTP and 0.1 µl [[alpha]-32P]dCTP (6000 Ci/mmol), 100 mM Tris-HCl (pH 8.3), 500 mM KCl, 1-5 mM MgCl2 (depending on optimum conditions for each primer pair) and 1 U Taq polymerase. Cycle conditions were the same as for the primary PCR. The resulting radioactive PCR products were then diluted in formamide loading buffer (1 ml in 10 ml buffer). Samples were denatured for 2 min at 95°C, immediately put on ice for 2 min, and 2.5 ml were loaded onto the SSCP gel. A commercial acrylamide, MDE Gel Solution (Flowgen Instruments Ltd.), specifically for SSCP analysis, was used in the preparation of the SSCP gels. SSCP gels were run overnight (14 h) in 6* TBE buffer at 4°C. Voltage was maintained at a constant 200 V. Gels were transferred to 3 MM Whatman filter paper (Alchem), dried at 80°C under vacuum, and exposed to autoradiograph film for 4 h or overnight at -70°C.
All of the point mutations described were engineered in pUC:HCBS (5 ). Mutations were confirmed by sequencing. Mutant CBS cDNAs were cloned into pHCBS[Delta] and introduced into yeast strain WK63yCBS[Delta] by gap repair as described (5 ). Glutathione (30 µg/ml) was used to supply cysteine to the cells.
Yeast nitrogen base with varying amounts of pyridoxine was made using the recipe in ref. 19 . Doubling times were determined as follows. Cells (100 ml) were grown to an OD600 of ~1 in SC-Trp media supplemented with glutathione. The cells were washed three times with dH2O and diluted in 5 ml of SC-Trp media with defined pyridoxine levels to a calculated OD600 of 0.25. Cells were incubated at 30°C for 8 h to deplete existing cellular pyridoxine pools, and then OD followed for an additional 12 h and doubling times were calculated.
Work performed by WDK and CEK was supported in part by grants from the Pew Chritable Trust, National Institutes of Health (HL57299-01), and an appropriation from the Commonwealth of Pennsylvania. Work performed by PMG and ASW was supported by Project Grants from the Irish Heart Foundation and the Irish Health Research Board, and a Unit Grant jointly funded by the Irish Heart Foundation and the Irish Health Research Board. Work performed by ABG, PMU and HR was supported by the Norwegian Council on Cardiovascular Disease and funded in part by EU Commission Demonstration Project Contract no. BMH4-CT95-0505.
*To whom correspondence should be addressed. W. Kruger Tel: +1 215 728 3030; Fax: +1 215 728 3574; Email: wd_kruger@fccc.edu or M. Tsai Tel: +1 612 626 3692; Fax: +1 612 625 6994; Email: tsaix001@maroon.tc.umn.edu
Human Molecular Genetics
Pages
Introduction
Results
Patient characteristics
Mutation screening
A novel G797A mutation in association with the 68 bp duplication
Single strand conformation polymorphism (SSCP)
Functional analysis of CBS mutations in yeast
Discussion
Materials And Methods
Patients and controls
Mutation identification
Testing mutations in yeast
Acknowledgements
References
Patient
Sex
Age at diagnosis (y)
Treatment daily dose
tHcya relative to treatment
B6 responsed
Complications/symptoms
before (µmol/l)
after (µmol/l)
N1
Male
17
15 mg FA, 450 mg B6
151.0
7.6
R
Lens luxation
N2
Female
12
15 mg FA, 900 mg B6 and 6-12 g betaine
152.0
104.1b
N
Lens luxation, marfanoid features, developmental delay
N3
Male
7
15 mg FA, 900 mg B6 10 g betaine
pos
32.7c
N
Lens luxation, convulsions, developmental delay
N4a
Male
25
40 mg B6
pos
11.6
R
Lens luxation
N4b
Male
20
40 mg B6
pos
11.6
R
Lens luxation, marfanoid features, thromboembolic events, severe psychiatric illnesss responding to B6
N5
Female
20
10 mg FA, 80 mg B6 and 12 g betaine
273.2
111.4
N
Small cerebrovascular event
N6a
Female
33
0.4 mg FA, 40 mg B6
245.0
8.0
R
None
N6b
Male
27
0.4 mg FA, 40 mg B6
130.0
18.0
R
Severe psychiatric illness not responding to treatment
N7
Male
19
5 mg FA, 600 mg B6
315.0
230.0
Nc
Lens luxation, marfanoid features
N8
Female
9
60 mg FA, 600 mg B6
pos
6.2
R
Lens luxation, thromboembolic event, cerebral oedema, fully recovered after therapy
N9
Male
23
40 mg B6
221.0
15.8
R
Lens luxation
N10
Male
42
15 mg FA, 450 mg B6 12 g betaine
336.0
50.9
N
Lens luxation, small cerebrovascular event
Patient
68 bp duplication
Mutations
Clinical pyridoxine responsiveness
N1
68/68
G797A/G797A
Yes
N2
wt/wt
C785T/C785T
No
N3
wt/68
G919A/T959C
No
N4a/N4b
68/68
G797A/G797A
Yes
N5
wt/68
G919A/T959C
No
N6a/N6b
wt/wt
T833C/C1105T
Yes
N7
wt/wt
T833C/?
No
N8
68/wt
G797A/?
Yes
N9
68/68
G797A/G797A
Yes
N10
wt/wt
G919A/G919A
No
Mutation
Primers
Buffera
Annealing temp. (°C)
Enzyme
Sizes (bp)
G919A
5'-ATCATTGGGGTGGATCCCGA-3'
J
55
PvuII
112-wt
5'-ACCGTGGGGATGAAGTCGCAG-3'
81, 32-mut
68bp dup
5'-GTCCCCAAAGGCTCTGCTGC-3'
B
62
None
519-dup
5'-GTGGGGATGAAGTCGTAGCC-3'
451-wt
BsrI
451-wt
342, 168-dup
T833C
Same as above
B
62
BsrI
451-wt
342, 100-mut
C785T
5'-GGGCACGGGCGGCACCACCA-3'
E
68
NcoI
193-wt
5'-AACACCTCCCAGGCAGCGCA-3'
172,21-mut
G797A
5'-AAGCTGGACATGCTGGTGGC-3'
C
65
BstNI
98-mut
5'-CCACTCACCCTGCATCGAGG-3'
57, 41-wt
T959C
1st PCR
D
65
5'-CCAGGTGGCACAGGCAGGGA-3'
5'-CATCGTTGCTCTTGAACCACTTGGCC-3'
146-mut
2nd PCR
C
65
MscI
122-wt
5'-GAGACCTCTGGGGTCCTACC-3'
5'-CATCGTTGCTCTTGAACCACTTGGCC-3'
C1105T
5'-GTGGCAGTGCTGGCAGCACG-3'
A
60
HaeII
107-mut
5'-ATGTAGTTCCGCACTGAGTC-3'
68, 38-wt
Exon
Oligo
Sequence
Size
(bp)Annealing
temp. (°C)[MgCl2]
(mM)
1 (a)
S: I21
5'-tgaaccgacgcctctctcctt-3'
98
60
1.0
A: 36
5'-atgctccgagcaggtgcacct-3'
1 (b)
S: 35
5'-agaggataaggaagccaagga-3'
132
60
1.5
A: I22
5'-tgtgatcaaaagcaggactta-3'
2 (a)
S: 3
5'-ggcaaaatctccaaaaatct-3'
127
55
2.5
A: I9
5'-ctccaaagccagggcact-3'
2 (b)
S: I8
5'-ataattgtggactcctct-3'
75
55
1.5
A: 4
5'-gtgtccccgattttcttcag-3'
3
S: 3F
5'-ggtcccctctgtgattcatactct-3'
222
55
1.5
A: 3R
5'-cggcatgggtaggggacaac-3'
4
S: I11
5'-ctctcaccctctgtgtgccctca-3'
82
60
2.0
A: 6
5'-tggcatcacgatgatgcagc-3'
5
S: I24
5'-aaggtgcaggccaccgcttt-3'
166
60
1.5
A: 37
5'-ctggtctaggatgtgagaatt-3'
6
S: 32
5'-taccgcaacgccagcaacccc-3'
100
55
1.0
A: I13
5'-aacgcaatcaagatggacagag-3'
7
S: 7F
5'-ccaggcagggacccaagaat-3'
170
60
1.5
A: 7R
5'-ccactccgcactgtccctct-3'
8
S: I3
5'-gcagttgttaacggcggtat-3'
248
55
2.5
A: I7
5'-ggctctggactcgaccta-3'
9
S: I17
5'-ctgacgggctgtggtggggtcc-3'
115
55
1.0
A: 33
5'-cgcacagcagcccctcttgcgc-3'
10
S: 10F
5'-gcacgtgcacaattcatgcata-3'
278
60
1.5
A: 10R
5'-ggctgccggttctcaggtga-3'
12
S: 12F
5'-gcgagagcgtttgtccttat-3'
~230
60
1.5
A: 12R
5'-ggcagacagaacccaggact-3'
REFERENCES
+Present address: Department of Neuropathology, Beaumont Hospital, Dublin 9, Ireland
This page is maintained by OUP admin. Last updated Sat Oct 18 13:29:50 BST 1997. Part of the OUP Journals World Wide Web service.
Copyright
Oxford University Press, 1997