| Human Molecular Genetics | Pages |
©1999 Oxford University Press |
Molecular basis of congenital Lp(a) deficiency: a frequent apo(a) `null' mutation in Caucasians
Introduction
Results
Identification and population frequencies of a donor splice site mutation in the K IV type 8 intron
Characterization of the splicing defect
Effect of the splice site mutation on plasma Lp(a) concentration
Cloning and expression of the alternatively spliced cDNA analogue in HepG2 cells
Analysis of plasma apo(a) from mutation carriers
Discussion
Materials And Methods
DNA samples
Sequencing of the introns
PCR and DGGE mutation analysis of K IV type 8 exon 1
PCR-based restriction assay of K IV type 8 exon 1
Direct sequencing of PCR products
RNA preparation, first-strand apo(a) cDNA synthesis and cDNA amplification
Expression plasmids
Cell culture and transient transfection
Immunoprecipitation, apo(a) size fractionation and immunodetection
Acknowledgements
References
Molecular basis of congenital Lp(a) deficiency: a frequent apo(a) `null' mutation in Caucasians
Received June 7, 1999; Revised and Accepted July 26, 1999
High plasma concentrations of lipoprotein(a) [Lp(a)], a covalent low-density lipoprotein-apolipoprotein(a) [apo(a)] complex, are associated with coronary heart disease and stroke. Heritability of Lp(a) levels is high and the major locus determining Lp(a) concentrations is the apo(a) gene. We here demonstrate that a G->A substitution at the +1 donor splice site of the apo(a) kringle (K) IV type 8 intron occurs with a high frequency (~6%) in Caucasians but not in Africans and is associated with congenital deficiency of Lp(a) in plasma. This mutation alone accounts for a quarter of all `null' apo(a) alleles in Caucasians. RT-PCR analysis based on apo(a) illegitimate transcription in lympho- blastoid cells demonstrated that the donor splice site mutation results in an alternative splicing of the K IV type 8 intron and encodes a truncated form of apo(a). Expression of the alternatively spliced cDNA analogue in HepG2 cells showed that the truncated apo(a) form is secreted but is unable to form the covalent Lp(a) complex. Immunoprecipitated plasma apo(a) from homozygotes for the mutation was almost completely fragmented. Taken together, our data indicate that a failure in complex formation followed by fast degradation in plasma of the truncated free apo(a) is one mechanism which underlies the null Lp(a) type associated with the donor splice site mutation.
INTRODUCTION
Lipoprotein(a) [Lp(a)] is a macromolecular complex in human plasma which represents a quantitative genetic trait with a heritability of 0.7 or higher (1,2). High concentrations of Lp(a) are associated with atherothrombotic disease i.e. myocardial infarction, stroke, peripheral vascular disease and childhood thromboembolism (for a review see ref. 3). The covalent Lp(a) complex is assembled in plasma from low-density lipoprotein (LDL) and apolipoprotein(a) [apo(a)] (4), a plasminogen-related glycoprotein (5). The apo(a) gene has been demonstrated to be the main determinant of the quantitative Lp(a) trait (1,2,6-8). Two types of variation related to Lp(a) plasma levels have been identified. Variation in the number of plasminogen (plg)-like kringle (K) IV type 2 tandem repeats in the apo(a) gene is inversely correlated to Lp(a) plasma levels in all populations studied so far (for a review see ref. 9). Cell biology studies have shown that the effect of this transcribed variable number of tandem repeats (VNTR) is causal and is due to differences in the secretion rate of the apo(a) isoforms (10-12). Other identified polymorphisms in the apo(a) gene with relevance to the Lp(a) concentration include a pentanucleotide repeat polymorphism (PNP) in the promoter (13,14) and a +93 C/T polymorphism in the untranslated region (15) which introduces a new ATG start codon (16). None of these polymorphisms or other known sequence variations in the apo(a) gene, either alone or in combination, explain the enormous 1000-fold differences in Lp(a) levels among individuals and the several-fold difference across populations, which is one remarkable feature of the Lp(a) trait. Average Lp(a) levels are ~3-fold higher in Africans than in Caucasians and their distribution is close to Gaussian, e.g. in Khoi-San, whereas it is extremely positively skewed in Caucasians, some Asians and Inuit (for reviews see refs 3,9). The reason for this is presently unknown. It may be explained by random differences of apo(a) allele frequencies between populations or by selective forces.
Alleles with the same number of K IV repeats may segregate with Lp(a) concentrations that differ by a factor of 200 (17). Moreover, `operational null alleles' [defined by absence of apo(a) isoforms from immunoblots which depends on the sensitivity of the method] exist over the whole spectrum of apo(a) size alleles but are more frequent within large apo(a) alleles (18,19). Studies in baboon hepatocytes (11) demonstrated that operational apo(a) `null' alleles are either `transcript negative' or `transcript positive'. Cox et al. (20) reported that a splice mutation in the protease domain of apo(a) underlies one of the baboon apo(a) `transcript positive null' alleles and results in intracellular accumulation and degradation of apo(a) which is not secreted. The molecular basis of operational apo(a) `null' alleles in humans is unknown. Some may simply represent alleles with extremely high numbers of K IV type 2 repeats which cause very low secretion rates from cells and intracellular degradation (10-12).
In contrast to apo(a) K IV type 2, the K IV types 3-10 are non-repetitive and this domain of the gene is invariable in size (21,22). In vitro studies with recombinant apo(a) deletion mutants identified the unique K IV types 5-10 as being responsible for all functional properties of apo(a), e.g. lysine/fibrin(ogen) binding and Lp(a) particle formation (23-26), yet analysed. The ability of apo(a) to bind to lysine residues of fibrin(ogen) and to interfere with the fibrinolytic system at several levels is supposed to mediate the pathogenicity of Lp(a) (27-29).
Lp(a) assembly includes a non-covalent interaction between apo(a) and apolipoprotein B-100 (apoB-100) of LDL followed by single disulphide bridge formation realized by K IV type 9 of apo(a) and apoB-100 (23,25,26). The efficiency of Lp(a) formation is supposed to be a factor which might influence Lp(a) levels. Apo(a) is almost lacking from plasma in disorders associated with absent LDL, e.g. abetalipoproteinemia (30).
Because of the presumed functional importance of the apo(a) unique kringles, in vivo mutations and truncated apo(a) forms related to this domain are expected to result in loss of the respective functions. Identification and study of such natural apo(a) variants should therefore be of great interest and may help to understand the still unknown function of apo(a)/Lp(a). They may also clarify whether homozygotes for true-molecularly defined-`null' alleles indeed have no clinical phenotype.
We report here a donor splice site mutation in the intron of apo(a) K IV type 8 which results in a truncated form devoid of K IV types 8, 9 and 10, K V and the protease domain. This mutation is common in Caucasians, accounting for 25% of `null' alleles, but is not present in Africans. We performed cloning and expression of the observed in vivo apo(a) splice variant to specify the mechanism underlying the complete absence of Lp(a) in plasma caused by the mutation.
RESULTS
Identification and population frequencies of a donor splice site mutation in the K IV type 8 intron
The introns of the unique apo(a) kringles were partially sequenced to be used for primer design (data not shown).We then screened both exons of the invariant apo(a) K IV repeats types 6-10 and the flanking intron sequences for mutations by denaturing gradient gel electrophoresis (DGGE) in population samples from Africa (Khoi-San, Black Africans) and Europe (Tyroleans). Apo(a) exons exhibiting variant DGGE patterns were sequenced.
In the course of the study a G->A transition in the donor splice site (position 1) of the 6 kb intron (31) separating the two exons of K IV type 8 was identified (Fig. 1) in individuals from the Tyrolean population sample. In addition we observed the recently described A->C polymorphism at position 66 of the same exon (32). Specific intron-located primers used for amplification of exon 1 of K IV type 8 were designed after partial sequencing of the surrounding introns as described in Materials and Methods. The sequence data are available from GenBank (http://www.ncbi.nlm.nih.gov/Genbank/GenbankOverview.html ).
Figure 1. DGGE analysis of the first exon of apo(a) K IV type 8. Lane 1, a heterozygote with a wild-type allele and the A->C substitution in position 66 (32); lane 2, a heterozygote for the wild-type and the newly identified +1G->A mutation in the K IV type 8 intron; lane 3, homozygous wild-type DGGE pattern.
Although the splice site mutation has a relatively high frequency (q = 0.053) in the Tyrolean sample (n = 113), the respective DGGE pattern was not present in 200 analysed Africans. Since the mutation produces an NdeI restriction site, a PCR-based restriction assay was used to determine the mutation frequency in Finns (n = 126) representing another Caucasian sample. The mutation was found to occur with a frequency of 0.0635 in Finns. In both Tyroleans and Finns, the alleles of the splice site polymorphism were in Hardy-Weinberg equilibrium. Approximately 11% of the Europeans are heterozygous and 0.3% homozygous for the G->A splice site mutation.
Characterization of the splicing defect
The mutation at position +1 of the K IV type 8 intron disrupts the first invariant GT dinucleotide of the donor splice site sequence and predicts a splicing defect. The apo(a) gene is physiologically expressed in liver, and in minor amounts in brain and testis (33,34). These organs cannot be analysed from healthy individuals and we therefore took advantage of the illegitimate (ectopic) transcription phenomenon defined as minor expression of any gene in any cell type (35,36). Correctly spliced illegitimate transcripts as well as alternatively spliced mRNAs resulting from splicing defects have been reported in non-specific cells as bone fide versions of those existing in specific tissues in all situations analysed so far (for a review see refs 37,38). Therefore we used lymphocytes to analyse the possible processing abnormalities of the apo(a) mRNA caused by the splice site mutation in the K IV type 8 intron.
Part of the apo(a) cDNA containing K IV types 7, 8 and 9 (Fig. 2) was amplified by reverse transcription (RT) and subsequent nested PCR using total RNA from EBV transformed lymphocytes from an individual homozygous for the splice site mutation and from a subject with the wild-type sequence. HepG2 cells stably transfected with apo(a) cDNA (K18C cell line) were used as a positive control. An apo(a) cDNA-specific band (846 bp) was visible after the first RT-PCR run with primers 1F/1R from HepG2 RNA but not from lymphocyte RNA. Following the second `nested' PCR amplification with primer pair 2F/2R, a fragment of the expected size of 649 bp was obtained from the lymphoblastoid cell line without the splice mutation and the HepG2 cells (Fig. 3). The sequence of the fragments corresponded exactly to the published apo(a) cDNA sequence (5). In contrast, the cDNA amplicon from the homozygote for the splice site mutation was 64 bp shorter. Sequence analysis revealed a deletion starting from position 97 of the first K IV type 8 exon until the beginning of the second K IV type 8 exon. This is explained by alternative splicing of these exons (Fig. 4). Apparently, the dinucleotide GT in position 97/98 of the first exon is activated as an alternative 5[prime] donor splice site resulting in skipping of the second part of the exon and a change of the reading frame. The third codon following the new exon junction is a UGA stop codon. Therefore a truncated form of apo(a) lacking part of K IV type 8, K IV types 9 and 10, K V and the protease domain is predicted.
Figure 2. Scheme of a section of the apo(a) gene illustrating the positions of the primers for genomic PCR amplification of K IV type 8 exon 1 (8-1F/8-1R) and for nested cDNA PCR (1F/1R and 2F/2R). Primer 1R used for reverse transcription is indicated by an asterisk. The size of the introns is given according to Mihalich et al. (31). The sequence of the exon 1-intron boundary of K IV type 8 is shown under the corresponding part of the gene and the mutated G->A at position 1 of the intron is shown in bold.
Figure 3. Ethidium bromide stained gel showing the apo(a) cDNA fragments obtained by RT-PCR from illegitimate transcripts in lymphocytes. The PCR products result from the second round of the nested cDNA amplification with primer pair 2F/2R. Lane 1, 649 bp RT-PCR amplicon from HepG2 cell line K18C stably transfected with apo(a) cDNA, which served as a positive control; lane 2, 585 bp fragment obtained from a lymphoblastoid cell line from an individual homozygous for the splice site mutation; lane 3, 649 bp fragment corresponding to apo(a) cDNA with correctly spliced exons of K IV types 7, 8 and 9 obtained from a lymphoblastoid cell line from an individual with the wild-type splice site sequence; lane 4, negative control for the RT-PCR.
Figure 4. Apo(a) cDNA sequence resulting from normal and alternative splicing of the K IV type 8 intron. (A) Part of the cDNA sequence of K IV type 8 starting at position 78 (upper line) and the corresponding amino acid sequence (lower line) according to McLean et al. (5). The beginning of exon 2 is indicated by an arrow. The dinucleotide GT at position 97/98 of the wild-type sequence, which serves as an alternative donor splice site, is shown in bold. (B) cDNA sequence resulting from alternative splicing of the K IV type 8 intron (upper line) and the predicted protein sequence (lower line). The changed codons are underlined.
Effect of the splice site mutation on plasma Lp(a) concentration
The type of mutation identified in apo(a) suggests that it may either affect the apo(a) protein abundance (via decreased translation or secretion) and/or impair the Lp(a) formation and hence the Lp(a) plasma concentration. Information on the apo(a) K IV type 2 repeat polymorphism in the total population sample was available at the DNA (PFGE) and protein (SDS-AGE) level and Lp(a) plasma concentrations were also known (19; unpublished data). Therefore, it was possible to estimate by different approaches whether or not the splice site mutation affects Lp(a) levels.
We first noted that all individuals heterozygous for the splice site mutation had only one or no detectable apo(a) isoform suggesting that the mutation results in a true null type. [As an operational `null' apo(a) allele we defined here an allele that is associated with <0.01 mg/dl plasma Lp(a) since this is the lower limit of detection of our SDS-AGE immunoblotting (19). As true `null' alleles we refer to molecularly defined apo(a) null alleles.]
Next we performed an analysis of Lp(a) concentrations in homozygotes for the wild-type allele in comparison with hetero- and homozygotes for the splice site mutation. Lp(a) concentrations were significantly lower in G/A heterozygotes than in G/G homozygotes and lowest in a splice site mutation (A/A) homozygote (Table 1) indicating a clear gene dosage effect. This result did not change when the effect of the K IV type 2 VNTR on the Lp(a) concentration was considered. This was done by calculating expected individual Lp(a) concentrations in the three groups from the average K IV type 2 allele associated Lp(a) concentrations using multiple linear regression analysis as described (14,15). The difference between measured and expected Lp(a) level was defined as [Delta]Lp(a). Comparison of [Delta]Lp(a) levels of the wild-type and mutant genotypes showed significantly lower than expected Lp(a) levels in carriers of the A-alleles (Table 1) demonstrating that the effect of the splice mutation on Lp(a) concentration is not caused by random differences in K IV repeat allele frequencies.
Table 1. Mean Lp(a) and [Delta]Lp(a) levels associated with the G->A splice site mutation and with the wild-type sequence in the pooled Caucasians
| n | Mean Lp(a) (mg/dl) | Mean [Delta]Lp(a)a | |
| Total population | 223 | 10.57 | -0.62 |
| G/G (wild-type) | 198 | 11.6 | 0.23 |
| G/A | 24 | 2.49 | -6.94 |
| A/A | 1 | 0.03 P = 0.0000b |
-15.09 P < 0.02b |
bBy Kruskal-Wallis test.
The ELISA assay is designed to measure Lp(a) concentrations (39). The value of 0.03 mg/dl in the subject A/A results from immunoreactivity produced by the apo(a) fragments present in the splice mutation homozygotes (Fig. 9).
Third, we performed family studies where the cosegregation of apo(a) haplotypes carrying the splice variant with the non-expressed isoforms representing `null' apo(a) alleles was followed (Fig. 5). Haplotypes were defined by the K IV type 2 repeat, 5[prime] PNRP and the Met/Thr polymorphism in K IV type 10. In all five families analysed (data not shown for three of the families) the splice site mutation segregated with the `null' allele. The number of K IV repeats associated with the mutant haplotype was different for each family (26, 28, 30, 33 and 35, respectively).
Figure 5. Co-segregation of the splice site mutation haplotypes and apo(a) isoforms in two families. (A) Pedigrees together with the haplotypes defined by four apo(a) polymorphisms. From top to bottom: 5[prime] PNRP (given by the number of repeats); K IV type 2 VNTR (given by the number of K IV repeats); M/T polymorphism in K IV type 10; and +1G->A splice site mutation K IV type 8 intron. The mutant haplotypes are shaded. (B) apo(a) K IV type 2 genotyping of the family members performed by PFGE/Southern blotting. The size alleles which co-segregate with the splice site mutation are indicated by an asterisk. Markers with known repeat numbers are given in the first and last lanes. (C) Apo(a) phenotyping of the family members performed by SDS-AGE/western blotting. Isoform size was determined from the standard for apo(a) isoforms (Immuno, Vienna, Austria) (lane S). The expected positions of the isoforms corresponding to the splice mutation alleles are indicated by an asterisk.
The distribution of the splice site mutation on K IV type 2 size alleles in comparison to all `null' alleles in the Austrian and Finnish population samples is graphically represented in Figure 6. As mutant alleles we considered the non-expressed alleles in the heterozygous splice mutation carriers with one expressed allele. The distribution of the splice site mutation on the K IV type 2 alleles is broad which may reflect the dynamics of the K IV type 2 polymorphism or indicate that the mutation is old.
Figure 6. Histogram showing the frequency distributions of all apo(a) `null' alleles (total bars) and of the splice site mutation in K IV type 8 (shaded) in relation to the K IV type 2 repeat polymorphism in Caucasians.
Cloning and expression of the alternatively spliced cDNA analogue in HepG2 cells
In order to analyse the consequences of the mutation underlying the null apo(a) type in vivo we cloned and expressed the alternatively spliced apo(a) variant in HepG2 cells. One reason for the null type could be an intracellular retention and degradation of the truncated protein, as was recently observed for a truncated apo(a) form in baboon (20). Alternatively, occurrence of a premature termination codon (PTC) as a result of alternative splicing may lead to a decrease in the mRNA abundance (for a review see ref. 40). To test the mechanism(s) underlying the `null' apo(a) type, we constructed an expression plasmid pCMV-A18-AS encoding an mRNA corresponding exactly to the identified alternatively spliced messenger, i.e. the PTC created by the frameshift was followed by the complete downstream sequence up to the normal stop codon and 5[prime]-untranslated region (Fig. 7). The wild-type expression plasmid pCMV-A18 determined r-apo(a) with eight K IV type 2 repeats, all unique K IV types, K V and the protease domain (23). As an additional control for r-apo(a) molecular weight and properties we used the deletion mutant pCMV-A18[Delta]V-P which also codes for r-apo(a) with eight K IV type 2 repeats but is devoid of K V and the protease domain (24).
Figure 7. Scheme of the mutant apo(a) cDNA carried by the construct pCMV-A18-AS. K IV types are numbered from 1 to 10. SP, signal peptide; V, kringle V; PD, protease domain. The AatII/MscI fragment derived by RT-PCR and used to replace the corresponding part of the wild-type plasmid pCMV-A18 is indicated by a shaded area. The alternatively spliced K IV type 8 is shown by darker shading. The cDNA is flanked at the 5[prime] end by the CMV enhancer/promotor and a heterologous intron and at the 3[prime] end by an SV40 polyadenylation site (23).
HepG2 cells known to produce apoB-100 containing lipoproteins (e.g. LDL), but not apo(a) or Lp(a) (41), were transiently transfected with the wild-type and mutant plasmids. The expressed apo(a) variants were immunoprecipitated from cell lysates and media with a polyclonal anti-apo(a) antibody and size-fractionated by reducing SDS-PAGE (Fig. 8A). The r-apo(a) encoded by the splice mutant construct showed the expected decrease in the molecular weight (Fig. 8A). High level of expression demonstrated that, under experimental conditions, the wild-type and alternatively spliced mRNA templates did not differ significantly in stability.
Figure 8. Immunoblotting of r-apo(a) from transfected HepG2 cells. Constant amounts of cell lysates (300 µl) or cell medium (600 µl) from HepG2 cells transiently transfected with pRc/CMV (mock control, lane 1), wild-type pCMV-A18 (lane 2), splice mutant pCMV-A18-AS (lane 3) and deletion mutant pCMV-A18[Delta]V-P (lane 4) were subjected to immunoprecipitation with polyclonal anti-apo(a) antibody (A) or with anti-apoB-100 antiserum (B). The immunoprecipitates were analysed by reducing SDS-PAGE followed by immunoblotting with the monoclonal anti-apo(a) antibody 1A2. The positions of the precursor [pr-apo(a)] and mature [m-apo(a)] forms are indicated. Cell lysate and cell medium from K18C cells which stably express apo(a) served as a positive control for the immunoprecipitations (lane +).Cell medium from K18C cells was immunoprecipitated with rabbit pre-immune serum as a negative control for the immunoprecipitations (lane -). Human plasma with apo(a) isoforms of known K IV repeat length was applied as apo(a) size standard (lane S).
Like the wild-type protein, the mutant form was present in the cell lysate mainly as a precursor and in small amounts as a mature, glycosylated form (42) suggesting no difference in the protein processing rate. Fragments which might indicate intracellular degradation of the mutant form were not observed. Apo(a) ELISA (data not shown) and immunoblotting of the culture supernatants (Fig. 8A) demonstrated that the wild-type and truncated r-apo(a) are secreted in comparable amounts. Therefore decreased secretion cannot underlie the minimal apo(a) concentration determined by the splice site mutation in vivo.
The truncated apo(a) form resulting from the donor splice site mutation in the K IV type 8 intron is lacking K IV type 9 and therefore we predicted that it will be unable to bind covalently to apoB-100 and to form a stable Lp(a) particle. To confirm this we immunoprecipitated the apoB-100-containing lipoproteins from the transfected cell media and subjected the precipitate to anti-apo(a) immunoblotting (Fig. 8B). Apo(a) was co-precipitated from the media of cells expressing wild-type protein or apo(a) deletion mutant (pCMV-A18[Delta]V-P) containing K IV types 1-10 indicating the presence of r-Lp(a) particles. In contrast, the truncated apo(a) generated by alternative splicing of K IV type 8 was not co-immunoprecipitated by the anti-apoB antibody. This indicates that it exists in the cell supernatant only as an LDL-unbound, free form as predicted.
Analysis of plasma apo(a) from mutation carriers
Only minimal amounts of apo(a) immunoreactivity [<0.03 mg/dl `Lp(a)'] were detected by ELISA in plasma from the identified homozygous individuals (sibs II.2 and II.4 from family 1; Fig. 5). To determine whether the predicted truncated forms circulate in trace amounts in plasma we immunoprecipitated apo(a) from a homozygous mutation carrier and for comparison from controls with low Lp(a) levels and analysed the immunoprecipitates by reducing SDS-PAGE.
A major apo(a) immunoreactive protein with a size of ~95 kDa was detected in plasma from the homozygote for the splice site mutation (Fig. 9, subject C). Additionally, eight larger apo(a) fragments that ranged in size from ~120 to 600 kDa were visible after prolonged exposure of the gel. An approximate molecular weight of 600 kDa is calculated for the truncated apo(a) form of subject C consisting of 24 K IV type 2 repeats and unique K IV types 1-7. Therefore the faint band of the highest molecular weight (~600 kDa) most likely represents the intact mutant apo(a). However, the observed pattern of apo(a) immunoreactive fragments indicates that the truncated apo(a) is subjected to extensive degradation since the intensity of the bands increased with the reduction of the size and the smallest fragment (~95 kDa) was the main one present in the plasma of this individual. Identical apo(a) degradation was seen in plasma from the sib of subject C carrying the same mutant apo(a) alleles (data not shown).
Figure 9. Immunoblot analysis of plasma apo(a) from a homozygous for the splice site mutation subject C and wild-type individuals (A and B). Apo(a) was immunoprecipitated from plasma aliquots (100 µl from subject A, 300 µl from subject B, 600 µl from subject C) and size fractionated by reducing SDS-PAGE. The subsequent immunoblot was performed with the anti-apo(a) antibody 1A2. The sizes of the apo(a) fragments common for all individuals are indicated and the intact isoforms are marked by arrows. The position of the ~600 kDa immunoreactive protein observed in subject C is indicated by an arrow and an asterisk. Lp(a) levels and genotypes and phenotypes of theapo(a) K IV type 2 VNTR polymorphism of the analysed individuals are given in the text.
In contrast, the apo(a) banding pattern of control individuals (subjects A and B) consisted of the major full-length isoforms and a series of discrete fragments which decreased in size and intensity. The smallest apo(a) fragment detected in the controls migrated also to a position of ~95 kDa but was only present in trace amounts.
The anti-apo(a) K IV type 2 immunoreactive fragments with sizes of ~95, ~120, ~140 and ~160 kDa were found to be common for the mutation carrier and the wild-type individuals (Fig. 9). They differed in size by ~20-25 kDa, which corresponds to the estimated molecular mass of a single kringle and apparently originate from the NH2-domain of the protein.
DISCUSSION
We describe here a splice site mutation in the human apo(a) gene which occurs with an overall allele frequency of ~6% in two tested European populations (Austrians and Finns). This mutation has a profound decreasing effect on Lp(a) plasma concentration and is the first molecularly defined apo(a) allele resulting in an Lp(a) null phenotype in humans. Family studies demonstrated co-segregation of the mutation with non-expressed apo(a) alleles [non-detectable intact apo(a) on immunoblots]. Homozygotes completely lacked Lp(a) in their plasma and only minimal apo(a) immunoreactivity was detectable by apo(a) ELISA.
The G->A mutation reported here disrupts the consensus GT dinucleotide at the donor splice site of the 6 kb intron of apo(a) K IV type 8 and activates a cryptic splice site in the first exon of K IV type 8. This results in skipping of the last 64 bases of the exon, a new exon junction, and a frameshift leading to a stop codon. The predicted apo(a) truncated forms consisting of K IV types 1-7 were, however, not detected by standard SDS-AGE phenotyping in plasma from splice site mutation carriers (Fig. 5) but only after immunoprecipitation and in minimal amounts (Fig. 9).
Cloning of the alternatively spliced apo(a) messenger and further expression in hepatocarcinoma cell line HepG2 indicated that the protein deficiency in plasma is not solely due to reduced mRNA abundance. The expression experiments also demonstrated that the null apo(a) type in plasma is not due to defective secretion of the truncated protein. Our in vitro experiments, however, do not rule out that PTC-mediated RNA decay (40) contributes to the null Lp(a) type. As for truncating apoB mutations (43), reduced mRNA levels may be one factor contributing to the reduced apo(a) concentration.
The truncated apo(a) is lacking K IV type 9 which realizes the covalent binding of apo(a) to apoB-100 (23). One reasonable explanation for the observed null Lp(a) type is, therefore, that in addition to possible PTC-mediated mRNA instability the mutation also prevents the formation of the covalent apo(a)-apoB-100 complex followed by rapid degradation of the uncomplexed apo(a). The primary non-covalent docking of apo(a) to LDL which is believed to involve K IV type 8 sequences and precedes the disulphide bridge formation might also be impaired. Some authors (26,44) have suggested that LDL-unbound apo(a) is cleared rapidly from the plasma. Indeed, free apo(a) represents a small fraction of the total plasma apo(a) from control individuals and was shown to be fragmented rather than intact (45).
Immunoprecipitation experiments with r-apo(a) expressed in HepG2 cells indeed demonstrated that the truncated r-apo(a) resulting from alternative splicing of K IV type 8 was not complexed with apoB-100 in the culture medium. However, it was intact. In contrast, analysis of immunoprecipitated apo(a) from plasma of homozygous mutation carriers demonstrated predominantly fragments with low molecular weight which suggests extensive degradation of the mutant form. Fragments with the same size were detected also in control individuals though their quantitative distribution was different (Fig. 9). Apo(a) fragments in plasma have been shown to be derived from the N-terminal domain of the protein and to contain a variable number of K IV type 2 repeats (45). Apparently the peptides of small size common for all tested subjects are derived by a general apo(a) degradation pathway that is promoted for the truncated mutant form. The degradation of the LDL-free mutant apo(a) is in agreement with the hypothesis that unassembled apo(a) is rapidly removed from plasma. Neither truncated nor wild-type r-apo(a) were found to be fragmented in culture medium from HepG2 cells indicating that factors in human blood or tissues are responsible for the apo(a) fragmentation, probably by proteolytic cleavage.
Because of their high molecular weight, the apo(a) fragments are supposed to be excreted into the urine by secretion rather than by filtration (45). We found accumulation of the smallest described fragment representing the end degradation product of apo(a) in plasma. This may suggest that the apo(a) fragments enter the urine via a specific secretion mechanism with limited capacity.
Usually mutations that disrupt the structure and expression of human genes and impair their function result in-or predispose to-genetic disease. In contrast, studies in mice have identified numerous genes where homozygous null mutant phenotypes appear to be normal, which is explained by genetic redundancy. However, this seems not to be a likely explanation for null mutation in the human apo(a) gene because no lipoprotein with characteristics similar to Lp(a) is known to exist in mammalian plasma. The situation may therefore be more complex for the human apo(a) gene which is the major gene controlling concentrations of the Lp(a) complex, high concentrations of which are associated with atherothrombotic disease (myocardial infarction and stroke) (for a review see ref. 3). On the contrary, individuals with very low Lp(a) in plasma, as these first molecularly defined individuals with congenital Lp(a) deficiency identified here, appear to be healthy. This has led to the belief that Lp(a) has no vital function. In line with this, no physiological function of apolipoprotein(a) or its complex with LDL is presently known.
Lp(a) may exert its function only in certain situations, e.g. when challenged by environmental factors like pathogens. In such situations low or absent Lp(a) may represent a susceptibility state, and high Lp(a) may be protective. Hence association of congenital Lp(a) deficiency with a clinical phenotype may be difficult to detect and may even not or only rarely occur in some ethnic groups or geographic areas. Such a scenario could explain the lower Lp(a) levels and the presence of the splice site mutation in Caucasians as opposed to Africans.
Random changes in the frequencies of alleles associated with different apo(a) concentrations during the migration and expansion of modern humans are not expected to result in the observed systematic directional changes in the Lp(a) concentration distribution, i.e. from high and almost Gaussian in Khoi-San to low and extremely positively skewed distributions in most non-Africans and occurrence of null Lp(a) phenotypes. `Loss of concentration' mutations may have accumulated in Caucasians due to loss of selective pressure (e.g. by certain pathogenes) that maintains Lp(a) in Africans at high levels.
The described null mutation in the apo(a) gene is not present in Khoi-San, an evolutionary `old' human ethnic group, and in Black South Africans. On the contrary, it was found in two independent Caucasian samples (Austrians and Finns). This suggests that the splice site mutation occurred after the split of Africans and non-African populations. Nevertheless the mutation was associated with a broad range of K IV type 2 repeats (from 21 to 38) which indicates that the size polymorphism of the apo(a) gene has undergone considerable expansion/contraction during human evolutionary history. More data will allow the study of the dynamics of this process. Also the true reason for the difference in Lp(a) levels between ethnic groups, e.g. the type of selection operating in Africa, still needs to be identified.
MATERIALS AND METHODS
DNA samples
The Tyrolean (n = 113) and African (Black South Africans, n = 100; Khoi-San, n = 100) population samples in addition to 30 White South African families used in this study have been described previously (19,46). All individuals in the population samples were unrelated. Additionally, DNA containing agarose plugs were prepared from EDTA blood as described (19) from 126 unrelated Finnish individuals collected as a random subset of the population and from members of Tyrolean and Finnish families of probands with the identified splice mutation. Lp(a) concentration, geno- and phenotypes of the apo(a) K IV type 2 repeat polymorphism and other intragenic polymorphisms were known for the Tyrolean and African samples (14,15,19) or determined as described elsewhere (14,19).
Sequencing of the introns
The introns surrounding the first exon of K IV type 8 were amplified from YAC clone (ID no. ICRFy900C0535) containing the human apo(a) gene (22) using described primers and PCR conditions (31). The PCR products were partially sequenced by cycle sequencing (see below) with the following primers in exon 1 of K IV type 8: 8-1seq F, 5[prime]-TCAGTCTTGGTCGTCTATG-3[prime] and 8-1seq R, 5[prime]-CCTCGATAACTCTGTCCAT-3[prime].
PCR and DGGE mutation analysis of K IV type 8 exon 1
The first exon of K IV type 8 was amplified by PCR with a set of primers situated in the introns: 8-1F, 5[prime]-CTTTTCTAGGCAATA- CTGAGC-3[prime] (positions -24 to -44) and 8-1R, 5[prime]-GTTCCTTTTATGGCTAACATG-3[prime] (positions +11 to +31). For DGGE, GC reach sequences (GC clamps) were included at the 5[prime] ends of the primers (50 bases to the forward primer and 20 bases to the reverse one) to obtain the highest sensitivity of mutation detection (47). The melting profile of the fragment was analysed by the computer algorithm Melt 87 (48). The PCR was carried out in 50 µl final volume as described (49). The reaction was performed within 35 cycles with the following steps: 94°C for 40 s, 57°C for 40 s and 72°C for 40 s. The PCR started with an initial denaturation at 94°C for 5 min and ended with the final extension at 72°C for 7 min. The PCR products (with size of 305 bp including 70 bp GC-clamps) were concentrated by lyophilization and diluted in 14 µl DGGE loading buffer (20% Ficoll 400, 10 mM Tris-HCl pH 7.5, 1 mM EDTA, 0.25% xylene cyanol and 0.25% bromophenol blue). Prior to the DGGE run the probes were denatured at 98°C for 3 min and allowed to reanneal at room temperature for 10 min. DGGE was run in 9% polyacrylamide gels for 12-16 h at 80 V (6 V/cm) in 1× TAE buffer. The electrophoretic equipment was provided by C.B.S. Scientific (Del Har, CA), model no. DGGE-4000. The denaturing conditions were created by 60°C constant temperature of the electrophoresis chamber and urea/formamide gradient in the gel (2.1-4.9 M urea and 12-28.8% formamide). After the run the gels were stained with ethidium bromide and analysed under UV light.
PCR-based restriction assay of K IV type 8 exon 1
PCR amplification of exon 1 of K IV type 8 with primer pair 8-1F/8-1R was carried out at the described conditions. Seven microlitres of the reaction were incubated with 10 U NdeI and 1× restriction buffer R+ (MBI Fermentas, Amherst, NY) in 20 µl final volume for 6 h at 37°C. NdeI digestion of K IV type 8 exon 1 amplicons containing the splice site mutation resulted in two fragments with sizes of 204 and 31 bp. The electrophoretic separation was performed on a 2% agarose gel.
Direct sequencing of PCR products
The PCR products of interest were purified on Qiaquick columns (Qiagen, Hilden, Germany) and then served as a template for a single strand cycle sequencing using Big Dye Terminator Cycle Sequencing Ready Reaction kit (Perkin-Elmer, Norwalk, CT). Sequencing reactions were carried out in 40 cycles according to the manufacturer's protocol. After the reaction the probes were purified on Sepharose columns (Perkin-Elmer) and were analysed on an ABI PRISM 310 Genetic Analyzer. Each PCR product was sequenced in the forward and reverse direction with the primers used for the corresponding PCR.
RNA preparation, first-strand apo(a) cDNA synthesis and cDNA amplification
Total cellular RNA was isolated from human apo(a) stably transfected HepG2 cell line K18C (21) and EBV transformed lymphocytes using TRIzolTM Reagent (Gibco BRL Life Technologies, Paisley, UK). Twenty milligrams of total RNA was reverse transcribed with primer 1R, 5[prime]-CTGATG- CCAGTGTGGTATC-3[prime] using AMV Reverse Transcriptase (Boehringer Mannheim, Mannheim, Germany) according to the manufacturer's protocol. cDNA-nested PCR amplification was performed in two subsequent rounds as described (50). Four microlitres of the RT reaction served as a template for the first PCR round with primers 1F, 5[prime]-GCGACGTCCACGGCTGT-3[prime] and 1R. Two microlitres of the amplification reaction were used further as a template for the second nested PCR run with primer pair 2F, 5[prime]-CACACTGGCATCAGAGAAC-3[prime] and 2R, 5[prime]-GTGACAGTGGTGGAGGATA-3[prime]. Both PCR rounds were carried out in 40 cycles at the following steps: 94°C for 40 s, 61°C for 40 s, 72°C for 1 min. The RT reaction and the cDNA-PCR amplification were controlled for RNA/cDNA contamination by RNA-free negative control for the RT served further as a `template' for the nested PCR. In addition, each PCR round was also carried out with a second DNA-free negative control.
Expression plasmids
The apo(a) expression plasmid pCMV-A18 and the deletion mutant pCMV-A18[Delta]V-P have been described previously (23,24). The pCMV-A18-AS (Fig. 7) carrying the alternatively spliced cDNA analogue was constructed as follows: total RNA isolated from splice mutant lymphoblastoid cell line was reverse transcribed with primer 3R, 5[prime]-TAAAACACCAAGGGCCTGT-3[prime] as described above. The RT was followed by nested PCR with primer pairs 3F, 5[prime]-TTATACCATGGATCCCAAT-3[prime] and 3R for the first round, and 4F, 5[prime]-CAAGTGTCCTTGCGACGTC- CACGGCTGT-3[prime] and 4R, 5[prime]-ATGCCGGTGTGGTGTCATG-3[prime] for the second round. For the PCR high fidelity Pwo DNA Polymerase (Boehringer Mannheim) was used. The amplified cDNA (1132 bp) covered the alternatively spliced K IV type 8 and the neighbour K types 6, 7, 9 and 10. This fragment was digested with AatII/McsI (New England BioLabs, Beverly, MA) and used to replace the corresponding restriction fragment of the wild-type plasmid pCMV-A18 via several restriction and subcloning steps using standard cloning techniques (51). The complete cloned sequence derived from RT-PCR and the restriction/ligation sites used for subcloning were verified by direct automated sequencing in the mutant plasmid pCMV-A18-AS.
Cell culture and transient transfection
Untransfected human hepatocarcinoma cells (cell line HepG2) and the HepG2 cell line K18C stably transfected with pCMV-A18 were cultured as described (10,23). Transient transfection of HepG2 cells was performed by lipofection. Cells (0.7 × 106) were seeded in [empty] 60 mm culture dishes 24 h before transfection. The DNA transfection with Lipofectamine reagent (Gibco BRL) was carried out according to the manufacturer's protocol. After overnight incubation the transfection mixture was replaced by 2 ml growth medium and the cells were allowed to recover for 48 h prior to analysis of gene expression.
Immunoprecipitation, apo(a) size fractionation and immunodetection
Human plasma aliquots stored no longer than 2 months at -80°C and fresh cell lysate/cell medium preparations adjusted to end concentration of 0.6% SDS and 1% Triton X-100 in HEPES-buffered saline (HBS) were treated with proteinase inhibitors (1 mM PMSF, 5 µg/ml of each aprotinin and leupeptin) and subsequently precleared with 100 µl Pansorbin (Calbiochem, San Diego, CA) under gentle agitation at 4°C for 30 min. The tubes were centrifuged for 10 min at 12 000 g and the supernatants were mixed with 40 µl 10% Protein A Sepharose (Pharmacia, Uppsala, Sweden) and 4 µl of a monospecific polyclonal rabbit anti-apo(a) antibody or a rabbit anti-apoB-100 antiserum raised against purified human LDL (Behringwerke, Marburg, Germany). Incubation with a rabbit pre-immune serum was performed as a negative control for the immunoprecipitation. After 16 h incubation under gentle agitation at 4°C the samples were centrifuged as described. The pellet was resuspended in 1 ml washing buffer (0.2% SDS, 1.25% Triton X-100, 1 mM PMSF, 5 µg/ml of each aprotinin and leupeptin in HBS) followed by centrifugation at 10 000 g for 2 min. This procedure was repeated twice. The final pellet was solubilized in 15 µl of sample buffer and subjected to reducing SDS-PAGE on commercially available 4-12% Bis-Tris gels according to the manufacturer's protocol (NOVEX, San Diego, CA). The immunoblotting was performed as described (23) with the monoclonal antibody 1A2 recognizing K IV type 2 of apo(a) (52).
ACKNOWLEDGEMENTS
We are grateful to Dr Hans Georg Kraft for helpful advice on statistics. The excellent technical assistance of Susanne Rauchenwald is acknowledged. We thank Dr Christian Ehnholm and Dr Esa Tahvanainen (National Public Health Institute, Finland) for providing the Finnish population and family samples, and Marita Kotze (Stellen-bosch, South Africa) for the South African family samples. This work was supported by grant P11695 from the Austrian Fonds zur Förderung der wissenschaftlichen Forschung and by grant PL951678 from the Federal Ministry for Science and Transport (BIOMED 2 Project) to G.U.
REFERENCES
+To whom correspondence should be addressed. Tel: +43 512 507 3450; Fax: +43 512 507 2861; Email: gerd.utermann{at}uibk.ac.at
This page is run by Oxford University Press, Great Clarendon Street, Oxford OX2 6DP, as part of the OUP Journals
Comments and feedback: jnl.info{at}oup.co.uk
Last modification:
Copyright© Oxford University Press, 1999.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  L. A. Cox, S. Birnbaum, M. C. Mahaney, D. L. Rainwater, J. T. Williams, and J. L. VandeBerg Identification of Promoter Variants in Baboon Endothelial Lipase That Regulate High-Density Lipoprotein Cholesterol Levels Circulation, September 4, 2007; 116(10): 1185 - 1195. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. M. Luke, J. P. Kane, D. M. Liu, C. M. Rowland, D. Shiffman, J. Cassano, J. J. Catanese, C. R. Pullinger, D. U. Leong, A. R. Arellano, et al. A Polymorphism in the Protease-Like Domain of Apolipoprotein(a) Is Associated With Severe Coronary Artery Disease Arterioscler Thromb Vasc Biol, September 1, 2007; 27(9): 2030 - 2036. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J-P Chretien, J Coresh, Y Berthier-Schaad, W H L Kao, N E Fink, M J Klag, S M Marcovina, F Giaculli, and M W Smith Three single-nucleotide polymorphisms in LPA account for most of the increase in lipoprotein(a) level elevation in African Americans compared with European Americans J. Med. Genet., December 1, 2006; 43(12): 917 - 923. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Rubin, H. J. Kim, T. A. Pearson, S. Holleran, R. Ramakrishnan, and L. Berglund Apo[a] size and PNR explain African American-Caucasian differences in allele-specific apo[a] levels for small but not large apo[a] J. Lipid Res., May 1, 2006; 47(5): 982 - 989. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Ogorelkova, H. G. Kraft, C. Ehnholm, and G. Utermann Single nucleotide polymorphisms in exons of the apo(a) kringles IV types 6 to 10 domain affect Lp(a) plasma concentrations and have different patterns in Africans and Caucasians Hum. Mol. Genet., April 1, 2001; 10(8): 815 - 824. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Klose, F. Fresser, S. Kochl, W. Parson, A. Kapetanopoulos, J. Fruchart-Najib, G. Baier, and G. Utermann Mapping of a Minimal Apolipoprotein(a) Interaction Motif Conserved in Fibrin(ogen) beta - and gamma -Chains J. Biol. Chem., December 1, 2000; 275(49): 38206 - 38212. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||