Biochbnica et Biophysica Acta, ! !27 ( i 992) 199-207

© 1992 Elsevier Science Publishers B.V. All rights reserved 0005-2760/92/$05.0U

199

BBALIP 53956

c D N A sequence and alternative m R N A splicing of surfactant-associated protein C (SP-C) in rabbit lung * !an Conne!ly and Fred Possmayer MRC Group in Fetal and Neonatal Health and Det'elopment and Departments of Biochemistry and Obstetrics and Gynaecology, University of Western Or!ratio, London (Canada)

(Received 15 October 1991) (Revised manuscript received 25 February 1992)

Key words: Neonatal respiratory distress syndrome; Surfact~mt associated protein C; Alternative splicing; Pulmonary surfactant; Lung; Phospholipid An 784 base pair (bp) copy DNA (eDNA) for the low molecular weight hydrophobic surfactant-associated protein C (SP-C) has been isolated from a Agtll eDNA librat~¢ constructed fr:ml fetal rabbit lung mRNA. The eDNA, which coded fi~r a 193 amino-acid proprotein with 6 bp 5' and 193 bp 3' untranslated segments, possesses considerable nucleic acid and predicted amino-acid homology with previously reported SP-C cDNAs. The predicted amino-acid sequence of the 35 amino-acid mature polypeptide shares 94-97% identity with human, rat and mouse SP-C and is 88-91% homologous to the mature proteins from bovine, porcine and canine lung. The last 12 amino acids of mature SP-C are highly hydrophobic and invariant. Alignment of the rabbit and human nucleic acid sequetici~s required introduction of a 27 bp gap in the rabbit sequence at a site corresponding to the exon-intron junction of the 5th exon of the human genomic sequence. Since previous studies have identified differential splicing at the 5' and 3' ends of the human 5th exon, we investigated the potential existence of alternative splicing of rabbit SP-C mRNA. Reverse transcription (RT) of total RNA followed by polymerase chain reaction (PCR) was used to establish the relative abundance of alternative splicing products from fetal and adult hmg and from rabbit kidney, placenta and liver. The relative abundance of the 250, 280 and 350 bp bands observed was the same in lung and other tissues. PCR amplification of genomic rabbit DNA indicated that the 350 bp fragment corresponds to the unspliced nascent transcript. The lack of developmental or tissue-specific abundance patterns implies the absence of secondary influences on SP-C mRNA polymorphism. Indeed, frec energy of formation calculations predicted the presence of hairpin structures favouring formation of the more abundant 250 bp form. These observations plus the absence of any effect of alternative splicing on SP-C protein structure led u~ to conclude a physiological role is unlikely.

Introduction Pulmonary surfactant is a complex mixture of phospholipids and surfactant apoproteins which stabilizes the lung by reducing surface tension in the terminal airways. Insufficient surfactant results in significant pulmonary pathology as occurs with the respiratory distress syndrome (RDS) of premature infants [1,2].

Correspondence to: F. Possmayer, Dept. of Obstetrics and Gynaecology, The University of Western Ontario, Room 9-OF 12, 339 Windermere Road, London, Ontario, Canada N6A 5A5. * The sequence data reported in this paper have been submitted to the EMBL/Genbank Data Libraries under the accession number X65078. Abbreviations: bp, base pair; PC, phosphatidylcholine; PCR, polymerase chain reaction; RT, reverse transcription; SP-, surfactant-associated protein.

Four distinct surfactant-associated protein have been identified and designated SP-A, SP-B, SP-C and SP-D [3,4]. SP-A and SP-D are collagen-like, calcium-dependent ieetins [5,6]. SP-B and SP-C are low molecular weight hydrophobic proteins with M r values of 8000 and 3500, respectively, produced through amino- and carboxy-terminal proteolytic processing of larger proproteins, which contribute to the adsorption and spreading of surfactant phospholipids to form the surface monolayer [3]. SP-B is also involved in the squeeze-out of unsaturated phospholipids, such as phosphatidylglycerol which results in a monolayer enriched in dipalmitoylphosphatidylcholine [7]. The present report describes the cloning of SP-C from a fetal rabbit lung A g t l l eDNA library and its sequencing. During the alignment of the nucleic acid sequence with the human eDNA, a 27 bp gap in the rabbit sequence was noted at a site corresponding to the exon-intron junction of the 5th exon of the human

200 genomic SP-C sequence. Alternative splicing has been reported for the 5th exon of human SP-C mRNA [8,9]. We therefore i:~vestigated how this might affect rabbit SP-C mRNA. Materials and Methods

Materials. All chemicals used were of reagent grade. Reagents for polyacrylamide gel electrophoresis came from BDH Chemicals (Toronto, ON). Radiochemicals were from Dupont/New England Nuclear (Mississauga, ON). Restriction enzymes, buffers and random primer labelling reagents were from Pharmacia Canada (Bale d'Urfe, PQ). T7 RNA polymerase, rNTPs, Taq polymerase and buffers were from Promega/Fisher Scientific (Toronto, ON). dNTPs and M-MuLV reverse transcriptase were from Boehringer-Mannheim (Dorval, PQ). Subcloning and sequencing procedures were conducted with the Bluescript vector (Stratagene, Palo Alto, CA) and the host strain JM109 (recA.(recAl, endAl, gyrAg6, thi, hsdRlT, SupE4 +, relAl, A-, A(lac-proAB), (F', traD36, proAB, laclQ, lacZ AglS))), Preparation ofprobes. The human SP-C eDNA was a kind gift of Dr. Jeffrey Whitsett, Department of Pediatrics, University of CJnclnnad Cotl,~ge of Medicine, Cincinnati, OH [10]. The eDNA probes were prepared using a random primer oligo labelling kit from Pharmacia Canada and ['~2P]a-dCTP (50 ~Ci; 3000 Ci/mmol). Unincorporated label was removed using a Pharmacia Nick column. Probes which incorporated less than 80% of the label were not used. Library screening, The human SP-C eDNA was used to screen a rabbit hgtll eDNA libraw constructed in this laboratow [11] using standard techniques [12] and Nytran membranes (Scheicher and Schuell, Keene, NH). Sequencing. Single-stranded DNA for sequencing was prepared from the Bluescript vector using the R408 ~elper phage and the manufacturer's protocol (Stratagene, Palo Alto, CA). The MI3 universal primer was used for sequencing directed from the vector into the eDNA (Fig. 1). The synthetic primers used in the sequencing directed from within the eDNA were: 1. S '

-

CCTCAAACGTCTTCTC

-

31

2. 51

-

CCTGAGTCCTAGATGT

-

31

3, 51 - T G G G A C T G A G C G A G T . 31 Sequence data were compared and analyzed using the PC/Gene program (lntelligenetics, Palo Alto, CA). Nucleic acid sequence alignments used the method of Myers and Miller [13]. RNA secondary structures were predicted using the method of Zuker [14]. The short 6 bp 5' noncoding portion showed highly variable conservation with these other species. The 3' nontranslatod portions possessed less than 50% identity. The free energy determinations used the free energy values compiled by Salser [15].

Nucleic acid isolation. New Zealand White rabbits of known gestation (day of breeding assigned day 0) were purchased from Riemans Fur Ranch, St. Agatha, ON. Fetal lung tissue was dissected from rabbits of 22-30 days gestation as well as adult lungs. The tissue was rinsed briefly in cold PBS and then frozen at -80°C until it was used. RNA was isolated from 100 mg to 200 mg wet weight tissue samples using the technique of Auffray and Rougeon [16]. RNA concentration was determined by measuring the absorbance at 260 nm [12]. Total adult rabbit kidney DNA was isolated using proteinase K/phenol extraction, as described previously [ 12]. Characterization of rabbit SP-C m ~ A alternative splicing. 1 /~g of total rabbit RNA, isolated as described below, was reverse transcribed in 20/~i with 1 /~g of oligo [dT], 10 units of M-MuLV reverse transcriptase, 50 mM Tris (pH 8.3), 6 mM MgC! 2, 40 mM KCI, 10 mM dithioerythritol, 0.5 mM dATP, 0.5 mM dCTP, 0.5 mM dGTP and 0.5 mM dTTP at 37°C for 1 h. The reaction was terminated by heating at 95°C for 10 min. 1 /~1 of the RT reaction was used as the substrate for a PCR which was specific for a section of the rabbit SP-C mRNA bracketing the putative 5th exon-intron junctions. The PCR reaction was conducted in a volume of 100/~! which contained 5 units Taq DNA polymerase (Promega/Fisher Scientific), the manufacturer's buffers and 100 pmol of each of the following primers: 1. 2.

5'-GGAAGAATTCGCTGCTACCTCATGAAGAT-3' 5'-GGTTGAATTCTTGACGGAAGAACCCTTCTGC-3 m

The first 10 bases of these primers constitute EcoRl restriction sites; the remaining bases are complimentaw to sequences flanking the 5' and 3' ends of the putative rabbit SP-C 5th exon. The PCR reaction was annealed at 50°C, extended at 72°C and denatured at 95°C for 60 cycles. The final cycle extended at 72°C for 12 min. PCR amplification of genomic DNA used 50 ng of genomic kidney DNA and the conditions described above. 60 PCR cycles were required, since 40 cycles of PCR produced only a small quantity of product from the 22-day RNA samples and did not produce detectable products from the kidney, lung or placenta RNA samples. The mineral oil phase of the PCR reaction was removed oy extraction with chloroform. The aqueous phase was precipitated with ethanol and sodium acetate [12]. The PCR products were analyzed using electrophoresis through a 6% (28:2) polyacrylamide, 50 mM Tris, 50 mM boric acid and 1 mM ethylenediaminetetraacetic acid gel. The bands were stained with ethidium bromide and photographed using Polaroid type 55 Professional film (Polaroid, Cambridge, MA).

201

Results

TABLE !

The rabbit SP-C cDNA sequence 40000 plaque-forming units from a lambda gtll eDNA library, constructed with mRNA isolated from fetal rabbit lungs of 30 days gestation, were screened with an 858 bp human SP-C eDNA [10] using standard techniques [12]. Over 100 positives were obtained, of which 10 were selected for further screening. After three more rounds, six positives remained. The colony with the strongest signal was isolated as a putative rabbit SP-C eDNA. The lambda eDNA insert was subcloned into the EcoRI site of the Bluescript vector (Stratagene, Palo Alto, CA) for sequencing as diagrammed in Fig. 1. Single stranded DNA was prepared from the Blueseript vector using the R408 helper phage and the manufacturer's protocol (Stratagene, Palo Alto, CA). The resultant nucleic acid sequence of the rabbit SP-C eDNA (Fig. 2) was aligned with the mouse and human SP-C eDNA sequences using the method of Zuker [14]. The short 6 bp 5' noncoding portion showed highly variable conservation with these other species. 'The 3' nontranslated portions possessed less than 50% identity. The percentage identity among the sequences is tabulated in Table I. The translated portions of the rabbit, rat, human and mouse SP-C cDNAs shared over 70% sequence identity. Alignment of the human and rabbit eDNA sequences required insertion of gaps in the rabbit sequence. Interestingly, two of these gaps corresponded to the exon-intron splice junctions sites occurring at the 5' and 3' ends of the 5th exon of the human gene. Previous studies have demonstrated that the 5th exon of the human exhibits alternative splicing [8]. We therefore examined the possibility that the rabbit SP-C is alternatively spliced, similarly to the human SP'C. For convenience, the locations in the rabbit SP-C eDNA sequence where gaps were required to align the rabbit and human sequences will be referred to as the putative rabbit SP-C 5th exon splice junctions. The amino-acid seqnences reported for the 34-35 amino-acid canine, bovine and porcine mature SP-C showed 88-91% identity with the sequence predicted

Comparison of the known SP-C cDNA sequences h~ different ~pecies

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

5 I Eco RI

I Psi i

~i

I Eco RI

Fig. 1. Strategy for sequencing the rabbit SP-C eDNA. The rabbit SP-C eDNA was sequenced with the modified T4 DNA polymerase and the dideoxy technique. Sequences which were derived with the universal primer are depicted with solid arrows. Dashed arrows indicate sequences derived with the custom synthesised primers described in the text. The Pst I site was used in conjunction with the Pst l site in the Biuescript multiple cloning region to subclone the 3' end of the SP-C eDNA for sequencing with the universal primer.

Species

Percentage identity with rabbit SP-C 5' noncoding (%)

primary translation product

mature peptide (%)

3' noncoding (%)

84 78 77

90 90 86

27 25 48

(%) Human [8] Murine [9] Rat [19]

83 50 17

for the rabbit protein while the human, mouse and rat mature polypeptides exhibited 94-97% homology with the rabbit sequence (Table 11). All predicted sequences maintained the length and hydrophobicity of the central polyvaline tract as well as maintaining absolute conservation of the final 12 amino acids in the active peptide. Characterizalion of the splicing of the putative SP-C 5th exon in the rabbit The potential alternative splicing of the predicted rabbit SP-C 5th exon splice junctions was examined by analyzing the products after PCR amplification of reverse transcribed total rabbit RNA and genomic DNA. Two primers were chosen which were complementary to the sequences flanking the splice junctions of the putative rabbit SP-C 5th exon (Fig. 2). Genomic rabbit kidney DNA as well as total rabbit RNA from fetal lung, adult lung, kidney, placenta, and liver were analyzed. The products of the PCR reactions were analyzed with polyacrylamide gel electrophoresis and stained with ethidium bromide (Fig. 3). 6n PCR cvcles were used, since 40 cycles of PCR produ,. ~4 only a small quantity of product from the. 22-day RNA samples and did not produce detect:ble products from the kidney, liver or placenta RNA samples. 60 cycles of PCR amplify a single copy approx. 25000-fold more than the 100 pmol ef primer wiU allow, assuming 100% efficiency in each cycle Therefore, if substrate is present, 60 cycles should deplete all available primer. Using 60 cycles, bands of 350, 280 and 250 bp were found in all RT-PCR reactions. The three RT-ICR bands were of approx, the same relati~,e size and intensity in all samples. A PCR reaction was also performed using the same primers and genomic rabbit kidney DNA as a control reaction which would indicate which of the RT-PCR products was the product of unspliced nascent transcript. The DNA-PCR reaction produced two major bands of 350 bp and 320 bp, and th:ee minor bands of 340, 290 and 250 bp. The minor bands were not detectable when the DNA-PCR reactions were annealed at 55"C, as opposed to 50°C. Each of the three bands round in the acrylamide gel of the RT-PCR products was excised, subcloned into the Bluescript vector and sequenced. The sequence of

202 __

-

RSPC

CG

,

,,,,,~

CJ~.G&TGG&~TGGGCAGC~G&GGCCTTG&TGG

I

:|::|====

.*:$~:1|=:::||::

:

-37

::|:||

HSPC

- GGAG&GCATAGCACCTGCAGCAAGATGG~TGTGGGCAGCAAAGAGGTCCTG&TGG

-55

RSPC

-

HSPC

- AGAGCCCGCCGGACTACTCCGCAGCTCCCCGGGGCCGATTTGGCATTCCCTGCTG - 1 1 0

RSPC

-

RGRGCCCACCGG&CTACTCAGC&GCTCCCCGGGGCCGCTTCGGCATCCCCTGCTG - 9 2 lllIIII

IlIIlIIllII

:IllIll¢IIlllIlll

II

lllll

IIlllil8

CCCAGTGCACCTCAAACGTCTTCTCATCGTGGTCGTCGTGGTGGTCCTCGTGGTC IllIIIlIllll

IIlll

11811

81181III

II

811811111881

8

-147

III

HSPC

-

CCCAGTGCACCTGAAACGCCTTCTT&T~TGGTGGTGGTGGTGGTCCTCATCGTC - 1 6 5

RSPC

-

()T~A2".r~'I'GGGGGCCCS~CTC&TG~GCT~'.&CA2~&GCCAGA,~r.AC&CC~ lill1811118118

I1881181188111

81111111188188811118111

-202

I

HIP(:

-

GTGGTGA~"J~GTGGG&~CTCATGGGTCTCCACATGAGCCAGAAAr.ACACGG

RSP¢

-

&OATGGTC~fAOAO~,TGAGCAT~GG~'~GCCGG&GGTCCAGCJ~GCGCCI~;GCACI ' - 2 5 7

HSPC

-

~O&TGG~CTGG&GATGAGCAT~GCGCCGGAAGCCCAGCAACGCCTGGCCCT

RIPC

-

G&GCG>GGG~"GGGCACC&CGGCCACTTTCCCCATCG~CTCC~CCGGCATTGTC - 3 1 2

XSPC

-

G>O&GCACCTGGTTACCACTGCCACCTTCTCCATCGGCTC¢&CTGGCCTCGTG

Illtlll

Ill

II

Iillllillll

Ill

88

Illllllllll

lllll

811ll

I

Ill

IIIIll

IIIIiIl~

Illllllil1818

-220

II

181

8

-275

II

-330

PI 28PC

-

&CCTGCG&C~&CCAGCGGCTCCTG&TCGCCT&TAAGCCAGCCCCGGGAACC2~.~;,~ 8

iISlllllll

Ill

8llllllIIll

8llllllllll

lI

-367

lIIIIII

HSP¢

-

RSPC

- GCTACC'I~ATGAAGATGGCTCC&GACAGCATCCCCAGTCTGGAGGCTCTGGCT&G - 4 2 2

HSPC

"

GCT&CATCATGIa.AGAT&GCTCCAG&G&GCATCCCCAGTCTTGAGGCTCTCAATAG - 4 4 0

RSP¢

-

AAAATTCC& 8888 88~8

GTG?ATGACTA(:r-,AGr.J~GCTGCTGATCGCCTAC&AGCCAGCCCC'J~',GC&CCTGCT

Illll

lllllllllI

llllllll

llIllIlllIIlIl

IlllIIIl

-385

IIl

HSPC

-

GGCCAACCCTGCAGAGCCT - 4 5 0 ;8t88i c t 88~8 8888 AA&AGTCCACAACTTCCAGATGGAATGCTCTCTGCAGGCCAAGCCCCCAGTGCOT - 4 9 5

RSPC

-

CCCACTCAGCGGGGCCAGGACAAGGGG(:r.,AGCCGCACGGCt~3GCGTCC~CCGGAG I

II

~

Ill

llIIllll

lillll

18

818

l

l

81

-505

111888811

XSPC

-

ACGTCTAAGC~;GCCAGGr.,AG&GGGGO~AGATGr.,AGGCTC&GCACCCTCCGG&G - 5 5 0

RSP¢

-

G&O&GCTGGC~~CGCAGC~GTGAGCACCCTG'JL'G~3GGG&GGTGCCTCT :

XIPC

-

II

I

:IlllllIIllli

lllllll

lI:I:llllll:

IIIlI:ll

-560

I:

GGGACCCGGCCTTCCTGGGCA~GCCGTGAACACCCI~TGTGGCG&GG~CCGCT

-605

_t.. RSPC

-

C&TCT&CATCT&GG&CTC . . . . . . . . . . . |

H81~

-

RSl:~

-

I|8|||~I||||I

-

5

:|||18

|

1:8||185=

=

-604

Ill

QTACTAr..NI~f&GGACG~CCGGTG&GCAGGGTCAG~AAG¢CCr.,AACGGGAA - 6 6 0 P2 r-,AGG&GGG&CC~GCCAGr-AGAAGGGT . . . . . . CTTCCGTr~AG&GGr..AGGACGCT - 6 5 3 t

RSPC

&GGGTCTGCGGAAGCCCCG&GGGGGT

:

8

|

::tl

:8

:

8

|88

:

88:8

8¢

:8|

GCGTCTGCCCGCACC'I~CAGGGCCGGCCGTGGGCG&CCGG&GCTGCGGGGAG&GG It

:llll:l

:::I

:

l:

:

I:

Ill

l

:1

-708

::IlllllI

HSP¢

-

GC'F~CTGCCCACACCGCAGGG&CA/~CC~TGG&GAP.ATGGG&GCT~AGAGG

-770

RSI~

-

CAGCCCCr..AGGG. . . . . . . . . .

-753

I

XSP¢

-

II

¢Cr-,AGGGG'J~3GGGGG&CCCCCGCCACAGGAG~ lllll:ll

III

II

II

llIlll

IIII

&'I~GGAG'I~3GGr.AGAGGTGGCACCCAGGGGCCCGGGAACTCCTGCCAr.jt~CAGAA - 8 2 5

L R81:¢

-

TAAAGr.J~GC'JL~ACCGAAAA~ |IIIIII

IIII

IIII

-784

IlllIIIIII

XSP~

* TAAAGCAGCC/~ATTGAAA--AN~AA,qA

IdentLt¥

~ 617

-854

(78.7q)

Fig, 2, Nucleic acid sequence of a rabbit SP-C cDNA (RSPC) and human SP-C cDNA (HSPC) [20]. The first overlined sequence is the Kozak consensus sequence for eukaryotic translation initiation [23]. The second overlined sequence is the sequence which codes for the mature SP-C protein, The third overlined sequence is an 18 bp pol]~norphism in the splicing of the 5' end of the human SP-C 5th exon. The fourth overlined sequence codes for the translation stop codon. The fifth overlined sequence is an 8 bp polymorphism in the splicing of the 3' end of the human SP-C 51h exon, The sixth overlined sequence is the polyadenylation signal sequence. The underlined sequences P1 and P2 are the locations for the PCR primers used in the analysis of the alternative splicing of exon 5.

203 each band was confirmed by sequencing the products from at least two separate RT-PCR reactions. The relationship among the three sequences is diagrammed in Fig. 4. The acceptor, donor and predicted internal branch points of the different splicing patterns closely match reported consensus sequences [17]. No significant differences were observed in the strengths between the consensus sequence and the donor branch point and acceptor sequences of the alternate splice patterns (Table III). However, using the method of Zuker [14], two hairpin loops were predicted for the primary transcript. Using the free energy values tabulated by Salser [15], the hairpin loops have free energies of formation of - 6 3 . 3 kcal and - 18.3 kcal. Using the free energy calculations of Freier [18], the hairpin loops have free energies of formation of -14.1 kcal and -5.1 kcal, respectively (Fig. 5). In each case the internal branch site of the minor splicing product is within the stronger hairpin loop and the predicted internal branch site of the major product is in the weaker hairpin.

350> 28o. 5o

:C t3

tV

350>11

Discussion

Conservation of the nucleic acid and predicted amino-acid sequences of the rabbit SP-C eDNA was examined by comparison with the published SP-C sequences of other species. The translated portion of the rabbit SP-C eDNA possesses 77-84% identity with the corresponding nucleic acid sequences of the human, rat and mouse [9,19,20]. The portion of the rabbit SP-C eDNA sequence which codes for the mature SP-C peptide exhibits over 85% identity with the published eDNA sequences from three other species (Table I). The predicted amino acid sequence of the mature SP-C peptide has over 85% identity with the sequences of the mature SP-C peptides of the cow, pig and dog, and 94-97% identity with the :at, mouse and human sequences [9,19-22]. The last 12 amino acids in the predicted mature rabbit SP-C peptide possess 100% identity with all known species (Table 11). The hydro-

2ao. 5O ll Fig. 3. Acrylamide gel electrophoresis of the reverse transcribed and PCR amplified SP-C 5th exon from rabbit RNA and DNA. The total RNA used in the RT-PCR reactions were isolated from adult rabbit lung (lane A2 and BI), liver (lane A3) and kidney (lane A4), as well as from placenta (lane AS) and 22-day fetal lung (lane B2). Genomic kidney DNA was PCR amplified as a control (lane Al). The reaction products were separated on a 6% acrylamide IXTBE gel and stained with ethidium bromide. Gel A compares the relative abundance of the PCR products in different tissues, while Gel B compares die ontogenic expression of the PCR products. phobicity of the middle 20 amino acids, including the presence of a unique polyvaline tract, is also conserved across all species. Over 70% of the amino-acid substitutions observed between all species occur in the first

TABLE 11 Comparison of known mature SP-C amino acid sequences

Species

Rabbit Human [I0,21] Murine [9] Bovine [22] Porcine [22] Canine [23] Rat [19]

Amino-acid sequence ('-' indicates identitywith rabbit sequence)

Homology with rabbit (%)

10 20 30 F61PCCPVHLKRLLIVVVVVVLVVVV IV6ALLM6L

!oo

-R .

.

.

.

.

.

I-

-R--L ...... -R ...... ---CF-SS QR

NI N .....

V

............. ..........

- m

94 97 88 91 88 94

204 3.• ATCTAGGACTCA~AGCAGGTGGAGGAGGGCTTGGGCACCGCGA . . . CC'~.,CGGCGGACCC~.,CGClGGGACCCACGCAGGCTCACTCCCTCTCCCTCTCCAT... T, [

\ \

•

]

" " "CGCGCACCAGI~GA~CCTTTGCCTGTCACG/~GGTCTGCGGAAG ATCTAGGACTCAGGACAKCGTGCCTTTGCCTGTCACGCAGGGTCTGCGGAAG ATCTAG~~~TCTGCGGAAG

.

Fig, 4. The alternate splicing arrangement of the rabbit SP-C 5th exon, Part of the nuclcotide sequences obtained by sequencin~ the 350, 280 and ~ 0 bp RT-PCR bands are given. All three sequences start at nucleotide 568 and terminate at position 592 in Fig, 2. Sequence I corresponds to part of the unspliced primary transcription. Sequence 2 depicts part of the minor splicing product. Sequence 3 depicts part of the major splicing product corresponding to nucleotides 568-592 of Fig, 2, The solid slanting lines indicate the initiation and termination points for the excised sequences of the putative 5th intron which create the major and minor products, The bracketed overlined sequences in the genomic sequence are the predicted hairpin loops depicted in Fig, 5.

10 amino acids of the mature peptide. These sequence comparisons demonstrate that SP-C has considerable phylogenetic conservation. Inspection of the 5' end of the rabbit SP-C eDNA revealed the presence of two inframe A'FG start sites of positions 7 and 13 (Fig. 2). Only the initial inframe ATG site corresponding to position 7 is present in the human eDNA sequence, suggesting it is this site which is involved in initiating protein synthesis. However, it should be noted that the second ATG site, which is also present in the rat and mouse cDNAs, is located within a more favoured Kozak consensus sequence [23]. Therefore the possibility that the second ATG start site, not present in the human, could be partially or predominantly used in these other species cannot be eliminated. Alternative intron splicing is used in a number of systems to regulate th~ abundance of mature mRNA. The formation of proteins such as troponin, myosin, calcitonin, and ultrabithorax are known to arise through regulation of gene expression by alternative splicing [24]. Alternative splicing can allow func-

tionally different proteins to arise from the same gene in a developmentally- or tissue-dependent manner [25]. Previous studies have demonstrated that the human SP-C is alternatively spliced at the 3' and 5' ends of the 5th exon [8]. The alternative splicing which occurs at the 5' end of the 5th exon removes 18 nucleotides, thereby changing the human proprotein by six amino acids. The alternative splicing of the 3' end of the 5th exon, which is in the Y non-translated region of the mRNA, inserts or removes eight nucleotides. A function has not yet been ascribed to the alternative splicing of the human SP-C 5th exon [8]. Since alternative intron splicing can be used as a regulatory mechanism [25], the possible presence of differential intron splicing in rabbit SP-C mRNA was investigated. PCR analysis of genomic rabbit DNA produced two major bands (Fig~ 3). The 350 bp band produced by PCR of genomic DNA corresponds in size to a 350 bp band found in the RT-PCR, implying that the 350 bp species is the unspliced primary transcript. The RTPCRs of the liver, lung and kidney RNA samples

TABLE !!! Comp~rLq~n of the consensltv splice sequence with the sequences of tile minor and major rabbit SP-C alternative splicing sequence Sequence

Donor site

Internal branch point

Accepter site

Percent identity

Con~nsussequence*

AeGUAAGU . . . .

YNYRAY . . . .

YYYYYYYYYYYYYYYNCAGG

100

Minorproduct

AGGUGAG¢ . . . .

CTCCAT . . . .

UCCAUCGCGCACCAGACAGG

73

Majorproduct

AGGUGAGC . . . .

GACAAC . . . .

GCCUUUGCCUGUCACGCAGG

79

* From Re£ 17, Y is p]aimidine, R is purine, N is any base.

205 produced bands of 350, 280 and 250 bp (Fig. 3). The 280 bp and 250 bp RT-PCR species represent products of the alternative splicing of the putative rabbit SP-C 5th intron. The 250 bp band was ideaticai in sequence to the original rabbit SP-C eDNA isolated from the lambda library. The 280 bp product corresponds to an SP-C mRNA transcript, which is the product of the alternative splicing of the 3' end of the putative rabbit 5th exon. The alternative splicing of rabbit SP-C mRNA occurs in the non-translated sequence of the mRNA, eliminating the possibility of this alternatively spliced product resulting in a structurally different SP-C protein. Tissue specific RNAs can be amplified from non-expressing tissues [26] due to a low level of spurious transcription of all genes which is not believed to have physiological significance [27]. Previous studies have demonstrated that SP-C mRNA expression is lungspecific [6,27]. SP-C mRNA was not detec:ed in the

rabbit liver and kidney RNA samples by Northern blotting [28]. The low level of spurious transcription of SP-C mRNA in liver and kidney were used as examples of the rabbit SP-C default splicing pattern. The same pattern was observed in fetal lung, adult lung, kidney and liver samples. If the alternative splicing of SP-C has a regulatory role, a difference in the relative abundance of the products of splicing would be anticipated between lung and other tissues. There were no differences in the relative abundance of the RT-PCR products, either between lung samples of different gestational ages or between the different tissues (Fig. 3). The levels of the mRNAs for surfactant proteins increase markedly between 24 days and term [28]. The profile observed with RNA from fetal lung at 22-days gestation and adult lung was also obserJed at term (data not shown). Absence of any change in the SP-C mRNA pattern during this period suggests the pattern is not influ-

CAGG U G

A.

Cu°' o O.c : u

C,OoO,

.c A.uc

cuc:a o

o,.,Oo.o

Oo,O S. ooo ,ooo. c oo ° CG

O ®C A A U U

AGG G~ U

C AG

A C A ACGU

B • U Co cGC

G

• •

G

A,, AC.AoUG' "c : UUucC

O e O -. A~_ A%Ju ° --o

CUG ~

GG

AU

Acoc Fig. 5. The predicted structure of the hairpin loops in the rabbit SP-C 5th intron. Segments A and B over!:ncd in sequence I of Fig. 4 are arranged in hairpin structures as predicted by the algorithm of Zuker [14], The free energies of formation .,i ,.c hairpin loops were calculated using the methods and values of Frier [18],whichconsider the penalties for internal loops and bulges.

206 enced by the time and humeral influences involved in preparing fetal lung for extrauterine life. Thus, although all possible gestational ages and tissues were not included in this survey, it appears unlikely that the alternative splicing of the rabbit SP-C 5th intron plays a regulatory role in the regulation of rabbit SP-C mRNA abundance. This observation, in combination with the lack of an effect of alternative splicing on the SP-C amino-acid sequence, suggests that the alternative splicing probably does not have a physiological role. This conclusion is supported by the observation that alternative splicing has not been detected in the murine SP-C mRNA [9]. However, it must be stressed that a possible effect of alternate splicing on relative stability cannot be completely eliminated. A regulatory role for the alternative splicing of SP-C would lead us to expect the presence of a regulating transacting factor which could influence the splicing. In the absence of a regulatory role, and therefore any hypothetical transacting factor, the cause of the alternative splicin: should be restricted to the sequence of the 5th intron. There are no significant differences between the strengths of the predicted consensus splice sites of either the major or minor splicing product (Table Ill) [17]. However, examination of the potential secondary structures predicted by the method of Zoker [14] could explain the differential use of the two possible splice sites. The internal accepter site of the minor splicing product lies within a hairpin loop which has a free energy of formation of - 63.3 kcai using the values of Salser [15] and - 14.1 kcal using the values of Freier [18] (Fig. 5). The predicted accepter site of the major splicing product is also within a hairpin loop whk:h has a free energy of formation of only - 18.3 kcal using the values of Salser [15] and -5.1 kcal using the values of Freier [18] (Fig. 5). The values calculated using the method and values of Freier [18] take into account the penalty for internal loops and bulges in the hairpin, which are not taken into account in the more popular method and values of Salser [15]. We speculate that the major hairpin loop could interfere with the assembly of the spliceosome at the first splice site, thereby shifting the equilibrium of spliceosome formation to the next available consensus splice site. If the alternative splicing of the putative rabbit 5th intron is due to secondary structure in the 5th intron, then it would be expected that the mouse SP-C 5th intron, which does not exhibit alternative splicing [9], either would not have a predominant secondary structure or else would have a single large hairpin encompassing the entire int~on region, in contrast, the human 5th intron would be expected to have secondary structure similar to the rabbit. This appears to be the case since, using the values of Salser [15] and the method of Zuker [14], a single hairpin loop of -95.8 kcal is predicted for the mouse '~P-C 5th intron, while

the human 5th intron has two hairpin loops similar in organization to the rabbit, with free energies of formation of -60.1 kcal and -18.1 kcal. Although secondary structure has been shown to affect alternative splicing in viruses and in vitro [24,25], (its effect on the splicing patterns of eukaryotie mRNA has not been reported previously). The occurrence of non-functional alternative splicing is plausible, but to our knowledge has not been previously observed in a eukaryotic system. The SP-C alternative splicing observed with the rabbit does not have any apparent metabolic cost nor does it have an effect on the gene product. In evolutionary terms the alternative splicing would be neutral and therefore not subjected to selection pressure. In summary, rabbit SP-C has considerable phyiogenetic conservation of nucleic acid and amino-acid sequence. The predicted rabbit 5th intron is alternatively spliced such that 27 bases are either present or absent from the mature transcript. The alternative splicing may be due to the presence of secondary structures in the 5th intron, which interferes with spliceosome formation. The alternate splicing does not appear to have any physiological significance.

Acknowledgments These studies were supported by a Natural Science and Engineering Research Council (Canada) Strategic grant and by grants from the Medical Research Council of Canada. We would like to express our gratitude to Ms. Carol Richardson and Ms. Carol Ford for assistance during the initial stages of this work. We would also like to thank Dr. Jeff Whitsett, Dept. of Pediatrics, University of Cincinnati College of Medicine for providing the human SP-C eDNA clone and Drs. Geoff Hammond, George Mackie, and Mr Kevin Inchley, University of .Western Ontario for helpful advice. Ms. Barbara Lowery contributed excellent editorial assistance.

References i Robertson, B. and Lachmann B. (1986) Exp. Lung Dis. 14 279-310. 2 Jobe, A. and Ikegami, M. (1987) Am. Rev. Resp. Dis. 136, 1256-1275. 3 Possmayer, F. (1988) Am. Rev. Resp. Dis. 138, 85-89. 4 'Persson, A., Chang, D. and Crouch, E. (1990)J. Biol. Chem. 265, 5755-5760. 5 Hawgood, S. and Shifter, K. (1991) Annu. Rev. Physiol. 53, 375-394. 6 Weaver, T.E. and Whitsett, J.A. (1991) Biochem. J. 273,249-264. 7 Yu, S. and Possmayer, F. (1990) Biochim. Biophys. Acta 1046, 233-241. 8 Glasser, S.W., Korfhagen, T.R., Perme, C.M., Pilot-Matias, T.J., Kister, S.E. and Whitsett, J.A. (1988) J. Biol. Chem. 263, 1032610331.

207 9 Glasser, S.W., Korfhagen, T.R., Biuno, M.D., Dey, C. and Whitsett, J.A. (1990) J. Biol. Chem. 265, 21986-21991. !0 Glasser, S.W., Korfhagen, T.R., Weaver, T., Pilot-Matias, T., Fox, J.L. and Whitsett, J.A. (i~87) Proc. Nat!. Acad. Sci. USA 84, 4007-4011. ! 1 Xu, J., Richardson, C,, Ford, C., Spencer, T., Yao, L., Mackie, G., Hammond, G. and Possmayer, F. (1989) Biochem. Biophys. Res. Commun. 160, 325-332. 12 Sambrook, J., Fritsch, E.F. and Maniatis, T. (1989) Cold Spring Harbour Press, Cold Spring Harbour, New York. 13 Myers, E.W and Miller, W. (1988)CABIOS 4, 11-17. 14 Zucker, M. and Stiegler, P. (1981) Nucleic Acids Res. 9, 133-148. 15 Salser, W. (1977) Cold Spring Harbor Symp. Ouant. Biol. 42, 985-1002. 16 Auffray, C. and Rougeon, F. (1980) Eur. J. Biochem. 107, 303314. 17 Sharp, P.A. (1987) Science 235, 766-771. 18 Freier, S.M., Kierzek, R., Jaeger, J.A., Sugimoto, N., Caruthers, M,H., Neilson, T. and Turner, D.H. (1986) Proc. Natl. Acad. Sci. USA 83, 9373-9377.

19 Fisher, J.H., Shannon, J.M., Hofmann, T. and Mason, R.J. ~1989) Biochim. Biophys. Acta 995, 225-230. 20 Warr, R.G., Hawgood, S., Buckley, D.I.. Crisp, T.M., Schilling, J., Benson, B.J., Ballard, P.L., Clements, J.A. and White, R.T. ~.1987) Proc. Natl. Acad. Sci. USA 84. 7915-7919. 21 Johansson, J., Jornvall, H., Eklund, A., Christensen, N., Robertson, B. and Curstedt, T. (1988) FEBS Lett. 232, 61-64. 22 Johansson, J., Persson, P., Lowenadler, B., Robertson, B., Jornvail, H. and Cursedt, T. x,~,1,¢1°c~1~FEBS l. .,~tt . . 281~ 119-122. 23 Kozak, M. (1983) Nucleic Acids Res. 12, 857-872. 24 Andreadis, A., Gallego, M.E. and NadaI-Ginard, B. (1987) Annu. Rev. Cell Biol. 3, 207-242. 25 Smith, C.W.J., Patton, J.G. and NadaI-Ginard, B. (1989) Annu. Rev. Genet. 23, 527-577. 26 Sarkar, G. and Sommer, S.S. (1989) Science 244, 331-334. 27 Gross, I. (1990) Am. J. Physiol. 259, L337-L344. 28 Connelly, I,H,, Hammond, G.L., Harding, P.G.R. and Possmayer. F. ( 1991 ) Endocrinology 129, 2583-2591.