HUMAN MUTATION 1:55-62 (1992)
RESEARCH ARTICLE
Lethal Perinad Osteogenesis Pmperfecta Due to a Detected by Chemical Cleavage of an mRNAxDNA Sequence Mismatch John F. Bateman,’ Ingrid Moeller, Mamie Hannagan, Danny Chan, and William G. Cole Department of Paediacrics, University of Melbourne, Royal Children’s Hospital, Parkuilk, Victoria 3052, Australia Communicated by H. H. Dahl
A single base mismatch was detected by a chemical cleavage method in heteroduplexes formed between patient mRNA and a control collagen aZ(1) cDNA probe in a case of osteogenesis imperfecta type 11. The region of the mRNA mismatch was amplified using the polymerase chain reaction, cloned and sequenced. A heterozygous point mutation of G to C at base pair 1,774of the collagen a 2 ( I ) mRNA resulted in the substitution of glycine with arginine at amino acid position 457 of the helix. Type I collagen of al(1)-and a2(I)-chains from the patient migrated slowly on electrophoresis due to increased levels of posttranslational modification of lysine. The parents’ fibroblast collagen did not contain the mRNA mismatch and the collagens showed normal electrophoretic behaviour. Twodimensional electrophoresis of the CNBr peptides from the patient’s collagen confirmed the excessive posttranslational modification of the al(1)-and a2(I)-chains in the CNBr peptides N-terminal to the mutation due to disruption of the obligatory Gly-X-Y triplet repeat of the helix. The mutation led to reduced procollagen secretion and helix destabilization as evidenced by a decreased thermal stability. These data lend further support to the accumulating evidence that type I collagen a2(1) glycine substitution mutations result in the same spectrum of clinical severity as those in the al(1)-chain. The disruptive effect of the glycine mutations seems to be largely dependent on the nature of the substituted amino acid, the position in the a-chain of the mutation, and the nature of the local surrounding amino acid sequence, rather than whether the mutation resides in the al(1)-or a2(I)-chain. 0 1992 Wiley-Liss, Inc. KEY WORDS:
Type I collagen, Point mutations, Glycine substitutions, Connective tissue disorders, Bone d’iseases
INTRODUCTION
lie the “brittle bone disease,” osteogenesis imperfecta ( 0 I ) I (reviewed by Byers, 1990; Vuorio and De Crombrugghe, 1990; Kuivaniemi et al., 1991). These mutations cover the spectrum of possible gene alterations including deletions, insertions, rearrangements, and point mutations. While it is now clear that collagen I mutations lead to the disease in at least 90% of cases, the current challenge is to gain a better understanding of the mo-
lecular pathology and correlate the clinical phenotype with the biochemical abnormality. The relatively mild form of 0 1 (01 type I, Sillence et al., 1979) is characterised by a reduced amounts of normal type I collagen in tissues and produced by cell cultures (Wenstrup et al., 1990). This reduced production may result from mutations that produce a nonfunctional allele of COLlA1, aberrant mRNA processing, or production of structurally abnormal prooll (I) chains that cannot associate and form procollagen molecules and are degraded. The clinical consequences result
‘The abbreviations used are: 01, osteogenesis imperfecta; PCR, polymerase chain reaction; bp, base pair; CNBr, cyanogen bromide.
Received Februay 19, 1992; accepted March 10,1992. *To whom reprint requestskorrespondence should be addressed.
Mutations of the al(1)and ol2(I) chains of type I collagen have been consistently shown to under-
0
1992 WILEY-LISS.INC.
56
BATEMAN ET AL.
from the diminished type 1 collagen content (Bonadio et al., 1990). In the more severe forms of 0 1 structurally abnormal collagen chains are incorporated into collagen molecules and, coupled with increased collagen breakdown, lead to impaired tissue collagen function (Wenstrup et al., 1990; Byers, 1990; Kuivaniemi et al., 1991). A relatively common mutation found in 0 1 that leads to destabilization of the collagen triple helix is the substitution by another amino acid of the obligatory glycine residue of the Gly-X-Y triplet repeat sequence of the triple helix. A pattern is emerging that glycine substitutions toward the C-terminus of the al(1) or a2(I) chains are more clinically severe than those toward the N-terminus. This is consistent with the model that since helix formation propagates from the C- to N-terminus, those mutations toward the C-terminus will be more disruptive. Because of the [al(I)],aZ(I) molecular composition it is also predicted that mutations in the a2(I) chain would be less clinically severe than those in the al(1). However, it is also becoming apparent that these generalizations must be modified by considering the nature of the glycine substitution and the surrounding amino acid sequence (Byers, 1990; Bachinger and Davis, 1991) and many more mutations spanning the collagen helical domain will need to be characterized before any strict correlation can be made between the biochemical and clinical phenotypes. In this paper we report the substitution of G l ~ - 4 5 7by~ Arg in the a2(I) chain of type I collagen in a child with the perinatal lethal form of
01.
Cetus Corp., USA. Sequenase sequencing kits were purchased from United States Biochemical Corp., USA. All other chemicals were commercially available analytical grade.
Clinical Summary The patient (0145) was born at 36 weeks gestation and died shortly after birth. Autopsy revealed the beaded ribs and short crumpled long bones typical of lethal perinatal 0 1 type IIA) (Sillence e t al., 1979; Cole, 1988). T h e parents were normal and unrelated. Dermal fibroblast cultures were established from the parents, patient, and controls and maintained as described previously (Bateman et al., 1984, 1986). All tissue biopsies were obtained with informed consent and approval of the Ethics Committee of this hospital.
Collagen Biosynthetic Labelling Confluent fibroblast cultures were biosynthetically labelled as previously described (Bateman et al., 1984, 1986). After culture for three days in growth medium containing 0.25 mM sodium ascorbate, labelling with 10 pCi/ml ~ [ 5 - ~ H ] p r o line was carried out for 18 hr in medium containing 10% (vlv) dialysed fetal calf serum, 0.25 mM sodium ascorbate, and 0.1 mM P-aminopropionitrile fumarate. The cell layer and medium fractions were separated for analysis and the procollagens, precipitated with 25% saturated (NH4),S04, were subjected to limited proteolysis with pepsin (Bateman e t al., 1984, 1986).
SDS-Polyacrylamide Gel Electrophoresis MATERIALS AND METHODS
~-[5-~H]Proline (30 Ci/mmol) and [3ZP]dCTP (3,000 Ci/mmol) were purchased from Amersham Australia Pty. Sydney, NSW, Australia. Dulbecco modified Eagle's medium and fetal calf serum were purchased from Flow Laboratories Australia, Stanmore, NSW, Australia. Pepsin, sodium ascorbate, P-aminopropionitrile fumarate, trypsin, and chymotrypsin were purchased from Sigma Chemical Co., St. Louis, MO. Restriction endonucleases and nick-translation kits were obtained from Boehringer Mannheim, Germany. cDNA synthesis kits and the M13mp8 vector were purchased from Amersham Australia Ltd., Sydney, Australia, and Tag polymerase was obtained from the Perkin Elmer-
'Amino acid positions are numbered by the standard convention in which the first glycine of the triple helical domain of the a-chain is number one.
Collagen a-chains were resolved on 5% (w/v) separating gels containing 2 M urea (Bateman et al., 1984, 1986). Collagens prepared from cell cultures and tissues were also digested with CNBr (Scott and Veis, 1977) and analysed by two-dimensional electrophoresis (Cole and Chan, 1981). The a-chains and peptides were detected by fluorography (Bonner and Laskey, 1974).
Collagen Thermal Stability The freeze-dried [3H]proline-labelled collagens were dissolved in 0.4 M NaCVO.l M Tris-HC1 buffer, pH 7.4, at 4°C. The samples were warmed stepwise (l"C/min) from 34 to 43°C and at 0.5"C intervals samples were taken and digested with a mixture of trypsin and chymotrypsin as described by Bruckner and Prockop ( 1981). The proportion of the collagen that was resistant to proteolysis at each temperature interval was determined by scin-
&(I) GLY457 TO ARG SUBSTITUTIONIN OSTEOGENESIS IMPERFECTA
57
tillation counting of collagen a-chain bands that were resolved by electrophoresis and identified by fluorography (Bateman et al., 1988a). The thermal denaturation temperature, T,, was defined as the temperature at which half of the collagen was degraded. Formation and Chemical Cleavage mRNA:cDNA Heteroduplexes
Total RNA was isolated from confluent fibroblast cultures (Lui et al., 1979; Wake et al., 1985). mRNA:cDNA heteroduplexes were formed in a volume of 50 ~1 containing approximately 5 ng (50,000-100,000 dpm) of labelled cDNA probe, 5-10 Fg of total RNA, 80% (v/v) formamide, 40 mM PIPES, pH 6.4, 1 mM EDTA, and 0.4 M NaC1. The mixture was denatured at 80°C for 5 min, incubated at 60°C for 2 hr, and ethanol precipitated. Chemical modification of mismatched nucleotides with hydroxylamine or osmium tetroxide, cleavage with piperidine, and electrophoresis of the products on denaturing 7 M urea/5% ( w h ) acrylamide gels have been described (Cotton et al., 1988; Dahl et al., 1989; Lamande et al., 1989; Bateman et al., 1989). Collagen protein analyses defined the apparent site of the mutation in the collagen chains (see Fig. 2), and this was used as a guide to the selection of suitable control cDNAs for heteroduplex formation. Control al(1) and a2(I) cDNA probes (Bateman et al., 1991), chosen to span the abnormal region, were purified and end-labelled with [cI-~’P]~CTPby the fill-in reaction using the Klenow fragment of DNA polymerase I (Sambrook et al., 1989). Amplification and Sequencing of cDNA
First-strand cDNA was synthesized from total
RNA using a cDNA synthesis kit primed with oligo(dT). Approximately 50 ng of cDNA was amplified by the PCR (Saiki et al., 1988) through 30 cycles using Taq polymerase (Kogan et al., 1987). Each cycle consisted of denaturation at 92°C for 1.5 min, annealing of primers at 63°C for 1.5 min, and primer extension at 72°C for 3 min (Lamande et al., 1989; Bateman et al., 1989). A 357-bp fragment corresponding to bases 1,522-1,879 of the a2(1) cDNA was amplified with the primers 5 ’-GGAAAAGAAGGTCCTGTC-3’ and 5‘GGAGACCAAACTCACCAT-3’ (De Wet et al., 1987).3 3The base pairs are numbered from the start of transcription of the prouZ(1)mRNA (DeWet et al., 1987).
FIGURE 1. Electrophoresis of
pepsin-digestedfibroblast collagens. Fibroblast cultures were labelled for 18 hr with [3H]prolineand the collagens were pepsin-digested and analysed without reduction on SDS/polyaqlamide gels (5%) (see the Experimental section for details). C, cell layer fraction; M. medium fraction. The migration positions of type 1 collagen al(1)and a2(I)chains and type 111 collagen [al(lll)]3 are shown.
Amplification products of the predicted size were purified, treated with T4 polynucleotide kinase (Sambrook et al., 1989), and cloned into a SmaI cut, dephosphorylated M13mp8 vector. Multiple clones from two independent amplification reactions were sequenced using a Sequenase kit. RESULTS Type I Collagen Protein Analysis
The al(1)-chain of 0145 fibroblast type I collagen showed slow electrophoretic migration compared to control fibroblast al(1)-chains (Fig. 1). This migrational abnormality was also evident in al(1)- and a2(1)-chains of type I collagen extracted from the bone and dermis of the patient and was due to overhydroxylation of lysine (Bateman et al., 1986). The overmodified collagen produced by 0145 fibroblasts was secreted poorly, at 60% the level of control fibroblasts (Fig. 1; Bateman et al., 1986). Electrophoretic analysis of the collagen a-chains produced by the parents’ fibroblasts showed only normally migrating species which were efficiently secreted (data not shown). Two-dimensional electrophoresis of the type I collagen CNBr peptides demonstrated that only the al(I)CB8, al(I)CB3, and the a2(I)CB4 peptides
58
BATEMAN ET AL.
the mRNA derived from the parents (Fig. 3b). Treatment with osmium tetroxide/piperidine did not reveal any T mismatches in 0145 mRNA and no mismatches were detected by hydroxylamine or osmium tetroxide treatment of heteroduplexes formed between 0145 mRNA and control al(1) cDNA (data not shown).
Characterization of the mRNA Mutation
FIGURE 2. Two-dimensionalelectrophoresis of the CNBr peptides of 0145 fibroblast collagen. Fibroblasts were labelled in cell culture and collagens in the medium fraction were digested with CNBr and the products analysed by two-dimensional PAGE (see the Experimental section for details). CNBr peptides are identified as follows: 3.5, a2(1)CB3.5; 4, a2(I)CB4 8. al(l)CB8; 7, al(l)CB7; 6, al(I)CB6 and 3, al(I)CB3. The “tailing” of al(I)CB3, al(I)CB8. and a2(I)CB4due to increased lysine posttranslational modification is indicated by arrows.
showed ‘(tailing” of the proximal edges resulting from the overmodification of the lysine residues (Fig. 2). The C-terminal a l ( 1 ) peptide, al(I)CB6, showed n o evidence of abnormal migration due to overmodification. T h e C-terminal peptide of the a2(I) chain, a2(I)CB3.5, also had an apparently normal migration although the resolution of this large peptide by two-dimensional electrophoresis may not be adequate to enable detection of such subtle changes in electrophoretic migration (Fig. 2). This peptide map result clearly indicated that the overmodification of lysine residues was restricted to the N.termina1 half of the al(1) and a2(I) chains and suggested that the mutation resided in the central region of the triple helix.
Detection of an Abnormal mRNA Sequence by Chemical Cleavage Heteroduplexes were formed using control a1(I) and a2(I) cDNAs, which spanned the central portion of the collagen helix, and mRNA extracted from cultured dermal fibroblasts of 0145 and controls. Hydroxylamine/piperidine treatment cleaved a 220-bp fragment from the 311-bp TUL+NCOI fragment of a2(I) cDNA. This finding demonstrated that there was a mismatched C at approximately base pair 1,773-1,775 of the control a2(I) cDNA (Fig. 3a). This mutation was not present in
To define the mutation, a short length of firststrand d ( I ) cDNA was amplified by the PCR using unique oligonucleotides chosen to span the region containing the mismatch. Sequencing of M13 subclones of the amplified products identified the mutation as a single base substitution at base pair 1,774 which converted the codon for glycine (GGT) to CGT (arginine) at amino acid position 457 of the helix (Fig. 4). The abnormal sequence was confirmed by sequencing of multiple clones from two separate PCR reactions. Clones containing the normal sequence (GGT) were also identified indicating that the patient was heterozygous for the a2(I) mutation. The 0145 mutant clones also contained a G to C change at position 1,780, converting Ala-459 to Pro. This sequence variant was identified in three other individuals (Bateman et al., 1991) and probably represents a neutral sequence variant. Collagen Melting Temperature Estimation of the thermal denaturation temperature (T,) of type I collagen synthesized by 0145 and control fibroblasts was made by measuring its resistance to proteinase digestion after heating to different temperatures (Fig. 5). The T , of 0145 type I collagen was approximately 40.9”C compared with 41.2”C for control type I collagen. Since 0145 is heterozygous this probably represents an underestimate of the reduction in the stability of the mutant type I molecules. The T , of type 111 collagen from 0145 was the same as control values (data not shown).
DISCUSSION The mutation in this case of lethal perinatal 01 was a heterozygous point mutation of G to C at base pair 1,774 of the 012(I) cDNA resulting in the substitution of Gly-457 by Arg. Since this point mutation occurs within the sequence encoded by exon 28, it must represent a point mutation in one allele of COLlA2. mRNA from both parents did not contain the mutation suggesting that the base change arose as a sporadic new mutation in the patient. Somatic and gonadal mosaicism, which
&(I) GLY-457TO ARG SUBSTITUTION IN OSTEOGENESIS IMPERFECTA
59
FIGURE 3. Mutation detection by chemical modification and cleavage. Total RNA extracted from 0145 and control fibroblasts (a) and 0145 and parent’s fibroblasts (b)were hybridized to a 311-bp TaqI-XhoI 32P-labelledfragment of the uZ(1) cDNA and reacted with hydroxylamine for 0 to 60 min as indicated, treated with piperidine. and the resultant frag-
ments resolved on a 5% (w/v) polyacrylamide gel (see the Experimental section for details). The labelled a2(1) cDNA fragment produced by cleavage a t mismatched Cs by hydroxylamine (220 bp) and the uncleaved probe (311 bp) are indicated. The molecular weight standards used were 32Plabelled Haelll fragments of $X174 DNA.
has been described in the parents of several cases of 0 1 (Constantinou et al., 1989; Wallis et al., 1990a; Cohn et al., 1990), was not excluded as the source of the mutation in 0145. Increased levels of lysine modification in the al(1)and a2(1)chains have been characterized in collagen extracted from bone and dermis in 0145 (Bateman et al., 1986). Two-dimensional CNBr peptide mapping studies demonstrated that the overmodified residues were restricted to a1 (I) and a2(I)chains regions toward the N-terminus of the helix from the mutation. Tha charge-change was not able to be directly demonstrated at the protein level by two-dimensional peptide mapping because of the lack of resolution of the multiple charged forms of the large a2(I)CB3.5 peptide. In the case of Gly to Arg mutations in the al(1) chain the
well-resolved oll (I) CB peptides have allowed the direct detection of the mutation at the protein level (Bateman et al., 1987a,b). The correspondence of the regional localization of overhydroxylation with the position of mutation in 0145 and in all other cases studied by us (Bateman et al., 1991) has confirmed that the demonstration of the overmodified domains protein is a reliable indicator of the location of the primary mutation. These data and the detection of a small reduction in the thermal stability of type I collagen in 0145 support the model that this glycine substitution both destabilises the triple helix and disrupts helix propagation from the C- to Nterminus, exposing the regions toward the N-terminus of the helix to excessive enzymatic posttranslational modification.
60
BATEMAN ET AL.
FIGURE4. DNA sequence of mutant and normal cDNA clones.
cDNA clones of the PCR-amplified mRNA region containing the mismatched base were sequenced (see the Experimental section for details). The 0145 mutant clone contains a G to C substitutionat base position 1774 converting Gly-457 to arginine. In mutant clones base position 1,780 was C, so that amino acid 459 was proline rather than alanine (De Wet et al., 1987). Nucleotides different to the published sequence are boxed.
Tissue analysis of 0145 dermis and bone showed that, in common with other cases of lethal perinatal 01, there was a marked reduction in tissue collagen content (Bateman et al., 1986). In 0145, tissue collagen levels were approximately half that of control values. Fibroblast collagen secretion was 60% of control values (Bateman et al., 1986) and it is likely that there was also increased degradation of the mutant collagen as determined in other cases of lethal perinatal 0 1 (Bateman et al., 1984). This increased collagen degradation results, in part, from the incorporation of the mutant proa2(I) into type I procollagen molecules, disrupting helix formation and stability and leading to “protein suicide” (Prockop and Kivirikko, 1984). However, some mutant collagen was secreted by the fibroblasts and the presence of this abnormal collagen in the extracellular matrix dramatically amplified the deleterious effect of the mutation.
T h e secretion of mutant collagens into the extracellular matrix seems to be the determining factor for severe 01 phenotypes, since when collagen production was reduced to 50% without the production and secretion of mutant collagen, a milder 0 1 type I results (Bonadio et al., 1990). The mechanism whereby the mutant species has such a catastrophic effect on bone matrix organisation is currently unknown. However, in one case studied in detail, the mutation causes a conformational change to the collagen triple helix which delays fibril formation and alters fibril morphology (Kuivaniemi et al., 1991). Such alterations to collagen fibril architecture may interfere with the normal bone mineralisation process. With an increasing number of 0 1 mutations now defined a model relating the position and type of the glycine substitution mutation to the clinical severity has emerged (Byers, 1990). This is best documented for the al(1) chain and suggests that point mutations toward the C-terminus of the helix will, in general, lead to more severe clinical consequences. However, this simple model must be modified t o consider the sequence context and nature of the substitution. The position-specific effect of the mutations seen in 0 1 may result from disturbances to specific functional sequences involved in collagen-collagen, collagen-matrix, or collagen-cell interactions. The local sequence may also play a critical role in how the collagen triple helix is affected by the mutation by providing domains of relatively low or high helix stability (Bachinger and Davis, 1991) defined largely by the presence or absence of stabilizing hydroxyproline residues (Burjanadze, 1979). This model proposes that mutations in regions of low local stability will be less deleterious to helical stability. In addition, if sequences aminoterminal to the mutation are of high stability renucleation of helix formation will occur readily and the interruption to helix propagation caused by the mutation will be minimal (Bachinger and Davis, 1991). In 0145 the Gly-457 to Arg occurs in a “loose” part of the helix, and taken with the small apparent reduction in the T, of 0.3”C, compared with T, reductions of more than 1°C in other cases of lethal perinatal01 (Baker et al., 1989; Kuivaniemi et al., 1991), suggests that this mutation should have a relatively mild effect on helix stability. However the lack of an adjacent stabilising aminoterminal nucleation site may lead to a more pronounced effect of the mutation on helix propagation.
&(I) GLY-457 TO ARG SUBSTITUTION IN OSTEOGENESIS IMPERFECTA
61
-
O 903
80-
3 70-60C 3 500 u
.-m 2 0
g
-E 0 2 0
n
‘0-0
O\
\ o
t\
40- 0-OControl 30-- 0-001 45 2010-
0T
4
FIGURE
5. Thermal denaturation of type I collagen. (3H]Proline-labelledcollagens were prepared from the medium of 0145 and control fibroblasts. They were treated with pepsin and after inactivation of the pepsin the collagens were warmed slowly from 34 to 43°C. Serial samples were taken and digested with trypsidchymotrypsin and the resistant
chains were resolved by electrophoresisand the melting temperature of the helix was determined by the relative proportion of the a-chains that were resistant to proteolysis (see the Experimental section for details). (4Type I collagen from 0145 [a2(1)Gly-457 to Arg); (0)type I collagen from control fibroblasts.
The predominant factor in generating the lethal phenotype in 0145 may be the nature of the substituting amino acid, arginine. Glycine to arginine mutations have been characterized in the al(1) chain at amino acid positions 154 (Starman et al., 1989), 391 (Bateman et al., 1987a), 550 (Byers, 1990), 667 (Bateman et al., 198713, 1988b), 847 (Wallis et al., 1990b), and 976 (Lamande et al., 1989). All of these lead to the lethal form of 0 1 except the most amino terminal mutation at Gly154 which results in 0 1 type 111. These data suggest that substitutions of al(1) glycine by the bulkier arginine are extremely deleterious along the majority of the helix leading to the lethal phenotype. In the other reported a2(I) Gly to Arg substitution, Gly-1012 to Arg leads to 0 1 type IV (Wenstrup et al., 1988). However, this may represent a special case since the mutation is in the last helical triplet and may not, therefore, have the same effect on helix stability and/or propagation. It has also been suggested that a2(I) glycine substitutions may be less deleterious than corresponding mutations in the al(1) chain because of the [c~l(I)]~a2(1) stoichiometry of the type I collagen molecule (Vuorio and De Crombrugghe, 1990). However, accumulating data demonstrate that a2(I) substitutions of Gly-457 by Arg, Gly586 by Val (Lamande et al., 1989) and five cases of Gly to Asp in the C-terminal third of the helix (Byers, 1990) lead to the lethal form of 01. This suggests that as further 012(I) mutations are characterized a similar spectrum of clinical severity to
that seen in al(1) will be displayed and that the expression of clinical phenotype will depend not only on the type, position, and sequence surrounding the collagen mutation, but also on how the abnormality affects the complexities of collagen biosynthesis, fibril deposition, and structural and regulatory interactions within the extracellular matrix.
ACKNOWLEDGMENTS This work was supported by grants from the National Health and Medical Research Council of Australia and the Royal Children’s Hospital Research Foundation.
REFERENCES Bachinger H-P, Davis ]M (1991) Sequence specific thermal stability of the collagen triple helix. Int J Macromolec 13:152156. Baker AT, Rarnshaw JAM, Chan D, Cole WG, Bateman ]F (1989) Changes in collagen stability and folding in lethal pennatal osteogenesis imperfecta. The effect of al(l) glycine to arginine mutations. Biochem J 261:253-257. Bateman JF, Mascara T, Chan D, Cole, WG (1984) Abnormal type I collagen metabolism by cultured fibroblasts in lethal perinatal osteogenesis imperfecta. Biochem J 2 17:103-1 15. Bateman JF, Chan D, Mascara T, Rogers JG, Cole WG (1986) Collagen defects in lethal perinatal osteogenesis imperfecta. Biochem J 240:699-708. Bateman JF, Chan D, Walker ID, Rogers JG, Cole WG (1987a) Lethal perinatal osteogenesis imperfecta due to the substitution of arginine for glycine at residue 391 of the al(1)-chain of type 1 collagen. J Biol Chern 262:7021-7027. Bateman JF, Mascara T, Chan D, Cole WG (1987b) A structural
62
BATEMAN ET AL.
mutation of the collagen al(1) CB7 peptide in lethal perinatal osteogenesis imperfecta. J Biol Chem 262:4445-4451. Bateman IF, Harley V, Chan D, Cole WG (1988a) Comprehensive analysis of collagen metabolism in vitro using 3H/14Cproline dual-labeling and polyacrylamide gel electrophoresis. Anal Biochem 168:171-176. Bateman IF, Lamande SR, Dahl H-HM, Chan D, Cole W G (1988b) Substitution of arginine for glycine 664 in the collagen u l (I) chain in lethal perinatal osteogenesis imperfecta: demonstration of the peptide defect by in vitro expression of mutant cDNA. J Biol Chem 263: 11627-1 1630. Bateman JF, Lamande SR, Dahl H-HM, Chan D, Mascara T, Cole WG (1989) A frameshift mutation results in a truncated nonfunctional carboxyterminal proal(1) propeptide of type 1 procollagen in osteogenesis imperfecta. J Biol Chem 264: 10960- 10964. Bateman JF, Lamande SR, Hannagan M, Moeller I, Dahl H-HM, Cole WG (1991) A chemical cleavage method for the detection of RNA base changes: Experience in the application to collagen mutations in osteogenesis imperfecta. Am J Med Genet, in press. Bonadio J , Saunders TL, Tsai E, Goldstein SA, Morris-Wiman J , Brinkley L, Dolan DF, Altschuler RA, Hawkins JE, Bateman IF, Mascara T,Jaenisch R (1990) A transgenic mouse model of the mild dominant form of osteogenesis imperfecta. Proc Natl Acad Sci USA 87:7145-7149. Bonner WM, Laskey RA (1974) A film detection method for tritium-labelled proteins and nucleic acids in polyacrylamide gels. Eur J Biochem 46:83-88. Bruckner P, Prockop DJ (1981) Proteolytic enzymes as probes for the triple helical corformation of procollagen. Anal Biochem 110:360-368. Burjanadze TV (1979) Hydroxyproline content and location in relation to collagen thermal stability. Biopolymers 18:93 1938. Byers PH (1990) Brittle bones-fragile molecules: disorders of collagen gene structure and expression. Trends Genet 6:293 -300. Cohn DH, Starman BJ, Blumberg B, Byers PH (1990) Recurrence of lethal osteogenesis imperfecta due to parental mosaicism for a dominant mutation in a human type I collagen gene (COLIAI). Am J Hum Genet 46:1034-1040. Cole WG, Chan D (1981) Analysis of the heterogeneity of human collagens by two-dimensional polyacrylamide-gel electrophoresis. Biochem J 197:377-383. Cole WG (1988) Osteogenesis imperfecta. Bailliere's Clin Endocrinol Metab 2:243-265. Constantinou CD, Nielsen KB, Prockop DJ (1989) A lethal variant of osteogenesis imperfecta has a single base mutation that substitutes cysteine for glycine 904 of the ul(1) chain of type I procollagen. The asymptomatic mother has an unidentified mutation producing an overmodified and unstable type I procollagen. J Clin Invest 83:574-584. Cotton RGH, Rodrigues NR, Campbell RD (1988) Reactivity of cytosine and thymine in single-base-pair mismatches with hydroxylamine and osmium tetroxide and its application to the study of mutations Proc Natl Acad Sci USA 85:4397-4401. Dahl H-HM, Lamande SR, Cotton RGH, Bateman JF (1989) Detection and localization of base changes in RNA using a chemical cleavage method Anal Biochem 183:263-268. De Wet W, Bernard M, Benson-Chanda V, Chu M-L, Dickson L,
Weil D, Ramirez F (1987) Organization of the human prou2(1) collagen gene. J Biol Chem 26216032-16036. Kogan SC, Doherty M, Gitschier J (1987) An improved method for prenatal diagnosis of genetic diseases by analysis of amplified DNA sequences: Application in hemophilia A. N Engl J Med 317:985-990. Kuivaniemi H, Tromp G , Prockop DJ (1991) Mutations in collagen genes: causes of rare and some common diseases in humans. FASEB J 5:2052-2060. Lamande SR, Dahl H-HM, Cole WG, Bateman JF (1989) Characterization of point mutations in the collagen COLlAl and COLl A2 genes causing lethal perinatal osteogenesis imperfecta. J Biol Chem 264:15809-15812. Lui CP, Slate OL, Gravel R, Ruddle FH (1979) Biological detection of specific mRNA sequences by microinjeciton. Proc Natl Acad Sci USA 76:4503-4506. Prockop DJ, Kivirikko KI (1984) Heritable diseases of collagen. N Engl J Med 311:376-386. Saiki RK, Gelfand DH, Stoffel S,Scharf SJ, Higuchi R, Horn GT, Mullis KB, Erlich HA (1988) Primer-directed enzymatic amplification of DNA with a thermostable DNA polymerase. Science 239:487-491. Sambrook J, Fritsch EF, Maniatis T (1989) Molecular Cloning, a Laboratory Manual, 2nd ed. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory. Scott PG, Veis A (1976) The cyanogen bromide peptides of bovine soluble and insoluble collagens. Connect Tissue Res 4: 107-116. Sillence DO, Senn AS, Danks DM (1979) Genetic heterogeneity in osteogenesis imperfecta. J Med Genet 16:lOl-116. Starman BJ, Eyre D, Charbonneau H, Hanylock M, Weis MA, Weiss L, Graham JM, Byers PH (1989) Osteogenesis imperfecta. The position of the substitution for glycine by cysteine in the triple helical domain of the proul(1) chains of type I procollagen determines the clinical phenotype. J Clin Invest 84:1206-1214. Vuorio E, De Crombrugghe B (1990) The family of collagen genes. Annu Rev Biochem 59:837-872. Wake SA, Mercer JFB (1985) Induction of metallothionein mRNA in rat liver and kidney after copper chloride injection. Biochem J 228:425-432. Wallis GA, Starman BJ, Zinn AB, Byers PH (1990a) Variable expression of osteogenesis imperfecta in a nuclear family is explained by somatic mosaicism for a lethal point mutation in the a l ( I ) gene (COLlAl) of type I collagen in a parent. Am J Hum Genet 46:1034-1040. Wallis GA, Starman BJ, Schwartz MF, Byers PH (1990b) Substitution of arginine for glycine at position 847 in the triple helical domain of the u l ( l ) chain of type I collagen produces lethal osteogenesis imperfecta. Molecules that contain one or two abnormal chains differ in stability and secretion. J Biol Chem 265:18628-18633. Wenstrup RJ, Cohn D H, Cohen T , Byers PH (1988) Arginine for glycine substitution in the triple helical domain of the products of the a2(1) collagen allele (COLLAZ) produces the osteogenesis imperfecta type IV phenotype. J Biol Chem 263:77347740. Wenstrup RJ, Willing MC, Starman BJ, Byers PH (1990) Distinct biochemical phenotypes predict clinical severity in non-lethal variants of osteogenesis imperfecta. Am J Hum Genet 46:975982.
RESEARCH ARTICLE
Lethal Perinad Osteogenesis Pmperfecta Due to a Detected by Chemical Cleavage of an mRNAxDNA Sequence Mismatch John F. Bateman,’ Ingrid Moeller, Mamie Hannagan, Danny Chan, and William G. Cole Department of Paediacrics, University of Melbourne, Royal Children’s Hospital, Parkuilk, Victoria 3052, Australia Communicated by H. H. Dahl
A single base mismatch was detected by a chemical cleavage method in heteroduplexes formed between patient mRNA and a control collagen aZ(1) cDNA probe in a case of osteogenesis imperfecta type 11. The region of the mRNA mismatch was amplified using the polymerase chain reaction, cloned and sequenced. A heterozygous point mutation of G to C at base pair 1,774of the collagen a 2 ( I ) mRNA resulted in the substitution of glycine with arginine at amino acid position 457 of the helix. Type I collagen of al(1)-and a2(I)-chains from the patient migrated slowly on electrophoresis due to increased levels of posttranslational modification of lysine. The parents’ fibroblast collagen did not contain the mRNA mismatch and the collagens showed normal electrophoretic behaviour. Twodimensional electrophoresis of the CNBr peptides from the patient’s collagen confirmed the excessive posttranslational modification of the al(1)-and a2(I)-chains in the CNBr peptides N-terminal to the mutation due to disruption of the obligatory Gly-X-Y triplet repeat of the helix. The mutation led to reduced procollagen secretion and helix destabilization as evidenced by a decreased thermal stability. These data lend further support to the accumulating evidence that type I collagen a2(1) glycine substitution mutations result in the same spectrum of clinical severity as those in the al(1)-chain. The disruptive effect of the glycine mutations seems to be largely dependent on the nature of the substituted amino acid, the position in the a-chain of the mutation, and the nature of the local surrounding amino acid sequence, rather than whether the mutation resides in the al(1)-or a2(I)-chain. 0 1992 Wiley-Liss, Inc. KEY WORDS:
Type I collagen, Point mutations, Glycine substitutions, Connective tissue disorders, Bone d’iseases
INTRODUCTION
lie the “brittle bone disease,” osteogenesis imperfecta ( 0 I ) I (reviewed by Byers, 1990; Vuorio and De Crombrugghe, 1990; Kuivaniemi et al., 1991). These mutations cover the spectrum of possible gene alterations including deletions, insertions, rearrangements, and point mutations. While it is now clear that collagen I mutations lead to the disease in at least 90% of cases, the current challenge is to gain a better understanding of the mo-
lecular pathology and correlate the clinical phenotype with the biochemical abnormality. The relatively mild form of 0 1 (01 type I, Sillence et al., 1979) is characterised by a reduced amounts of normal type I collagen in tissues and produced by cell cultures (Wenstrup et al., 1990). This reduced production may result from mutations that produce a nonfunctional allele of COLlA1, aberrant mRNA processing, or production of structurally abnormal prooll (I) chains that cannot associate and form procollagen molecules and are degraded. The clinical consequences result
‘The abbreviations used are: 01, osteogenesis imperfecta; PCR, polymerase chain reaction; bp, base pair; CNBr, cyanogen bromide.
Received Februay 19, 1992; accepted March 10,1992. *To whom reprint requestskorrespondence should be addressed.
Mutations of the al(1)and ol2(I) chains of type I collagen have been consistently shown to under-
0
1992 WILEY-LISS.INC.
56
BATEMAN ET AL.
from the diminished type 1 collagen content (Bonadio et al., 1990). In the more severe forms of 0 1 structurally abnormal collagen chains are incorporated into collagen molecules and, coupled with increased collagen breakdown, lead to impaired tissue collagen function (Wenstrup et al., 1990; Byers, 1990; Kuivaniemi et al., 1991). A relatively common mutation found in 0 1 that leads to destabilization of the collagen triple helix is the substitution by another amino acid of the obligatory glycine residue of the Gly-X-Y triplet repeat sequence of the triple helix. A pattern is emerging that glycine substitutions toward the C-terminus of the al(1) or a2(I) chains are more clinically severe than those toward the N-terminus. This is consistent with the model that since helix formation propagates from the C- to N-terminus, those mutations toward the C-terminus will be more disruptive. Because of the [al(I)],aZ(I) molecular composition it is also predicted that mutations in the a2(I) chain would be less clinically severe than those in the al(1). However, it is also becoming apparent that these generalizations must be modified by considering the nature of the glycine substitution and the surrounding amino acid sequence (Byers, 1990; Bachinger and Davis, 1991) and many more mutations spanning the collagen helical domain will need to be characterized before any strict correlation can be made between the biochemical and clinical phenotypes. In this paper we report the substitution of G l ~ - 4 5 7by~ Arg in the a2(I) chain of type I collagen in a child with the perinatal lethal form of
01.
Cetus Corp., USA. Sequenase sequencing kits were purchased from United States Biochemical Corp., USA. All other chemicals were commercially available analytical grade.
Clinical Summary The patient (0145) was born at 36 weeks gestation and died shortly after birth. Autopsy revealed the beaded ribs and short crumpled long bones typical of lethal perinatal 0 1 type IIA) (Sillence e t al., 1979; Cole, 1988). T h e parents were normal and unrelated. Dermal fibroblast cultures were established from the parents, patient, and controls and maintained as described previously (Bateman et al., 1984, 1986). All tissue biopsies were obtained with informed consent and approval of the Ethics Committee of this hospital.
Collagen Biosynthetic Labelling Confluent fibroblast cultures were biosynthetically labelled as previously described (Bateman et al., 1984, 1986). After culture for three days in growth medium containing 0.25 mM sodium ascorbate, labelling with 10 pCi/ml ~ [ 5 - ~ H ] p r o line was carried out for 18 hr in medium containing 10% (vlv) dialysed fetal calf serum, 0.25 mM sodium ascorbate, and 0.1 mM P-aminopropionitrile fumarate. The cell layer and medium fractions were separated for analysis and the procollagens, precipitated with 25% saturated (NH4),S04, were subjected to limited proteolysis with pepsin (Bateman e t al., 1984, 1986).
SDS-Polyacrylamide Gel Electrophoresis MATERIALS AND METHODS
~-[5-~H]Proline (30 Ci/mmol) and [3ZP]dCTP (3,000 Ci/mmol) were purchased from Amersham Australia Pty. Sydney, NSW, Australia. Dulbecco modified Eagle's medium and fetal calf serum were purchased from Flow Laboratories Australia, Stanmore, NSW, Australia. Pepsin, sodium ascorbate, P-aminopropionitrile fumarate, trypsin, and chymotrypsin were purchased from Sigma Chemical Co., St. Louis, MO. Restriction endonucleases and nick-translation kits were obtained from Boehringer Mannheim, Germany. cDNA synthesis kits and the M13mp8 vector were purchased from Amersham Australia Ltd., Sydney, Australia, and Tag polymerase was obtained from the Perkin Elmer-
'Amino acid positions are numbered by the standard convention in which the first glycine of the triple helical domain of the a-chain is number one.
Collagen a-chains were resolved on 5% (w/v) separating gels containing 2 M urea (Bateman et al., 1984, 1986). Collagens prepared from cell cultures and tissues were also digested with CNBr (Scott and Veis, 1977) and analysed by two-dimensional electrophoresis (Cole and Chan, 1981). The a-chains and peptides were detected by fluorography (Bonner and Laskey, 1974).
Collagen Thermal Stability The freeze-dried [3H]proline-labelled collagens were dissolved in 0.4 M NaCVO.l M Tris-HC1 buffer, pH 7.4, at 4°C. The samples were warmed stepwise (l"C/min) from 34 to 43°C and at 0.5"C intervals samples were taken and digested with a mixture of trypsin and chymotrypsin as described by Bruckner and Prockop ( 1981). The proportion of the collagen that was resistant to proteolysis at each temperature interval was determined by scin-
&(I) GLY457 TO ARG SUBSTITUTIONIN OSTEOGENESIS IMPERFECTA
57
tillation counting of collagen a-chain bands that were resolved by electrophoresis and identified by fluorography (Bateman et al., 1988a). The thermal denaturation temperature, T,, was defined as the temperature at which half of the collagen was degraded. Formation and Chemical Cleavage mRNA:cDNA Heteroduplexes
Total RNA was isolated from confluent fibroblast cultures (Lui et al., 1979; Wake et al., 1985). mRNA:cDNA heteroduplexes were formed in a volume of 50 ~1 containing approximately 5 ng (50,000-100,000 dpm) of labelled cDNA probe, 5-10 Fg of total RNA, 80% (v/v) formamide, 40 mM PIPES, pH 6.4, 1 mM EDTA, and 0.4 M NaC1. The mixture was denatured at 80°C for 5 min, incubated at 60°C for 2 hr, and ethanol precipitated. Chemical modification of mismatched nucleotides with hydroxylamine or osmium tetroxide, cleavage with piperidine, and electrophoresis of the products on denaturing 7 M urea/5% ( w h ) acrylamide gels have been described (Cotton et al., 1988; Dahl et al., 1989; Lamande et al., 1989; Bateman et al., 1989). Collagen protein analyses defined the apparent site of the mutation in the collagen chains (see Fig. 2), and this was used as a guide to the selection of suitable control cDNAs for heteroduplex formation. Control al(1) and a2(I) cDNA probes (Bateman et al., 1991), chosen to span the abnormal region, were purified and end-labelled with [cI-~’P]~CTPby the fill-in reaction using the Klenow fragment of DNA polymerase I (Sambrook et al., 1989). Amplification and Sequencing of cDNA
First-strand cDNA was synthesized from total
RNA using a cDNA synthesis kit primed with oligo(dT). Approximately 50 ng of cDNA was amplified by the PCR (Saiki et al., 1988) through 30 cycles using Taq polymerase (Kogan et al., 1987). Each cycle consisted of denaturation at 92°C for 1.5 min, annealing of primers at 63°C for 1.5 min, and primer extension at 72°C for 3 min (Lamande et al., 1989; Bateman et al., 1989). A 357-bp fragment corresponding to bases 1,522-1,879 of the a2(1) cDNA was amplified with the primers 5 ’-GGAAAAGAAGGTCCTGTC-3’ and 5‘GGAGACCAAACTCACCAT-3’ (De Wet et al., 1987).3 3The base pairs are numbered from the start of transcription of the prouZ(1)mRNA (DeWet et al., 1987).
FIGURE 1. Electrophoresis of
pepsin-digestedfibroblast collagens. Fibroblast cultures were labelled for 18 hr with [3H]prolineand the collagens were pepsin-digested and analysed without reduction on SDS/polyaqlamide gels (5%) (see the Experimental section for details). C, cell layer fraction; M. medium fraction. The migration positions of type 1 collagen al(1)and a2(I)chains and type 111 collagen [al(lll)]3 are shown.
Amplification products of the predicted size were purified, treated with T4 polynucleotide kinase (Sambrook et al., 1989), and cloned into a SmaI cut, dephosphorylated M13mp8 vector. Multiple clones from two independent amplification reactions were sequenced using a Sequenase kit. RESULTS Type I Collagen Protein Analysis
The al(1)-chain of 0145 fibroblast type I collagen showed slow electrophoretic migration compared to control fibroblast al(1)-chains (Fig. 1). This migrational abnormality was also evident in al(1)- and a2(1)-chains of type I collagen extracted from the bone and dermis of the patient and was due to overhydroxylation of lysine (Bateman et al., 1986). The overmodified collagen produced by 0145 fibroblasts was secreted poorly, at 60% the level of control fibroblasts (Fig. 1; Bateman et al., 1986). Electrophoretic analysis of the collagen a-chains produced by the parents’ fibroblasts showed only normally migrating species which were efficiently secreted (data not shown). Two-dimensional electrophoresis of the type I collagen CNBr peptides demonstrated that only the al(I)CB8, al(I)CB3, and the a2(I)CB4 peptides
58
BATEMAN ET AL.
the mRNA derived from the parents (Fig. 3b). Treatment with osmium tetroxide/piperidine did not reveal any T mismatches in 0145 mRNA and no mismatches were detected by hydroxylamine or osmium tetroxide treatment of heteroduplexes formed between 0145 mRNA and control al(1) cDNA (data not shown).
Characterization of the mRNA Mutation
FIGURE 2. Two-dimensionalelectrophoresis of the CNBr peptides of 0145 fibroblast collagen. Fibroblasts were labelled in cell culture and collagens in the medium fraction were digested with CNBr and the products analysed by two-dimensional PAGE (see the Experimental section for details). CNBr peptides are identified as follows: 3.5, a2(1)CB3.5; 4, a2(I)CB4 8. al(l)CB8; 7, al(l)CB7; 6, al(I)CB6 and 3, al(I)CB3. The “tailing” of al(I)CB3, al(I)CB8. and a2(I)CB4due to increased lysine posttranslational modification is indicated by arrows.
showed ‘(tailing” of the proximal edges resulting from the overmodification of the lysine residues (Fig. 2). The C-terminal a l ( 1 ) peptide, al(I)CB6, showed n o evidence of abnormal migration due to overmodification. T h e C-terminal peptide of the a2(I) chain, a2(I)CB3.5, also had an apparently normal migration although the resolution of this large peptide by two-dimensional electrophoresis may not be adequate to enable detection of such subtle changes in electrophoretic migration (Fig. 2). This peptide map result clearly indicated that the overmodification of lysine residues was restricted to the N.termina1 half of the al(1) and a2(I) chains and suggested that the mutation resided in the central region of the triple helix.
Detection of an Abnormal mRNA Sequence by Chemical Cleavage Heteroduplexes were formed using control a1(I) and a2(I) cDNAs, which spanned the central portion of the collagen helix, and mRNA extracted from cultured dermal fibroblasts of 0145 and controls. Hydroxylamine/piperidine treatment cleaved a 220-bp fragment from the 311-bp TUL+NCOI fragment of a2(I) cDNA. This finding demonstrated that there was a mismatched C at approximately base pair 1,773-1,775 of the control a2(I) cDNA (Fig. 3a). This mutation was not present in
To define the mutation, a short length of firststrand d ( I ) cDNA was amplified by the PCR using unique oligonucleotides chosen to span the region containing the mismatch. Sequencing of M13 subclones of the amplified products identified the mutation as a single base substitution at base pair 1,774 which converted the codon for glycine (GGT) to CGT (arginine) at amino acid position 457 of the helix (Fig. 4). The abnormal sequence was confirmed by sequencing of multiple clones from two separate PCR reactions. Clones containing the normal sequence (GGT) were also identified indicating that the patient was heterozygous for the a2(I) mutation. The 0145 mutant clones also contained a G to C change at position 1,780, converting Ala-459 to Pro. This sequence variant was identified in three other individuals (Bateman et al., 1991) and probably represents a neutral sequence variant. Collagen Melting Temperature Estimation of the thermal denaturation temperature (T,) of type I collagen synthesized by 0145 and control fibroblasts was made by measuring its resistance to proteinase digestion after heating to different temperatures (Fig. 5). The T , of 0145 type I collagen was approximately 40.9”C compared with 41.2”C for control type I collagen. Since 0145 is heterozygous this probably represents an underestimate of the reduction in the stability of the mutant type I molecules. The T , of type 111 collagen from 0145 was the same as control values (data not shown).
DISCUSSION The mutation in this case of lethal perinatal 01 was a heterozygous point mutation of G to C at base pair 1,774 of the 012(I) cDNA resulting in the substitution of Gly-457 by Arg. Since this point mutation occurs within the sequence encoded by exon 28, it must represent a point mutation in one allele of COLlA2. mRNA from both parents did not contain the mutation suggesting that the base change arose as a sporadic new mutation in the patient. Somatic and gonadal mosaicism, which
&(I) GLY-457TO ARG SUBSTITUTION IN OSTEOGENESIS IMPERFECTA
59
FIGURE 3. Mutation detection by chemical modification and cleavage. Total RNA extracted from 0145 and control fibroblasts (a) and 0145 and parent’s fibroblasts (b)were hybridized to a 311-bp TaqI-XhoI 32P-labelledfragment of the uZ(1) cDNA and reacted with hydroxylamine for 0 to 60 min as indicated, treated with piperidine. and the resultant frag-
ments resolved on a 5% (w/v) polyacrylamide gel (see the Experimental section for details). The labelled a2(1) cDNA fragment produced by cleavage a t mismatched Cs by hydroxylamine (220 bp) and the uncleaved probe (311 bp) are indicated. The molecular weight standards used were 32Plabelled Haelll fragments of $X174 DNA.
has been described in the parents of several cases of 0 1 (Constantinou et al., 1989; Wallis et al., 1990a; Cohn et al., 1990), was not excluded as the source of the mutation in 0145. Increased levels of lysine modification in the al(1)and a2(1)chains have been characterized in collagen extracted from bone and dermis in 0145 (Bateman et al., 1986). Two-dimensional CNBr peptide mapping studies demonstrated that the overmodified residues were restricted to a1 (I) and a2(I)chains regions toward the N-terminus of the helix from the mutation. Tha charge-change was not able to be directly demonstrated at the protein level by two-dimensional peptide mapping because of the lack of resolution of the multiple charged forms of the large a2(I)CB3.5 peptide. In the case of Gly to Arg mutations in the al(1) chain the
well-resolved oll (I) CB peptides have allowed the direct detection of the mutation at the protein level (Bateman et al., 1987a,b). The correspondence of the regional localization of overhydroxylation with the position of mutation in 0145 and in all other cases studied by us (Bateman et al., 1991) has confirmed that the demonstration of the overmodified domains protein is a reliable indicator of the location of the primary mutation. These data and the detection of a small reduction in the thermal stability of type I collagen in 0145 support the model that this glycine substitution both destabilises the triple helix and disrupts helix propagation from the C- to Nterminus, exposing the regions toward the N-terminus of the helix to excessive enzymatic posttranslational modification.
60
BATEMAN ET AL.
FIGURE4. DNA sequence of mutant and normal cDNA clones.
cDNA clones of the PCR-amplified mRNA region containing the mismatched base were sequenced (see the Experimental section for details). The 0145 mutant clone contains a G to C substitutionat base position 1774 converting Gly-457 to arginine. In mutant clones base position 1,780 was C, so that amino acid 459 was proline rather than alanine (De Wet et al., 1987). Nucleotides different to the published sequence are boxed.
Tissue analysis of 0145 dermis and bone showed that, in common with other cases of lethal perinatal 01, there was a marked reduction in tissue collagen content (Bateman et al., 1986). In 0145, tissue collagen levels were approximately half that of control values. Fibroblast collagen secretion was 60% of control values (Bateman et al., 1986) and it is likely that there was also increased degradation of the mutant collagen as determined in other cases of lethal perinatal 0 1 (Bateman et al., 1984). This increased collagen degradation results, in part, from the incorporation of the mutant proa2(I) into type I procollagen molecules, disrupting helix formation and stability and leading to “protein suicide” (Prockop and Kivirikko, 1984). However, some mutant collagen was secreted by the fibroblasts and the presence of this abnormal collagen in the extracellular matrix dramatically amplified the deleterious effect of the mutation.
T h e secretion of mutant collagens into the extracellular matrix seems to be the determining factor for severe 01 phenotypes, since when collagen production was reduced to 50% without the production and secretion of mutant collagen, a milder 0 1 type I results (Bonadio et al., 1990). The mechanism whereby the mutant species has such a catastrophic effect on bone matrix organisation is currently unknown. However, in one case studied in detail, the mutation causes a conformational change to the collagen triple helix which delays fibril formation and alters fibril morphology (Kuivaniemi et al., 1991). Such alterations to collagen fibril architecture may interfere with the normal bone mineralisation process. With an increasing number of 0 1 mutations now defined a model relating the position and type of the glycine substitution mutation to the clinical severity has emerged (Byers, 1990). This is best documented for the al(1) chain and suggests that point mutations toward the C-terminus of the helix will, in general, lead to more severe clinical consequences. However, this simple model must be modified t o consider the sequence context and nature of the substitution. The position-specific effect of the mutations seen in 0 1 may result from disturbances to specific functional sequences involved in collagen-collagen, collagen-matrix, or collagen-cell interactions. The local sequence may also play a critical role in how the collagen triple helix is affected by the mutation by providing domains of relatively low or high helix stability (Bachinger and Davis, 1991) defined largely by the presence or absence of stabilizing hydroxyproline residues (Burjanadze, 1979). This model proposes that mutations in regions of low local stability will be less deleterious to helical stability. In addition, if sequences aminoterminal to the mutation are of high stability renucleation of helix formation will occur readily and the interruption to helix propagation caused by the mutation will be minimal (Bachinger and Davis, 1991). In 0145 the Gly-457 to Arg occurs in a “loose” part of the helix, and taken with the small apparent reduction in the T, of 0.3”C, compared with T, reductions of more than 1°C in other cases of lethal perinatal01 (Baker et al., 1989; Kuivaniemi et al., 1991), suggests that this mutation should have a relatively mild effect on helix stability. However the lack of an adjacent stabilising aminoterminal nucleation site may lead to a more pronounced effect of the mutation on helix propagation.
&(I) GLY-457 TO ARG SUBSTITUTION IN OSTEOGENESIS IMPERFECTA
61
-
O 903
80-
3 70-60C 3 500 u
.-m 2 0
g
-E 0 2 0
n
‘0-0
O\
\ o
t\
40- 0-OControl 30-- 0-001 45 2010-
0T
4
FIGURE
5. Thermal denaturation of type I collagen. (3H]Proline-labelledcollagens were prepared from the medium of 0145 and control fibroblasts. They were treated with pepsin and after inactivation of the pepsin the collagens were warmed slowly from 34 to 43°C. Serial samples were taken and digested with trypsidchymotrypsin and the resistant
chains were resolved by electrophoresisand the melting temperature of the helix was determined by the relative proportion of the a-chains that were resistant to proteolysis (see the Experimental section for details). (4Type I collagen from 0145 [a2(1)Gly-457 to Arg); (0)type I collagen from control fibroblasts.
The predominant factor in generating the lethal phenotype in 0145 may be the nature of the substituting amino acid, arginine. Glycine to arginine mutations have been characterized in the al(1) chain at amino acid positions 154 (Starman et al., 1989), 391 (Bateman et al., 1987a), 550 (Byers, 1990), 667 (Bateman et al., 198713, 1988b), 847 (Wallis et al., 1990b), and 976 (Lamande et al., 1989). All of these lead to the lethal form of 0 1 except the most amino terminal mutation at Gly154 which results in 0 1 type 111. These data suggest that substitutions of al(1) glycine by the bulkier arginine are extremely deleterious along the majority of the helix leading to the lethal phenotype. In the other reported a2(I) Gly to Arg substitution, Gly-1012 to Arg leads to 0 1 type IV (Wenstrup et al., 1988). However, this may represent a special case since the mutation is in the last helical triplet and may not, therefore, have the same effect on helix stability and/or propagation. It has also been suggested that a2(I) glycine substitutions may be less deleterious than corresponding mutations in the al(1) chain because of the [c~l(I)]~a2(1) stoichiometry of the type I collagen molecule (Vuorio and De Crombrugghe, 1990). However, accumulating data demonstrate that a2(I) substitutions of Gly-457 by Arg, Gly586 by Val (Lamande et al., 1989) and five cases of Gly to Asp in the C-terminal third of the helix (Byers, 1990) lead to the lethal form of 01. This suggests that as further 012(I) mutations are characterized a similar spectrum of clinical severity to
that seen in al(1) will be displayed and that the expression of clinical phenotype will depend not only on the type, position, and sequence surrounding the collagen mutation, but also on how the abnormality affects the complexities of collagen biosynthesis, fibril deposition, and structural and regulatory interactions within the extracellular matrix.
ACKNOWLEDGMENTS This work was supported by grants from the National Health and Medical Research Council of Australia and the Royal Children’s Hospital Research Foundation.
REFERENCES Bachinger H-P, Davis ]M (1991) Sequence specific thermal stability of the collagen triple helix. Int J Macromolec 13:152156. Baker AT, Rarnshaw JAM, Chan D, Cole WG, Bateman ]F (1989) Changes in collagen stability and folding in lethal pennatal osteogenesis imperfecta. The effect of al(l) glycine to arginine mutations. Biochem J 261:253-257. Bateman JF, Mascara T, Chan D, Cole, WG (1984) Abnormal type I collagen metabolism by cultured fibroblasts in lethal perinatal osteogenesis imperfecta. Biochem J 2 17:103-1 15. Bateman JF, Chan D, Mascara T, Rogers JG, Cole WG (1986) Collagen defects in lethal perinatal osteogenesis imperfecta. Biochem J 240:699-708. Bateman JF, Chan D, Walker ID, Rogers JG, Cole WG (1987a) Lethal perinatal osteogenesis imperfecta due to the substitution of arginine for glycine at residue 391 of the al(1)-chain of type 1 collagen. J Biol Chern 262:7021-7027. Bateman JF, Mascara T, Chan D, Cole WG (1987b) A structural
62
BATEMAN ET AL.
mutation of the collagen al(1) CB7 peptide in lethal perinatal osteogenesis imperfecta. J Biol Chem 262:4445-4451. Bateman IF, Harley V, Chan D, Cole WG (1988a) Comprehensive analysis of collagen metabolism in vitro using 3H/14Cproline dual-labeling and polyacrylamide gel electrophoresis. Anal Biochem 168:171-176. Bateman IF, Lamande SR, Dahl H-HM, Chan D, Cole W G (1988b) Substitution of arginine for glycine 664 in the collagen u l (I) chain in lethal perinatal osteogenesis imperfecta: demonstration of the peptide defect by in vitro expression of mutant cDNA. J Biol Chem 263: 11627-1 1630. Bateman JF, Lamande SR, Dahl H-HM, Chan D, Mascara T, Cole WG (1989) A frameshift mutation results in a truncated nonfunctional carboxyterminal proal(1) propeptide of type 1 procollagen in osteogenesis imperfecta. J Biol Chem 264: 10960- 10964. Bateman JF, Lamande SR, Hannagan M, Moeller I, Dahl H-HM, Cole WG (1991) A chemical cleavage method for the detection of RNA base changes: Experience in the application to collagen mutations in osteogenesis imperfecta. Am J Med Genet, in press. Bonadio J , Saunders TL, Tsai E, Goldstein SA, Morris-Wiman J , Brinkley L, Dolan DF, Altschuler RA, Hawkins JE, Bateman IF, Mascara T,Jaenisch R (1990) A transgenic mouse model of the mild dominant form of osteogenesis imperfecta. Proc Natl Acad Sci USA 87:7145-7149. Bonner WM, Laskey RA (1974) A film detection method for tritium-labelled proteins and nucleic acids in polyacrylamide gels. Eur J Biochem 46:83-88. Bruckner P, Prockop DJ (1981) Proteolytic enzymes as probes for the triple helical corformation of procollagen. Anal Biochem 110:360-368. Burjanadze TV (1979) Hydroxyproline content and location in relation to collagen thermal stability. Biopolymers 18:93 1938. Byers PH (1990) Brittle bones-fragile molecules: disorders of collagen gene structure and expression. Trends Genet 6:293 -300. Cohn DH, Starman BJ, Blumberg B, Byers PH (1990) Recurrence of lethal osteogenesis imperfecta due to parental mosaicism for a dominant mutation in a human type I collagen gene (COLIAI). Am J Hum Genet 46:1034-1040. Cole WG, Chan D (1981) Analysis of the heterogeneity of human collagens by two-dimensional polyacrylamide-gel electrophoresis. Biochem J 197:377-383. Cole WG (1988) Osteogenesis imperfecta. Bailliere's Clin Endocrinol Metab 2:243-265. Constantinou CD, Nielsen KB, Prockop DJ (1989) A lethal variant of osteogenesis imperfecta has a single base mutation that substitutes cysteine for glycine 904 of the ul(1) chain of type I procollagen. The asymptomatic mother has an unidentified mutation producing an overmodified and unstable type I procollagen. J Clin Invest 83:574-584. Cotton RGH, Rodrigues NR, Campbell RD (1988) Reactivity of cytosine and thymine in single-base-pair mismatches with hydroxylamine and osmium tetroxide and its application to the study of mutations Proc Natl Acad Sci USA 85:4397-4401. Dahl H-HM, Lamande SR, Cotton RGH, Bateman JF (1989) Detection and localization of base changes in RNA using a chemical cleavage method Anal Biochem 183:263-268. De Wet W, Bernard M, Benson-Chanda V, Chu M-L, Dickson L,
Weil D, Ramirez F (1987) Organization of the human prou2(1) collagen gene. J Biol Chem 26216032-16036. Kogan SC, Doherty M, Gitschier J (1987) An improved method for prenatal diagnosis of genetic diseases by analysis of amplified DNA sequences: Application in hemophilia A. N Engl J Med 317:985-990. Kuivaniemi H, Tromp G , Prockop DJ (1991) Mutations in collagen genes: causes of rare and some common diseases in humans. FASEB J 5:2052-2060. Lamande SR, Dahl H-HM, Cole WG, Bateman JF (1989) Characterization of point mutations in the collagen COLlAl and COLl A2 genes causing lethal perinatal osteogenesis imperfecta. J Biol Chem 264:15809-15812. Lui CP, Slate OL, Gravel R, Ruddle FH (1979) Biological detection of specific mRNA sequences by microinjeciton. Proc Natl Acad Sci USA 76:4503-4506. Prockop DJ, Kivirikko KI (1984) Heritable diseases of collagen. N Engl J Med 311:376-386. Saiki RK, Gelfand DH, Stoffel S,Scharf SJ, Higuchi R, Horn GT, Mullis KB, Erlich HA (1988) Primer-directed enzymatic amplification of DNA with a thermostable DNA polymerase. Science 239:487-491. Sambrook J, Fritsch EF, Maniatis T (1989) Molecular Cloning, a Laboratory Manual, 2nd ed. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory. Scott PG, Veis A (1976) The cyanogen bromide peptides of bovine soluble and insoluble collagens. Connect Tissue Res 4: 107-116. Sillence DO, Senn AS, Danks DM (1979) Genetic heterogeneity in osteogenesis imperfecta. J Med Genet 16:lOl-116. Starman BJ, Eyre D, Charbonneau H, Hanylock M, Weis MA, Weiss L, Graham JM, Byers PH (1989) Osteogenesis imperfecta. The position of the substitution for glycine by cysteine in the triple helical domain of the proul(1) chains of type I procollagen determines the clinical phenotype. J Clin Invest 84:1206-1214. Vuorio E, De Crombrugghe B (1990) The family of collagen genes. Annu Rev Biochem 59:837-872. Wake SA, Mercer JFB (1985) Induction of metallothionein mRNA in rat liver and kidney after copper chloride injection. Biochem J 228:425-432. Wallis GA, Starman BJ, Zinn AB, Byers PH (1990a) Variable expression of osteogenesis imperfecta in a nuclear family is explained by somatic mosaicism for a lethal point mutation in the a l ( I ) gene (COLlAl) of type I collagen in a parent. Am J Hum Genet 46:1034-1040. Wallis GA, Starman BJ, Schwartz MF, Byers PH (1990b) Substitution of arginine for glycine at position 847 in the triple helical domain of the u l ( l ) chain of type I collagen produces lethal osteogenesis imperfecta. Molecules that contain one or two abnormal chains differ in stability and secretion. J Biol Chem 265:18628-18633. Wenstrup RJ, Cohn D H, Cohen T , Byers PH (1988) Arginine for glycine substitution in the triple helical domain of the products of the a2(1) collagen allele (COLLAZ) produces the osteogenesis imperfecta type IV phenotype. J Biol Chem 263:77347740. Wenstrup RJ, Willing MC, Starman BJ, Byers PH (1990) Distinct biochemical phenotypes predict clinical severity in non-lethal variants of osteogenesis imperfecta. Am J Hum Genet 46:975982.
