Atherosclerosis, 86 (1991) 239-250 ( 1991 Elsevier Scientific Publishers ADONIS 002191509100074Q
ATHERO
239 Ireland.
Ltd. 0021-9150/91/$03.50
04593
Molecular basis of lipoprotein lipase deficiency in two Austrian families with type I hyperlipoproteinemia Bernhard Paulweber ‘, Heiko Wiebusch ‘, Gero Miesenboeck 3, Harald Funke ‘, Gerd Assmann 2, Bertram Hoelzl I, Manfred J. Sippl 4, Walter Fried1 ‘, Josef R. Patsch 3 and Friedrich Sandhofer 1 ’Ftrst Department of Medicine, Landeskrankenanstalten
-’Department
Salzburg Salzburg (Austria). ’ institute for Clinical Chemistry and Laboratory Medicine, Miinster (F. R. G.). of Medtcine, Division of Clinical Atherosclerosis Research, Unrversity of Innsbruck, Innsbruck
’ Instttutefor General Biology, Biochemistry
and Biophysics, Department
of Biochemistv.
lJnroersit_v of Sal:burg,
(Austria). Salzburg (Austrra)
(Received 28 September, 1990) (Accepted 22 October, 1990)
Summary
To determine the molecular basis for type I hyperlipoproteinemia in two Austrian families, the lipoprotein lipase (LPL) gene of two patients exhibiting LPL deficiency was analyzed by Southern blotting and by direct genomic sequencing of DNA amplified by polymerase chain reaction (PCR). All exons of the LPL gene except part of the noncoding region of exon 10, all splice donor and acceptor sites, as well as 430 basepairs of the 5’-region including the promotor were sequenced. A homozygous substitution of adenine for guanine in the fifth exon at cDNA position 818 of the LPL gene was found in both patients. Our sequencing strategy largely ruled out a linkage disequilibrium of the identified nucleotide change with another defect potentially causing the clinical phenotype. The base change described abolishes a normally present AuaII restriction site allowing the identification of carriers of the mutant allele by AvaII digestion of PCR fragments of exon 5; three members of the two families were homozygous for this mutation and ten members were heterozygous. The activity of LPL in postheparin plasma was almost completely absent in homozygotes and about half normal in heterozygotes. The loss of activity was related to LPL protein structure. This mutation alters the amino acid sequence at residue 188 from Gly to Glu. The conformational preferences of the protein chain around position 188 were calculated with the use of a knowledgebased computerized method. The most probable conformation is a beta-turn formed by residues 189-192.
Correspondence to: Dr. F. Sandhofer. First Department of Medicine. Landeskrankenanstalten Salzburg, A-5020 Salzburg. Austria. Tel. (662)31581-2800; Fax: (662)387255.
Abbreuiations: LPL. lipoprotein lipase; HL, hepatic lipase: PL, pancreatic lipase; VLDL, very low density lipoproteins; HDL, high density lipoprotein; FFA. free fatty acids; apo C-II. apolipoprotein C-II; apo B. apolipoprotein B; PCR. polymerase chain reaction.
240 The mutation alignment.
Key words:
seems
to destabilize
the beta-turn
and/or
Lipoprotein lipase gene; Lipoprotein lipase genomic sequencing; Lipase protein structure
Introduction Lipoprotein lipase (LPL; EC 3.1.1.34) hydrolyses dietary and endogenous triglycerides transported in plasma chylornicrons and very low density lipoproteins [l-3]. Familial deficiency of LPL is the most frequent defect underlying type I hyperlipoproteinemia with an estimated prevalence of less than 1 case per million population [4]. The defect is due to either the synthesis of a structurally defective enzyme or its complete absence. Functional LPL deficiency can result also from a deficiency of the specific activator protein, apo C-II [5], or the presence of an LPL inhibitor within the circulation [6]. The mode of inheritance of LPL deficiency is autosomal recessive. Homozygotes exhibit postheparin plasma LPL activities below three standard deviations under the mean for normal subjects and severe fasting hypertriglyceridemia [7]. Heterozygotes show reduced LPL activities which can overlap those of unaffected control subjects and variable degrees of triglyceride elevation [8]. The LPL gene is located on chromosome 8 [9]; it extends over 30 kb, and consists of 10 exons that are interrupted by 9 introns [lo-121. Exons l-9 are of average size (105-276 bp) while exon 10 is 1948 bp long consisting of the very last base of the stop codon together with the entire 3’-untranslated region. The LPL gene codes for a protein of 475 amino acids including a 27-residue signal peptide. A few variants of the LPL gene in patients with familial LPL deficiency have already been reported. For example using Southern blot analysis Langlois et al. [13] found a 2 kb insertion caused by a partial duplication of exon 6 [14]. Beg et al. [15] detected a guanine for adenine substitution at cDNA position 781 of the LPL gene in a patient with absent postheparin plasma LPL activity.
a yet
deficiency;
larger
domain
critical
Polymerase-chain
for substrate
reaction;
Direct
The aim of this study was to identify the molecular defect in patients with type I hyperlipoproteinemia from two Austrian families. To this end, we have sequenced all exons, the splice sites, and the promotor region of the LPL gene after PCR amplification of genomic DNA from two affected subjects. Characterization of molecular defects in the LPL gene are not only required to explain the underlying metabolic defect of familial LPL deficiency, but may also provide better insight into the structure-function relationship of this important enzyme in lipoprotein metabolism. Experimental procedures Patients Three male patients with type I hyperlipoproteinemia from 2 families living in a secluded valley near the city of Salzburg, Austria, were studied. The index cases, aged 42, 34 and 33 years, were known to suffer from type I hyperlipoproteinemia from early childhood on. They presented with marked hepatosplenomegaly and had a history of recurrent pancreatitis. None of them had xanthomas. The other relatives studied showed no clinical abnormalities. Plasma lipids Blood was drawn into EDTA after an overnight fast of at least 12 h. Total serum cholesterol and triglyceride levels were determined enzymatically using the Monotest@ Cholesterin and the Triglyceride GPO-PAP kit (Boehringer Mannheim, Mannheim, F.R.G). HDL cholesterol was determined enzymatically after phosphotungstateMgCl, precipitation of apo B containing lipoproteins. Measurement of LPL activity For determination of LPL activity blood was collected on ice 10 min after i.v. injection of 60 IU
241 Heparin NOVO~ (Novo Industri A/S, Copenhagen, Denmark) per kg body weight. Samples were centrifuged at 4°C and the plasma was stored at - 70°C. Prior to the LPL assay, plasma samples were incubated on ice for 2 h with 0.5 volumes of a solution of goat IgG against human hepatic lipase, sufficient to completely suppress hepatic lipase activity in postheparin plasma [16,17]. For preparation of substrate, tri[9,10-3H]oleoylglycerol (Amersham. Arlington Heights, IL, U.S.A.) was incorporated into 10% Intralipidm (AB Kabi Vitrum, Stockholm, Sweden) with the use of a Branson Sonifier 250 [18]. Apart from the sample components, the complete assay mixture contained 5 mg/ml triacylglycerol emulsified with 0.6 mg/ml phosphatidylcholine; 60 mg/ml fatty acid-free bovine serum albumin (Sigma Chemical Corp.. St. Louis, MO, U.S.A.); 0.15 M Tris-HCl buffer. pH 8.5: 0.1 M NaCl; 15 IU/ml Heparin Nova@; and 12.5% (v/v) heat-inactivated human serum in a final volume of 200 ~1. Samples of lo-20 pl were assayed in triplicate; incubations were performed in a shaking water bath at 25 o C. To terminate the reactions, the fatty acids released were extracted as described by Spooner et al. [19] and radioactivity was quantified in a Beckmann LS 1801 liquid scintillation counter. Two human postheparin plasma standards, one human preheparin plasma standard. and one bovine skim milk standard were included in each assay to correct for interassay variation. Fatty acid release in our assay is linear over all incubation times and enzyme concentrations employed. LPL activities are expressed in nmol fatty acids released (FFA) per ml plasma per min at 25°C. DNA preparation and Southern blot Blood samples were collected into EDTA and stored at -70°C. Total genomic DNA was prepared from leukocytes of 10 ml blood with the Triton X-100 method [20]. Five pg of genomic DNA was digested with restriction enzymes BarnHI, BstNI, HindIll, NcoI, PuuII, P.stI and StuI (Boehringer Mannheim, Mannheim, F.R.G.) under conditions recommended by the manufacturer. Fragment separation on agarose gel electrophoresis, Southern transfer, and hybridization were performed according to standard procedures [21,22]. For hybridization a full length cDNA
probe of the LPL gene [ll] was isolated from a pUC 18 plasmid and purified by electrophoresis on low melting point agarose (Sigma Chemical Comp.. St.Louis, MO. U.S.A.) and labelled with [” P]dCTP (800 Ci/mmol. Amersham International, Arlington Heights. IL, U.S.A.) using the random oligonucleotide primer method [23]. Preparation of PCR fragments For PCR [24] the DNA Thermal Cycler (Perkin-Elmer/Cetus Corp.. Norwalk. CT. U.S.A.) was used. Eleven fragments of the LPL gene containing the entire coding sequence. part of the introns. the 5’-flanking region including the promotor. and the 3’-untranslated region of exon 10 were amplified (Fig. 1). For each PCR about 300 ng of genomic DNA was used. In the reaction mixture the final concentration of each dNTP was 200 PM and of oligonucleotide primers was 100 nM. Ohgonucleotides were synthesized using the Gene Assembler (Pharmacia LKB Biotechnology, Uppsala. Sweden). For each PCR 2.5 U of Taq polymerase (Amplitaq”. Perkin-Elmer/ Cetus Corp.. Norwalk, CT, U.S.A.) were used. The thermocycle file was set for a total of 30 cycles as follows: 1 min at 96°C for denaturing, 1 min at 60°C for annealing, and 3 min at 72°C for extending. For fragments longer than 2 kb the extension time was prolonged to 10 min. For fragments primed with oligonucleotides shorter than 20 bases the annealing temperature was lowered to 5355” C. PCR fragments shorter than 1 kb were isolated by electrophoresis on 4’% NuSieve agarose (FMC BioProducts. Rockland, ME, U.S.A.), and fragments longer than 1 kb on 1.5% agarose gels (Sigma Chemical Comp.. St. Louis, MO, U.S.A.). They were visualized by ethidium bromide staining (0.5 g/ml in gels and buffer). Areas containing the desired fragments were cut from the gel. and DNA was eluted using the biotrap electroseparation system (Schleicher and Schiill, Dassel, F.R.G.). Eluted DNA was desalted and concentrated to a final volume of 75 ~1 using Centricon 100 microconcentrators (Amicon, Danvers, MA, U.S.A.). Sequence analysis For sequence analysis single-stranded templates were produced from double-stranded PCR fragments with only one of the primers used for
242 kb
5
IO
15
25 I
20
5'yI'
CODON
30
188
NORMAL
ALLELE
666
----)
GLYCINE
HUTANT
ALLELE
6A6
---+
GLUTAHIC
ACID
Fig. 1. Structure of the human LPL gene. Introns are represented by white boxes, exons by solid boxes and their noncoding portions by shaded areas. Bars below the diagram of the gene show the fragments amplified by PCR. Arrows indicate direction and extent of sequencing. Detailed sequence and annealing positions of the primers are given in Table 2. In the lower part of the figure the codon in exon 5 containing the mutation is shown.
double-strand production. The single-strand was made manually in 10 cycles (melting at 96 o C for 1 min, annealing at 60°C for 1 mm, extending at 72°C for 2 mm) using 3 water baths. Singlestranded DNA was separated by gel electrophoresis, then electroeluted, extracted with phenol/ chloroform, and precipitated with LiCl/ ethanol [25]. After dissolving the DNA in water, the dideoxy-chain termination sequencing reaction was performed [26] using the T7 Sequencing kit from Pharmacia (Uppsala, Sweden) and [ 35S]dATP (500 Ci/mmol) (Du Pont, NEN Products, Boston, MA, U.S.A.) for labelling. The sequencing reaction recommended by the supplier was modified as follows: labelling was performed at room temperature for 90-120 s. Aliquots of the reaction mixture were transferred to fresh tubes containing 2.8 ~1 of each of the respective ddNTP mixture. The termination reaction was at 37 QC for 3 min. Subsequently, 1.5-2 ~1 of the mixtures was loaded on a sequencing gel system Macrophor (Pharmacia, Uppsala, Sweden) for separation of ddNTP terminated fragments. Wedge shaped (0.2-0.7 mm) sequencing gels contained 4-58 acrylamide/ bisacrylamide (19 : 1) and 8 M urea. Electrophoresis was performed at 60” C and at a constant voltage of 2000 V. After fixation the gels were dried and exposed to a Kodak X-OMAT AR film (Eastman Kodak Comp., Rochester, NY, U.S.A.) for 8 h to 5 days.
The sequencing strategy is illustrated in detail in Fig. 1. To synthesize a first set of PCR primers and sequencing primers (lu, 2u, 3u, 4u, 5u, 7u, 8u, 11, 21, 31, 41, 51, 61, 81, 91, 101, lu(s), lOu(sl), and 101(s) in Table 2) we relied on cDNA sequence information [ll] and data from Deeb and Peng [lo] regarding the sequences of intron-exon boundaries. Using these primers exons l-9, 600 bases of the 5’ and 300 bases of the 3’ end from the noncoding region of exon 10, and parts of introns 3, 4, 5, 7, and 8 were sequenced. The sequence information from introns 5 and 7 was used to synthesize primers 6u and 71; sequence information from intron 8 was used to construct primer 9u, which in conjunction with the lower strand primer 101(i) gave access to intron 9 sequence. This sequence information was utilized to synthesize primer 10~ which gave sequence information regarding the exon 10 splice acceptor site. Upper and lower strands of exon 5 were sequenced twice: beginning with the fragment spanning the region from exon 4 to exon 6 followed by the fragment containing only exon 5. As demonstrated in Fig. 1, the whole sequence from exon 3 to exon 10 including the introns was covered by PCR fragments allowing complete analysis of all exon-intron boundaries. The size of introns 1 and 2 precluded their amplification by simple PCR procedures. Since only 25 bp of the introns at the boundaries are known [lo] the following strategy
243 was used for analysis of the first 4-7 flanking bases of these introns: Primers 11(b), 2u(b), 21(b), and 3u(b), designed to anneal to these boundaries, were shortened on their 3’-end by 4-7 bases to a base total of 17 to 21 bases, while the corresponding primers lu, 2u, 21, and 31 for the opposite strand were of usual length (22-24 bases) (Table 2).
U.S.A.) and brought to a final volume of 30 ~1 in a Speed Vat centrifuge. The PCR fragments were digested with 7 U AuaII (Boehringer Mannheim. F.R.G.) for 2 h at 37°C in a total volume of 50 ~1. Fragments were separated on 5% NuSieve agarose (FMC BioProducts, Rockland, ME, U.S.A.) gels and visualized by the addition of 0.5 g/ml ethidium bromide to gel and buffer.
AuaZZ restriction fragment analysis Fragments containing exon 5 were amplified from genomic DNA by PCR using the primers 5u(r) and 51(r) (Table 2) as described above. The PCR mixtures were desalted using Centricon 100 microconcentrators (Amicon, Danvers. MA,
Statistical analysis The significance of differences between LPL activities of subjects heterozygous for the mutant allele and normal controls was tested using Student’s t-test.
TABLE
1
PLASMA
LIPIDS
AND
POSTHEPARIN
Proband
Genotype
Homozygous
a
Heterozygous
a
M.A. M.F. O.J. MS. M.L. M.M. M.T. M.K. P.B. P.A. O.G. 0,s. O.A. O.M.
PLASMA
Age (years)
Sex
41 34 32 68 12 38 8 61 49 65 20 56 6 5
M M M M M M M F F F F M F F
LPL ACTIVITIES
OF MEMBERS
TG
Chol.
HDL-chol.
(mg/dI)
(mg/dl)
(mg/dl)
4970 2740
223 198 299 195 150 185 142 224 243 358 196 242 201 n.d.
13 11 14 36 32 39 64 30 37 62 56 45 47 nd.
106 61 109 33 139 113 170 108 176 76 nd.
Mean f SD Normal
OF BOTH
FAMILIES LPL-activity (nmol FFA~min~~‘.ml-‘) 0 0 128 nd. 107 209 87 121 158 110 126 141 nd. 131.9+
M.J. M.B. R.A. M.C. M.G. O.C. O.R.
61 31 32 11 34 45 31
83 n.d. 49 51 n.d. nd. 100
247 n.d. 170 165 n.d. n.d. 161
60 n.d. 39 47 n.d. nd. 35
35.3 h
321 nd. 256 269 n.d. n.d. n.d.
Mean f SD
279.0 & 29.3 h
Controls (n = 50) Mean f SD
287.3 rt 71 .O h
n.d. = not determined; TG = triglycerides; Chol. = cholesterol; HDL-Chol. = HDL cholesterol. ’ For the mutation at cDNA position 818. h The difference between mean LPL activity of heterozygotes for the mutation and normal family members was statistically significant (P < 0.002). The difference between mean LPL activity of heterozygotes for the mutation and 50 unrelated controls was statistically significant (P < 0.001).
244 Prediction of the protein conformation in the region of interest
H.J H.S.
Probable backbone conformations in the vicinity of the Glu for Gly substitution at position 188 were calculated using the Boltzmann device, a knowledge-based computerized method for the prediction of local structures in oligopeptides [27]. In brief, the method models the energetic interactions between atomic groups using intramolecular potentials derived from a data base of three known dimensional protein structures. These potentials are used to calculate the most probable conformations for a given amino acid sequence from a large number of possible alternatives.
H.K. P.0 PA.
HA.
n.6. R.A.H.F. il.tl.
ll.T. H.C.
H.L
Results Plasma lipid analysis and LPL activities of various individuals from 2 Austrian families
Plasma lipids are shown in Table 1. The probands M.A., M.F. and O.J. exhibited marked hypertriglyceridemia and little if any elevations of cholesterol. HDL cholesterol levels were below the 5th percentile of values in a normal population [28]. All other family members had normal triglyceride levels.
0.6.
0.J.
$,
0.a.
c C
c /
c
\
T 6 A+ 6 A 6 A C C
Fig. 2. Autoradiogram showing the sequencing pattern of a patient homozygous for the A to G mutation at cDNA position 818. The position of the mutation is marked by an asterisk.
Fig. 3. Pedigrees and AvaII restriction fragment patterns of the 2 families. W, homozygotes for the mutation at cDNA 0, homozygotes for the wild position 818, ? ,?heterozygotes, type allele. Homozygotes are characterized by 2 restriction fragment bands (296 bp, 87 bp), heterozygotes by 3 bands (296 bp, 209 bp, 87 bp), homozygotes for the wild type allele by 2 bands (209 bp, 87 bp). In the right lower part of the figure the AuaII restriction fragments of exon 5 are shown. At cDNA position 730 an invariant restriction site is located. The restriction site which is abolished by the mutation at cDNA position 818 is marked by an asterisk. Arrows indicate the primers used for PCR.
The LPL activity from postheparin plasma of all homozygotes was virtually zero, while LPL activities of unaffected family members fell within normal limits. In heterozygous family members LPL activities were reduced and ranged from 87 to 209 nmol FFA. ml-’ *n-k-’ (mean 131.9, SD 35.3) (Table 1). These values were significantly
245 lower than the LPL activities observed in 50 controls (mean 287.3, SD 71.0, range 165-454 nmol FFAeml-’ .min-‘; P < 0.001) and the LPL activities of the 3 normal family members (321, 256, and 269 nmol FFA . ml-’ . min-‘; P < 0.002). Alterations in the LPL gene As a first approximation to detect any major alterations in the structure of the LPL gene from
TABLE
these family members, we digested DNA from the various individuals with a battery of restriction enzymes followed by Southern blot analysis using a full length cDNA probe for LPL. Our results showed that no major structural changes were present in the DNA from these individuals. Therefore we proceeded to sequence various segments of the gene to identify the molecular basis for the LPL deficiency. Fig. 1 illustrates the sequencing
2
POSITION
AND SEQUENCE
Primer no. *
OF PRIMERS
USED
Position b
SEQUENCING
Sequence ’
5’-caagctgggacgcaatgtgtgtcc-3’ 3’-cattcaaaacgcgcgtttgagggg-5’ 5’-ctcatatccaatttttcctttccag-3’ 3’-cattccctccgagaaaccccttc-5’ 5’-aagcttgtgtcatcatcttcagG-3’ 3’-Ccattctgaccctcttcctctg-5’ 5’-ccttcattttctttttcttccaaagG-3’ 3’-ccattctttcgttaaagcaacc-5’ 5’-cctgcttttttcccttttaagGCC-3’ 3’-CTCcatttataataaatcttcgcttaa-5’
41 5u 51
-25-l-22+1-2% -I-21+3-
-452’ -24 -1 -23 +1 -21 +1 -22 +3 -24
hu
-89-
-63’
5’-ccacatctcacctattttagacatgcc-3’
hl
+l-
-23
3’-Ccatccgacctctgacaacattta-5’
7u
-25s
-1
5’-catgttcgaatttcctccccaacag-3’
71
-23-
-41’
3’-gtagtaccgtggtcagggagaggac-5’
XU
-2%
lu 11 2u 21 3u 31 4U
-475%
FOR PCR AND/OR
-l-
+2
5’-CcaaatttattgcttttttgtttagGC-3’
XI
+6-
-23
3’-CTTTTTcattaatttacataaaaagaagg-5’
9u
-ll+l-
-50
91 IOU
- 147-
101 101(i) IUS)
+4+ 1607+70-
e
-25
5’-cctgacagaactgtacctttgtgaacag-3’ 3’-Ccactcgtaagacccgatttcgactg-5’
- 122 e
5’-cacatctcccctgggtttattctcac-3’
-22
3’-GGTGtgattcagtaataaaacataga-5’
+ 1632 d
3’-CCCGCTTAGATGTCTTGTTTCTTGCC-5’
+95d
5’-CCTTTAAAGGGCGACTTGCTCAGCGC-3’
lOu(s1)
+ 1976-
+ 1997 d
5’-GAATTCTGGATCTTTCGGACTG-3’
101(s)
+1827-
+1843d
3’-GAGGTTGCAATTTTCTG-5’
11(b)
-8-
-25
3’-aacgcgcgtttgagggga-5’
-8
5’-ctcatatccaatttttcc-3’
-25
3’-ctccgagaaaccccttctc-5’
20)
-25-
20)
-7-
3u(b)
-22-
-6
5’-aagcttgtgtcatcatc-3’
5u(r)
-7o-
-45
S’gagcagtgacatgcgaatgtcatacg-3’
51(r)
-53-
-19
3’-gggttataggatgagtcatcgaagttc-5’
a The number
h
’ d ’
of the primer refers to the corresponding exon, u = upperstrand, 1 = lower-strand, (i) = primer used for PCR and sequencing of the fragment spanning exon 9, intron 9 and the first part of exon 10, (s) = primer used only for sequencing, (b) = short primer used for sequencing of the boundary of exon 2 or 3. (r) = primer used for amplification of the exon 5 containing fragment, which was digested by AuaII. The numbers indicate the position of the primers referring to the 5’ (upper strand primers) or the 3’ (lower-strand primers) boundary of the corresponding exon. + indicates a position within an exon, - a position within an intron. Exceptions from this numbering are marked by d (see below). Capital letters represent bases of exons, small letters represent bases of introns. The position is indicated following the numbering of Deeb and Peng [17]. Primers were designed according to the information obtained by own sequencing.
246 strategy. In the two probands M.A. and O.J. a homozygous substitution of adenine for guanine was detected in exon 5 at position 818 (Fig. 2) [lo]. The sequence of the promotor region, the entire coding region including the sequence coding for the signal peptide, and the exon-intron boundaries were identical to the sequences re-
ported by Wion et al. [ll] and by Deeb and Peng
[wThe single base substitution at position 818 alters the sequence required for recognition by the restriction enzyme AuaII. This allowed us to test all family members for the presence of the mutation by AuaII digestion of PCR fragments of exon
A
Fig. 4. Stereo pair drawings of the proposed structures for: (A) sequence 186-196 (Thr-Arg-Gly-Ser-Pro-Gly-Arg-Ser-lle-Gly-lie) of LPL. Hydrogen atoms are not shown. The only side chain pointing into the loop is Ile 194. (B) sequence 186-196 with Gly 188 substituted by Glu. Dots represent approximate van der WaaIs radii of the GIu 188 and Ile 194 residues. In this orientation the carboxyl group of Glu 188 is in close contact with the Ile 194 side chain and the carbonyl oxygen of Ser 193. (C) sequence 186-196 with GIy 188 substituted by Glu. Dots represent van der Waals radii of the GIu 188 and Arg 192 residues. In this orientation the side chain of Glu 188 is in close contact with the side chain of Arg 192.
247 5. Fig. 3 shows the pedigrees of the 2 families. Because the primers used for amplification of exon 5 (5u(r) and 51(r) in Table 2) annealed to sequences in introns 4 and 5, the sequence spanned by the restriction fragments (383 bp) was longer than exon 5 (235 bp). As shown in Fig. 3, the wild type allele is cleaved into 3 fragments by AuaII (2 fragments with identical lengths of 87 bp and 1 fragment of 209 bp). The mutant allele yields only 2 AuaII fragments (87 bp and 296 bp). The patients with type I hyperlipoproteinemia (M.A.. M.F., and O.J.) were homozygous for the mutation, subjects M.S., M.L., M.M., M.T., M.K., P.B., P.A., O.G., O.S., O.A., and O.M. were heterozygous, and subjects M.J., M.B., R.A., M.C., M.G., O.C., and O.R. were normal (homozygous for the wild type allele). The mutation could not be detected in 45 controls (data not shown), thus making it unlikely that this base change is a common DNA polymorphism. Conformation
of the protein backbone
in the affected
region of LPL
The mutation causes a substitution of Glu for Gly at position 188 of the mature protein (Fig. 1). To investigate the possible structural cause for LPL deficiency, the probable conformations in the vicinity of position 188 were calculated for both the wild type and the mutated LPL protein [27]. In the wild type protein the most probable backbone conformation is a tight beta-turn formed by residues 189-192 (Ser-Pro-Gly-Arg) which is flanked by extended strands (hair pin loop) (Fig. 4A). With the exception of Ile 194 the side chains of all residues point away from the hair pin loop. The backbone conformation predicted for the mutant protein is very similar to the structure predicted for the wild type protein. However, the side chain of Glu 188 points into the hair pin loop approaching closely the hydrophobic part of the side chain of Arg 192 and Ile 194. Although the Glu side chain can be accommodated in the proposed structure, the replacement of Gly 188 by Glu results in a very tight packing of the hair pin loop. The side chain of Glu 188 with its negative charge can be oriented in two different ways with both yielding unfavorable interactions with neighboring residues. In one variant the negative charge is in close proximity to the carbonyl oxygen of Ser
193 carrying a negative partial charge and the side chain of Ile 194 (Fig. 4B). In the second variant. the negative charge approaches the hydrophobic part of the Arg 192 side chain (Fig. 4C). Therefore, in both variants the substitution of Glu for Gly at position 188 appears to destabilize the conformation of the beta-turn. Discussion Our studies within 2 Austrian families have demonstrated a clear cosegregation of the mutation with reduced (heterozygote state) or absent (homozygote state) postheparin plasma LPL activity. Because of the extensive sequencing strategy employed, linkage disequilibrium with a second causative mutation could be largely excluded, thus indicating that the mutation at cDNA position 818 was indeed responsible for the deficiency of LPL activity. This notion is supported by a mutation at cDNA position 781 in a patient from a Bethesda kindred [15]. Since no mutation was found in the exon-intron boundaries of the gene a splicing defect could also be largely excluded. Our studies are in good agreement with a recent study by Emi et al. [29] in which the same mutation was identified as responsible for LPL deficiency in a family of Northern European origin. Their study differs from ours in that cDNA rather than genomic fragments were employed to detect the mutation and therefore only the coding portion of the LPL gene was examined. Nevertheless, expression of normal and mutant cDNA clones allowed to confirm in an in vitro system that the mutation indeed generates an inactive enzyme. Finding the same base substitution in families with LPL deficiency derived from different ethnic backgrounds suggests that it is a widespread mutation. It might account for a considerable portion of mutations underlying lipoprotein lipase deficiency. The predicted backbone conformation in the vicinity of the mutation is a tight beta-turn flanked by extended strands. Several experimental results support the suggested conformation around Gly 188. Nuclear magnetic resonance studies on oligopeptides indicate that the motif Pro-Gly has a strong preference to form chain reversals or beta-turns [30]. Even more direct evidence is pro-
248
vided by the recent X-ray analysis of human PL [31]. Since this protein has strong sequence homology to LPL [32], the conformation of human PL is an excellent model for the LPL conformation. Gly 188 of LPL corresponds to Ile 209 of human PL and the sequence of human PL that corresponds to the predicted turn of LPL is ValPro-Lys-Leu [31]. In the X-ray structure of human PL this segment forms a tight turn similar to the conformation predicted for the corresponding segment of LPL. When speculating as to how the substitution of Glu for Gly at residue 188 of the LPL protein abolishes enzyme function, one has to consider the location of the substitution in a region which is highly conserved among LPL, hepatic lipase (HL), and pancreatic lipase (PL) [11,32-341. However, the affected residue itself is not conserved nor has a well-definded function been attributed to the region around Gly 188. In this context it appears appropriate to consider new insight in the structure and function of lipases which belong to the group of serine-type esterases [31,35]. Results of the above mentioned X-ray crystallography provide evidence that Ser 152, Asp 176, and His 263 form the catalytic triad of human PL [31]. Sequence alignment [34] indicates that this triad corresponds to the triad formed by Ser 132, Asp 156, and His 241 in human LPL. Therefore it seems unlikely that the mutation at position 188 directly affects the lipolytic site of LPL. However, loss of enzyme function could be due to improper folding of the LPL protein as a result of the Gly to Glu mutation. If the predicted structure around Gly 188 is correct, the introduction of a Glu residue results in tight packing and in unfavorable interactions of the negative charge with neighboring hydrophobic groups, destabilizing thus the loop conformation. It may be of interest that human HL has a Glu residue homologous to the mutated Glu residue in the LPL of our patients. There are, however, a number of additional different residues in the corresponding segment of human HL, which may compensate for the steric strain and unfavourable interactions. The mutation could also affect enzyme function by another mechanism. X-ray crystallography shows that in human PL the catalytic triad is covered by a large surface loop formed by residues
237-261 which was called the flap region [31]. Repositioning of this surface loop is required for the substrate triglyceride to gain access to the active site. Assuming a similar structure for human LPL, the flap region is in direct contact with the hair pin loop carrying the mutated Glu residue. Introduction of a negative charge by glutamic acid in this region may interfere with the conformational change of the flap region on interfacial activation and/or with the accommodation of a bulky hydrophobic triglyceride molecule [31]. Substitution of a single amino acid with sharply contrasting physicochemical characteristics has led to a complete loss of LPL activity. None of the regions which have been considered to form functional domains of the enzyme have been mapped to the region containing Gly 188. These regions are the active site (Ser 132, Asp 156, His 241). the interface region for alignment of lipid substrates (Asn 125 to Gly 142) the activator binding site (Lys 147, Lys 148), and the glycosaminoglycan binding site (Leu 276 to Lys 304) [11,31,33,36]. As discussed above, the mutation described in this report and the loss of enzyme activity caused by this mutation clearly assign to the region around Gly 188 a crucial role for enzyme function. Acknowledgements
The technical assistance of Miss C. Talman, Miss E. Meisl, Miss C. Garstenauer, Miss A. Reckwerth, and Miss J. Roessler is gratefully acknowledged. The cDNA clone of LPL was a kind gift of R.M. Lawn and K. Fischer (Genentech, South San Francisco, CA). The antibodies against human hepatic lipase were kindly provided by Dr. Gunilla Bengtsson-Olivecrona (University of Ume& Sweden). This work was supported by grants S-46/11 to F. Sandhofer and S-46/06 to J.R. Patsch of the Austrian Fonds zur Fijrderung der wissenschaftlichen Forschung, grant LH-27341 from the National Institute of Health to J.R. Patsch, and grants to H. Funke and G. Assmann from Bundesministerium fur Forschung und Technologie, F.R.G. References 1 Garfinkel,
AS. and Schotz, M.C.. In: Gotto, A.M.. Jr. (Ed.), Plasma Lipoproteins, Elsevier Science Publishing Co., New York, 1987, pp. 335-357.
249 2 Olivecrona, T. and Bengtsson-Olivecrona. G.. In: Borensztajn. J. (Ed.), Lipoprotein Lipase. Evener Press, Chicago. IL, 1987, pp. 15-58. 3 Jackson, R.L., In: Boyer. P.D. (Ed.). The Enzymes, Vol. 16. Academic Press, Orlando. FL, 1983, pp. 141-181. 4 Brunzell. J.D.. In: Striver. C.R.. Beaudet. A.L.. Sly. W.S. and Valle. D. (eds.). The metabolic basis of inherited disease, 6th Ed. McGraw-Hill. New York. 1989. pp. 116% 1180. 5 Breckenridge. W.C.. Little. J.A.. Steiner. G.. Chow. A. and Poapst. M.. Hypertriglyceridemia associated with deflciency of apolipoprotein C-II. N. Engl. J. Med., 298 (1978) 1265. 6 BrunLell. J.D.. Miller. N.E.. Alaupovic. P.. St. Hilaire, R.J., Wang. C.S.. Sarson. D.L.. Bloom, S.R. and Lewis. B.. I‘amihal chylomicronemia due to a circulating inhibitor of lipoprotein lipase activity. J. Lipid Res., 24 (1983) 12. 7 Have]. R.S.. Jr. and Gordon. J.R.. Idiopathic hyperlipemia: metabolic studies in an affected family, J. Clin. Invest.. 39 (1960) 1777. X Babirak, S.P.. Iverius. P.H., Fujlmoto. W.Y. and Brunzell, J.D.. Detection and characterization of the heterozygote state for lipoprotein lipase dehciency, Arteriosclerosis. 9 (19X9) 326. 9 Sparha. R.S.. Zollner. S.. Klisak, I.. Kirchgessner. T.G.. Komaromy. M.C.. Mohandas. T.. Schotz, M.C. and Lusis. A.J.. Human genes involved in lipoplysis of plasma lipoproteins: mapping of loci for lipoprotein lipase to 8~22 and hepatlc lipase to 15~21. Genomics. 1 (1987) 138. 10 Deeb. S.S. and Peng, R.. Structure of the human lipoprotein hpaae gene, Biochemistry. 2X (1989) 4131: correction: Biochemistry. 28 (1989) 6786. 1 I Wion. K.L.. Kirchgessner. T.G.. Lusis. A.T., Schotz. M.C. and Lawn, R.M.. Human lipoprotein lipase complementary DNA sequence. Science. 235 (1987) 1638. 12 Kirchgessner. T.G.. Svenaon. K.L.. Lusis. A.J. and Schotz. M.C.. The sequence of cDNA encoding lipoprotein lipase. J. Biol. Chem.. 262 (19X7) 8463. 13 Langlols. S.. Deeb. S.. Brunzell. J.D., Kastelein. J.J. and Hayden. M.R.. A major Insertion accounts for a significant proportion of mutation, underlying human lipoprotein Ilpa.\e deficiency. Proc. Nat]. Acad. Sci. USA.. X6 (1989) 948. I4 Devhn, R.H.. Deeb, S.. Brunzell, J. and Hayden. M.R.. Partial gene duplication involving exon-alu interchange rehults in lipoprotein lipase deficiency, Am. J. Hum. Genet.. 46 (1990) 11’. 15 Beg. O.U.. Meng. MS.. Skarlatos. S.I.. Previato, L.. Brun/ell. J.D.. Brewer, H.B.. Jr. and Fojo. S.S.. Lipoprotein Iipaae(Bethesda): a bingle amino acid substitution (Ala-176 1 Thr) leads to abnormal heparin binding and loss of enLym,c activity, Proc. Nat]. Acad. Sci. USA., X7 (1990) 3414. 16 Zaidan. H., Dhanireddy. R.. Hamosh. M.. BengtssonOhvecrona. G. and Hamosh. P.. Lipid clearing in premature infants during continuous heparin infusion: role of circulating lipases. Pediatric Res.. 19 (1985) 23.
17 Patsch, J.R.. Prasad. S.. Gotto, A.M., Jr. and Patsch. W.. High density lipoprotein2: relationship of the plasma levels of this lipoprotein species to its composition, to the magnitude of postprandial lipemia and to the activities of Iipoprotem lipase and hepatic lipase. J. Clin. Invest.. X0 (19x7) 341. 18 Peterson. J.. Olivecrona. T. and Bengtsson-Olivecrona. G.. Distribution of lipoprotein lipase between plasma and tiasues: effect of hypertriglyceridemla. Biochim. Biophys. Acta. X37 (1985) 262. 19 Spooner. P.M., Garrison. M.M. and Scow. K.O.. Regulatlon of mammary and adipose tissue lipoprotein lipase and blood trlacylglycerol in rats during late pregnancy: effect of probtaglandins. J. Clin. Invest.. 60 (1977) 702. 20 Kunhel. L.M.. Smith, K.D.. Bayer. S.H.. Borgoankar. S.D.. Wachtel. S.S.. Miller. O.J.. Breg. W.R.. Jones. HW. and Rary. J.M.. Analysis of human Y-chromosome-.\peclfic reIterated DNA in chromosome variant\. Proc. Natl. Acad. Sci. LISA., 74 (1977) 1745. 21 Southern. E.. Detection of specific sequences among DNA fragments separated hy gel electrophoresls. J. Mol. Biol.. YX (1975) 503. 22 Sambrook, J.. Fritsch. E.F. and Maniaus. 1. Molecular Clomng: A Laboratory Manual. 2nd Ed. Vol. 2. Cold Spring Harbor L.ahoratory Press. Nev, York, 19x9. pp. Y.31mY.62. 23 Feinherg, A.P. and Vogelstein, B.A., A technique for radmlabeling DNA restriction endonucleaae fragment5 to high \peciflc activity. Anal. Biochem.. 132 (19X3) 6. 24 Saiki. R.K.. Gelfand. D.H.. Stoffel. S.. Scharf. S.J., Higuchl. R.. Horn. G.T.. Mullia, K.B. and brlich. HA.. Primer-directed enzvmatic amplification of DNA with a thrrmosrahle DNA polymerase. Science, 239 (19xX) 4X7. 25 Samhrook. J., Fritsch. E.F. and Maniatlb. ‘I-., Molecular Cloning. A Lahoratory Manual. 2nd edn.. Vol. 2, Cold Sprmg Harbor Laboratory Presh. Neu I’ork. 1YXY. pp. E3% El 4. 26 Sanger. S.. Nicklen. S. and C‘oul~on. A.R.. DNA brquencing with chain-terminating inhlhitors. Proc. Nat]. Acd. Sci. USA.. 74 (1977) 5463. 27 Sippl. M.J.. Calculation of conformauonal ensembles from potentials of mean force. An approach to the Lnowledgehased prediction of local structurea in globular proteina, .I. Mol. Biol.. 213 (1990) 859. 2X Assmann, G. and Schulte, H.. Procam Trial. Pansclentia Verlag, Hedingen. Ziirich. 19X6. 29 Fmi. M.. Wilson, D.E., Iverius. P.-H.. Wu. I_.. Hata. A.. Hegele. R.. Williams, R.R. and Lalouel, J.M.. Missense mutation (Gly + GlulXX) of human hpoprotein lipase imparting functional deficiency. J. B~ol. Chem.. 765 (1990) 5910. 30 Dyson, H.J.. Rance. M.. Houghton. R.A., Lerner, R.A. and Wright. P.E., Folding of immunogenic peptide fragments of prote1n.s in water solution. 1. Sequence requirements for the formation of a reverse turn. J. Mol. Blol., 201 (1988) 161. 31 Winkler, F.K.. D’Arcy, A. and HunLiker. W. Structure of the human pancreatic lipahe. Nature. 343 (1990) 771.
250 32 Kirchgessner. T.G., Chuat, J.-C., Heinzmann, C., Etienne, J., Guilhot, S., Svenson, K., Ameis, D., Pilon, K., D’Auriol, L., Andalibi, A., Schotz, M.C., Galibert, F. and Lusis, A.J., Organization of the human lipoprotein lipase gene and evolution of the lipase gene family, Proc. Natl. Acad. Sci. USA.. 86 (1989) 9647. 33 Persson, B., Bengtsson-Olivecrona, G., Enerbaeck, S., Olivecrona, T. and Joernvall, H., Structural features of lipoprotein lipase: lipase family relationships, binding interactions, non-equivalence of lipase cofactors, vitellogenin similarities and functional subdivision of lipoprotein lipase, Eur. J. Biochem., 179 (1989) 39.
34 Datta, S.. Luo, C.-C., Li, W.-H., Tuinen, P.V., Ledbetter, D.H., Brown, M.A., Chen, S.-H., Liu, S.-w. and Chan, L. Human hepatic lipase: cloned cDNA sequence, restriction fragment length polymorphisms, chromosomal localization and evolutionary relationships with lipoprotein lipase and pancreatic lipse, J. Biol. Chem., 263 (1988) 1107. 35 Blow, D., More of the catalytic triad, Nature, 343 (1990) 694 (Editorial). 36 Yang, C.-Y., Gu, Z.W., Yang, H.X., Rohde, M.F., Gotto, A.M., Jr. and Pownall, H.J., Structure of bovine milk lipoprotein lipase, J. Biol. Chem., 264 (1989) 16822.