Single gene disorders can be studied at the level of 1) the mutant gene product, or protein; 2) the message, or messenger RNA [mRNA]; and 3) the genome, or DNA. Study at the protein level provides an understanding of the biochemical pathologic features underlying the disorder. Immunologically based assays, such as the Western blot, determine the size and amount of protein in cell extracts or clinical samples [1]. Other methods of protein analysis (such as amino acid sequencing, x-ray crystallography, enzymatic assays, cell-binding assays, and so on) allow the protein's structure and function to be determined. Analysis of mRNA provides information on gene expression. Although somatic cells contain the same genomic DNA code, complex regulatory mechanisms determine which genes are actively expressed and transcribed into mRNA in specific tissues. Northern blotting determines the size and amount of a particular mRNA isolated from a tissue; it involves separating the mRNA by electrophoresis according to its size, transferring the mRNA to a membrane, and hybridizing the membrane with a labeled nucleic acid probe. Autoradiography can then visualize the mRNA [2]. The mRNA can also be reverse-transcribed into complementary DNA (cDNA). The cDNA reflects the gene's exon sequences, which simplifies analysis of large genes. The structural basis of a genetic disorder also can be determined by studying the genomic DNA.

Methodologic advances in molecular genetics in the last decade have facilitated the study of inherited diseases at the DNA level. In particular, the polymerase chain reaction (PCR) allows selective in vitro amplification of specific gene segments from genomic DNA or cDNA [3, 4]. The characteristics of the amplified segment can be analyzed by various techniques to detect mutations or polymorphisms [5].

We reviewed some methods for DNA analysis used to study the apolipoprotein E (apo E) gene and type III hyperlipoproteinemia, an inherited disorder of lipid metabolism that involves a variant form of apo E (Figure 1).

View larger version (21K): [in this window] [in a new window]   Figure 1. Methods of DNA analysis of the apolipoprotein E (apo E) gene. The polymerase chain reaction (PCR)-based methods shown were used to detect and characterize in detail the apo E gene mutations and to determine apo E genotype. ASO = allele-specific oligonucleotide; RFLP = restriction fragment length polymorphism; DGGE = denaturing gradient gel electrophoresis.

Methods for DNA Analysis

DNA Preparation

Human genomic DNA consists of 3 x 10⁹ nucleotide bases and includes the coding (exons) and noncoding (introns) regions of all human genes. Because genomic DNA is present in all nucleated cells, any tissue can be used for genomic DNA analysis. For example, DNA for prenatal diagnosis is obtained from fetal cells in amniotic fluid or from chorionic villus cells. In adults, peripheral blood leukocytes are an abundant, accessible DNA source. The buffy coat from 5 mL of blood yields enough DNA for most types of genetic testing.

Several methods for extracting DNA from leukocytes have been described. Cell lysis by boiling the blood sample for a few minutes releases the DNA for analysis [6, 7]. However, the quality and yield of DNA obtained by this method is inconsistent. To ensure the recovery of high-molecular-weight DNA, more time-consuming and costly enzymatic methods are used [8, 9]. Recently introduced nonenzymatic methods provide milligram quantities of high-quality DNA from 10 mL of whole blood in less than 24 hours [10]. Once purified, DNA can be preserved indefinitely in aqueous solution; DNA is remarkably resistant to degradation, as indicated by its use in anthropologic and forensic investigations [11-13].

Another source of DNA for analysis is cDNA synthesized using mRNA isolated from a particular tissue as a template. For example, if a protein is produced preferentially in the liver, mRNA can be isolated from a liver biopsy specimen and transcribed into cDNA using the enzyme reverse transcriptase [14]. The cDNA contains only the coding sequences of the genes expressed in that tissue because the introns are excised during post-transcriptional mRNA processing. Standard methods of DNA analysis can be applied to the cDNA.

Whatever the source of the DNA, sufficient quantity of the region to be studied must be available. Polymerase chain reaction has most revolutionized DNA analysis, allowing it to begin to have a substantial effect in clinical medicine.

Polymerase Chain Reaction

Polymerase chain reaction, an in vitro method of DNA amplification, relies on the repetitive cycling of three events at different temperatures. It can amplify DNA fragments and yield quantities sufficient for analysis from very small samples, including tissue in paraffin block sections, cytologic smears, hair, urine, and even single cells [13, 15-18]. The required reagents are 1) a DNA template [usually leukocytic genomic DNA]; 2) two single-stranded synthetic DNA oligonucleotides with sequences complementary to the areas flanking the region to be amplified; 3) the four deoxyribonucleotide triphosphates [deoxyadenosine triphosphate, deoxycytidine triphosphate, deoxyguanosine triphosphate, and deoxythymidylic acid triphosphate]; and 4) a DNA polymerase [19].

The PCR cycles consist of 1) denaturation of the double-stranded DNA template to two single strands at 94 °C; 2) annealing of the primers to the single-stranded DNA templates at a specific temperature, usually at about 60 °C; and 3) extension of the primers in both directions along the template, typically at 72 °C [19, 20]. The cycles are repeated about 30 times. Newly synthesized fragments in turn serve as templates for a "chain reaction" that results in the exponential amplification of the target sequence, producing millions of copies. Fragments of DNA ranging from 40 to more than 2000 base pairs can be amplified in a few hours [3].

Polymerase chain reaction has been facilitated by the introduction of Thermus aquaticus (Taq) DNA polymerase, a heat-stable enzyme that can endure the high temperatures required for DNA denaturation [20]. Programmable automated thermocyclers have further improved the efficiency and reproducibility of this method. The normal sequence of the DNA flanking the area being studied must be known to synthesize the oligonucleotide primers, which is a limitation of PCR [19]. However, certain modifications of the method permit amplification of DNA segments when sequences are unknown [21-23]. Another limitation of PCR is the rare, artifactual introduction of mutations into the product by replication errors [24]. Therefore, most mutations detected within PCR products must be verified by independent experiments. Polymerase chain reaction is used in microbiology experiments to amplify viral or bacterial DNA that would otherwise be insufficient to detect [25]. Polymerase chain reaction is also used in oncology research to study genes related to tumorigenesis [26]. Residual disease after therapy for certain leukemias and lymphomas may be detected using PCR [27, 28].

Before the use of PCR, laborious cloning procedures or radioisotopes or both were needed to isolate the region of the genome to be analyzed. In vitro amplification of DNA has facilitated the approach to genetic diseases at the DNA level. Structural variation within the PCR-amplified DNA segment can be studied directly by DNA sequencing or indirectly using single-strand conformational polymorphism, denaturing gradient gel electrophoresis, allele-specific oligonucleotide hybridization, or restriction fragment length polymorphism analysis (Figure 1).

DNA Sequencing

The most direct method for DNA analysis is DNA sequencing because it provides the sequence of nucleotide bases in the DNA segment being studied. The most commonly used procedure, developed by Sanger and colleagues [29], is a polymerization reaction in which the essential reagents include 1) the template, or DNA fragment to be sequenced, which has been amplified either by cloning or PCR to provide enough substrate for the reaction; 2) the four deoxyribonucleotides; 3) a radiolabeled nucleotide; 4) a DNA polymerase; 5) oligonucleotide primers to initiate polymerization; and 6) chemical analogs of the four nucleotides that inhibit the DNA polymerase (dideoxyribonucleotides).

Four reactions are performed simultaneously in separate tubes, one for each dideoxyribonucleotide mixture. As DNA synthesis occurs along the template, the radiolabeled nucleotide is incorporated, making the synthesized strands radioactive. The dideoxyribonucleotide analogs compete with the normal nucleotides and are incorporated at random positions along the complementary strands, where they terminate DNA synthesis [29].

The strands generated in each of the four reaction tubes are separated by gel electrophoresis and visualized by autoradiography. DNA molecules are negatively charged and travel through the gel toward the anode at a speed largely determined by their size, with smaller fragments migrating faster. A series of radioactive bands varying in length according to the position of chain termination is identified. Because the four reactions are run through the gel side by side, with termination at either adenine (A), cytosine (C), guanine (G), or thymine (T), the sequence can be read stepwise from the radiographic film (Figure 2).

View larger version (128K): [in this window] [in a new window]   Figure 2. DNA sequence of a portion of the apolipoprotein E gene. The sequence within brackets comprises codons 250 to 259 in exon four from a mutant allele identified by DNA sequencing. Reading stepwise from 5' to 3', the sequence is ATA GGC CTG CAG GCC GAG GCC TTC CAG GCC. The missense mutation involves a G-for-C substitution at codon 251, which causes the substitution of glycine for arginine. A = adenine; C = cytosine; G = guanine; T = thymine.

The PCR product can be sequenced directly after amplification or after first being cloned into bacteria [30, 31]. Automated DNA sequencing with fluorescent dyes is used in large-scale projects [32].

Rapid Screening for Coding Sequence Mutations

Single-strand conformational polymorphism and denaturing gradient gel electrophoresis are used to scan PCR fragments for the presence of mutations. In single-strand conformational polymorphism analysis, the PCR fragment is radiolabeled, made single-stranded, and electrophoresed under specified conditions [33]. Single base changes can be detected as mobility shifts by autoradiography because the mobility of a single-stranded DNA fragment in a nondenaturing gel is determined by its size and sequence [33]. Mutations also can be detected by denaturing gradient gel electrophoresis because double-stranded DNA fragments differing by a single base pair become denatured at different temperatures and concentrations of a chemical denaturant [34, 35]. Once the mutation-containing region is identified by single-stranded conformational polymorphism analysis or denaturing gradient gel electrophoresis, the mutation then can be characterized further by DNA sequencing [36].

Allele-specific Oligonucleotide Hybridization

In the 1970s, Southern [37] described a technique that hybridized a radiolabeled nucleic acid probe to a specific sequence among fragments of genomic DNA that had been generated by cleavage with restriction endonucleases and blotted onto a filter. Although Southern blotting can identify the presence of a DNA sequence in the genome and can detect large deletions or insertions within a gene, it cannot be used to detect a single base pair change, unless this alters a restriction enzyme site [37].

The allele-specific oligonucleotide hybridization method is analogous to Southern blotting and relies on the recognition of a specific DNA sequence by a radiolabeled synthetic DNA oligonucleotide probe. The probe hybridizes to the complementary region of target DNA and makes it visible after autoradiography. This method can identify a mutation within genomic DNA regardless of whether it changes a restriction site. Synthetic oligonucleotide probes are smaller than cloned probes and hybridize only with perfectly complementary DNA sequences. For diagnosis, probes are designed to hybridize with the normal or the mutant allele. Therefore, this method can be used only to detect known mutations [6, 7]. The fragment of DNA containing the target sequence may be prepared by PCR amplification to enrich it over the rest of the genomic DNA and improve the specificity of the hybridization signal [6, 7].

Restriction Fragment Length Polymorphism

Restriction fragment length polymorphism (RFLP) analysis uses restriction endonucleases that cleave double-stranded DNA at specific recognition sequences. Endonucleases are produced by bacteria as resistance factors against bacteriophage infection [38, 39]. They degrade phage DNA but do not affect bacterial DNA [38]. Each enzyme recognizes a specific short DNA sequence that is usually palindromic (that is, it reads the same from 5' to 3' in both complementary strands) [38]. Restriction endonucleases are named according to the bacterial species from which they are derived (for example, EcoRI is derived from Escherichia coli RY13, HhaI from Haemophilus haemolyticus, and so on) [38].

Traditional Southern blotting to detect RFLPs used nonamplified genomic DNA. DNA amplified by PCR also can be subjected to restriction endonuclease digestion. The DNA fragments generated are separated by gel electrophoresis according to their size. The DNA bands then can be transferred by Southern blotting to a membrane where they are hybridized to a labeled probe. Alternatively, the segments may be radiolabeled during PCR amplification before endonuclease digestion [37, 40]. In each case, the radioactive bands are visualized by autoradiography. However, in the preferred method for detecting RFLPs, the fragments are electrophoresed through a gel that is then stained with ethidium bromide, which intercalates between adjacent bases and fluoresces under ultraviolet light, making the DNA bands visible without radioactivity [41].

DNA mutations that abolish or create a cleavage site change the number and size of the fragments generated. Insertion or deletion mutations between restriction sites change the fragments' sizes. However, only rarely does a single gene disorder result from a mutation that affects a restriction site. This limits the applicability of a specific restriction enzyme in RFLP analysis when screening for DNA mutations.

The distribution of restriction enzyme sites within coding and noncoding regions of the genome vary greatly among persons studied. Such differences reflect the natural genomic variation that creates RFLPs, which usually have no physiologic relevance [42]. However, RFLP analysis is useful in both association and linkage studies of genetic disorders [42, 43]. Restriction fragment length polymorphisms at a candidate gene can be used to mark an allele that can be studied in relation to disease states.

Genetic association studies compare the frequency of marker alleles (usually RFLPs) in contrasting patient cohorts, such as persons with the phenotype of interest and healthy controls [43]. An important difference in allele frequency between affected persons and controls suggests that a causative mutation in the candidate gene and the RFLP defining the allele occur together more frequently than would be expected due to chance alone. For example, certain apolipoprotein gene RFLPs have been associated with atherosclerosis using this method [44, 45]. Association studies, like case–control studies in general, are vulnerable to bias, and associations may be related to social, geographic, or ethnic differences and not to the specific gene being studied. Linkage analysis uses extended families and tests the assumption that a particular gene marker (such as an RFLP) is located close to the gene causing the disease, so that the marker and the disease cosegregate. If linkage between the genotype and phenotype is established with high statistical probability, the RFLP marker can be used for prenatal or presymptomatic screening, even if the actual defect causing the disease is not known [42, 43, 46]. Linkage data can be used to map the causative gene in affected families [42, 43].

Often the gene responsible for a disorder has been mapped and sequenced, but the disease shows genetic heterogeneity because various mutations result in the same phenotype. Some mutations are restricted to a limited geographic area and some have been described only in a single family. In such cases, RFLP-based linkage analysis can be used only within the kindred. Duchenne muscular dystrophy, hemophilia A, and cystic fibrosis are diseases with genetic heterogeneity that are not amenable to RFLP analysis in the general population [47-51].

We know the genetic basis of several familial disorders of lipid metabolism [52-54]. Some are very rare. Others, such as familial hypercholesterolemia [52], have marked genetic heterogeneity and cannot be studied easily by RFLP analysis. Type III hyperlipoproteinemia is unique because RFLP analysis of the apo E gene allows accurate typing of the isoform associated with susceptibility for the disease.

Type III Hyperlipoproteinemia

Overview of Normal Lipid Metabolism

Cholesterol and triglycerides are transported within plasma lipoproteins. The four major lipoprotein classes are chylomicrons, very low-density lipoprotein (VLDL), low-density lipoprotein (LDL), and high-density lipoprotein (HDL). Lipoproteins differ and can be separated by ultracentrifugation or electrophoresis Table 1, and each has characteristic cholesterol, triglyceride, and protein contents [55]. During normal metabolism, the lipoproteins lose, acquire, and exchange their lipid and protein constituents and dynamically change their composition. Lipoproteins transport lipids to the liver for bile acid synthesis and to peripheral tissues for storage and synthesis of cell membranes and steroid hormones. Lipoprotein transport occurs through the exogenous and the endogenous pathways summarized in Figures 3 and 4 [55].

View larger version (23K): [in this window] [in a new window]   Figure 3. Exogenous lipid transport. Dietary fats are absorbed as chylomicrons, which are composed mainly of triglycerides (TG) and a small proportion of cholesterol esters (CE). Their main apolipoprotein is apo B48. As chylomicrons enter the circulation, they acquire apo E and apo CII. Apolipoprotein CII is a cofactor for lipoprotein lipase (LPL), an enzyme that hydrolyzes triglycerides into fatty acids and glycerol. The lipolytic action of lipoprotein lipase makes chylomicrons progressively smaller, depletes triglycerides, and enriches cholesterol ester. Chylomicron remnants are taken up by the liver through the remnant receptor, for which apo E is a high-affinity ligand. VLDL = very low-density lipoprotein.

View larger version (40K): [in this window] [in a new window]   Figure 4. Endogenous lipid transport. Very low-density lipoproteins (VLDL) secreted by the liver are rich in triglycerides (TG) and contain apo B100, apo E, and the apo Cs. The lipolytic action of lipoprotein lipase (LPL) generates remnant particles that are enriched in cholesterol ester (CE) and apo E and are depleted of triglycerides. Some VLDL remnants are absorbed by endocytosis mediated by the low-density lipoprotein (LDL) receptor and the remnant receptor. Other VLDL remnants are hydrolyzed by hepatic lipase (HL) to yield LDL rich in cholesterol ester and containing apo B100. Most LDL undergoes LDL receptor-mediated endocytosis in the liver or peripheral tissues. The remaining LDL is absorbed in the tissues by non-LDL receptor-mediated processes (scavenger receptor).

Apolipoproteins are the structural protein components of lipoproteins [56]. Some apolipoproteins are ligands for receptor-mediated endocytosis. For example, apo B binds to the LDL receptor and apo E binds to the LDL receptor and LDL receptor-related protein [55]. Other apolipoproteins are activators of enzymes. For example, apo AI activates lecithin cholesterol acyltransferase, and apo CII activates lipoprotein lipase [55].

Genetic defects of apolipoproteins can cause various abnormalities in lipid metabolism [57]. Rare defects of the apo CII gene cause severe hypertriglyceridemia [58, 59]. Rare defects of the apo AI gene cause very low HDL cholesterol levels [57]. Some mutations of the apo B gene cause very low LDL cholesterol levels [60]. Familial defective apo B is caused by a mutation in the apo B gene that impairs binding to the LDL receptor and increases plasma LDL cholesterol levels [53, 60]. Susceptibility to type III hyperlipoproteinemia is associated with an apo E isoform that binds poorly to cell receptors [61].

Characteristics of Type III Hyperlipoproteinemia

Type III hyperlipoproteinemia is a hereditary dyslipidemia characterized by the accumulation of remnant particles of triglyceride-rich lipoproteins [62]. These particles collectively constitute ß-VLDL, a fraction with a density less than 1.006 g/mL that migrates in the ß rather than the usual pre-ß position on agarose gel electrophoresis [63].

Patients with type III hyperlipoproteinemia typically have combined hypercholesterolemia and hypertriglyceridemia, along with detectable ß-VLDL and a ratio of VLDL cholesterol to total triglyceride exceeding 0.30 [64]. Approximately one half of the patients have tuberous and palmar xanthomas [62, 63]. Palmar xanthomas are yellow streaks in the palmar creases and are pathognomonic of this disorder [63, 65]. Type III hyperlipoproteinemia is associated with premature atherosclerosis, and the ß-VLDL particles are considered atherogenic [62].

The prevalence of type III hyperlipoproteinemia in the general population is approximately 1 in 10 000 [66]. The prevalence of homozygosity for the E2 isoform (E2/2) in the general population is 1 in 100. Therefore, genetic or environmental secondary factors are required for the full expression of the phenotype in susceptible persons with the E2/2 genotype. Other rare variants of apo E that bind defectively to the LDL receptor have been described in type III hyperlipoproteinemia Table 2 [67-78]. Deficiency of apo E in one kindred with type III hyperlipoproteinemia also has been reported [79, 80]. Although this dyslipidemia has a definite genetic basis, environmental factors affect its full phenotypic expression. For example, severe hyperlipidemia usually manifests in adulthood. The disorder is more common in men and is rare in women before menopause. Obesity, diabetes, and hypothyroidism contribute to type III hyperlipoproteinemia [62, 63].

Apolipoprotein E

Apolipoprotein E is a 299-amino acid protein found in chylomicrons, VLDL, intermediate-density lipoprotein, and HDL [81]. It plays an important role in the metabolism of these lipoproteins by binding to the LDL receptor in hepatic and extrahepatic tissues and a putative apo E receptor or LDL receptor-related protein [81-83].

Protein Polymorphism

The polymorphic nature of apo E was first described in the 1970s [84]. Later, three common isoforms (E4, E3, E2) were identified. The alleles are determined by a single gene and are expressed codominantly, generating six possible phenotypes: E4/4, E3/3, E2/2, E4/3, E3/2, and E4/2 [85-87]. The three isoforms are distinguished by their charge by isoelectric focusing on polyacrylamide gels in a pH gradient Figure 5 [87]. E4 is the most positively charged and E2 is the most negatively charged isoform [87]. Amino acid sequencing showed that the differences result from cysteine to arginine interchanges at residues 112 and 158 [88, 89]. E2 has cysteine at both positions, E4 has arginine at both positions, and E3 has cysteine at position 112 and arginine at position 158 [89]. E3 is the normal form because it is the most prevalent in all populations studied [90].

View larger version (103K): [in this window] [in a new window]   Figure 5. Apolipoprotein E (apo E) phenotype determined by isoelectric focusing. These are the isoelectric focusing gels of three homozygous persons. Apolipoprotein E isoforms are distinguished on the basis of charge by polyacrylamide gel electrophoresis. E2 has two cysteines (Cys) and is the most negatively charged isoform. E4 has no cysteine and is the most positively charged isoform. The lighter bands constitute minor sialated forms. The top of the gel is negatively charged and the bottom is positively charged. apo C = apolipoprotein C; Arg = arginine.

Population Studies of Apolipoprotein E Polymorphism

Apolipoprotein E allelic frequencies vary among populations. In most studies of white persons, E3 had an allelic frequency of about 0.77, E2 was the least common, with an allelic frequency of about 0.08, and E4 comprised the rest [90]. The frequency of E4 is low in Japan and high in Finland, a country with a high prevalence of heart disease [90]. Apolipoprotein E allele frequency differences might contribute to variation in the prevalence of coronary heart disease among different populations [90].

The mechanism of the association between apo E and atherosclerosis is not clear. E4 could be associated with atherosclerosis in the general population, but study results conflict [90]. Severe dyslipidemia and vascular lesions develop in mice made apo E deficient by gene targeting, whereas transgenic mice that overexpress apo E seem to be protected [91, 92].

A recent meta-analysis of 14 799 persons from 45 population samples in 17 countries revealed a consistent relation between plasma cholesterol and apo E isoforms [93]. Persons with the E2 and E4 alleles had about 5% lower and 5% higher levels of plasma cholesterol, respectively, than did E3/3 homozygotes [93]. This study also showed that those with the E2 and E4 alleles had higher levels of plasma triglycerides than did E3/3 homozygotes [93]. E4 also has been associated with increased levels of Lp(a), a highly atherogenic lipoprotein [94].

The mechanism by which LDL is increased in persons with the E4 allele may be related to the faster catabolic rate of E4 compared with E3 or E2 [95, 96]. In vivo studies showed that E4-containing lipoproteins are internalized and catabolized more quickly by the liver [61, 95-97]. The resulting increase in intracellular cholesterol concentration could down-regulate LDL receptor expression and reduce LDL uptake [96].

Apolipoprotein E isolated from persons with E2/2 alleles binds defectively to LDL receptors in vitro, regardless of their plasma lipid levels [98]. The overall association between lower plasma cholesterol levels and the E2 allele in large populations without type III hyperlipoproteinemia has been explained by up-regulation of the LDL receptor secondary to a lower rate of cholesterol-rich remnant particle transport into the liver [96, 97].

Evidence suggests that apo E variation affects the serum cholesterol response to diet. Persons with the E4 allele may be more sensitive to dietary manipulation of LDL cholesterol levels [99, 100]. The E2 allele may enhance LDL cholesterol lowering in response to dietary fiber [101]. The role of apo E in mediating the response to lipid-lowering drugs has not been determined [102-104].

Protein Structure

Apolipoprotein E is divided into two independently folded domains with different functions [81, 105]. The amino-terminal region contains the receptor-binding domain, whereas the carboxy-terminal region mediates binding to the surface of lipoproteins [81, 105]. The receptor-binding domain has been characterized by combining the information obtained from several approaches [81, 106, 107]. Naturally occurring variants with defective receptor-binding activity have been very informative. Almost all of the apo E mutations reported in persons with type III hyperlipoproteinemia have abnormal binding; and involve residues between 136 and 146 (Table 2); (Figure 6). This information helped to localize the receptor-binding domain in the vicinity of residues 140 to 160. X-ray crystallography determined the three-dimensional structure of the amino terminal domain of apo E [105]. The receptor-binding domain was delineated to span residues 136 to 150, part of a helix with positive electrostatic potential [105].

View larger version (22K): [in this window] [in a new window]   Figure 6. Location of apolipoprotein E (apo E) mutations associated with type III hyperlipidemia with respect to the putative receptor-binding domain. In this schematic diagram, the numbers indicate apo E amino acid residues. The seven arrowheads 3' of position 121 represent the 7-amino acid insertion of the apo E Leiden mutation. See Table 2..

However, amino acids further removed from residues 136 to 150 also play an important role in receptor binding. The E2 isoform has a receptor-binding affinity that is 1% to 2% that of E3 and differs from E3 only at position 158 [81]. The defect in the apo E Leiden variant consists of a 7-amino acid insertion at residue 121 and results in binding activity that is 25% of the normal E3 isoform [67]. Studies of truncated variants of apo E suggest that residues in the 171 to 183 region are necessary for normal binding [106]. These amino acids might contribute to the higher-order structure of the region and stabilize or align the residues in the actual receptor-binding domain [105-107].

Gene Expression

Unlike the other important apolipoproteins, which are synthesized primarily in the liver and small intestine, apo E is produced in various tissues, suggesting that it may also be involved in processes other than traditional lipid transport [81]. The largest quantity of apo E mRNA is found in hepatic tissue [2]. In addition, patients who have received liver transplants acquire the serum apo E phenotype of the donor, suggesting that most circulating apo E is synthesized in the liver [108]. Apolipoprotein E is also produced in the adrenal gland and kidney [1, 2]. The expression of the apo E gene is regulated by multiple positive and negative elements within its promoter region [109].

The second largest concentration of apo E mRNA is found in the brain [2, 81]. The cerebrospinal fluid apo E phenotype of liver transplant recipients does not change, implying that the apo E found is synthesized locally [108]. An increased frequency of the E4 allele was found in patients with Alzheimer disease [110], a discovery that holds promise for future Alzheimer disease research.

Gene Analysis

The gene for apo E has been cloned, sequenced, and mapped to chromosome 19 [111, 112]. It is part of an apolipoprotein gene cluster that includes apo CI and apo CII [113]. Its structure, like that of other apolipoprotein genes, consists of four exons, with most of the protein-coding sequence contained in exon four [56, 111, 112]. The three major apo E alleles in the general population vary by single nucleotide substitutions at codons 112 and 158, coding for arginine or cysteine, as previously described [112].

Determination of apo E genotype can be more accurate than protein phenotyping. Various structural changes at the DNA level alter the isoelectrophoretic appearance of apo E, and separating apo E isoforms based solely on their charge may not be sufficient to identify these structural differences (Table 2). In addition, post-translational modification of apo E with carbohydrate chains containing sialic acid causes charge shifts on isoelectric focusing similar to those that result from cysteine substitutions [87, 114].

Apolipoprotein E genotyping has been simplified by RFLP analysis of the PCR-amplified portion of the gene that encompasses codons 112 and 158. This method of restriction isotyping [41] uses the endonuclease HhaI to generate DNA fragments of characteristic lengths for the E4, E3, and E2 alleles. The fragments are separated by gel electrophoresis and visualized after staining with ethidium bromide Figure 7 [41].

View larger version (96K): [in this window] [in a new window]   Figure 7. Apolipoprotein E (apo E) genotype determined by restriction isotyping. A. Electrophoretic separation of HhaI digested fragments of a polymerase chain reaction-amplified portion of apo E exon four, encompassing codons 112 and 158. The six common apo E genotypes are shown. A standard DNA molecular weight marker with fragment sizes specified in base pairs is included for reference. B. HhaI restriction map for the three common apo E alleles. HhaI cleavage sites are indicated by arrows. The length of the fragments generated is specified on top of the lines. Codons 112 and 158 are marked.

Restriction isotyping is rapid, relatively inexpensive, and ideal for testing the many patients required in epidemiologic studies to determine the biological relevance of apo E polymorphism. DNA extracted from 5 mL of whole blood can be PCR amplified, digested with HhaI, electrophoresed, and visualized with ethidium bromide in less than 1 day. In our laboratory, the total cost to determine a single apo E genotype is approximately $7.00. This compares favorably to the labor-intensive procedure of isoelectric focusing, which requires about 4 days and is estimated to cost $36.00 per sample. Although newer methods of apo E isoform typing, such as immunoblotting, have been described [115], these methods are still laborious and prone to technical artifacts.

In addition to identifying persons homozygous for the E2 allele, restriction isotyping can help identify new apo E mutations if the mutation occurs within the amplified region of the apo E gene. The endonuclease used (HhaI) recognizes a 4-base pair sequence (GCGC) and cleaves the E4 allele at six positions, after the same region of exon four described in the original method is amplified [41]. Single-base substitution changes at any of these 24 bases would alter the restriction pattern. Insertions or deletions between cleavage sites would also be detected. Some of the mutant apo E forms that have been described in association with type III hyperlipoproteinemia can be identified by this technique. Apolipoprotein E Leiden, apo E Christchurch, R136C, and the arginine 142cysteine mutation listed in Table 2 have characteristic patterns by restriction isotyping [68, 69, 71]. Theoretically, other restriction enzymes can be used to identify mutations at other locations when these are suspected. The mutation of apo E Philadelphia displays characteristic AvaI and BbvI RFLPs [78].

Conclusions

The ease and accessibility of PCR-based methods facilitates the study of inherited disorders at the DNA level. Apolipoprotein E genotype can be determined by restriction isotyping. Restriction isotyping can be used to assess patients who are susceptible to type III hyperlipoproteinemia. Rare mutations of apo E also can be detected by restriction isotyping. They can be characterized further by DNA sequencing. Screening the family members of affected persons can help to identify other persons who are at risk for developing the disorder. These persons can be targeted to modify atherosclerosis risk factors. Apolipoprotein E restriction isotyping may become an adjunct in assessing the risk for hyperlipidemia or atherosclerosis or both in specific patients and possibly in predicting their response to dietary or pharmacologic intervention.

Abbreviations

apo E: apolipoprotein E

HDL: high-density lipoprotein

LDL: low-density lipoprotein

PCR: polymerase chain reaction

VLDL: very low-density lipoprotein