Eur. J. Biochem. 174, 287-295 (1988) 0 FEBS 1988 Plastid-localised seed acyl-carrier protein of Brassica napus is encoded by a distinct, nuclear multigene family Richard SAFFORD, John H. C. WINDUST, Clare LUCAS, Jacquie DE SILVA, Christopher M. JAMES, Amanda HELLYER, Colin G. SMITH, Antoni R. SLABAS and Stephen G. HUGHES Biosciences Division, Unilever Research, Colworth House, Sharnbrook (Received December 9, 1987/February 19, 1988) - EJB 87 1375 Acyl-carrier protein (ACP) is a key component involved in the regulation of fatty acid biosynthesis in plants. cDNA clones encoding ACP from Brassica napus (oil seed rape) embryos have been isolated using oligonucleotide probes derived from heterologous ACPs. Analysis of the DNA sequence data, in conjunction with N-terminal amino acid sequence data, revealed ACP to be synthesized from nuclear DNA as a precursor containing a 5 1 -amino-acid N-terminal extension. Immunocytochemical studies showed ACP to be localised solely within the plastids of B. nupus seed tissue and it would therefore appear that the N-terminal extension functions as a transit peptide to direct ACP into these organelles. Analysis of several cDNA clones revealed sequence heterogeneity and thus evidence for an ACP multigene family. From ten cDNA clones, six unique genes, encoding five different mature ACP polypeptides, were identified. Northern blot hybridisation studies provide evidence that the seed and leaf forms of rape ACP are encoded by structurally distinct gene sets. De novo synthesis of fatty acids is catalysed by fatty acid synthetase which consists of seven or eight catalytic domains. In animals [l] and yeast [2] the domains are present on one or two multifunctional polypeptide chains (type I fatty acid synthetase), which are localised within the cytoplasm. In contrast, in plants [3] the fatty acid synthetase domains exist as discreet, monofunctional activities (type 11) which are organellar in location. Some insight into the genetic regulation of type I1 fatty acid synthetase systems has recently been obtained through cloning of genes from both yeast [4] and mammals [5]. Our interest lies in understanding the genetic control of fatty acid biosynthesis in plants, in particular within developing oil seeds. As a first step towards that objective we report here the molecular cloning of cDNA encoding seed-expressed acyl- carrier protein (ACP) from Brassica napus (oil seed rape). ACP is a key component of the plant Fatty acid biosyn- thetic machinery, serving both as a component of fatty acid synthetase and also as an acyl donor in desaturation and acyl- transfer reactions [6]. Recent studies have shown two major ACP isoforms to be expressed in leaf tissue [7, 81, but appar- ently only one major isoform in seeds [8]. To date, characteri- sation of plant ACP has been largely confined to the leaf- expressed forms. Thus, spinach [S] and barley [7] isoforms have been purified and N-terminal analysis suggests that, in both species, the isoforms are products of distinct genes. Using ACP as a representative marker protein, the site of fatty acid biosynthesis in leaves has been identified as the chloroplast [9]. In developing soybean seeds, ACP levels increase in close correlation with storage lipid synthesis [lo] suggestive of a regulatory role for ACP in this process. Despite this important role, ACP has not previously been localised within, or purified from, a seed source. This paper provides the first insight into the origin, struc- ture and expression of genes co-ordinating fatty acid biosyn- thesis in oil-bearing seeds. It reveals seed ACP to be localised within plastid bodies and to be encoded in nuclear DNA, being synthesised as a precursor containing an N-terminal extension sequence which presumably directs import of the protein into the plastids. Seed ACP was found to be encoded by a multigene family which is not expressed in leaf tissue. MATERIALS AND METHODS RNA isolation Total RNA was extracted from Brassica napus (var. Jet Neuf) embryos (18 - 25 days post-anthesis) essentially as de- scribed in [26]. For B. napus leaf RNA, tissue was ground in 8 M guanidinium hydrochloride, 20 mM Mes pH 7.0,20 mM EDTA, 50 mM dithiothreitol[27], followed by phenol/chloro- form extraction and ethanol precipitation. RNA was selectively precipitated with 2 M LiC1. Poly(A)-rich RNA was isolated using poly(U)-Sephadex (BRL). cDNA lihrary construction B. napus embryo poly(A)-rich RNA was size-fractionated Research Laboratory, Colworth House, Sharnbrook, Bedfordshire, On a 15-60y0 gradient in o.l NaC1j mM Tris/ MK44 lLQ, England HCI pH 7.5, 10 mM EDTA. A low-molecular mass fraction enriched in ACP mRNA was identified by immunoprecipita- saline citrate. tion [28] of rabbit reticulocyte lysate in vitro translation prod- Correspondence to R. Safford, Biosciences Division, Unilever Abbreviations, ACp, acyl-carrier protein; NaCl/Cit, standard
Transcript
Eur. J. Biochem. 174, 287-295 (1988) 0 FEBS 1988
Plastid-localised seed ac yl-carrier protein of Brassica napus is encoded by a distinct, nuclear multigene family Richard SAFFORD, John H. C. WINDUST, Clare LUCAS, Jacquie DE SILVA, Christopher M. JAMES, Amanda HELLYER, Colin G. SMITH, Antoni R. SLABAS and Stephen G. HUGHES Biosciences Division, Unilever Research, Colworth House, Sharnbrook
(Received December 9, 1987/February 19, 1988) - EJB 87 1375
Acyl-carrier protein (ACP) is a key component involved in the regulation of fatty acid biosynthesis in plants. cDNA clones encoding ACP from Brassica napus (oil seed rape) embryos have been isolated using oligonucleotide probes derived from heterologous ACPs. Analysis of the DNA sequence data, in conjunction with N-terminal amino acid sequence data, revealed ACP to be synthesized from nuclear DNA as a precursor containing a 5 1 -amino-acid N-terminal extension.
Immunocytochemical studies showed ACP to be localised solely within the plastids of B. nupus seed tissue and it would therefore appear that the N-terminal extension functions as a transit peptide to direct ACP into these organelles. Analysis of several cDNA clones revealed sequence heterogeneity and thus evidence for an ACP multigene family. From ten cDNA clones, six unique genes, encoding five different mature ACP polypeptides, were identified. Northern blot hybridisation studies provide evidence that the seed and leaf forms of rape ACP are encoded by structurally distinct gene sets.
De novo synthesis of fatty acids is catalysed by fatty acid synthetase which consists of seven or eight catalytic domains. In animals [l] and yeast [2] the domains are present on one or two multifunctional polypeptide chains (type I fatty acid synthetase), which are localised within the cytoplasm. In contrast, in plants [3] the fatty acid synthetase domains exist as discreet, monofunctional activities (type 11) which are organellar in location.
Some insight into the genetic regulation of type I1 fatty acid synthetase systems has recently been obtained through cloning of genes from both yeast [4] and mammals [5]. Our interest lies in understanding the genetic control of fatty acid biosynthesis in plants, in particular within developing oil seeds. As a first step towards that objective we report here the molecular cloning of cDNA encoding seed-expressed acyl- carrier protein (ACP) from Brassica napus (oil seed rape).
ACP is a key component of the plant Fatty acid biosyn- thetic machinery, serving both as a component of fatty acid synthetase and also as an acyl donor in desaturation and acyl- transfer reactions [6]. Recent studies have shown two major ACP isoforms to be expressed in leaf tissue [7, 81, but appar- ently only one major isoform in seeds [8]. To date, characteri- sation of plant ACP has been largely confined to the leaf- expressed forms. Thus, spinach [S] and barley [7] isoforms have been purified and N-terminal analysis suggests that, in both species, the isoforms are products of distinct genes. Using ACP as a representative marker protein, the site of fatty acid biosynthesis in leaves has been identified as the chloroplast
[9]. In developing soybean seeds, ACP levels increase in close correlation with storage lipid synthesis [lo] suggestive of a regulatory role for ACP in this process. Despite this important role, ACP has not previously been localised within, or purified from, a seed source.
This paper provides the first insight into the origin, struc- ture and expression of genes co-ordinating fatty acid biosyn- thesis in oil-bearing seeds. It reveals seed ACP to be localised within plastid bodies and to be encoded in nuclear DNA, being synthesised as a precursor containing an N-terminal extension sequence which presumably directs import of the protein into the plastids. Seed ACP was found to be encoded by a multigene family which is not expressed in leaf tissue.
MATERIALS AND METHODS
RNA isolation
Total RNA was extracted from Brassica napus (var. Jet Neuf) embryos (18 - 25 days post-anthesis) essentially as de- scribed in [26]. For B. napus leaf RNA, tissue was ground in 8 M guanidinium hydrochloride, 20 mM Mes pH 7.0,20 mM EDTA, 50 mM dithiothreitol[27], followed by phenol/chloro- form extraction and ethanol precipitation. RNA was selectively precipitated with 2 M LiC1. Poly(A)-rich RNA was isolated using poly(U)-Sephadex (BRL).
cDNA lihrary construction
B. napus embryo poly(A)-rich RNA was size-fractionated Research Laboratory, Colworth House, Sharnbrook, Bedfordshire, On a 15-60y0 gradient in o.l NaC1j mM Tris/ MK44 lLQ, England HCI pH 7.5, 10 mM EDTA. A low-molecular mass fraction
enriched in ACP mRNA was identified by immunoprecipita- saline citrate. tion [28] of rabbit reticulocyte lysate in vitro translation prod-
Correspondence to R. Safford, Biosciences Division, Unilever
Abbreviations, ACp, acyl-carrier protein; NaCl/Cit, standard
288
ucts with rabbit anti-(spinach leaf ACP I) antibodies. The enriched poly(A)-rich RNA was used to synthesize cDNA using the RNase H procedure [29]. After tailing with dCTP, the cDNA was annealed with dG-tailed Pstl-cleaved pBR322 (BRL) and used to transform Escherichia coli HBlOl [30]. 3500 tetracycline-resistant clones were selected.
Oligonucleotide probes
Two sets of oligonucleotides, derived from consensus amino acid sequences of E. coli [13], spinach [I41 and barley [7] leaf ACPs, were synthesized as gene probes. The mixed oligonucleotides, complementary to the coding sequences, were synthesized by the phosphoamidate method [31] at Unilever Research (Vlaardingen, Netherlands). Thus S’RAAYTCYTCYTCNA3’ (64 x 14-mer) was derived from the amino acid sequence Leu-Glu-Glu-Glu-Phe (LEEEF) and 5’-CATNACRATYTCNAC3’ (96 x 15-mer) from the se- quence Val-Glu-Ile-Val-Met (VEIVM). N = A, C, G or T; R = AorG; Y = TorC. Oligonucleotides were 5’-end-labelled using [p3 ’PI ATP and T4 polynucleotide kinase.
cDNA library screening
Colony blots of chloramphenicol-amplified recombi- nant clones were prepared on Biodyne A membrane (Pall) and hybridised to [32P]-oligonucleotides in 5 x NaCl/Cit, 2 x Denhardts, 1% SDS, 100 pg/ml denatured salmon sperm DNA at 30°C (LEEEF probe) or 35°C (VEIVM probe). Washing stringency was 5 x NaCl/Cit, 1 YO SDS for 30 min at the hybridisation temperature.
Northern blot analysis
Poly(A)-rich RNA was glyoxal-denatured [32], excluding dimethylsulphoxide, separated by electrophoresis in 1 % agarose gels containing 10 mM sodium phosphate pH 6.5 and transferred to Genescreen membrane (NEN). Hybridisation to [32P]oligonucleotides was under similar conditions to those described for colony blot screening. Hybridisation to nick- translated cDNA insert from ACP clone 28F10 was carried out at 48°C in 5 x NaCl/Cit, 1 YO SDS, 2 x Denhardts, 100 pg/ ml denatured salmon sperm DNA.
Southern blot analysis
Plasmid DNA was isolated from recombinant clones by the alkaline lysis method [33] and purified by CsCl gradient cenlrifugation. cDNA inserts were excised with PstI, electro- phoresed in 1 Yo agarose gels and blotted onto nitrocellulose [34]. Hybridisation with [32P]oligonucleotides was carried out as described for colony blots.
DNA sequence analysis
Restriction fragments of cDNA inserts were subcloned into MI3 vectors and sequenced using the dideoxy method [12]. Analysis of sequence data was performed using pro- grammes developed by Staden [36] and DNA Star Inc.
ACP pur fication and amino acid sequence analysis
ACP was purified from B. napus seed material (25-50 days post-anthesis) as the [3H]palmitoyl derivative [12]. The preparation was subjected to sequential Edman degradation
Fig. 1. ACP levels ( 0 ) and in vivofatty acid syntlzetase activity ( during B. napus seed developmenr. ACP levels were determined using the malonyl-CoA exchange reaction [37] and in vivo fatty acid synthetase activity was calculated from the increase in total fatty acid content [38]
on an Applied Biosystem (ABI) 470 amino acid sequencer. Phenylthiohydantoin derivatives were identified and quan- tified using an ABI 120 PTH analyser.
Immunocytochemical localisation of’ ACP
B. napus seed material was fixed with 1% paraformal- dehyde/0.05% glutaraldehyde in 0.05 M sodium phosphate buffer pH 6.8 for 2 h at 0°C. Following dehydration through graded ethanol, the sample was embedded in LRG/GMA resin and polymerised by ultraviolet light for 24 h at room temperature. Ultrathin sections, mounted on nickel grids, were immunostained using rabbit anti-(spinach ACP I) anti- bodies followed by goat anti-(rabbit 1g)-colloidal-gold (1 5 nni diameter).
RESULTS
A C P induct ion kinetics
The induction kinetics of ACP activity in relation to fatty acid synthetase activity in developing B. napus seed are shown in Fig. 1. The increase in activity is seen to precede the increase in fatty acid biosynthesis, suggestive of a regulatory role for ACP in storage lipid synthesis in oil seed rape.
ACP cDNA cloning
Poly(A)-rich RNA, isolated from B. napus embryos (1 8 - 25 days post-anthesis), was translated in vitro and the products were immunoprecipitated with rabbit-anti-(spinach ACP I) antibodies. Analysis of the immunoprecipitate by SDS/poly- acrylamide gel electrophoresis (SDS-PAGE) and fluoro- graphy (Fig. 2A, lane 2) showed a single ACP polypeptide (shown below to represent the precursor form of ACP) migrat- ing with an apparent molecular mass of 20 kDa (ACPs are known to migrate with anomalously high molecular masses on SDS-PAGE, owing to their highly acidic, non-hydrophobic character [Ill). Since ACP mRNA was found to be of such low abundance, embryo poly(A)-rich RNA was subjected to
289
Fig. 2. ( A ) Fluorograph of rape embryo in vitro translation products. ( B ) Northern blot of B. napus embryo poly(A)-rich R N A hyhridised with two mixed synthesis oligonucleotide probes derived from conserved regions of heterologous ACPs (see Materials and Methods). (A) Lane 1, translation products from unfractionated embryo poly(A)-rich RNA; lane 2, lane 1 products immunoprecipitated with rabbit anti-(spinach ACP I) antibodies; lane 3, translation products from a low-molecular-mass sucrose gradient fraction of embryo poly(A)-rich RNA; lauc 4, lane 3 products immunoprecipitated with rabbit anti-(spinach ACP I) antibodies. 14C-labelled molecular mass (kDa) markers are shown. (B) Lane 1 , LEEEF sequence probe; lane 2, VEIVM sequence probe. Size markers nucleotides are A HindIII/EcoRI fragments
sucrose gradient centrifugation and a fraction enriched in ACP mRNA identified by immunoprecipitation of in vitro translation products (Fig. 2A, lanes 3 and 4). The enriched mRNA was used to construct a 3500 clone cDNA library.
To screen for ACP clones, two sets of mixed oligo- nucleotides, derived from consensus sequences of hetero- logous ACPs (see Materials and Methods), were used. Initially the oligonucleotides were used in Northern blot experiments with B. napus embryo poly(A)-rich RNA to identify ACP mRNA as an approximately 750-nucleotide species (Fig. 2 B). Both oligonucleotide probes were found specifically to hybrid- ise to ACP mRNA and were used, sequentially, to screen the cDNA library. A total of seven cross-hybridising clones were identified, the smallest ofwhich (05E01,350 bp) was subjected to DNA sequence analysis. Translation of the resulting nucleotide sequence revealed an amino acid sequence highly homologous to that of barley and spinach leaf ACP sequences, both the VEIVM and LEEEF probe sequences being present, the former as part of the highly conserved pantetheine-binding domain of ACP. Having thus established that clone 05E01 encoded ACP (an estimated 55 amino acids), the cDNA insert from this clone was labelled, via nick-translation, and used to rescreen the cDNA library under stringent conditions. Ten positive clones were identified including six of the seven orig- inally identified by oligonucleotide screening. The cDNA in- serts of these ten clones, in the approximate size range of 350-700 bp, were DNA sequenced and, from the derived amino acid sequences, were all found to encode ACP (see Fig. 6).
ACP cDNA unulysis
The complete nucleotide and derived amino acid sequence of one of the longest ACP cDNAs (28F10) is shown in Fig. 3. The 670-bp cDNA [exclusive of the poly(A) tail] contains an open reading frame initiating at nucleotide 61 and terminat-
ing at nucleotide 462, thereby encoding a 134-amino-acid polypeptide (Mr 14700). Definitive interpretation of the open reading frame required N-terminal amino acid sequence data and to obtain this it was necessary to develop a novel pro- cedure in order to purify ACP from B. napus seed [12]. N-terminal analysis of purified ACP yielded the sequence AAKPETVEKV- and this enabled the mature form of ACP to be defined as an 83-amino-acid polypeptide (Mr 9200) corresponding to nucleotides 214-462 of the cDNA (Fig. 3). It is therefore apparent that ACP is synthesized in a precursor form containing a 51-amino-acid N-terminal extension se- quence, a finding suggestive of an organellar location for seed ACP. Direct confirmation of this suggestion was obtained when thin sections of B. napus seed tissue, immunostained with anti-ACP antibodies/colloidal gold conjugate, revealed ACP to be localised solely within the plastids of the cotyledon cells (Fig. 4). We therefore conclude that seed ACP is encoded in nuclear DNA and synthesized as a cytoplasmic precursor containing a transit sequence, which directs import of the protein into plastids. Since ACP is an essential component of plant fatty acid synthetase activity, it is implicit from our immunocytochemical studies that, within seed tissue, the plas- tid is the sole site of fatty acid synthesis.
ACP inultigene family Comparison of the DNA sequences obtained from individ-
ual ACP cDNA clones revealed interclone heterogeneity, thus providing evidence that seed ACP is encoded by a multigene family. Fig. 5 shows the nucleotide sequences of the coding regions of the ten cDNA clones and it appears that six differ- ent ACP genes are represented (Southern blot hybridisation of B. nupus DNA with a seed-expressed ACP cDNA indicated about 20 ACP genes/haploid genome; data not shown). The ten sequences can be classified into two subfamilies, the first containing cDNAs 28F10, 10Hl1, 11D11, 34F12, 04F05 and
Fig. 3. Nucleotide and predicted amino acid sequence of B. n a p s embryo ACP determinedfrom clone 28FIO cDNA. The precursor cleav- age site is marked by an arrow. The putative polyadenylation signal sequence is underlined
05E01 and the second containing cDNAs 34C02, 10C04, 22C01 and 29C08. The major structural features dis- tinguishing the two subfamilies relate to two sequences, each of nine bases, which occur in the transit peptide region. Thus the 28F10 subfamily contains the first nine-base sequence (nucleotides 94- 102) but not the second (nucleotides 133 - 141), whilst the 34C02 subfamily contains the second squence, but not the first. Within the first subfamily, cDNAs 28F10, 10H11 and 11D11 have an identical nucleotide sequence. cDNAs 34F12, 04F05 and 05E01 also share a common se- quence which shows eleven nucleotide substitutions when compared to the 28F10 sequence. In the second subfamily a greater degree of divergence exists between individual mem- bers, each of the four cDNA sequences in fact representing a different ACP gene. Two members of this subfamily (10C04/ 22C01) were found to be incomplete at their 5’ ends, thus making definitive comparisons impossible. However, notwith- standing this limitation, it appears that cDNAs 34C02 and 1 OC04 are most closely related (7/342 nucleotide substitutions plus one extra C-terminal codon for 10C04). 22C01 has 15/ 290 variant nucleotides when compared to 34C02 and 17/290 compared with 1OCO4. 29C08 is the most divergent of all the ACP cDNAs, not only when compared to the first subfamily
(e. g. 41/414 substitutions compared to cDNA 28F10), but also within the same subfamily (e.g. 33/414 substitutions with respect to 34C02).
Mature ACP amino acid sequence heterogeneity Fig. 6 shows the amino acid sequences of the ACP precur-
sors derived from the ten cDNA clones. Eight of the ten clones were found to encode a complete sequence for the mature form of ACP (clones 04F05/05E01 being incomplete at the C terminus). Of these eight clones, seven encoded an 83-amino- acid polypeptide, whilst 1 (10C04) encoded 84 amino acids (one additional C-terminal residue being present). Thus, in molecular mass terms, essentially a single isoform of ACP was found to be present in rape embryos, consistent with previous immunoblot studies with spinach and castor bean seeds [8]. However, comparison of amino acid sequences derived from individual cDNA clones revealed the existence of sequence heterogeneity such that five different mature rape ACP poly- peptides were identified. The ACP amino acid sequences con- form to the same subfamily classification as deduced for the ACP nucleotide sequences. The six cDNAs comprising the 28F10 subfamily were all found to encode an identical poly- peptide sequence (allowing for the incomplete 04F05/05E01 sequences). The four cDNAs of the 34C02 subfamily were found to encode four unique ACP polypeptides, all of which differed from the 28F10 sequence.
In total, 13 of the 83 ACP amino acid residues are found to be variant. Some of the amino acid substitutions are con- servative (e.g. V/A, V/I, E/D, A/L) but others (e.g. E/K, G/D, Q/P, Q/K) are not and these may be functionally more significant. Interestingly, up to five of the eight Gln residues present in 28F10 are found to be replaceable, being variously substituted by Pro, Lys, Glu, Leu and Asn. In general, the variable residues in the rape ACP sequences occur in regions which are not highly conserved between heterologous ACPs (see Fig. 7). No sequence variation was found in the pantetheine-binding regions (residues 89 - 101, Fig. 6) from individual rape ACP cDNA clones, a predictable finding in view of the high degree of interspecies conservation that exists for this domain (see below).
Comparison of rape ACP with heterologous ACPs Comparison of the derived rape ACP amino acid sequence
(as typified by clone 28F10) with ACP sequences obtained from E. coli [13], barley leaf [7] and spinach leaf [14] reveal respective homologies of 38%, 61% and 62% (Fig. 7). The rape ACP sequences contain a high proportion (z 24%) of acidic residues (D/E) in common with other plant ACPs so far studied. The domain surrounding the pantetheinylation site (Ser-39) of rape ACP is highly conserved with respect to E. coli, consistent with the observation that E. coli acyl synthetase can acylate the pantetheine residue of rape ACP [12]. The plant ACPs are most highly diverged from the E. coli sequence at the N terminus and this may reflect the compartmentalised nature of fatty acid biosynthesis in plants and the consequent evolution of domains to facilitate trans- port of the protein across intracellular membranes.
The amino acid sequences of the transit peptides deduced from individual cDNAs are represented within the ACP pre- cursor sequences (Fig. 6). As previously discussed, the se-
29 1
Fig. 4. Imrnunocytochernical localisation of ACP in B. napus seed using rabbit anti- [spinach ACP I ) antibodies [see Materials und Methods). ob = oil body; sg = starch grain; cw = cell wall; pl = plastid; nuc = nucleus. Scale bar = 1 pm. In control experiments, preincubation of anti-ACP antibodies with rape seed ACP was found completely to abolish localisation with the second antibody
quences can be classified into two subfamilies (allowing for the three clones with incomplete N termini). Both subfamilies have the same overall length of transit peptide, namely 51 amino acids. The two major distinguishing features concern two distinct tripeptide sequences, NHG (residues 32 - 34) and SIP (residues 45 -47). The 28F10 subfamily contains the NHG sequence but not the SIP sequence, whilst the 34C02 subfamily contains the SIP sequence but not the NHG se- quence. The functional significance, if any, of these differences is not clear, although interestingly both of the tripeptide sequences contain a residue (G/P), which is known to disrupt ordered secondary structure [15]. It is also of note that there is sequence conservation immediately adjacent to these two sites, such that NHG is surrounded by the sequence -S/ RTNLSF- and SIP by the sequence -R/TRLSI-. Such a con- served sequence, lying in different positions within the two classes of transit peptide, might imply that it has a specific role in targetting ACP to the appropriate intracellular location.
The ACP transit peptides, in striking contrast to the ma- ture ACPs, contain no acidic residues and have an overall net positive charge due to the presence of 5/6 strongly basic residues (R/K). The transit peptides of the 28F10 subfamily have the same amino acid sequence except for one conserva- tive substitution (A-V) at residue 17 in cDNAs 34F12 and 04F05/05E01. Since cDNAs 10C04 and 22C01 have incom- plete N termini, a complete comparison of the members of the 34C02 subfamily is not possible. However, comparison of 34C02 with 29C08 reveals eight variant residues, seven in the N-terminal region. Most of these substitutions are of a conservative nature. The level of amino acid divergence within the transit peptide sequences is greater than within the mature ACP sequence, consistent with observations made, for ex- ample, with the multigene family encoding the chloroplast- imported protein, small subunit of ribulose-bisphosphate car- boxylase [ 161.
Comparison of the transit peptide sequences of ACP with those of chloroplast-imported proteins reveals two common features, namely an overall basic character and an uncharged, hydroxyamino-acid-rich N terminus. Further similarities,
however, are not evident; no sequences are found, for in- stance, corresponding to the homology blocks recently pro- posed [17] to be important in the uptake of chloroplast pro- teins. All the ACP precursors contain the cleavage site CJA, a sequence not previously reported in chloroplast precursor proteins, although recently identified [18] at the cleavage site of the precursor of the maize WX’ protein, a polypeptide imported into amyloplasts.
5’ and 3‘ non-coding regions On the basis of an analysis of the longest cDNA clones,
the 5‘ and 3’ non-coding regions of ACP mRNA are estimated at approximately 50 and 200 nucleotides respectively. Seven of the ten cDNA clones were found to possess either partial or ‘complete’ 5‘ non-coding sequences (see Fig. 8A). Of these seven cDNAs, five belong to the 28F10 subfamily and their 5’ non-coding sequences are highly homologous (four variant nucleotides in total). Of the other two cDNAs, 29C08 has a distinctive divergent feature in the form of a eleven-base insertion in the middle of the 5’ non-coding region. It is of interest that four different translation initiation sequences [ 191 are represented within the seven cDNA clones. This variability could give rise to differential mRNA translation efficiency, although all the clones contain the highly conserved A residue at the -3 position relative to the ATG, a feature postulated to be important for efficient mRNA translation [19].
Eight of the ten cDNA clones possess partial or complete 3’ non-coding sequences (cDNAs 04F05/05E01 being incom- plete at the 3‘ end). It is readily apparent from Fig. 8B that considerable variation exists in the length of 3’ non-coding regions from individual cDNA clones. This variation could be a consequence of multiple polyadenylation sites within ACP genes or else the result of 3‘ deletions during the cloning procedure (see below). The cDNAs of the 28F10 subfamily (28F10, 10H11/11D11, 34F12) share a high degree of ho- mology in their 3’ untranslated regions (one variant base between 28F10 and 10H11/11D11 and 13 variant bases be- tween 28F10 and 34F12). However, the members of the 34C02
Fig. 5. Comparison of the nucleotide sequences of the coding regions of ten rape embryo ACP cDNAs. The cDNA sequences are given in comparison to cDNA 28F10; only nucleotide replacements are shown, dashes represent identical nucleotide to 28F10 sequence. The cDNA sequences 10H11 and 1 1 D11 are identical as are 04F05 and 05E01. cDNAs 34F12,10C04 and 22C01 are incomplete at the 5’ end and cDNAs 04F05/05E01 are incomplete at the 3’ end. cDNA 10C04 has an additional codon at the 3’ end. The sequences can be classified into two subfamilies, the first consisting of cDNAs 28F10, 10H11/11D11, 34F12 and 04F05/05E01 and the second consisting cDNAs 34C02, 10C04, 22C01 and 29C08. Nucleotides 94- 102 are absent from the first subfamily sequences and nucleotides 133-141 are absent from the second subfamily sequences
subfamily show considerable divergence in this region. shown not only the presence of multiple polyadenylation sites cDNAs 34C02 and 10C04 are most closely related in this within plant genes, but also that plant polyadenylation signal subfamily, but these two sequences still have 33 variant bases. sequences differ significantly from the animal consensus cDNAs 29C08 and 22C01 show considerable divergence from sequences, AATAAA. Examination of the regions 10 - 30 each other and from the other ACP cDNAs. nucleotides upstream of the poly(A) tails of the ACP cDNA
Assignation of polyadenylation signal sequences within clones (Fig. 8 B) reveals no consensus polyadenylation signal the ACP cDNAs is not clear cut. Recent studies [20] have sequence. The most likely signal sequences for the ‘full-length’
293
70 f 6o 10 20 30 6 0
28F10 10H11/11D11
34F12 04F05/OSE01
34C02 10C04 22c01 29C08
28F10 lOHll/1lD11
34F12 04F05/05E01
34C02 1OC04 22c01 29C08
80 90 100 110 120 130 137
Fig. 6. Comparison of the predicted amino acid sequences of the coding regions of ten rape embryo ACP cDNAs. The amino acid sequences are given in comparison to cDNA 28F10; only amino acid replacements are shown, dashes represent amino acids identical to the 28F10 sequence. cDNAs 34F12, 10C04 and 22C01 are incomplete at the N terminus and cDNAs 04F05/05E01 are incomplete at the C terminus. cDNA 10C04 has one additional C-terminal amino acid. The precursor cleavage site is denoted by an arrow. Amino acids 45 -47 are absent from cDNAs 28F10, 10H11/1lDll, 34F12,04F05/05E01 and amino acids 32-34 are absent from cDNAs 34C02, lOC04.22C01 and 29C08
Rape 28F10
Spinach I
B a r l e y I
Sp inach I1
B a r l e y I1
E. c o l i
Rape 28F10
Spinach I
B a r l e y I
E. c o l i
Rape 28F10
Spinach I
Bar l ey 1
E. c o l i
30
S T I E M R
40 50
20
S L K D D Q Q nun
D N A
T
Fig. I . Comparison of ACP amino acid sequences from rape embryo (deduced from cDNA 28FlO) E. coli [13] , spinach (141 leaf and barley [7] leaf. Sequences are aligned via Ser-39, the phosphopantetheine-attachment site. Homologous regions are boxed. Only N-terminal sequence data are available for spinach and barley leaf ACP I1
clones are AATGAA (28F10, 34F12), AATAA (34C02). The remaining 'shorter' cDNA clones could represent transcripts polyadenylated from a signal sequence 5' to those used in clones 28F10, 34F12 and 34C02 or they may have arisen as a result of 3' deletions during cloning.
Leaf and embryo ACP gene families
At the amino acid level there is a high degree of sequence conservation between leaf and embryo forms of ACP (e.g. 62% homology between spinach leaf and rape embryo ACPs). However, it would appear, at least on the basis of Northern
hybridisation studies, that this degree of conservation does not exist at the gene level, even within a single species. Fig. 9A shows parallel Northern blots of rape leaf (lane 1) and rape embryo (lane 2) poly(A)-rich RNA hybridised with the 680-bp cDNA insert from rape embryo ACP clone 28F10. Even under the low stringency conditions employed (48 "C, 5 x NaCl/Cit), no cross-hybridisation of embryo cDNA to leaf RNA is ob- served (as control, Fig. 9 B, a 14-mer oligonucleotide derived from a highly conserved region of ACP is seen to hybridise to ACP mRNA from both tissues). This finding strongly suggests that leaf and embryo forms of rape ACP are encoded by two gene sets which show considerable divergence.
294
A ACP28F.SBQ ACP1OH.SEQ ACPllD.SEQ ACP04F.SEQ ACPO5E.SEQ ACP34C.SEQ ACP29C. SEP
B ACP28F.SEQ ACP1OH.SEQ ACP11D.SEQ ACP34F.SEQ ACP34C.SEQ ACP1OC.SEQ ACP22C.SBQ ACP29C.SEQ
250v 260v 270v 280v 29ov 300v 310v 320v AT---TTCAGAATGAATTATGAATTTTCAAAAAAAAACCCCCCCCCCCC AAAAATTCAGAATGAAT-ATCA-TTTTCAAGACCCCCCCCCCCC - - A A A T T C A G U C T T G G - A A T G A A A A A C C C C C C C C C C C C
Fig. 8. ( A ) Comparison of nucleotide sequences of the 5’ untranslated regions of rape embryo ACP cDNAs. (B) Comparison ofnucleotide sequences of the 3’ untranslated regions of rape embryo ACP-cDNAs. (A) Nucleotide sequences are given in full and aligned for maximum hoinology via the ATG initiation codons. Dashes represent gaps in the sequencc created by the alignment procedure. cDNAs 34F12, 10C04 and 22C01 have no 5’ untranslated sequence. (B) Nucleotide sequences are given in full and aligned for maximum homology via the termination codons. Dashes represent gaps in the sequence created by the alignment procedure. cDNAs 04F05 and 05E01 have no 3’ untranslated sequence. Putative polyadenylation signal sequences are underlined
DISCUSSION
In conclusion, these studies have shown that seed plastid- localised ACP of B. napus is encoded in nuclear DNA, being synthesized from cytoplasmic poly(A)-rich RNA. ACP is synthesized as a precursor containing an N-terminal extension sequence which directs import of the protein into plastids. Since ACP, a key member of the fatty acid biosynthetic ma- chinery, is found to be localised exclusively within the seed plastids we conclude that this organelle is the sole site of de novo fatty acid biosynthesis in this tissue. It is thus apparent that an analogous compartmentalisation exists in the seed as in the leaf where the chloroplast has been identified as the sole site of de novo fatty acid synthesis [9].
B. napus seed ACP was found to be encoded by a multigene family, six unique genes, encoding five different mature ACP polypeptides, being represented in ten cDNA clones. Multiple ACP genes could provide a mechanism to regulate the levels of ACP in response to the highly variable cellular demands for fatty acid synthesis, which occur during oil seed development. Thus during seed development fatty acids are constitutively synthesized for membrane biosynthesis but, in addition, fatty acids destined for storage triglyceride are synthesized in a developmentally regulated fashion. We have shown that ACP levels increase dramatically during the oil synthesis phase of rape seed development. The presence of multiple ACP genes
would permit differential control of ACP expression to meet these requirements. Oligonucleotide probes, derived from variant regions of ACP cDNAs, may be used to investigate the differential expression of individual members or subgroups of the ACP gene family. This approach has been used [21] to demonstrate that expression of individual members of the petunia small subunit of ribulose-bisphosphate carboxylase (SSU Rubisco) multigene family within leaf tissue varies from 2-47% of total SSU Rubisco expression.
The multigenic nature of ACP may also reflect, to some extent, the multifunctional role that the protein performs. Thus, ACP participates not only as a component of fatty acid synthetase, but also in the desaturation of stearate [22], release of oleate by oleoyl-ACP thioesterase [23], and acyl transfer to glycerol 3-phosphate and monoacylglycerol3-phosphate [24]. Recent in vitro studies [25] have shown that the two major isoforms of spinach leaf ACP possess differential reactivity in respect of certain of the above reactions of fatty acid metab- olism. We have identified five different ACP polypeptides present in rape embryos, and it may be that certain of these polypeptides also preferentially carry out particular reactions. If this is so, then control of fatty acyl metabolism could be influenced by expression of particular members of the ACP multigene family.
The lack of hybridisation between gene sequences encoding leaf and embryo forms of rape ACP (see Fig. 9)
295
Fig. 9. Northern blot of B. napus embryo (lanes 2, 4) and leaf (lanes 1, 3) poly(A)-rich RNA. (A) Hybridisation with rape embryo ACP cDNA (clone 28F10,690 bp) at 48°C in 5 x NaCl/Cit. (B) Hybridisa- tion with mixed-synthesis 14-mer oligonucleotide derived from con- sensus LEEEF sequence (35"C, 5 x NaCl/Cit)
was a somewhat unexpected result, in view of the degree of sequence conservation which exists at the amino acid level between leaf and seed ACPs from heterologous sources. Earlier studies with spinach and castor bean [S] provided some evidence for tissue-specific expression of ACP, in that two major ACP isoforms were found to be present in leaves, but only a single isoform in seeds. Our results strongly suggest that the ACP isoforms present in rape seed are encoded by a multigene family which is not expressed in leaf tissue.
We wish to thank our colleagues at Unilever Research, Vbaardingen, Netherlands for synthesis of oligonucleotide probes and Dr. J. B. Ohlrogge for provision of rabbit anti-(spinach ACP I) anti- bodies.
REFERENCES 1. Tsukamoto, Y., Wong, H., Mattick, J. S. & Wakil, S. J. (1983)
J . Biol. Chem. 258, 15 31 2 - 15 322. 2. Schweizer, E. (1980) in Multifunctionalproteins (Bisswanger, H. &
Schmincke-Ott, E., cds) pp. 197-215, Wiley, New York. 3. Ohlrogge, J. B. (1982) Trends Biochem. Sci. 7, 386-387. 4. Schweizer, M., Roberts, L. M., Holtke, H.-J., Takabayashi, K.,
Hollerer, E., Hoffman, B., Muller, G., Kottig, H. & Schweizer, E. (1986) Mol. Gen. Genet. 203,479-486.
5. Witowski, A., Naggert, J., Mikkelsen, J. & Smith, S. (1987) Eur. J . Biochem. 165,601 -606.
6. Stumpf, P. K. (1984) in Fatty acid metabolism and its regulation (Numa, S., ed.) pp. 155- 179, Elsevier, Amsterdam.
7. H0j, P. B. & Svendsen, I. (1983) Carlsberg Res. Commun. 48,
8. Ohlrogge, J. B. & Kuo, T. M. (1985) J. Biol. Chem. 260, 8032- 285 - 305.
8037.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19. 20.
21.
22.
23.
24.
25.
26.
27
28
29 30 31
32
Ohlrogge, J. B.. Kuhn, D. N. & Stumpf, P. K. (1979) Proc. Nut1
Ohlrogge, J. B. & Kuo, T. M. (1984) Plant Physiol. 74, 622-
Rock, C. 0. & Cronan, J. E., Jr (1979) J . Bid. Chem. 254,9778 -
Slabas, A. R., Harding, J., Hellyer, A,, Roberts, P. & Bambridge,
Vanaman, T. C., Wakil, S. J. & Hill, R. L. (1968) J . Biol. Chem.
Kuo, T. M. & Ohlrogge, J. B. (1984) Arch. Biochem. Biophys.
Chou, P. Y. & Fasman, G. D. (1978) Adv. Enzymol. 47, 45-
Dean, C., van den Elzen, P., Tamaki, S., Black, M., Dunsmuir,
Karlin-Neumann, G. A. & Tobin, E. M. (1986) EMBO J. 5,
Klosgen, R. B., Gierl, A., Schwarz-Sommer, Z . & Saedler, H.
Kozak, M. (1984) Nucleic Acids Res. 12, 857-872. Dean, C., Tamaki, S., Dunsmuir, P., Favreau, M., Katayama, C.,
Dooner, H. & Bedbrook, J. (1986) Nucleic Acids Res. 14,2229- 2240.
Dean, C., van den Elzen, P., Tamaki, S., Dunsmuir, P. & Bedbrook, J. (1 985) EMBO J. 4, 3055 - 3061.
McKeon, T. A. & Stumpf, P. K. (1982) J. Biol. Chem. 257,
Ohlrogge, J. B., Shine, W. E. & Stumpf, P. K . (3978) Arch. Biochem. 189, 382-391.
Frentzen, M., Heinz, E., McKeon, T. A. & Stumpf, P. K. (1983) Eur. J. Biochem. 129,629 - 636.
Guerra, D. J., Ohlrogge, J. B. & Frentzen, M. (1986) Plant Physiol. 82, 448 -453.
Hall, T. C., Ma, Y., Buchbinder, B. V., Pyne, J. W., Sun, S. M. & Bliss, F. A. (1978) Proc. Natl Acad. Sci. USA 75, 3196- 3 200.
Kaulen, H., Schell, J. & Kreuzaler, F. (1986) EMBO J. 5, 1-8.
Safford, R., de Silva, J., Lucas, C., Windust, J. H. C., Shedden, J., James, C. M., Sidebottom, C. M., Slabas, A. R., Tombs, M. P. & Hughes, S. G. (1987) Biochemistry26, 1358-1364.
Gubler, U . & Hoffman, B. J. (1983) Gene 25,263 -269. Mandel, M. & Higa, A. (1970) J . Mol. Biol. 53, 159 - 162. Adams, S. P., Kauka, K. S., Wykes, E. J., Holder, S. B. & Gallupi,
McMaster, G. K. & Carmichael, G. G. (1977) Proc. Natl Acad.
Acad. Sci. USA 76, 1194- 1198.
625.
9785.
H. E. (1987) Biochim. Biophys. Acta, in the press.
243,6420 - 643 1 .
234,290 - 296.
149.
P. & Bedbrook, J. (1987) Mol. Gen. Genet. 206,465-474.
9-13.
(1986) Mol. Gen. Genet. 203, 237-244.
12141-12147.
G. R. (1983) J . Am. Chern. SOC. 105, 661 -663.
Sci. USA 74.4835-4838. 33. Birnboim, H. C. & Doly, J. (1979) Nucleic Acids Res. 7, 1513-
34. Southern, E. M. (1975) J. Mol. Biol. 98, 503-517. 35. Sanger, F., Nicklen, S. & Coulson, A. R. (1977) Proc. Natl Acad.
36. Staden, R. (1982) Nucleic Acids Res. 10,4731 -4751. 37. Majerus, P. W., Alberts, A. W. & Vagelos, P. R. (1969) in Methods
38. Slabas, A. R., Roberts, P. A., Ormesher, J. & Hammond, E. W.
