virus RNA 3 influence symptom phenotype on leaves - NCBI

0 downloads 0 Views 3MB Size Report
Oct 31, 1991 - BNYVV RNAs 1, 2, 3 and 4 are necessary for the natural infection process ... the virus within the root system while RNA 4 has been shown.
The EMBO Journal vol.11 no.2 pp.479-488, 1992

Two proteins encoded by beet necrotic yellow vein virus RNA 3 influence symptom phenotype on leaves

Isabelle Jupin1, H.Guilley, K.E.Richards and G.Jonard Institut de Biologie Moleculaire des Plantes du CNRS et de l'ULP, 12 rue du General Zimmer, 67000 Strasbourg, France 'Present address: Laboratory of Plant Molecular Biology, The Rockefeller University, 1230 York Avenue, New York, NY 10021, USA Communicated by A.L.Haenni

RNA 3 of the beet necrotic yellow vein virus (BNYVV) quadripartite RNA genome is not essential for virus multiplication on leaves of Tetragonia expansa but has dramatic effects on symptom expression. Virus isolates containing RNA 3 produce bright yellow local lesions while isolates lacking RNA 3 produce much milder symptoms. Using directed mutagenesis of cDNA clones followed by in vitro synthesis of biologically active transcripts, a 25 kDa open reading frame (ORF) of RNA 3 was shown to be responsible for the yellow local lesion phenotype. In addition, two deletion mutants of RNA 3 were found to elicit the appearance of severe necrotic local lesions. Analysis of one of these mutants revealed that necrosis was due to the overexpression of a second short ORF, N, overlapping the 3'-terminal portion of the 25 kDa ORF. As shown by gene fusion studies, gene N is not detectably expressed from full-length RNA 3 but is translationally activated by deletion of upstream sequences. Introduction of gene N into the genome of the unrelated DNA virus, cauliflower mosaic virus, elicits a necrotic response instead of the typical mosaic symptoms, demonstrating that gene N can induce necrosis outside of the context of a BNYW infection. Key words: furovirus/mutagenesis/rhizomania/symptomatology/transcript

Introduction Pathogens are responsible for severe diseases in plants resulting in major losses in many important crops. Symptoms may vary from local discolorations to severe perturbation of growth and development or even death of the plant. The induction of disease symptoms is likely to be the last step in a complex and little understood process in which the agent modifies the metabolism of its host at many different levels. Viruses are particularly well suited for studying the molecular basis of plant -pathogen interactions and symptom development because they have small well characterized genomes in which specific genetic loci can be readily modified by in vitro mutagenesis. Extensive work of this sort has been undertaken, notably with tobacco mosaic virus (TMV) and cauliflower mosaic virus (CaMV) as well as with satellite RNAs and viroids (for reviews, see Van Loon, 1987; Daubert, 1988). While such analysis ©D Oxford University Press

has permitted identification of viral genes or nucleotide sequences influencing symptoms, it appears that multiple and independent loci can also jointly determine symptomatology. As yet, no obvious explanation relating particular viral gene products or viral sequences to the resulting phenotype has been established. We have investigated the molecular basis of the symptoms induced by beet necrotic yellow vein virus (BNYVV). As described below, the high degree of compartmentalization of the BNYVV genome, the dispensibility of the 'small' viral RNAs for leaf infections, and in particular, the pronounced effect of RNA 3 on symptomatology, all make BNYVV an attractive system for analyzing the roles of viral genes in symptom production. BNYVV is a furovirus that is the causative agent of sugar beet rhizomania (Tamada, 1975). This disease is characterized by browning of vascular bundles in the tap root, leading to root stunting and loss of sugar content together with abnormal proliferation of secondary rootlets that often undergo necrosis. At a later stage of infection, the virus sometimes moves to the leaf system and induces the vein-associated yellowing and necrosis symptoms from which the virus draws its name. BNYVV has a multipartite genome consisting of four single-stranded plus-sense RNAs of 6.8, 4.7, 1.8 and 1.5 kb (Bouzoubaa et al., 1985, 1986, 1987), all of them possessing similarities at the 5' and 3' extremities (Bouzoubaa et al., 1987). BNYVV RNAs 1, 2, 3 and 4 are necessary for the natural infection process, which is defined as transmission of the virus to sugar beet roots by the fungal vector Polymyxa betae, proliferation of virus within the roots and production of rhizomania symptoms (Koenig et al., 1986; Lemaire et al., 1988; Tamada and Abe, 1989). RNAs 1 and 2 are involved in replication and packaging of the viral genome; RNA 3 appears to play a role in multiplication or translocation of the virus within the root system while RNA 4 has been shown to be important for efficient transmission of the virus by P. betae (Tamada and Abe, 1989). On the other hand, when BNYVV is mechanically inoculated to leaves of laboratory hosts such as Chenopodium quinoa or Tetragonia expansa, only RNAs 1 and 2 are needed to maintain the infection (Koenig et al., 1986; Quillet et al., 1989). RNAs 3 and 4 can be eliminated intentionally from the isolate, or, if present at the outset, they may disappear spontaneously or undergo extensive internal deletion in the course of propagation (Bouzoubaa et al., 1985, 1991; Koenig et al., 1986; Tamada et al., 1989). When BNYVV is mechanically inoculated to leaves, RNA 3 has striking effects on symptomatology (Tamada, 1975; Quillet et al., 1989; Tamada et al., 1989). Viral isolates containing full-length RNA 3 produce intense yellow local lesions (yellow spot or YS symptoms) on inoculated leaves while isolates in which RNA 3 is missing or has undergone intemal deletions generally produce a much milder phenotype (M symptoms) in which the lesions consist of faint rings or chlorotic spots. Another type of symptom on leaves has also

479

I.Jupin et al.

been noted: severe necrotic local lesions (NS or necrotic spot symptoms) (Tamada, 1975). A natural isolate, BNYVV-F15, which has been reported to contain an RNA 3 species bearing a long internal deletion (Bouzoubaa et al., 1988), is able to induce this necrotic lesion phenotype. Artificial BNYVV RNA 3 obtained by in vitro transcription of full-length cDNA clones is biologically active and produces the same symptoms as natural RNA 3 on all hosts tested (Ziegler-Graff et al., 1988; Quillet et al., 1989). RNA 3 carries an open reading frame (ORF) for a 25 kDa polypeptide (P25) which is synthesized in leaves in the course of BNYVV infection (Niesbach-Kl6sgen et al., 1990). Deletion mutagenesis of RNA 3 transcripts has shown that P25 is not required for in vivo multiplication of RNA 3 and that sequences active in cis in RNA 3 replication are confined to the extremities of the RNA (Jupin et al., 1990). RNA 3 also directs synthesis in infected tissue of a 500 nucleotide (nt) subgenomic RNA, RNA 3sub, which is 3'-coterminal with RNA 3 (Bouzoubaa et al., 1991). Finally, one ORF (gene N), which overlaps the 3'-terminal portion of the P25 ORF, will be shown in this paper to influence symptom phenotype. Here, we have used RNA 3 transcripts produced in vitro to investigate in greater detail the influence of RNA 3 on symptom expression on leaves. We show that numerous alterations in the RNA 3 sequence which truncate or eliminate the P25 ORF all eliminate the yellow spot lesion phenotype, demonstrating that P25 is responsible for these symptoms. In addition, the expression of ORF N, which is not detectably translated from full-length RNA 3, can induce necrotic local lesions when rendered accessible for translation by deletion of upstream sequences.

Results P25 is responsible for the induction of yellow spots In a first series of experiments various deletions were engineered into the cDNA insert of pB35 carrying the complete cDNA sequence of BNYVV RNA 3 (Ziegler-Graff et al., 1988) by eliminating the sequence between naturally occurring or purposely created unique restriction sites (see Figure 1 and Materials and methods). These deletion mutations were designed to leave intact the 5'-terminal 300 nt and the 3'-terminal 70 nt as these regions are essential in cis for RNA 3 replication (Jupin et al., 1990). Transcripts prepared from the mutant clones were co-inoculated to T. expansa leaves along with BNYVV RNAs 1 and 2 to provide helper functions and symptoms were scored 8 days post-inoculation. The successful establishment of infection and the accumulation of the mutated RNA 3 species in the inoculated leaves were analyzed by Northern hybridization using radioactive viral antisense RNAs as probes (data not shown). Of the various deletion mutants tested (Figure 1), only B13 and ASE, which contain deletions outside of the P25 ORF, can induce YS symptoms similar to those characteristic of infection with wild-type RNA 3 transcript t35 (Figure 2B). In contrast, all deletion mutations which disrupt the P25 ORF eliminate the YS phenotype. In most cases, mild symptoms were produced, indistinguishable from those induced by RNAs 1 and 2 alone (Figure 2A). Surprisingly, two exceptions were noted. Both AEA and 24AXA induced necrotic local lesions (NS). These mutants and the possible mechanisms by which necrosis is induced by each will be discussed in greater detail below. 480

In four additional mutants, the P25 ORF was interrupted frameshift mutations (mutants F 1 - F4 in Figure 1). In a fifth construct, 35CCG, the AUG initiation codon of the P25 ORF was mutated to CCG. The next in-frame AUG in the P25 ORF is 243 nt downstream (Figure 1). Translation initiation by scanning ribosomes at this point would produce an extensively N-terminally truncated and presumably non-functional P25. All the aforesaid mutations abolished YS symptoms (Figure 1). Since the frameshifts and the point mutation do not introduce major changes in the overall nucleotide sequences, these results represent strong evidence that the integrity of P25, rather than some cis-active feature of wild-type RNA 3 such as RNA secondary structure, is necessary for induction of the intense yellow local lesions. at different points by

Mutations which provoke NS symptoms: ORF N As mentioned, two different RNA 3 deletion mutants, AEA and 24AXA (Figure 1), were unexpectedly both able to induce necrotic spot (NS) lesions. In the case of SEA, the resulting NS lesions on leaves of T. expansa are shown in Figure 2C. NS symptoms were also induced by both mutants on C. quinoa as was necrosis of non-inoculated upper leaves of the systemic BNYVV host Spinacea oleacera (data not shown). The NS symptom phenotype was dominant with respect to YS as leaves inoculated with a mixture of equal amounts of zEA and t35 (plus RNAs 1 and 2) displayed only the NS response. The deletion in zEA has the effect of bringing a 180 nt ORF (nt 1052-1231 in full-length RNA 3), which normally overlaps the 3'-terminal portion of the P25 ORF, to a position nearer the 5' terminus (Figure 3). We shall refer to this small ORF as ORF N and to its predicted 59 amino acid (6.5 kDa) translation product as protein N. As a consequence of the deletion in mutant AEA, ORF N is placed in approximately the same position relative to the 5' terminus as is occupied by the translationally active P25 ORF in full-length RNA 3. This is not the case for the deletion mutants AEH and AEB, which provoke mild symptoms, because ORF N in these mutants is still separated from the original P25 start site by lengthy sequences which contain numerous AUGs (Figure 1). Therefore, we hypothesized that the novel more 5'-proximal location of ORF N in mutant AEA permits its translation and the resulting short gene product, protein N, is responsible for the formation of necrotic local lesions. ORF N is translated in mutant AEA In order to test the above hypothesis, we have fashioned gene fusion constructs in which the 3'-terminal 7 nt and the stop codon of ORF N (and some downstream sequences, see Materials and methods) of both full-length RNA 3 and mutant AEA were replaced with a DNA fragment containing the 3'-terminal 1856 residues of the f-glucuronidase (GUS) gene in both orientations (Figure 3A). In the constructs carrying the GUS gene in the plus-sense orientation, B13GUS(+) and B13GUS(+)AEA, ORF N and the GUS coding sequence are fused in-frame (Figure 3A) so that the fusion protein is composed of the first 57 amino acids of protein N and the last 602 amino acids of GUS protein. Note that the authentic GUS initiation codon was not carried by the GUS gene fragment so that any GUS activity is expected to result from an ORF N -GUS fusion protein. Inoculation of B13GUS(+) and B13GUS(+)AEA (plus

BNYVV RNA 3 proteins and symptom expression Miltant

SNzuptonill

Modification

P4.6 P -2 -

-

wvild typ t 35

1113

4;

t-Vtws;AI222-15-4w-

E

A3 14-376

A'II

.X380-569

AE.J13

A380-750--

AL'A

_fi A380-1033

AA AFS

A380-i 47{}

All 1

A570- 's A

AS

.

-

__

_

_

_

_

--

_

_

-

_

_

_

_

_

1

__

NS M

_ r

---

_ f I-

-

N1 .

NI

ABA

A755-1033

24AX,A

A933-1033

35_( ((;

.t G

1.

+

Sit

ait

5Tn

1F 2

+4iut

ait

-54

F3

+4nt

ait 932_

l4

.2 +2nllt

_

---

AH2A

-

_

-

N1 I., ,'., '-L 1e .,---.,, .~~~~~~~~~~~~~~.-

-. . -.--.--. .---. ..

I.-.~'-

(((

_= L...

NS

MI N1

-+ 4 rn I..I..

i-

Nl

I.. I.."

e

r

Fig. 1. Effect of different mutations in BNYVV RNA 3 transcripts, all co-inoculated with wild-type BNYVV RNAs 1 and 2, on the phenotype of symptoms on leaves of T.erpansa. The structure of wild-type RNA 3 and of the RNA 3-derived subgenomic RNA (3sub) are shown schematically at the top. Rectangles represent P25, ORF N and ORF P4.6 described in the text. In-frame AUGs in the P25 ORF are indicated by small circles. The 17 out-of-frame AUGs in the P25 ORF are not shown. Deletions in the RNA 3 sequence are represented by broken lines with the coordinates of each deletion (in nucleotides) shown to the left. C-terminal missense sequences resulting from frameshift mutations in the P25 ORF are represented by black rectangles. Symptom phenotypes were yellow spot (YS), mild (M) and necrotic spot (NS).

BNYVV RNAs and 2) to T. expansa leaves gave rise to YS- and mild-type symptoms, respectively. Northern

hybridization revealed that both GUS-containing RNA 3 species were replicated in the course of infection, although the GUS-containing RNA 3 constructs accumulated at least five times less efficiently than the constructs lacking the GUS gene (data not shown). This partial inhibition of replication of the mutants carrying the ORF N -GUS fusions may be related to their failure to produce NS symptoms. However,

we cannot strictly rule out the possibility that the replacement of sequences near the C terminus or downstream of gene N by the GUS gene (Figure 3A) could also influence symptom phenotype. No GUS activity was detected in the YS local lesions induced by B13GUS(+) suggesting that ORF N is silent or

only poorly expressed from full-length RNA 3 (Figure 3A). On the other hand, the lesions obtained from infection with B 13GUS(+))AEA contained GUS activities >60-fold above 481

I.Jupin et al.

Fig. 2. Symptoms provoked on leaves of T.expansa 8 days after inoculation with BNYVV RNAs 1 and 2 (A), RNAs 1 and 2 plus the wild-type RNA 3 transcript t35 (B), RNAs 1 and 2 plus AEA transcript (C) and RNAs 1 and 2 plus AEATAA transcript (D). The very faint lesions in (A) and (D) are not readily visible in the photograph.

background (Figure 3A). Western blot analysis of soluble protein from such lesions using a GUS-specific antiserum revealed a major GUS-specific band of 72 kDa (data not shown), rather than the 66 kDa of native GUS. This 6 kDa increase in size is expected if the GUS enzyme, in mutant B13GUS(+)AEA, is synthesized as an ORF N-GUS fusion protein rather than by some other mechanism (e.g. initiation further downstream within the GUS sequence). Infection with the mutants B13GUS(-) and B13GUS(-)AEA containing the GUS gene in the other orientation did not produce significant amounts of GUS activity (Figure 3A). From these experiments we conclude that ORF N is expressed only

482

weakly if at all from full-length RNA 3 but becomes efficently translated upon removal of upstream sequences. Expression of protein N induces necrosis In order to show that novel expression of ORF N in the AEA mutant is responsible for induction of necrosis we have converted a UCA codon near the beginning of ORF N (at the position corresponding to amino acid 23 of protein N) to a UAA termination codon by mutagenesis of a single nucleotide (Figure 3B). The corresponding transcript ZEATAA was efficiently replicated in planta but no longer induced necrotic lesions (Figure 2D). Introduction of the

_-

'A Z _

i i I(I t N ) e

,

--L

... ..

_

(( .

k

I't N it %

2( s ±O ploe, +II}

..

'

..

B13

____E_S.____

I)I

:.

f}

I.

.~....

1 35

81361 S-i-

BNYVV RNA 3 proteins and symptom expression

-.

2_Zv+-

2

I

_ + II

B .l 5

~~~~~~ -._'-____

81361~~~~~

A13it

-

'|

0

-

BI3(;I St+.51K 4

F

B 8 3 (1 A FA

,-

_ . _+41-i

-

"L.