Transcription of the chicken anemia virus (CAV) genome and synthesis of its 52-kDa protein

Gene. I 18 (1992) 267-27 1 0 1992 Elsevier Science Publishers

GENE

B.V. All rights reserved.

261

0378-l 119/92/$05.00

06588

Transcription of the chicken its 52-kDa protein (RNA transcript;

polycistronic;

overlapping

Mathieu H.M. Noteborn”, Alex J. van der Eb a

anemia virus (CAV) genome and synthesis of

ORFs;

Northern

Onno Kranenburga,

blot; dot-blot

Alt Zantema”,

analysis;

S 1 mapping;

in vitro translation)

Guus Kochb, Gerben F. de Boerb and

Ii Laboratory for Molecular Carcinogenesi.~. Sylvius Laboratory. University of Leiden, Leiden, The Netherlands. Tel. 131-71)276115; and h Central Veterinar)j Institute.

Virology Department.

Received

by H. van Ormondt:

Leb:md.

The Netherlands.

21 January

Tel. (31-3200) 76611

1992; Revised/Accepted:

21 March,‘28

March

1992; Received

at publishers:

7 May 1992

SUMMARY

This paper describes the expression of the chicken anemia virus (CAV) genome, a recently characterized single-stranded circular-DNA virus of a new type [Noteborn et al., J. Virol. 65 (1991) 3131-31391. The major transcript from the CAV genome is an unspliced mRNA of about 2100 nucleotides (nt). Its transcription start point and poly(A)-addition site are located at nt 354 and 2317 of the CAV sequence, respectively. In vitro translation experiments provide evidence that the major CAV open reading frame encodes a 52-kDa protein by using the fifth AUG as a start codon of the unspliced CAV mRNA.

INTRODUCTION

Chicken anemia virus (CAV), formerly called chicken anemia agent, transiently causes severe anemia as a result of the destruction of erythroblastoid cells in bone marrow, and immunodeficiency due to the depletion of cortical thymocytes in young chickens (Jeurissen et al., 1989). CAV was first isolated in 1979 by Yuasa et al. (1979). CAV infections are encountered worldwide, and so far only one serotype has been described (McNulty, 1991). Southern analysis and restriction-enzyme mapping showed only

Correspondence

to: Dr. M.H.M.

Carcinogenesis,

University

The Netherlands.

Noteborn,

of Leiden,

Tel. (31-71)276113;

Laboratory

for Molecular

P.O. Box 9503, 2300 RA Leiden, Fax (31-71)276125.

Abbreviations: aa, amino acid(s); bp, base pair(s); CAV, chicken anemia virus; ds, double strand(ed); kb, kilobase or 1000 bp; MuLV, murine leukemia virus; nt. nucleotide(s); ORF, open reading frame; PA, polyacrylamide; Pollk, Klenow (large) fragment of E. coli DNA polymerase I; ss, single strand(ed);

tsp, transcription

start point(s).

minor differences between various field isolates (Noteborn et al., 1992). CAV is a small virus with a virion diameter of 23-25 nm (Gelderblom et al., 1989; Todd et al., 1990). Its genome is a circular ss DNA of about 2.3 kb. Only the minus-strand DNA is encapsidated (Noteborn et al., 1991). One (major) 50-kDa protein was detected in purified virus (Todd et al., 1990). CAV multiplies via a circular ds replicative intermediate. Recently, this 2.3-kb ds DNA was isolated from CAVinfected lymphoblastoid cell lines and cloned in a bacterial vector. The cloned material apparently represents the complete CAV DNA genome and contains all the elements required for the CAV replication cycle and pathogenicity (Noteborn et al., 1991). The cloned CAV genome was proven to be representative for CAV isolates from the field (Noteborn et al., 1992). The genome comprises three partially or completely overlapping ORFs encoding putative proteins of 51.6, 24.0, and 13.6 kDa. The CAV genome contains one obvious promoter upstream from the ORFs and only one poly(A)-addition signal downstream. All ORFs are located on the same DNA strand (Noteborn

268 et al., 1991).These data suggested that the use of the various ORFs may be regulated by alternative splicing, or that CAV encodes a polycistronic mRNA.

A

B

1

In this paper, we report that we detected only one (major) unspliced polyadenylated RNA transcript in CAV-infected cells and provide experimental evidence that the major CAV ORF encodes a 52-kDa protein by using the fifth AUG as start codon of the CAV mRNA.

hours

2 d

kb

-0 -8

4.0-16 2.0EXPERIMENTAL

AND DISCUSSION

-24

1 .o(a) Transcription of the CAV genome To study the transcription of the CAV genome,

0.5MDCC-

MSBl cells (Yuasa, 1983) were infected with CAV-Cux-1 (von Biilow et al., 1983). Two days after infection, when CAV-specific cytopathologic lesions. were visible, total RNA was precipitated from the cell lysates. One (major) CAV transcript of about 2.1 kb was detected in the poly(A)+ RNA fraction of CAV-infected cells. This CAVspecific RNA was absent in mock-infected MDCC-MSB 1 cells (Fig. 1A). Dot-blot analysis showed that CAV RNA was already weakly detectable 8 h after infection, but its level rose enormously at 32 h after infection (Fig. 1B). To investigate the location of the tsp and polyadenylation sites of CAV RNA, nuclease-S 1 mapping of the CAV RNA was performed. The tsp of the CAV RNA was located at nt 354 of the CAV sequence (Noteborn et al., 1991), as shown in Fig. 2A. This conclusion was confirmed by a primer-extension experiment using a 32P-end-labeled CAV DNA primer and MuLV reverse transcriptase (data not shown). Based on the nt sequence, we expected the tsp of the CAV RNA to be around nt 354. A complete set of promoter/enhancer elements is situated upstream from this tsp, such as the TATA box at nt 324. The polyadenylation site was located at nt 2317, which is 25 nt downstream from the only perfect polyadenylation signal (AAUAAA) in CAV RNA (Fig. 2B). The protected DNA smear with a length of 90 to 100 nt might be due to a G+C-rich region around nt 2245, which is predicted to form a stable hairpin and which complicates the nuclease-Sl assay by giving rise to additional protected fragments. In summary, the transcription from CAV DNA starts at nt 354, and terminates at nt 2317. CAV RNA is polyadenylated and has a length of approx. 2100 nt. This implies that the (major) CAV mRNA remains unspliced. (b) Synthesis of the CAV 52-kDa protein The largest ORF (nt 853-2200) encodes a 51.6-kDa protein of 449 aa. In our laboratory, we detected a single 52-kDa protein in CsCl-gradient-purified virus particles (S. Veldkamp, unpublished data). The N-terminal region of the largest ORF encodes an aa sequence which is highly sim-

-32

-48

Fig. 1. Characterization by the LiCl/urea

of CAV RNA synthesis.

method,

finity chromatography formed

as described

and polyadenylated

on oligo(dT)-cellulose. by Laird-Offringa

bridized with a ‘“P-labeled

et al., 1991). (Panel A) Northern RNA from noninfected

of total Cux-I

RNA,

MDCC-MSBl isolated

analysis.

MDCC-MSB

by af-

was per-

Three pg of poly-

1 cells (lane 1) or CAV-

cells (lane 2) were size-fractionated

9~ agarose gel. (Panel B) Dot-blot from MDCC-MSBl

nylon filter (Gelman

Sciences,

analysis.

cells infected

at 0, 8, 16, 24, 36 and 48 h after infection,

Biotrace-RP

analysis

et al. (1990). The RNA was hy-

adenylated

2.2 M formaldehyde-l

Northern

DNA probe derived from the plasmid pCAV/E

(Noteborn

Cux-1 infected

Total RNA was isolated RNA was isolated

with CAV-

was blotted

USA) according

on a Two pg onto a

to Maniatis

et al. (1982).

ilar to histone proteins. Histone proteins are known for their high Arg content and ability to bind and protect DNA (e.g., SwissProt data base, accession Nos. P14402 and PSOOOl). Thus, the N-terminal region of the 5 1.6-kDa CAV protein might have a DNA-binding function, perhaps within the virus capsid. The largest ORF is preceded by 499 nt harboring four AUG triplets. Two other ORFs lie upstream from the largest ORF. One of these terminates only 4 nt before the start codon of the largest ORF, and the other partially overlaps with it (Noteborn et al., 1991). Nuclease-Sl analysis of the potential splice sites in the upstream region of the largest ORF showed that these are not utilized (data not shown). These data imply that the 52-kDa protein is synthesized from the largest CAV ORF of the unspliced CAV RNA with the fifth AUG of the CAV RNA as the start codon. To establish whether the largest ORF is indeed used to produce the 52-kDa protein from the unspliced CAV RNA, we have carried out in vitro translation experiments on a series of plasmids specifying RNAs of increasing length. In vitro transcription of plasmid PEP- 1.96 should yield RNA with the largest ORF preceded by a 5’-upstream region containing four AUGs. From plasmid pEp-1.52, we ex-

269

A

petted RNA with a 62-bp 5’-noncoding no AUGs (Fig. 3). The RNAs derived

B

12CTAG

12CTAG nt

nt 430-

526-

167-

NCOI I

EcoRl LSP’ DNA probe protected

Fig. 2. Nuclease-Sl infected nuclease-S

or

*-

-*

mapping

of CAV RNA. Total RNA (20 pg) of CAV-

mock-infected

1 mapping.

S 1 and the protected

MDCC-MSBl

cells

were

analysed

by

The RNA/DNA

hybrids were treated with nuclease

DNA fragments

were fractionated

64; PA/8 M urea gel, with an nt sequence lengths of the probe and the expected

protected

DNA fragment

The

are indi-

an nt sequence

CAV using DNA primer CAV-1 (Noteborn

PEP-1.52, upon in vitro translation, both yielded a 52-kDa protein (Fig. 4A). This in vitro CAV protein migrates with the same velocity as a protein that could be synthesized in vitro from hybrid-selected RNA of CAV-infected cells, but not from hybrid-selected RNA of noninfected cells (Fig. 4, B and C). However, less 52-kDa protein is produced by RNA derived from PEP- 1.96 than by RNA from PEP- 1.52. Hybrid-selected CAV RNA also yielded at least two other CAV-specific products. Most likely, these products start at AUGs downstream from the start codon of the 52-kDa protein. This hypothesis was supported by comparing the translation products of RNA synthesized in vitro that either still carries the start codon at pos. 853 (Fig. 3: 1.52kb RNA; maximum protein 52-kDa) with RNA from which AUG-853 has been removed (1.44-kb RNA; maximum protein 44 kDa). Both revealed these faster migrating products (Fig. 4C). It is, however, unlikely that the downstream AUGs in the large ORF of CAV RNA serve as start codons in infected cells, since there only the 52-kDa product has been observed (Todd et al., 1990; our own unpublished data). For the translation of the CAV 52-kDa protein, an internal AUG is used as start codon. The use of an internal AUG codon for initiation of translation has been described for other RNAs. Several mechanisms may result in the choice of an internal start codon. For instance, in cases where the upstream AUGs occur in an unfavorable context for translation (Kozak, 1986) an internal start codon may be used. The four AUGs upstream from the CAV 52-kDa ORF, indeed, have unfavorable surrounding sequences, whereas the first AUG in the 52-kDa ORF is the only one in the CAV genome following the Kozak consensus: an A residue at -3, and a G at + 4. Ribosomes might scan the CAV RNA and then start preferentially at the fifth AUG.

on a denaturing

ladder as a length marker.

cated. Lanes C, T, A and G of each panel contain of cloned

ACCI I

*

* DNA

pA 1

region containing from PEP- 1.96 or

ladder

1111

1

IIll

1

I

I

I

1000

500

I,

I

I

I

II

1500

I

et al., 1991). The

DNA fragment

(nt l-43 1, see scheme) was 5’-end-labeled material was fractionated

sis: RNA from mock-infected (lane 2). The zsp is indicated the DNA fragment

and hybridized

cells (lane 1) and from CAV-infected by an asterisk.

cells

1.52-kb

5’-end of

(Panel B) Localization

of

to total RNA.

RNA from CAV-infected is indicated

Lanes:

1, RNA

cells. The position

from mock-infected

by pA. The 32P-labeled 3’-end of the DNA fragment

with an asterisk.

cells; 2,

of the poly(A)-addition

site

is marked

RNA

1.44-kbRNA

the poly(A)-addition site. The 525-bp NcoI-AccI DNA fragment (nt 2150356, see scheme) was 3’Jabeled with PolIk (Maniatis et al., 1982) and hybridized

bp

by electrophore-

by a bent arrow. The ‘*P-labeled

is indicated

I

TGA

sequence reactions were carried out according to Sanger (Maniatis et al., 1982). (Panel A) Localization of the 5’ terminus. The 431-bp EcoRI-EspI with total RNA. The protected

I

2000

Fig. 3. The position

of the largest ORF

CAV RNA relative to the

and the first five AUGs

( x ) of

tspof CAV RNA. The RNAs expressed from

the 1.96-kb, 1.52-kb and 1.44-kb DNA fragments cloned in the in vitro expression vector pEP40 are indicated by solid lines. The position of the

tspis indicated by a bent arrow and of the poly(A)-addition The longest CAV ORF is indicated by the stippled box.

site by ‘PA’.

270

A

B

kDa

’

the 52-kDa ORF might function as a ‘ribosome landing path’, as has been described for RNA of picornaviruses (Pelletier and Sonenberg, 1988) and a cellular mRNA, encoding the immunoglobulin heavy-chain-binding protein

C

123

*

(Macejak and Sarnow, 1991). Although the presence of such a ‘ribosome landing path’ would favor the use of the 5th AUG as start codon, it is rather unlikely in this case. If ribosomes would bind to an internal entry site, one would have expected that the translation efficiencies of the RNAs derived from the 1.96-kb and 1.52-kb DNAs were roughly equal. However, the 449-nt 5’-noncoding region clearly

-69 52-

impaired the translation efficiency of the larger transcript. Further investigations are needed to prove which of the above processes regulate the synthesis of the CAV protein(s). (c) Conclusions

Fig. 4. Synthesis

of the 52-kDa

1.52-kb (nt 792-2319) (xc

protein.

Fig. 3) were each cloned in plasmid

trol of the T7 promoter mids. PEPdownstream

1.96.kb

(Laird-Offringa

pEP40 to come under the conet al., 1989). The resulting

from the coding sequence and transcribed of Promega

(Madison,

cific CAV RNAs was accomplished pCAV/E

plas-

1 cells, essentially

UK). Proteins

et al.

lysate system

by electrophoresis

14”,, PA-O. 1”” SDS gels. Gels were fixed and prepared (Panel A) In vitro translation

or mock-

by McGrogan

in a rabbit-reticulocyte

were fractionated

of spe-

in

for fluorography.

of RNA derived from PEP-1.96

(lane 1) or

I .52 RNA (lane 2). The proteins were labeled with [ ‘JC]leucine

(Amersham, UK). In the control, no RNA was added to the reticulocytc ly sate (lane 3). (Panel B) In vitro translation of hybrid-selected CAV RNA from mock-infected 2). The proteins (Panel

(lane 1) and CAV-infected

MDCC-MSBl

were labeled with [“S]methionine

C) In vitro translation

RNA

We thank Dirk J. van Roozelaar, Ms. Claudia A.J. Verschueren and Sjoerd Veldkamp for excellent technical assistance, and Rob Hoeben and Christiaan Karreman for helpful discussions and computer analysis. This research was made possible with research grants from the Netherlands Ministry of Economic Affairs, and Aesculaap BV, Boxtel, The Netherlands.

cells (lane

(Amersham,

of PEP-1.52-derived

ACKNOWLEDGEMENTS

linearized

from CAV-infected

as described

accord-

selection

with nitrocellulose-bound

(1979). These RNAs were translated (N90, Amersham,

essentially

WI). Hybrid

DNA and RNA isolated

infcctcd MDCC-MSB

from PEP-

(nt 356-2319).

CAV DNA fragments

I .96, PEP- 1.52 and PEP-1 .44, were linearized at the EcoRI site

ing to the protocol plasmid

The

and 1.44-kb (nt 875-2319)

(1) The transcription from CAV DNA starts at nt 354 and terminates at nt 2317. The CAV RNA has a length of about 2100 nt, is polyadenylated and remains unspliced. (2) In vitro, the full-length CAV RNA specifies the 52kDa protein from its longest ORF. (3) Our data strongly suggest that also in vivo the 52kDa protein is synthesized from the largest CAV ORF of the unspliced CAV RNA with the fifth AUG from the 5’-end as the start codon.

UK).

(lane 2) and

PEP-1.44 RNA (lane 3). In the control. no RNA was added to the rcticulocytc lysate (lane 1). The proteins wcrc lab&d with [‘Hlleucine (Amer-

REFERENCES

sham, UK). The symbols

Gelderblom,

* and > denote proteins

derived from hybrid-

selected or in vitro synthcsired CAV RNA, respectively, migration mobility than the 52.kDa CAV protein.

that have a higher

H.. Kling,

Jeurissen,

S.H.M.,

Pal. J.M.A.

of cortical thymocytes (1989) 115-123.

Alternatively, the inefficient translation of the upstream ORFs might be due to the hairpin structure that is predicted by computer analysis between nt 369 and 520. Such an RNA-hairpin structure may impair the scanning of the CAV RNA by ribosomes, resulting in a suppression of initiation at the first two AUGs, as shown for other RNAs (Macejak and Sarnow, 1991). The 5’ region upstream from

S., Lurz,

Morphological characterization Virol. 109 (1989) 115-120.

Kozak,

M.: Bifunctional

R., Tischcr, ofchicken

I. and von Btilow, V.:

anemia agent (CAA). Arch.

and De Boer, G.F.: Transient

induced messenger

by chicken

anemia

depletion

agent. Thymus

RNAs in eukaryotes.

14

Cell 47 (1986)

48 l-483. Laird-Offringa,

I.A.. Elfferich,

P., Knaken,

H.J., De Ruiter, J. and Van

der Eb, A.J.: Analysis of polyadenylation site usage of the c-myc oncogene. Nucleic Acids Res. 17 (1989) 6499-65 14. Laird-Offringa,

I.A., De Wit, C., Elfferich,

P. and Van der Eb, A.J.:

Poly(A)-tail shortcning is the translation-dependent step in c-myc mRNA degradation. Mol. Cell. Biol. IO (1990)6132-6140.

271 Macejak,

D.G. and Sarnow,

P.: Internal initiation oftranslation

by the 5’ leader of a cellular mRNA. Maniatis,

T., Fritsch,

Laboratory Harbor, McGrogan,

E.F.

Manual.

bridization

Sambrook,

Cold Spring

NY, 1982. M., Spector,

H.J.: Purification

and

Nature

J.: Molecular

Harbor

mediated

Laboratory,

Cloning.

A

Cold Spring

D.J., Goldenberg,

C.J., Halbert,

of specific adenovirus

2 RNAs

D. and Raskas,

by preparative

hy-

and selective thermal elution. Nucleic Acids Res. 6 (1979)

M.S.: Chicken

anemia agent: a review. Avian Pathol. 20 (1991)

M.H.M.,

Kranenburg, Zantema,

O.,

De Boer, Vos,

A., Koch,

Characterization all elements

G.F.,

J.G.,

Roozelaar,

Jeurissen,

S.H.M.,

G., Van Ormondt,

of cloned chicken

for the infectious

D.J.,

Karreman, Hoeben,

C., R.C.,

H. and Van der Eb, A.J.:

anemia virus DNA that contains

replication

cycle. J. Virol. 65 (1991)

M.H.M.,

Verschueren,

C.A.J.,

Van Roozclaar,

kamp, S., Van der Eb, A.J. and De Boer, G.F.: Detection

D.J.,

Veld-

of chicken

and polymerase

chain reaction.

21 (1992) 107-118.

J. and Sonenberg,

karyotic

N.: Internal

initiation

of translation

of cu-

mRNA directed by a sequence derived from poliovirus

RNA.

334 (1988) 320-325.

Todd, D., Creelan, J.L., Mackic, D.P., Rixon, Purification and biochemical characterisation

F. and McNulty, M.S.: ofchicken anemia agent.

J. Gen. Virol. 71 (1990) 819-823. von Billow, V., Fuchs, B., Vielitz, B. and Landgraf, H.: Frilhsterblichkeitssyndrom bei Kilken nach Doppelinfektion mit dem Virus des Krankheit

(MDV) und einen Anamic-Errcger

(CAA). J.

Vet. Med. B 30 (1983) 742-750. Yuasa,

N.: Propagation

and infectivity

titration

of the Gifu-l

strain of

chicken

anemia

agent in a cell line (MDCC-MSBl)

derived

Marek’s

disease

lymphoma.

Q. 23 (1983)

Natl. Inst. Anim. Health

from

13-20. Yuasa,

3131-3139. Noteborn,

virus by DNA hybridization

Marekschen

187-203. Noteborn,

Pelletier, Nature

593-607. McNulty,

anemia

Avian Pathol.

353 (1991) 90-94.

N., Taniguichi,

of an agent inducing

T. and Yoshida,

I.: Isolation

and some properties

anemia in chicks. Avian Dis. 23 (1979) 366-385.