mitochondria uncoupling protein - NCBI

2 downloads 152 Views 1MB Size Report
Jan 26, 1987 - Communicated by J.E.Walker ... phosphate carrier protein (Kolbe and Wohlrab, 1985; Walker et ..... of 15 kd (Thomas and Halestrap, 1981).
The EMBO Journal vol.6 no.5 pp.1367-1373, 1987

Sequence of the bovine mitochondrial phosphate carrier protein: structural relationship to ADP/ATP translocase and the brown fat mitochondria uncoupling protein

Michael J.Runswick, Steven J.Powell, Pal Nyren' and John E.Walker Medical Research Council Laboratory of Molecular Biology, Hills Road, Cambridge CB2 2QH, UK 'Present address: The Department of Biochemistry, The Arrhenius Laboratory, University of Stockholm, Sweden Communicated by J.E.Walker

A cDNA encoding the precursor of the bovine mitochondrial phosphate carrier protein has been cloned from a bovine cDNA library using a mixture of 128 different 17-mer oligonucleotides as hybridisation probe. The protein has an N-terminal extension of 49 amino acids not present in the mature protein. This extension has a net positive charge and is presumed to direct the import of the protein from the cytoplasm to the mitochondrion. Comparison of the protein sequence of the mature phosphate carrier with itself, with ADP/ATP translocase and with the uncoupling protein from brown fat mitochondria shows that all three proteins contain a 3-fold repeated sequence 100 amino acids in length, and that the repeats in the three proteins are related to each other. This inplies that the three proteins have related threedimensional structures and mechaniims and that they share a common evolutionary origin. The distribution of hydrophobic residues in the phosphate carrier protein suggests that each repeated 100 amino acid element is composed of two membrane-spanning a-helices linked by an extensive hydrophilic domain. This model is similar to that first proposed for the ADP/ATP translocase and later for the brown fat mitochondria uncoupling protein. Key words: mitochondria/phosphate carrier/ADP/ATP translocase/cDNA clone -

from both porcine and bovine heart mitochondria (Bisaccia and Palmieri, 1984; Kolbe et al., 1984) and appears on polyacrylamide gels as two bands with similar mol. wts (-33-34 kd) and with the same N- and C-terminal sequences (Kolbe et al., 1984). The N-terminal sequence of the protein has been determined by direct sequence analysis (Kolbe and Wohlrab, 1985; Walker et al., 1986). As described below, we have used this information to design an oligonucleotide which has been used to isolate a cognate cDNA clone from a bovine library (Gay and Walker, 1985a). Its DNA sequence has been determined and the deduced amino acid sequence is related to those of the ADP/ATP carrier and the uncoupling protein from brown fat mitochondria, strongly suggesting that the three carrier proteins have related structures and mechanisms, and are derived from the same ancestral gene.

Results and Discussion Cloning and DNA sequence In the first screening of the cDNA library with the mixture of 128 oligonucleotides, four independent, positively hybridising clones were isolated. However, DNA sequence analysis showed that only one of them contained sequences that could be recognised as coding for the known N-terminal sequence of the phosphate carrier protein (Kolbe and Wohlrab, 1985; Walker et al., 1986). The sequence of the insert in this recombinant was determined completely in both senses of the DNA (see Figure 1 and Materials and methods). Severe difficulties were en0

IRL Press Limited, Oxford, England

600

800

1000

1200

=. 1U .1111 BamH l/EcoR

Introduction ATP is synthesised from ADP and inorganic phosphate by oxidative phosphorylation in mitochondria. Supply of substrates in the mitochondrial matrix is maintained by two proteins in the inner membrane, the ADP/ATP translocase and the phosphate carrier (Meijer and van Dam, 1974; La Noue and Schoolwerth, 1979). The former, a protein of Mr 32.8, exchanges external ADP against internal ATP (Klingenberg, 1985). Its amino acid sequence is known (Aquila et al., 1982) and has been shown to contain three repeated related sequences, each of which has been suggested to be folded into a structural element composed of two potential membrane-spanning segments linked by an extramembranous domain (Saraste and Walker, 1982). Related repeated sequences are found also in the uncoupling protein, a protein carrier in the inner membrane of brown fat mitochondria, which returns into mitochondria protons that have been removed by the respiratory chain, and so bypasses ATP-synthase (Aquila et al., 1985). The phosphate carrier mediates the uptake of phosphate into the mitochondrial matrix, either by proton cotransport or in exchange for hydroxyl ions. It has been purified

400

200

*

1347

9 3' _

EcoR

*

-

4-00

14-

Pst I 4---

BamH I/Sma I

4-

P

Sma l/EcoR I Fnu4H I

. 44

Fig. 1. Sequence analysis of a cDNA clone coding for the bovine phosphate carrier protein. The insert is flanked at its 5' end by a BamHI site (denoted by large vertical arrows) and at its 3' end by an EcoRI site (V) and contains two internal EcoRI sites. The sequences of the EcoRI-BamHI fragment (bases 1-579) and the EcoRI fragments (bases 580-627 and 628-1347) were determined with the aid of synthetic oligonucleotides (denoted by heavy bars) after recloning into M13 vectors. The overlaps through the internal EcoRl sites were established with PstI fragments (V, Pst sites) and the sequence of the G:C rich region (bases 140-188) was confirmed with fragments generated by SmaI (small vertical arrows) and Fnu4HI (vertical bars).

1367

M.J.Runswick et al. PstI

80

60

40

20

PstI

100

GGCCTCTGTGAGCTGCAGCCTTTCTGAGAGACCGGTTGTGTGTTCGCCATCTTAGGGAGAAGATGTACTCGTCCGTGGTGCACCTAGCGCGGGCGAACCCCTTCAACGCGCCTCACCTGC MY S S V V H L A R AN P F N A P H L -49

Q -30

-40 Mature

140

180

160

1

200

220

240

AGTTGGTGCACGATGGTCTCGCGGGGCCCCGCAGCGACCCCGCCGGGCCCCCGGGCCCACCCCGCCGCTCTCGCAACCTGGCAGCCGcC ECTGTAGAAGAGCAGTgTAGCTGTGACTATG A A A A V E E Q Y S C D Y L

V

H

D

G

L

A

G

P

R -20

S

D

P

A

G

P

P

P

G

R

R

S

*R

N

L

1

G

10

360

340

320

300

280

260

P -10

GATCTGGCAGATTCTTTATCCTTTGTGGACTTGGAGGAATTATTAGCTGTGGCACAACACATACAGCATTGGTTCCTCTAGATCTGGTTAAATGCAGGATGCAGGTGGACCCACAAAAGT I I P Y S

G

R

F

F

IL

C

G 20

L

G

S C

G

G

T

T 30

UT

A

L V

P

L

D LV 40

|

C R

N

Q

V

D

Q

S0

K

T1

Pstl

440

420

400

380

K

480

ACAAGAGCATATTTAATGGATTTTCAGTTACACTCAAAGAGGATGGTTTTCGTGGTTTGGCCAAAGGATGGGCTCCGACTTTCATTGGCTACTCCCTGCAGGGACTCTGCAAGTTTGGCT K

S I

F

N

G

F

S

V T 60

L

K

E D

G

F R

G

|T2

S00

L A 70

K G W A

P T

F 1

Y

G 80

S

L

Q

G

L

C

K

F

G 90

F

T3

Eco RI

560

540

520

600

TTTATGAAGTCTTCAAAGTTTTGTACAGCAACATGCTTGGAGAGGAGAATGCCTATCTGTGGCGCACATCACTGTATTTGGCTGCCTCTGCCAGTGCTGAATTCTTTGCTGACATTGCTC Y E

V

F

K

V

L

Y

S N 100

620

ML

G E E

L W

N AY

R

T

S LY

L A

A S

680

660

640

A

E

F

F

A

D I AL 130

120

110

Eco RI

A S

700

720

TGGCTCCTATGGAAGCTGCTAAGGTTCGAATTCAAACCCAGCCAGGTTATGCTAACACTCTGAGGGATGCAGCTCCCAAAATGTATAAGGAAGAAGGCTTAAAGGCGTTCTACAAGGGGG A

P M

E

A

A

K

V

R I

Q

T

Q

P G

Y

A N T

150

140

L

R D

A

A

P

K

NY

K 160

E

E

G

L

K

A

F

Y

KG 170

V

840 760 780 820 740 800 TTGCCCCTCTCTGGATGAGGCAGATACCATACACCATGATGAAGTTTGCCTGCTTTGAACGTACTGTTGAAGCATTGTACAAGTTTGTGGTTCCCAAGCCCCGAAGTGAATGTTCAAAGC A P L W M R Q I P Y TN N K F A C F E R T V E A L Y K F V V P K P R S E C S K P

180

200

190

210

960 880 900 940 860 920 CAGAGCAGCTGGTTGTCACATTTGTGGCAGGTTACATAGCTGGAGTCTTCTGTGCCATTGTTTCTCACCCTGCTGATTCCGTGGTGTCTGTGTTGAATAAAGAGAAGGGTAGCAGTGCCT E Q L V V T F V A G Y I A G V F C A I V S U P A D S V V S V L N K E K G S S A S 220 230 250 240 1080 980 1020 1000 1040 1060 CTGAGGTCCTGAAGAGGCTTGGATTTAGAGGTGTATGGAAGGGACTCTTTGCCCGTATCATCATGATCGGCACTCTGACTGCACTACAGTGGTTTATCTATGACTCTGTGAAGGTCTACT KR L G F R G V W KG L F A R I I N I G T L T A L Q W F I Y D S V K V Y F E V L 290 260 270 280

Hind III

1120

1100

1160

1180

1280

1300

1200

TCAGGCTCCCTCGCCC'TCCTCCACCTGAGATGCCAGAATCTCTGAAGAAAAAGCTTGGGTACACTCAATAGATAAATGAAAGCAAATATGGACTGAATCTGCGTGTTGATCAGTGTTTGA N P P P P E P E S L K L G Y T Q * R

L

P

N

K

300

1220

K-

310

1240

1260

1320

UGAAAATGCAGAAGGAACTTTTATATATTTGACAGTGTAGGAAGTTGTCTATTCCTAATATAATTACTGTAGTACTCTTGTCAAAAGCAAGAGTTTCAAACTTAATGTGGA

~

1340 AACTG TACATG AAAAAAAAAAAAAAAA

Fig. 2. DNA sequence of a cDNA encoding the precursor of the bovine mitochondrial phosphate carrier protein. The import sequence of the protein runs from amino acids -49 to -1 and the mature protein from 1-313. The box around the DNA sequence coding for residues 1-6 is the region to which the oligonucleotide probe hybridised. Restriction enzyme sites used in the DNA sequencing and for preparation of prime-cut probes are underlined. The box near to the 3' end of the DNA sequence is a recognition site for addition of poly(A) (Proudfoot and Brownlee, 1976). The dashes under the protein sequence represent protein sequences determined on the mature protein (Kolbe and Wohlrab, 1985; Walker et al., 1986) and on peptides TI -T3 isolated from a tryptic digest of the phosphate carrier protein.

countered in the region comprising bases

140-188, a sequence

composed almost entirely of G and C residues. The final sequence presented in Figure 2 is 1347 bases in length. It is terminated at its 3' end by the sequence (A)16 and so the 3' end of the mRNA is represented in the cDNA clone.

Fourteen bases to the 5' side of this run of A residues is the sequence AATAAA, which presumably serves as a polyadenylation signal (Proudfoot and Brownlee, 1976). By Northern blot 1368

RNA coranalysis (Figure 3) the mRNA in heart poly(A)+ to be 1470 bases responding to the cDNA clone was estimated may long and so the possibility remains that the cDNA clone species of hybridisations weaker not be entirely full length. (Much cannot be ex-

- 3 and 4 kb were also noted, but their presence plained at present.) However, DNA and not RNA standards were

used in this experiment and so the estimates of RNA size are not precise.

Mitochondrial phosphate carrier protein Table I. Amino acid composition of the mature phosphate carrier protein

-8 37

-

3 79

- 2 -22

43 25

l -77

1 69

go

Amino acid

From DNA sequence

Aspartic acid Asparagine Threonine Serine Glutamic acid Glutarnine Proline Glycine Alanine Valine Cysteine Methionine Isoleucine Leucine Tyrosine Phenylalanine Histidine Lysine Arginine Tryptophan

8 5 14 20 18 10 18 25 28 24 8 9 14 29 17 21 2 24 14 5

Total -

By amino acid analysisa

14 13 20 27 18 26 29 26 6 8 13 34 16 21 1 19 13 6

313

0-56 aFrom Kolbe et al. (1984).

of its mol. wt by polyacrylamide gel electrophoresis in the presence of SDS are 33-34 kd, and the mol. wt calculated from the sequence is 35 020. Secondly, the amino acid composition

Fig. 3. Northern blot analysis of poly(A)+ RNA from bovine heart with a probe derived from the cDNA clone for the bovine phosphate carrier protein. Poly(A)+ heart RNA (2.5 1zg) was probed with a prime cut probe corresponding to bases 629-1131 in Figure 2. The marker sizes are determnined with bacteriophage X DNA (3 /Ag) digested with NdeI and endlabelled with [cs-32P]ATP in the presence of the Kienow fragment of DNA polymerase. The sizes of the resulting fragments are shown in kb.

Identification of the clone The DNA sequence contains an open reading frame (bases 210-1148) that encodes the mature phosphate carrier protein. This conclusion is based upon direct protein sequences determined from the intact protein and from peptides derived from it (see Figure 2). These include the N-terminal residues, 1-47, sequenced partly by analysis of the intact protein and partly from a fragment (residues 1-47) produced by treatment of the protein with mild acid (Kolbe and Wohlrab, 1985). It should be noted that residue 48 as determined by DNA sequencing is proline, and that indeed the linkage Asp-47 -Pro-48 would be expected to be acid labile. Sequence analysis (by I.M.Fearnley and J.E.Walker) of peptides TI -T3 (isolated by H.Kolbe and H.Wohlrab) gave further support for the identity of the cDNA. A third piece of evidence is provided by carboxypeptidase digestion of the intact protein (Kolbe et al., 1984) which released firstly amino acids with elution times in an amino acid analyser consistent with the presence of glutamine and threonine (but unresolved from serine and asparagine), and later tyrosine and glycine, and finally leucine. These data fit with the C-terminal sequence LGYTQ deduced from the DNA sequence. Other data measured on the mature phosphate carrier protein are also in accord with the deduced sequence. Firstly, estimates

determined by amino acid analysis for the most part is very similar to that calculated from the sequence; the only significant deviations are in values for leucine and lysine (Table I). The mitochondrial import sequence The open reading frame extends in a 5' direction from the start of the mature protein and encodes a presequence 49 amino acids in length. The mol. wt of the precursor is calculated to be 40 140. The proposed initiator methionine is the only methionine in this region, and the presence of an in-phase stop codon (bases 9-11) excludes the possibility that the reading frame could extend beyond the 5' end of the isolated cDNA. It is assumed that the proposed presequence (amino acids 1-49), as with other nuclearcoded mitochondrial proteins, serves to direct the protein into the mitochondrion and that it is removed during import (Schatz and Butow, 1983). In common with other mitochondrial import sequences, the import sequence of the phosphate carrier has a net positive charge (three histidines, five arginines, two aspartic acids), which has been proposed to help to drive the precursors across the inner mitochondrial membrane, aided by the net negative electrochemical potential inside (Viebrock et al., 1982). The presence of acidic residues in an import sequence is unusual, but not unique; they are found also in the presequences of bovine P1 and P2 proteolipids (Gay and Walker, 1985b), yeast cytochrome cl (Sadler et al., 1984) and the fl-subunit of human F1-ATPase (Ohta and Kagawa, 1986). The length of the presequence is 49 amino acids, which falls within the rather wide range of observed sizes, from 22 amino acids in bovine cytochrome oxidase subunit IV (Lomax et al., 1984) and chicken aspartate aminotransferase (Jaussi et al., 1985) to 68 amino acids in bovine ATP synthase proteolipid P2 (Gay and Walker, 1985b) and yeast cytochrome c peroxidase (Kaput et al., 1982). Other imported mitochondrial proteins, including ADP/ATP translocases from Neurospora crassa (Arends and Sebald, 1984), yeast

1369

;r'l,-:~Ii

M.J.Runswick et al.

//s

Phosphate Carrier

/:

/*

$'7.'

/

1XA~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ *

lb

, / /~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~,

ADP/ ATP Translocase

>~~~~~~~~~~~~~~~~~. 0or

Uncoupler Protein

Uncoupler Protein

ADP/ATP

Phosphate

Translocase

Carrier

Fig. 4. Comparison of the protein sequences of the phosphate carrier protein with those of bovine ADP/ATP mitochondria uncoupler protein. The calculations were made with the DIAGON (Staden, 1982b) with a score division on the axes represents 10 amino acids.

translocase and the hamster brown fat of 270 and a window of 25 amino acids. One

~ ~ ~ ... ~ ~ ~ ......... ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~

Fr. *

S

j

-.

r

i.r p W.

r,

.!.

%. .(" F)

r..

c.

*~~~~~~~~~.

r

; !* ,

j,[ !oi ')

;2S

r

-i

-

,.

"C

HGFIl rO ^t (O , ; *;HrFt f S fS

.1 U Jg j~$

K

ii__-__ -_ ___ _

r

''jH

;.

..

--

F

-

LJ' F F H' f)

r,> Hi.lr; 1-. i1 P f-

----A ----6 6 - A

Is' P ;P ")

r

r

jT

__

-PHtiV-

U PF

__..L

1.

_

_

1.

Fig. 5. Alignment of repeated sequences in the uncoupler protein, ADP/ATP translocase and the phosphate carrier protein. These were made according to Figure 4. * represents a deletion. Identities and conservative substitutions present in at least five of the nine sequences are shaded. Each repeat contains a pair of hydrophobic segments (I and II, III and IV, V and VI respectively), joined by hydrophilic segments, A, B and C as first proposed by Saraste and Walker (1982) for the ADP/ATP translocase.

(Adrian et al., 1986) and cow (S.J.Powell and J.E.Walker, unpublished results), the brown adipose tissue uncoupling proteins from hamster (Aquila et al., 1985) and rat (Bouillard et al., 1986) 1370

and the d subunit of bovine ATP synthase (J.E.Walker and M.J.Runswick, unpublished results) have no processed Nterminal sequence (other than removal of initiator methionine and

Mitochondrial phosphate carrier protein

a)x

after an arginine residue. An arginine residue frequently is found in the vicinity of the cleavage site in import sequences. It may

Uncoupler u

410 2o

m

A

be at the site of cleavage, as for example in bovine cytochrome oxidase subunit IV (Lomax et al., 1984) and ATP synthase P1 and P2 proteolipids (Gay and Walker, 1985b), or up to six amino acids preceding the cleavage site as in human ornithine carbamyl transferase (Horwich et al., 1984).

n

c L-i

o

21 ffK t l M I W aJW ISequence repeats A

V7

fl 4 \

IfPairwise comparisons of the phosphate carrier with the bovine -o ADP/ATP translocase and the hamster brown fat mitochondria Ir - 2( uncoupling protein and comparisons of each of the three pro-I-A FL c{ i[]Lteins with itself, are very distinctive (Figure 4). Firstly, they show that the three proteins contain three internal repeats which have -4(0 . been noted previously for the translocase (Saraste and Walker, , ,1982) and for the uncoupling protein (Aquila et al., 1985). 50 l00 150 200.. 250 300 Secondly, the persistence of the parallel diagonal lines 100 amino acids apart across the matrix shows that all of the repeated elements in each of the three proteins are related to each other. O ADP/ATP Translocase This is confirmed by the alignment of the nine segments of se4( quence (Figure 5). The alignment shows that the most highly conserved residues are found in the proposed ca-helical segments (I-VI; see below) and also in the regions of the loops A, B and D. C that connect them to the a-helices. The central parts of loops ~0c A-C are poorly conserved as are the N- and C-terminal regions, H and the regions between the three repeats. It is likely that these A f \l tW W IIJ i regions contribute to the specificities of the carriers. T wV v v v Hydrophobic segments I..9 The similarity between the three proteins is further emphasised l FA-FiiIJ---]II-->B--FiVIJ--1jj-C-V~1 E3 by their hydrophobic profiles (Figure 6). In the case of translocase, this has been suggested to be indicative of the presence of six membrane-spanning segments (I -IV) in the pro-4C tein, although segment II is rather weakly hydrophobic (Saraste 50 100 150 and Walker, 1982). The lengths of the proposed membrane200 250 Residues spanning segments (-25 amino acids) are consistent with the view that they are probably folded into a-helices. These segments contain some charged residues (see span IV of the phosphate carcarrier 4C Phosphate rier). Charged residues in other membrane proteins such as bacteriorhodopsin may be structural, forming ion pairs with equal and opposite charges, or, possibly, as with buried acidic residue in the dicyclohexylcarbodiimide-reactive proteolipid of ATP syna) thase, they may have a function in the activity of the protein. c The regions between segments I and II, HI and IV, and V and VI, called A, B and C respectively (Figures 5 and 6) are relatively hydrophilic and were proposed to form extensive extra-membrane domains. Thus, the 3-fold sequence repeat in the translocase was Ir - go ____ proposed to have a structure comprising a pair of transmembrane a-helices ( and A, HIB and IV, and V and joined by the hydrojI B-----B F --IC philic elements i, and C respectively.Vl The sequence of the uncoupling protein can be interpreted in a similar way, although ... Aquila et al. (1985) preferred a slightly different interpretation 100 150 200 50 250 300 with five transmembrane c-helices and a transmembrane (-sheet, in the region designated as helix II in Figures 5 and 6. The hydrophobic profile of the phosphate carrier suggests that it also Fig. 6. C'omparison of the hydrophobic profiles of the phosphate carrier Iv

.

.

.

.

------V

.

.

.

.

.

.

.

.

..

.

.

-

.C

L-1

-]jf-A-E¶--.

.

.

.

.

protein w 'ith those of bovine ADP/ATP translocase and the hamster brown fat mitocl hondria uncoupler protein. The calculations were made with

amnino acids. The horizontal bar average hydrophobicity as defined by Kyte and Doolittle

HYDROIPLOT with a window of 11 represent s an (1982).

sometin nes concomitant acetylation of the N terminus); yet the proteinms enter the mitochondria. The site of cleavage of the import sequence of the phosphate

carrier is between two alanine residues and is five amino acids

wilhvsxmeba-pnigsget,adinhscseeg will have SiX membrane-spanning segments, and in this case segment II is somewhat more hydrophobic than in the translocase

and the uncoupling protein. It follows that the phosphate carrier also is constituted of three repeated domains with the same overall organisation as in both ADP/ATP translocase and the uncoupling protein. Common properties of mitochondrial carriers The inner membrane of mitochondria contains at least 12 different carrier proteins involved in the transport of anions and 1371

M.J.Runswick et al.

protons across the membrane (La Noue and Schoolwerth, 1979). Three of them, the ADP/ATP translocase, the uncoupler protein of brown fat mitochondria and the phosphate carrier have been well characterised. All three are single polypeptides with similar mol. wts ( - 33 kd) and related sequences, and it appears in at least the case of the ADP/ATP translocase that the active form is a dimer (Klingenberg, 1985). The a-ketoglutarate/malate carrier also has a similar mol. wt (Bisaccia and Palmieri, 1984) and preliminary sequencing studies indicate that it contains sequences related to the three carriers discussed in this paper (M.J.Runswick, J.E.Walker, F.Bisaccia and F.Palmieri, unpublished results). So it appears that four of the carrier proteins are derived from a common ancestor and is clear that at least three of them are derived by a process involving gene duplications from a smaller ancestral element corresponding to the 100 amino acid residue repeat. However, the pyruvate carrier appears to be smaller and is reported to have an apparent mol. wt of 15 kd (Thomas and Halestrap, 1981). A number of other similarities between carriers have also been noted. For example, both ADP/ATP translocase and the uncoupler protein bind purine nucleotides. However, their binding sites are very different; in the translocase it is mobile and involved in translocation, being accessible from both sides of the membrane. Whereas in the uncoupler protein it is more rigid and is involved in the regulation of translocation (Aquila et al., 1985). These sites have not been adequately mapped in the primary structure of the proteins, nor has the topology of the proteins been established. So the significance of observed structural repeats, which clearly must be of fundamental importance in the functioning of carriers, is obscure.

Materials and methods Oligonucleotide synthesis A mixture of 128 oligonucleotides with the sequence 3' GCNGTNGARGARCARTA 5'was synthesised by T.V.Smith using an Applied Biosystems 380B oligo-

nucleotide synthesiser. This corresponds to the protein sequence AVEEQY, residues 1-6 of the mature protein (Kolbe and Wohlrab, 1985; Walker et al., 1986). Twelve unique oligonucleotides, 16 or 17 bases in length, used as primers in the determination of the complete double-stranded DNA sequence were synthesised by the same procedure. Screening of the cDNA library and DNA sequencing The construction of the library in the pUC8 vector (Vieira and Messing, 1982) and the methods used to screen it with the mixed oligonucleotide were described earlier (Gay and Walker, 1985a,b). Hybridisations were performed for 18-24 h at 5°C below the minimum melting temperature of the mixed oligonucleotide (estimated to be 46°C). The inserted DNA in positively hybridising clones was released by digestion with EcoRI and BamHI. The products were fractionated by electrophoresis in the presence of ethidium bromide in a 1 % LGT agarose minigel. Bands were excised from the gel under u.v. light. DNA was recovered by phenol extraction (Wieslander et al., 1979) and cloned into appropriate M13mp8 and nmp9 vectors (Messing and Vieira, 1982). Sequences at the ends of the cloned fragments were determined by primed synthesis in the presence of chain terminators (Sanger et al., 1977; Biggin et al., 1983) with the universal primer, LMB2 (Duckworth et

al., 1981).

One of the four recombinants that hybridised with the oligonucleotide probe was found to contain sequences encoding the N-terminal region of the phosphate carrier protein. The insert in this recombinant had been released from the plasmid by digestion with EcoRI and BamHI. Two fragments were detected by ethidium bromide staining after agarose gel electrophoresis: an EcoRI -BamHI fragment (bases 1-579, Figure 2) and an EcoRI fragment (bases 628-1347, Figure 2) estimated to be 600 and 730 bp respectively. The sequences of the inserts were completed by synthesising appropriate oligonucleotides for use as primers in further sequencing experiments. After synthesis their concentrations were estimated by measurement of OD260. They were diluted in H20 at 0.2 pmol/,ul, and 2 ul of solution of the oligonucleotide was used in each sequencing reaction. In order to overlap these two sequences, a PstI restriction enzyme digest was investigated (see Figures 1 and 2 for restriction enzyme sites). Thereby it was found that the two original fragments were joined by a small EcoRI

1372

fragment (bases 579-628) that had not been detected by ethidium bromide staining after agarose gel electrophoresis. This small fragment was also cloned and sequenced. Compressions in sequencing experiments were resolved by substitution of deoxy-inosine triphosphate for dGTP in the sequencing mixtures, the products being separated by electrophoresis in a 6% gel (Mills and Kramer, 1979; Bankier and Barrell, 1983). Further confirmation of compressed sequences was obtained by cutting directly in these sequences with SmaI and Fnu4HI, and cloning and sequencing the resultant products. Computer methods DNA sequences were compiled into a data base with aid of the computer programs DBAUTO and DBUTIL (Staden, 1982a) and then analysed with ANALYSEQ (Staden, 1984). Protein sequences were examined with HYDROPLOT, a version of SOAP (Kyte and Doolittle, 1982) and DIAGON (Staden, 1982b). Northern analysis Total heart mRNA was extracted from fresh bovine heart tissue by the method of Chirgwin et al. (1979). As soon as possible after slaughter of the animal, cubes of tissue were frozen in liquid nitrogen, transported to the laboratory, and then pulverised under liquid nitrogen and homogenised for 60 s with a Polytron homogeniser in 4 M guanidinium thiocyanate containing 25 mM sodium citrate, pH 7.0, 0.1 M 2-mercaptoethanol and 0.5% sarkosyl. The homogenate was centrifuged (130 000 g, 18 h, 20°C) through a cushion of 5.7 M CsCl containing 0.1 M EDTA, pH 7.0. Poly(A)+ RNA was prepared by the method of Aviv and Leder (1972) except that NaCl was replaced by LiCl (Maniatis et al., 1982). Northern blot analysis was performed as follows. RNA samples were treated with glyoxal (Carmichael and McMaster, 1980), and fractionated on a 1.4% HGT agarose gel submerged in stirred 10 mM sodium phosphate, pH 7.0. RNA was transferred to Hybond N (Amersham International PLC, Amersham, UK) and cross-linked to the matrix by u.v. irradiation. Hybridisation was performed at 65°C with a 'prime cut' probe (Farrell et al., 1983) corresponding to the EcoRI-HindHl fragment (bases 629-1131, see Figure 2). Membranes were washed at 65°C four times in 6 x SSC, twice in 2 x SSC and twice in 0.2 x SSC (SSC is 0.15 M NaCl, 0.015 M Na citrate). Autoradiography was carried out for 3 days at -70°C with preflashed film and an intensifying screen.

Acknowledgements We thank Drs Kolbe and Wohlrab for sending peptides T1-T3 for sequence analysis, and Mr I.M.Fearnley for his help with these experiments. We are grateful to Mr T.V.Smith for synthesising oligonucleotides. P.N. was supported by a postdoctoral scholarship from the N.F.R.

References Adrian,G.S., McCammon,M.T., Montgomery,D.L. and Douglas,M.G. (1986) Mol. Cell Biol., 6, 626-634.

Aquila,H., Misra,D., Eulitz,M. and Klingenberg,M. (1982) Hoppe-Seyler's Z. Physiol. Chem., 363, 345-349. Aquila,H., Link,T.A. and Klingenberg,M. (1985) EMBO J., 4, 2369-2376. Arends,H. and Sebald,W. (1984) EMBO J., 3, 377-382. Aviv,H. and Leder,P. (1972) Proc. Natl. Acad. Sci. USA, 69, 1408-1412. Bankier,A.T. and Barrell,B.G. (1983) In Flavell,R.A. (ed.), Techniques in Nucleic Acid Biochemistry. Elsevier, Ireland, B508/1 - B508/31. Biggin,M.D., Gibson,T.J. and Hong,G.F. (1983) Proc. Natl. Acad. Sci. USA, 80, 3963-3965. Bisaccia,F. and Palmieri,F. (1984) Biochim. Biophys. Acta, 766, 386-394. Bouillard,F., Weissenbach,J. and Riquier,D. (1986) J. Biol. Chem., 261, 1487-1490.

Carmichael,G.G. and McMaster,G.K. (1980) Methods Enzymol., 65, 380-391. Chirgwin,J.M., Przbyla,A.F., MacDonald,R.J. and Rutter,W.J. (1979) Biochemistry, 18, 5294-5299. Duckworth,M.L., Gait,M.J., Goelet,P., Hong,G.-F., Singh,M. and Titmas,R.C. (1981) Nucleic Acids Res., 9, 1691-1706. Farrell,P.J., Deininger,P.L., Bankier,A. and Barrell,B.G. (1983) Proc. Natl. Acad. Sci. USA, 80, 1565-1569. Gay,N.J. and Walker,J.E. (1985a) Biochem. J., 225, 707-712. Gay,N.J. and Walker,J.E. (1985b) EMBO J., 4, 3519-3524. Horwich,A.L., Fenton,W.A., Williams,K.R., Kalonsek,F., Krans,J.P., Doolittle,R.F., Konigsberg,W. and Rosenberg,L.E. (1984) Science, 224, 1068-1074. Jaussi,R., Cotton,B., Juretic,P.C. and Schumperli,D. (1985) J. Biol. Chem., 260, 16060-16063.

Kaput,J., Goltz,S. and Blobel,G. (1982) J. Biol. Chem., 257, 15054-15058. Klingenberg,M. (1985) In Martonosi,A.N. (ed.), The Enzymes of Biological Membranes. Plenum Press, Vol. 4, pp. 511-553. Kolbe,H.V.J., Costello,D., Wong,A., Lu,R.C. and Wohlrab,H. (1984) J. Biol. Chem., 259, 9115-9120.

Mitochondrial phosphate carrier protein Kolbe,H.V.J. and Wohlrab,H. (1985) J. Biol. Chem., 260, 15899-15906. Kyte,J. and Doolittle,R.F. (1982) J. Mol. Biol., 157, 105-132. La Noue,K.F. and Schoolwerth,A.C. (1979) Annu. Rev. Biochem., 48, 871-922. Lomax,M.I., Bachman,N.J., Nasoff,M.S., Caruthers,M.H. and Grossman,L.I. (1984) Proc. Natl. Acad. Sci. USA, 81, 6295-6299. Maniatis,T., Fritsch,E.F. and Sambrook,J. (1982) Molecular Cloning. A Laboratory Manual. Cold Spring Harbor Laboratory Press, New York. Meijer,A.J. and van Dam,K. (1974) Biochim. Biophys. Acta, 346, 213-244. Messing,J. and Vieira,J. (1982) Gene, 19, 269-276. Mills,D.R. and Kramer,F.R. (1979) Proc. Natl. Acad. Sci. USA, 76, 2232-2235. Ohta,S. and Kagawa,Y. (1986) J. Biochem., 99, 135 -141. Proudfoot,N.J. and Brownlee,G.G. (1976) Nature, 263, 211-214. Sadler,I., Suda,K., Schatz,G., Kaudewitz,F. and Haid,A. (1984) EMBO J., 3, 2137-2143. Sanger,F., Nicklen,S. and Coulson,A.R. (1977) Proc. Natl. Acad. Sci. USA, 74, 5463-5467. Saraste,M. and Walker,J.E. (1982) FEBS Lett., 144, 250-254. Schatz,G. and Butow,R.A. (1983) Cell, 32, 316-318. Staden,R. (1982a) Nucleic Acids Res., 10, 4731-4751. Staden,R. (1982b) Nucleic Acids Res., 10, 2951-2961. Staden,R. (1984) Nucleic Acids Res., 12, 521-538. Thomas,A.P. and Halestrap,A.P. (1981) Biochem. J., 1%, 471-479. Viebrock,A., Perz,A. and Sebald,W. (1982) EMBO J., 5, 565-571. Vieira,J. and Messing,J. (1982) Gene, 19, 259-268. Walker,J.E., Feamley,I.M. and Blows,R.A. (1986) Biochem. J., 237, 73-84. Wieslander,L. (1979) Anal. Biochem., 98, 305-309. Received on November 24, 1986; revised on January 26, 1987

Note added in proof Since this paper was submitted, Aquila et al. [FEBS Lett., 212, 1-9 (1987)] have published a review containing a partial protein sequence for the bovine phosphate carrier. Residues 226-255 and 270-271, corresponding to a large segment of the third repeat, have not been determined. They find aspartic acid and glutamine at residues 87 and 107 respectively, whereas we find cysteine and asparagine at these positions.

1373