Atherosclerosis 141 Suppl. 1 (1998) S17 – S24
Molecular modelling and the biosynthesis of apolipoprotein B containing lipoproteins James Scott a,*, Naveenan Navaratnam b, Charles Carter c a
National Heart and Lung Institute, Imperial College School of Medicine, Hammersmith Hospital, London W12 ONN, UK b MRC Molecular Medicine Group, Collier Building, Hammersmith Hospital, Du-Cane Road, London W12 ONN, UK c Department of Biochemistry and Biophysics, CB 7260, Uni6ersity of North Caroline at Chapel Hill, Chapel Hill, NC 27599 -7260, USA
Abstract APOBEC-1 is the cytidine deaminase. We show by sequence alignment, molecular modelling and mutagenesis, that it is related in crystal structure to the cytidine deaminase of Escherichia coli (ECCDA). The two enzymes are both homodimers with composite active sites formed with loops from each monomer. In the sequence of APOBEC-1, three gaps compared to ECCDA match the size and contour of the minimal RNA substrate. We propose a model in which the asymmetric binding of one active site to the substrate cytidine which is positioned by the downstream binding of the product uridine and that this helps to target the other active site for deamination. © 1998 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Sequence alignment; Molecular modelling; Mutagenesis
1. Introduction Placental mammals use the two forms of apo B to transport cholesterol and triglyceride in the blood [1]. Full-length apo B-100 (512 kDa) is made in the liver and transports endogenously synthesized cholesterol and triglyceride in the circulation [2,3]. A shorter form, apo B-48 (241 kDa), is generated in intestinal absorptive cells by the tissue-specific editing of apo B-100 RNA and is used for dietary lipid absorption [4 –7]. The catalytic subunit of the apo B RNA editing enzyme (designated APOBEC-1 for apo B RNA editing cytidine deaminase subunit 1) is a 27 kDa member of the cytidine deaminase family of enzymes that act on monomeric nucleoside and nucleotide substrates [8–12]. APOBEC-1 on its own is not sufficient for RNA editing, but acts in concert with other auxiliary proteins The cytidine deaminase family includes the Escherichia coli cytidine deaminase (ECCDA). Crystal * Corresponding author. Tel.: +44-181-3838823; fax: + 44-1813832028; e-mail:
[email protected].
structures of ECCDA [10] complexed with various inhibitors [13,14] have been established. ECCDA is a homodimer of identical 31.5 kDa subunits. Each monomer is composed of a small, amino-terminal ahelical domain, and two larger core domains. The two ECCDA core domains have nearly identical tertiary structure, but little apparent amino acid sequence homology. They are connected by an extended chain, running the full length of the molecule from the outside of the first core domain to the outside of the second. Catalytic activity in ECCDA derives from a cluster of residues in the amino-terminal core domain, which bind zinc and activate a zinc-bound water molecule [10]. The two active sites of the homodimer are formed across the subunit interface. Zinc-binding, catalytic, and pyrimidine-binding sites derive from the amino-terminal core domain of one subunit. The rest of the active site comes from a homologous region in the carboxy-terminal core domain of the other subunit, which was described as a ‘pseudoactive site’ because of the structural homology to the active site.
0021-9150/98/$ - see front matter © 1998 Elsevier Science Ireland Ltd. All rights reserved. PII: S 0 0 2 1 - 9 1 5 0 ( 9 8 ) 0 0 2 1 3 - 5
S18
J. Scott et al. / Atherosclerosis 141 Suppl. 1 (1998) S17 – S24
The amino acid sequence and spacing of catalytic residues are conserved across the entire family of cytidine deaminases, including APOBEC-1 [8,10 – 12]. Presumably, the catalytic mechanism for deamination is also conserved [10,15 – 19]. However, the quaternary organisation of the deaminases allows them to be divided into two groups. The ECCDA and APOBEC-1 monomers are of similar molecular size. Like ECCDA, APOBEC-1 forms a homodimer [10,20], suggesting a closer phylogenetic and structural relationship between these two enzymes than between either of them and the group of smaller deaminases of 15 – 18 kDa, which lack the carboxy-terminal core domain found in ECCDA and, which is also apparently present in APOBEC-1 [10–12]. Sequences of the latter class of deaminases are homologous to the amino terminal core domain in ECCDA and they form homotetramers, each subunit contributing an apparently intact active site. Determinants for RNA editing are contained in a highly conserved, 26 – 30 nucleotide sequence, which can confer editing on other genes [21 – 25]. This sequence consists of six nucleotides upstream of the edited C in which most mutations enhance editing, and a mooring sequence at a fixed distance downstream in which most alterations reduce or abolish editing, which has been proposed to bind auxiliary editing factors [7,25 –30]. Mutagenesis to the RNA substrate and active site of APOBEC-1 in conjunction with UV cross-linking and competition studies indicate that APOBEC-1 binds through its active site to a U in an AU-rich sequence at the 3% end of the mooring sequence [17– 19,22,28]. This finding was especially provocative in light of the fact that APOBEC-1 can form a homodimer, suggesting that one active site in the dimer might bind to a U downstream of the edited C, and that this interaction could help position the C at nucleotide 6666 into the active site on the other monomer [6]. The structure of APOBEC-1 is as yet unknown. We show here how the major functional differences between ECCDA and APOBEC-1 can be specifically related to the creation in APOBEC-1 of a large cavity capable of binding an RNA tertiary structure of sufficient complexity to endow it with the requisite binding specificity for RNA editing. This new model is supported by extensive mutational analysis. It is consistent with our original hypothesis that the catalytically active form of APOBEC-1 is an asymmetric homodimer, one site of which is bound to a product, the downstream U, and that this supplemental recognition site helps insure that the appropriate substrate C is deaminated by the other active site.
2. Experimental Described in Ref. [31].
3. Results An initial alignment was obtained using the CLUSTAL computer program with the PAM250 residue weight table [32]. This alignment superimposed the active-site residues, as expected. However, as noted below, it also identified other regions of potential structural homology (Fig. 1). Long gaps (gap-1 and gap-2) in each core domain were evident in the initial alignment. Introducing a third, 18-residue gap (ECCDA 10-27, designated here as gap-0) shortened gap-1 and improved the sequence identity from three to six residues in the amino terminal domain, giving the alignment in Fig. 1. Gap-1 (23 residues ECCDA 79–101) immediately precedes the first zinc ligand H61, (ECCDA residue H102), whereas gap-2 (31 residues, ECCDA 190–220) begins immediately after the linker peptide, and hence deletes a different portion of the core domain than does gap-1. A conserved insertion of five residues precedes the second cluster of zinc ligands (APOBEC-1 residues 84SITWF88 between ECCDA and residues 123 and 124).
4. The gaps, substrate specificity, and the APOBEC-1 model Catalysis by ECCDA depends on binding of the enzyme to both the 3% and 5% hydroxyl groups of the ribose, which is completely sequestered from solvent [10], whereas APOBEC-1 must select a single cytidine from a specific RNA sequence. The contribution of the ribose to catalysis by ECCDA is expressed largely as an enhancement of kcat, implying that ribose binding contributes to the mechanism of transition-state stabilisation. The RNA backbone could itself provide APOBEC-1 with the mechanistic stabilisation given in ECCDA by enclosing the ribose itself. An important implication for our model is that the phosphodiester backbone of the RNA substrate bound to APOBEC-1 must fix the ribose of the edited C in a location close to that of the ribose in the ECCDA ligand complexes. A key criterion for the model was, therefore, that the principal differences between the two structures be related to their respective substrate specificities. What discriminates the dimeric ECCDA from the homologous tetrameric cytidine deaminases is that it has a composite active site, constructed with contributions from both monomers. Its two active sites are formed across the subunit interface, combining the catalytic and pyrimidine nucleoside-binding activities
J. Scott et al. / Atherosclerosis 141 Suppl. 1 (1998) S17 – S24
S19
Fig. 1. Comparison of primary sequence for ECCDA and APOBEC-1. (a) Alignment of amino acid sequences of ECCDA and APOBEC-1 was obtained with the CLUSTAL complex program. Manual adjustment created gap-0 and fore-shortened gap 1. The consensus sequence of APOBEC-1 was derived for human, rabbit, rat, mouse (data from Gene Bank numbers, respectively: L26234; U10695; L07114; U22264). Identities between APOBEC-1 and ECCDA are shaded. The domain organization of ECCDA (shown schematically in b) is superimposed on the aligned sequence of APOBEC-1, and corresponds to the amino terminal a-helical domain, active-site core domain, linker segment, and carboxy-terminal core domain (thin boundary lines). Catalytic residues conserved in all cytidine deaminases are enclosed by thick lines. The sequences are collinear except for four regions. Three gaps (0, 1, and 2) and a conserved insertion occur in the APOBEC-1 sequences. Leucine residues conserved in APOBEC-1 but not in ECCDA are indicated by asterisks. Leucine 218 is present in rat APOBEC-1, but is generally isoleucine in other species. (b) Line diagram showing the location of the three gaps, the insert in APOBEC-1 superimposed on the domain structure of ECCDA. Deletion and insertion mutants are shown in (c).
from the amino terminal core domain and contributions from the carboxy-terminal core domain from the other subunit, which cover 20% of the surface area of the ligand bound to the opposite monomer (Fig. 2). The structures and precise locations of gaps-1 and 2 in the ECCDA monomer suggest that this composite active site construction was instrumental in its evolutionary adaptation to a specific RNA substrate and can account for much of what is known about APOBEC-1. Gap-1 coincides with a loop (residues 79 – 101) that in ECCDA specifically recognises the 3%-hydroxyl group through residues Asn 89 and Glu 91, and is thus responsible for enveloping the ribose [10]. A significant portion of this loop would have to be absent or assume a different conformation to position any pyrimidine
base from an RNA substrate into the ECCDA active site. Gap-2 lies towards one end of the ECCDA monomer, away from the active site and dimer interface. The space left by removing gap-2, therefore, must be repositioned closer to the active site to use it in accommodating the RNA substrate. Removal of gaps-1 and 2 suggests adjustment of the remaining ECCDA structure to rejoin the loose ends. Loss of gap-1 leaves residues 78 and 102 (APOBEC-1 ˚ apart. A natural way to residues 60 and 61) 21.5 A rejoin these loose ends is to move the b-strand between residues 72 and 78 (APOBEC-1 residues 54–60) from its location in ECCDA, creating an anti-parallel b-loop at the surface, perhaps bounding the active site by wrapping round helical tertiary structure in the RNA substrate.
S20
J. Scott et al. / Atherosclerosis 141 Suppl. 1 (1998) S17 – S24
Fig. 2. E. coli cytidine deaminase provides a molecular model for apo B RNA Editing and a mechanism for substrate recognition. The separate monomers for E. coli cytidine deaminase and for the molecular model of APOBEC-1 are shown in dark blue and red. For each of the respective monomers gap-1 and gap-2 are shown in green for the blue monomer and yellow for the red monomer. The active site inhibitor, zebularine, is shown in the composite active site, formed by contributions from each of the monomers. In the APOBEC-1 model, the zinc ligands are shown in dark blue, together with co-ordinating amino-acid residues. In this model for APOBEC-1, gaps-1 and 2 have been removed from each monomer and form peptide mimic for the RNA substrate. Substrate C and product U binding are demonstrated in a 5% to 3% (see Fig. 1 and text).
The end left by gap-2 at residue 188 (APOBEC-1 residue 153) lies at the surface and can be rejoined to residue 221 (APOBEC-1 residue 154), by a modest rearrangement of the linker segment, without substantially disrupting the rest of the structure. The loss of gap-2 cleaves the carboxy-terminal core domain into two ‘split pieces’ by removing an internal b-strand (ECCDA residues 209 – 220) from the fourstranded b-sheet. The first split piece corresponding to ECCDA residues 220 – 249 (APOBEC-1 residues 154– 182) contributes to the active site on the opposite monomer through an unusual hydrogen-bond from the backbone carbonyl of residue A232 (APOBEC-1 residue 165) to the zinc ligand, H102 (APOBEC-1 residue H61) [10]. It is likely that this region remains essentially unchanged by removal of gap-2, to preserve the integrity of the active site. The bab crossover connection formed by carboxy-terminal residues 258– 294 (APOBEC-1 residues 191 – 227) can be repositioned readily. Its initial b-strand would replace the one removed with gap-2 forming a three-stranded b-sheet. Explicit precedent for this strand removal/insertion can
be found in the b-strand insertion in serine protease inhibitors following proteolysis [33]. The helix, residues 272–283 (APOBEC-1 residues 205–216), forms part of the dimer interface that covers the active site in ECCDA. Rolling these two helices back, away from the interface would open a large, deep, and continuous channel in the APOBEC-1 dimer, exposing the two active sites and effectively combining the space vacated by gap-2 and gap-1. This significant reshaping of the carboxy-terminal core domain in the APOBEC-1 monomer would provide access for the RNA substrate to both active sites of the APOBEC-1 dimer in our model.
5. Evaluation of the model The main features of the APOBEC-1 structure suggested by this homology model are the following: (1) the sequence alignment (Fig. 1) indicates that most aspects of the ECCDA tertiary structure could be preserved in APOBEC-1. (2) The dimeric structure and
J. Scott et al. / Atherosclerosis 141 Suppl. 1 (1998) S17 – S24
S21
Table 1 Summary of assay resultsa Interaction
b-gal activityb (wt.%)
Cross-linking
Editing
Wild type
Yes
100
Yes
Yes
Deletions N-7 N-10 N-14 1-224
Yes No No No
87 12 11 11
Yes No No No
Yes No No No
Insertions Gap-1 Gap-2
Yes No
80 15
Weak No
No No
97 103 88 44 34 78 40 90 80 10 92 11 95 67 51 92 80 70 73 60 41 65 70 79 19 92 94
Yes No No Noc Noc Yesc Noc Yes Yes No Yes No Yes Yes No No Yes No Yes Yes No Yes Yes Yes No Yes Yes
Yes No No Noc Noc Yesc Noc No Yes No Yes No Yes Yes No No Yes No Yes Yes No Yes Yes Yes No Yes Yes
Mutant
Point mutations R15 R16 R17 H61A E63A V64L C93A C96A D134A L135F R154H F156L A165S P168S P171S L173F L177F L180F L182F L187F L189F L193F L203F L210F L218F L223F L228F
ECCDA residue
L33 T34 G35 H102 E104 Q105 C129 C132 D169 L170 R221 F223 A232 P235 P238 L240 L244 L247 L249 Y254 D256 A260 L270 S277 C285 R290 –
Yes Yes Yes Weak Weak Yes Weak Yes Yes No Yes No Yes Yes Weak Yes Yes Yes Yes Yes Weak Yes Yes Yes No Yes Yes
a
The results of the homodimerization assay, UV-crosslinking assay and editing are shown for wild-type and deletion, insertion and point mutations. Interaction was demonstrated by the yeast two hybrid system. Interaction is shown quantitatively by the b-galactosidase assay. UV-crosslinking was either absent or present as was RNA editing. b Previously reported [19]. c Mean of three or more assays.
sequence homology to ECCDA imply that APOBEC-1 has composite active sites, and hence that homodimerization is crucial for activity. (3) The carboxy-terminal crossover connection is repositioned away from the dimer interface, and this modification combines the spaces vacated by gaps-1 and 2, opening the active site to accept a large RNA substrate. To test these aspects of the model, we have used assays for dimerization, RNA binding, and RNA editing to evaluate the effects of deletions, insertions, and site-directed mutagenesis of residues identical in the two sequences (Table 1). In summary, locations where the homology model preserves the dimerization contacts in ECCDA are all
sensitive to mutation, whereas an important part of the interface in ECCDA predicted not to be involved in dimer formation in the homology model is unaffected by mutation. All of these mutations affecting the dimer interfaces in both the ECCDA and in the APOBEC-1 homology model structures, therefore, have the expected behaviour.
6. A peptide mimic for the RNA substrate Together, the model and mutational analysis suggest that the RNA binding site is created at the dimer
S22
J. Scott et al. / Atherosclerosis 141 Suppl. 1 (1998) S17 – S24
interface by the loss of the gap peptides from ECCDA. This binding site, bounded across the dimer interface by the helix 206–216 is predicted to be reminiscent of the peptide-binding groove in the MHC molecules [34]. Previous studies have identified a highly conserved, 26 – 30 base region of apo B RNA that is necessary for RNA editing in vivo and in vitro [7,21 – 30]. This segment contains the edited C, and the downstream, AUrich RNA binding site for APOBEC-1. Both deamination and RNA binding require active site residues [17–19]. These results imply that the homodimeric APOBEC-1 utilises the active site in one monomer for AU-rich RNA binding through U, which is the deamination product, and that this binding somehow positions the targeted C for editing at the active site of the other monomer [6]. The distance between the ˚ , suggests two active sites in an ECCDA dimer, 21 A that in order for this to happen, the substrate segment must assume a tertiary structure that exposes both U and C bases to the active site at this distance apart. The predicted crevice formed in the APOBEC-1 model by removal of the two gaps and the refolding of the carboxy-terminal core domain consists of three interconnecting channels. The active site crevice created between the monomers by removing gap-2 is accompanied by two smaller tunnels to either side in the spaces left by gap-1. The volume and dimensions of this crevice plus the two tunnels are a remarkable match to those anticipated for the RNA substrate. The combined molecular weights of the four deleted and one inserted peptides (108−10= 98 residues/dimer) in APOBEC-1 is 10 kDa, whereas the corresponding weight for a 26-base RNA structure is also 10 kDa. The resulting active site channel has the rough dimensions 12×14× ˚ , and appears to be elliptical. The two smaller 45 A ˚ in diameter, are positioned tunnels, approximately 6 A to either side. Although we have not considered gap-0 thus far, it should be noted that this gap is nearly continuous with gap-1 in ECCDA, and could extend the gap-1 tunnels into the amino-terminal helical domain. The matching volume and dimensions suggest, in turn, that the four core domain gap peptides might be reassembled to resemble the shape of the minimal RNA substrate and its interaction with dimeric APOBEC-1. This structure fits snugly into the crevices of the APOBEC-1 model and would use the two pyrimidine ligands in ECCDA and their proximity to gap-1 to represent the two bases, U and C, bound to the two active sites of APOBEC-1. As noted above, gap-1 peptide includes residues 89 and 91 in ECCDA which make hydrogen-bonds to the 3% and 5% hydroxyl groups of the ribose [10], and could thus represent single-stranded RNA segments near the entry and exit of the substrate. The major channel between monomers appears large enough to accommodate a complex, multi-stranded
RNA tertiary structure whose purpose would be to position the two crucial bases, C6666 and a uridine in the neighbourhood of U6680, into the two active sites.
7. Discussion The main features of our model are: (1) conservation of ECCDA tertiary structure within domain 1 and much of domain 2; (2) remodeling in nonessential regions of both domains to accommodate the gaps in APOBEC-1 compared to ECCDA and provide access of the RNA substrate to the active sites; (3) conservation of a dimeric quaternary structure and exploitation of the composite dimeric active site organisation for specific substrate recognition. To test the model, a broad and representative sample of APOBEC-1 mutants suggested by the alignment and modeling were examined using biochemical assays for homodimerization, RNA binding, and RNA editing. Our mutagenesis strategy was designed to encompass the established features of the ECCDA structure, namely, catalytic residues and the structure of the active site, the domain organisation of the monomer, the configuration of the dimer, as well as the evident differences between APOBEC-1 and ECCDA, that is the gaps and the leucine-rich region. The analysis demonstrates that RNA substrate recognition by APOBEC-1 requires a homodimer, and provides detailed support in editosomal complex assembly for the model. The role of the auxiliary editing proteins and the order of assembly process remains to be established. There is a fundamental asymmetry to the proposed RNA recognition, which depends on binding a product to one of the two active sites while the other site catalyzes deamination of the substrate. Recent studies of ECCDA raise our curiosity about possibilities inherent in that asymmetry. The two ECCDA monomers are related by crystallographic symmetry in most of the crystals examined [10] and are therefore, presumably, functionally equivalent. However, analysis of the structural reaction profile of ECCDA-catalyzed deamination [10,14] has now demonstrated that the product and a substrate analog bind quite differently to the ECCDA active site, owing to the fact that the 4-keto oxygen of uridine interacts directly with the zinc [13], whereas the 4-NH2 group of the substrate analog deazacytidine does not [14]. Moreover, a new ECCDA crystal structure prepared at close to physiological temperatures revealed an asymmetric dimer in the crystallographic asymmetric unit [35]. The structural differentiation between substrate and product by ECCDA and the evidence for asymmetry suggest that the two ECCDA active sites may alternate between asymmetric dimer conformations during catalysis. This alternation would occur whenever one active
J. Scott et al. / Atherosclerosis 141 Suppl. 1 (1998) S17 – S24
site bound a product, uridine, while the other bound a substrate, cytidine, as in our model for RNA substrate recognition by APOBEC-1. Presumeably, deamination of nucleosides by ECCDA preserves two-fold symmetry only in time, cycling the two active sites through substrate and product binding. The detailed similarity documented in our homology model suggests that this broken symmetry of the ECCDA dimer may also be conserved in APOBEC-1. If so, then the downstream U might also serve as an allosteric effector, complementing its role in substrate recognition by signaling to the opposite monomer and activating it to catalyze deamination. Despite extensive mutagenesis, a specific U has not been identified as an essential requirement for RNA editing and UV crosslinking [7,17 – 19,22,26 – 28,30]. Rather, APOBEC-1 can probably bind to one of several downstream U residues. These and previous observations might, therefore, indicate that APOBEC-1 is tuned to search for a product in a U- or AU-rich context [21,22]. This tendency to search for a product might also explain the editing of multiple Cs in certain in vitro conditions and the hyperediting of multiple Cs in transgenic animals that overexpress APOBEC-1, with mass action being the driving force [36,37].
References [1] Kane JP. Apolipoprotein B: structural and metabolic heterogeneity. Annu Rev Physiol 1983;45:637–50. [2] Knott TJ, Pease RJ, Powell LM, Wallis SC, Rall SCJ, Innerarity TL, et al. Complete protein sequence and identification of structural domains of human apolipoprotein B. Nature 1986;323:734 – 8. [3] Yang CY, Chen SH, Gianturco SH, Bradley WA, Sparrow JT, Tanimura M, et al. Sequence, structure, receptor-binding domains and internal repeats of human apolipoprotein B-100. Nature 1986;323:738–42. [4] Powell LM, Wallis SC, Pease RJ, Edwards YH, Knott TJ, Scott J. A novel form of tissue-specific RNA processing produces apolipoprotein B-48 in intestine. Cell 1987;50:831–40. [5] Chen S-H, Habib G, Yang CY, Gu ZW, Lee BR, Weng SA, et al. Apolipoprotein B-48 is the product of a messenger RNA with an organ-specific in-frame stop codon. Science 1987;238:363 – 6. [6] Scott J. A place in the world for RNA editing. Cell 1995;81:833 – 6. [7] Smith HC, Sowden MP. Base-modification mRNA editing through deamination-the good, the bad and the unregulated. Trends Genet 1996;12:418–24. [8] Navaratnam N, Morrison JR, Bhattacharya S, Patel D, Funahashi T, Giannoni F, et al. The p27 catalytic subunit of the apolipoprotein B mRNA editing enzyme is a cytidine deaminase. J Biol Chem 1993;268:20709–12. [9] Teng B, Burant CF, Davidson NO. Molecular cloning of an apolipoprotein B messenger RNA editing protein. Science 1993;260:1816– 9. [10] Betts L, Xiang S, Short SA, Wolfenden R, Carter CWJ. Cytidine ˚ crystal structure of an enzyme: transitiondeaminase. The 2.3 A state analog complex. J Mol Biol 1994;235:635–56.
S23
[11] Bhattacharya S, Navaratnam N, Morrison JR, Scott J, Taylor WR. Cytosine nucleoside/nucleotide deaminases and apolipoprotein B mRNA editing. Trends Biochem Sci 1994;19:105–6. [12] Nakamuta M, Oka K, Krushkal J, Kobayashi K, Yamamoto M, Li WH, et al. Alternative mRNA splicing and differential promoter utilization determine tissue-specific expression of the apolipoprotein B mRNA-editing protein (APOBEC-1) gene in mice. Structure and evolution of APOBEC-1 and related nucleoside/nucleotide deaminases. J Biol Chem 1995;270:13042–56. [13] Xiang S, Short SA, Wolfenden R, Carter CW Jr. Transition state selectivity for a single hydroxyl group during catalysis by cytidine deaminase. Biochemistry 1995;34:4516 – 23. [14] Xiang S, Short SA, Wolfenden R, Carter CW Jr. Cytidine deaminase complexed to 3-deazacytidine. A ‘valence-buffer’ in zinc enzyme catalysis. Biochemistry 1996;35(5):1335– 41. [15] Driscoll DM, Zhang Q. Expression and characterization of p27, the catalytic subunit of the apolipoprotein B mRNA editing enzyme. J Biol Chem 1994;269:19843– 7. [16] Yamanaka S, Poksay KS, Balestra ME, et al. Cloning and mutagenesis of the rabbit ApoB mRNA editing protein. A zinc motif is essential for catalytic activity, and noncatalytic activity auxiliary factor(s) of the editing complex are widely distributed. J Biol Chem 1994;269:21725– 34. [17] Anant S, MacGinnitie AJ, Davidson NO. APOBEC-1, the catalytic subunit of the mammalian apolipoprotein B mRNA editing enzyme, is a novel RNA-binding protein. J Biol Chem 1995;270:14762– 7. [18] MacGinnitie AJ, Anant S, Davidson NO. Mutagenesis of APOBEC-1, the catalytic subunit of the mammalian apolipoprotein B mRNA editing enzyme, reveals distinct domains that mediate cytosine nucleoside deaminase, RNA binding, and RNA editing activity. J Biol Chem 1995;270:14768– 75. [19] Navaratnam N, Bhattacharya S, Fujino T, Patel D, Jarmuz AL, Scott J. Evolutionary origins of apo B mRNA editing: catalysis by a cytidine deaminase that has acquired a novel RNA-binding motif at its active site. Cell 1995;81:187 – 95. [20] Lau PP, Chen S-H, Wang JC, Chan L. A 40 kilodalton rat liver nuclear protein binds specifically to apolipoprotein B mRNA around the RNA editing site. Nucleic Acids Res 1990;18:5817– 21. [21] Bostrom K, Lauer SJ, Poksay KS, Garcia Z, Taylor JM, Innerarity TL. Apolipoprotein B-48 RNA editing in chimeric apolipoprotein EB mRNA. J Biol Chem 1989;264:15701– 8. [22] Davies MS, Wallis SC, Driscoll DM, Wynne JK, Williams GW, Powell LM, Scott J. Sequence requirements for apolipoprotein B RNA editing in transfected rat hepatoma cells. J Biol Chem 1989;264:13395– 8. [23] Driscoll DM, Wynne JK, Wallis SC, Scott J. An in vitro system for the editing of apolipoprotein B mRNA. Cell 1989;58:519–25. [24] Chen S-H, Li XX, Liao WS, Wu JH, Chan L. RNA editing of apolipoprotein B mRNA. Sequence specificity determined by in vitro coupled transcription editing. J Biol Chem 1990;265:6811– 6. [25] Hodges P, Scott J. Apolipoprotein B mRNA editing: a new tier for the control of gene expression. Trends Biochem Sci 1992;17:77 – 81. [26] Backus JW, Smith HC. Apolipoprotein B mRNA sequences 3% of the editing site are necessary and sufficient for editing and editosome assembly. Nucleic Acids Res 1991;19:6781 – 6. [27] Backus JW, Smith HC. Three distinct RNA sequence elements are required for efficient apolipoprotein B (apo B) RNA editing in vitro. Nucleic Acids Res 1992;20:6007 – 14. [28] Shah RR, Knott TJ, Legros JE, Navaratnam N, Greeve JC, Scott J. Sequence requirements for the editing of apolipoprotein B mRNA. J Biol Chem 1991;266:16301– 4. [29] Smith HC. Apolipoprotein B mRNA editing: the sequence to the event. Semin Cell Biol 1993;4:267 – 78.
S24
J. Scott et al. / Atherosclerosis 141 Suppl. 1 (1998) S17 – S24
[30] Backus JW, Schock D, Smith HC. Only cytidines 5% of the apolipoprotein B mRNA mooring sequence are edited. Biochim Biophys Acta 1994;1219:1–14. [31] Navaratnam N, Fujino T, Bayliss J, Jarmuz A, How A, Richardson N, et al. Escherichia coli cytidine deaminase provides a molecular model for apo B RNA editing and a mechanism for RNA substrate recognition. J Mol Biol 1998;275:695–714. [32] Higgins DG, Sharp PM. CLUSTAL: a package for performing multiple sequence alignment on a microcomputer. Gene 1988;73(1):237– 44. [33] Stein P, Chothia C. Serpin tertiary structure transformation. J Mol Biol 1992;221:615–21.
.
[34] Bjorkman PJ, Saper MA, Samraoui B, Bennett WS, Strominger JL, Wiley DC. Structure of the human class I histocompatibility antigen, HLA-A2. Nature 1987;329:506 – 12. [35] Kuyper, L., Carter, C.W.J., 1996. Resolving crystal polymorphisms by finding ‘stationary points’ from quantitative analysis of crystal growth response surfaces. J Crystal Growth. [36] Sowden M, Hamm JK, Smith HC. Overexpression of APOBEC-1 results in mooring sequence-dependent promiscuous RNA editing. J Biol Chem 1996;271:3011– 7. [37] Yamanaka S, Poksay KS, Driscoll DM. Hyperediting of multiple cytidines of apolipoprotein B mRNA by APOBEC-1 requires auxiliary protein(s) but not a mooring sequence motif. J Biol Chem 1996;271:11506– 10.