Can North American rose hybridizers safely use pollen from PNRSV infected roses?
By Henry Kuska
Tuesday 2 PM version, Draft, 2-11-2003
In the 1990s I ran a (free) e-mail rose breeding scientific literature “course”. One of the questions that I received was from a prominent North American rose hybridizer (of All American Rose caliber). He stated that, from what he had read, rose mosaic virus could not be spread by pollen. However, he had occasionally observed virused roses among his seedlings. He asked if I knew of any scientific rose literature on the subject.
First, I will define my terms. There are a number of viruses that historically have been included under the broad term “rose mosaic virus”. In this discussion, I will limit myself to what, at present, appears to be the most prevalent rose “mosaic” virus in North America, prunus necrotic ringspot virus (PNRSV). A more accurate scientific name is prunus necrotic ringspot ilarvirus, but I will use the more common prunus necrotic ringspot virus.
In the “Plant Viruses Online” database, the (general) modes of transmission for PNRSV are given as: “Transmitted by means not involving a vector. Virus transmitted by mechanical inoculation; transmitted by grafting; not transmitted by contact between plants; transmitted by seed (to over 80% in Prunus pennsylvanica but much less in peach); transmitted by pollen to the seed and transmitted by pollen to the pollinated plant.” See footnote (1)
The above is a general statement. It is possible (but not probable) that PNRSV could be transmitted in pollen of other infected plant species (to some non-zero degree) but not transmitted at all (zero) in roses. To see whether this remote possibility of zero transfer is actually the case, specific scientific studies of roses and PNRSV will be examined.
The earliest pertinent studies that I could find were three that were published in 1980.
One was by D. J. Barbara, East Malling Research Station, United Kingdom. She reported that: In rose, high levels of PNRSV virus were present in petals and stamens but only very low levels of virus were present in the sepals. Note, for the non scientific reader, pollen is released from the stamens. See footnote (2)
The second 1980 study appeared in 3 papers published by J. B Sweet, Long Ashton Research Station, University of Bristol,United Kingdom. See footnote (3). He reported that PNRSV was detected in the pollen of Peace and Queen Elizabeth roses, and in 1 % of two batches of seedling Rosa multiflora rootstocks. As is typical of scientific caution (or at least should be), he points out that finding the virus in the seedlings is not a definitive proof of transmission through the rose seed. However, his statement should not be interpreted that he did not find non grafted seedlings with virus; it means that he cannot be sure that the infected seedlings came from infected seeds (the virus may have infected the seedlings after they were planted (infected shears, infection from thrips, etc.)). What it does show is that there was transmission of PNRSV to non grafted seedlings in 2 different batches (unless there was experimental error - a “confirming” experiment has not yet been reported). Since multiflora seedlings are grown together in a field, I think that one can exclude the possibility that the seedlings were infected by “root grafts to an infected neighbor plant”. If one attempted to use this reasoning, one would be going in a circle logically i.e. one would have to explain how the infected seedling that supplied the root graft got infected.
This finding is so important, I think that we should look at it in more detail. In the Journal of Horticultural Science paper, Sweet states: "“Two plants from a batch of about 200 Rosa multiflora seedling rootstocks with mosaic symptoms on their leaves induced PNRSV-type symptoms in the woody indicator, whereas four symptom less plants did not. Virus was sap-transmitted from one of the mosaic-diseased R. multiflora seedlings which induced typical PNRSV symptoms in herbaceous test plants, and infected cucumber sap-precipated with NRSV-G but not with ApMV-P and PDV-B antisera.” Please note, these infected plants are not grafted to a rootstock; they are seedlings that were intended to be used as rootstocks . Also note, Sweet used a very crude detection systems (visual symptoms), we do not know how many more of the 194 seedlings were infected but did not show symptoms (as many reports have documented, often an infected plant will not have visual symptoms).
The third 1980 paper was published by B. J. Thomas. He found that both the gel immunodiffusion test and the latex test were unable to detect PNRSV in infected roses but that both ELSIA (enzyme-linked immunosorbent assay) and SSEM (serologically specific electron microscopy) methods were able to (in some cases ELSIA also failed). SSEM was twice as sensitive as ELSIA. Of particular importance here is that he was able to detect the virus in the rose stamens. See footnote (4).
The next two cited papers are also very important to this discussion as the first documents (and the second confirms) that it is possible to “transfer the virus by mechanical means”. This is an important finding because this removes the possibility that the virus is “too fragile” to be transferred by anything but a graft (living phloem cells of the plant). The “too fragile” theory seems to be one of the key assumptions of the “no transmission but by grafting” school of thought.
The first report of mechanical transfer was in 1981, by B. J. Thomas. He reported that his preliminary attempts to transfer the virus by mechanical means failed, but by adjusting his experimental procedure he was successful - including the use of infected rose petals and anthers (it is not clear if the petals and anthers were mixed together or tested separately). See footnote (5).
The ability to transfer the virus by mechanical means was confirmed in 1994. H. Baumgartnerova, Institute of Experimental Phytopathology and Entomology, Slovak Academy of Sciences, Czechoslovakia reported that the virus was positively mechanically transmitted from diseased leaves and pollen of roses. See footnote (6).
In fairness to possible overlooked articles, I should point out that my search may not be complete as computer abstracting searches are relatively recent; and there may be articles in relatively small scientific journals that have not yet been covered by the abstracting services nor footnoted in the articles that I have cited. However, I feel that the above cited research papers provide sufficient information such that North American rose hybridizers should be wary of using PNRSV diseased roses as pollen parents in their hybridizing program.
Examples of how others have interpreted the literature up to the date of publication of their own work are given next.
In the 1983 book “Compendium of Rose Diseases” by R. K. Horst, Professor Horst (Cornell University, Plant Pathology) wrote: “PNRSV is pollen-transmitted in fruit trees. Pollen transmission is suspected to occur in roses also, since spread in the field is slow.” See footnote (7).
Baldo Villegas (Associate Environmental Research Scientist (Entomologist),California Department of Food and Agriculture) consulted with the following plant pathologists with the California Department of Food and Agriculture’s Plant Pests Diagnostic Centre: Dan Opgenorth and Dennis Mayhew before preparing his web page article on the subject. He stated: " They have given me invaluable advice in preparing this article." Regarding spread, he stated: “Some pathologists suspect that mosaic may be pollen transmitted which could prompt removal if other roses in the garden are valuable and not already infected” See footnote (8).
In 1989 in an article titled “Incidence of Rose Viruses in Spain” M. Cambra, J.L. Martinez-Torres, M.J. Benaches, E. Camarasa, and M.T. Gorris studied 4,730 rose samples. They found 4.2% of the roses had PNRSV. The breakdown was: 44.0 % of the miniatures, 1.1 % of the hybrid teas, and 1.5 % of those budded on Manetti rootstocks. Cambra et.al. state: “The high rate of PNRSV contamination in miniature varieties seems to be associated to their long existence.” Later in another paragraph they say:“…since this virus is pollen transmitted (in addition to grafting).” They also suggest that one way to prevent the virus from spreading is to prevent the plants from flowering. See footnote (9).
Professor Gerald C. Adams (Department of Plant Biology and Department of Plant Pathology, Michigan State University) makes the following statements: “Rose Mosaic Virus in the Americas is most commonly caused by prunus necrotic ringspot virus which is a pollen transmitted virus but transmission by pollen is very low. When transmission by pollen does occur it is usually due to a high population of thrips. The thrips carry the pollen or infected sap and introduce it into a feeding scar on a leaf or petal.” Farther in the article he states: “Spread of PNRSV by pollen is measurable in cherry orchards and the virus has been found to exist at high titer in rose stamens. Never the less, field spread in rose nurseries has not been easily demonstrated and apparently is rare. In fact most spread of PNRSV is by grafting during vegetative propagation.” see footnote (10). I included the second quote for completeness, but please remember that the focus of the present paper is “spread in hybridizing”, not “field spread”.
Unfortunately, nature often is not as simple to understand as we would like. This appears to be one of those times as there are 2 papers that looked for PNRSV transfer to seedlings through pollen but did not find it.
The first paper is Sweet’s Journal of Horticultural Science report referred to earlier. He germinated seeds from the two infected plants (36 seedlings from one and 24 seedlings from the other. None showed visual symptoms and 5 random samples from each were grafted on P. Persica GF 305 seedlings (virus test plants) and did not test positive. There are several reasons that I do not consider this report definitive: 1) 10 grafting tests is too small a sample size, 2) the use of visual symptoms on 36 samples can now be considered as both too small a sample and to also be deficient in test sensitivity, and 3) if the infected two bushes were randomly positioned in the same field as 198 "healthy ones; it is probable that the pollen came from the healthy bushes (see discussion 2 paragraphs down).
The second virus-seed transmission test paper was published in 1984, by B. J. Thomas. See footnote (11). He reports on three different experiments that could provide some information regarding virus-seed transmission. In the first experiment he crushed 10 seeds from each laboratory infected virused plant with hips, and could not, using ISEM, detect any PNRSV in the seeds (he did not state the number of batches examined). In the second experiment he examined, with ISEM, 1067 seedlings grown from seeds harvested from the infected bushes and found no PNRSV infected plants. A possible explanation for the failure to detect PNRSV in the seeds or resulting seedlings is that the seeds may have also (like Sweet’s samples) been unintentionally produced from non-infected pollen. This possibility is mentioned because of the physical placing in the test plot that Thomas used. The following is Thomas’ description of the plantings: “Seed was collected for 3 yr. from virus infected plants planted in nematode-free soil 50 cm apart in two rows 2 m apart so that plants of the same species were opposite each other. A row of corresponding healthy roses were planted equidistant between these rows.” The roses utilized were all species roses, R. canina, R. canina, var. Brogs, R. corymbifera, R. multiflora, and R. rugosa If we accept the points made in the next 3 paragraphs, he should of been examining the seeds from the non virused bushes that were located between the two infected bushes. (His experiment may still be valid if the neighboring plants in the same “infected” row were of the same species and also had PNRSV - I have to add the “also had PNRSV” because he had plants with 3 different viruses in the study - the neighbors could have had one of the other 2 viruses. This information was not given. Also, except for R. rugosa, the other species roses are once bloomers, and the bloom periods for the same row infected different species neighbors may not have overlapped the bloom period of the diseased plant that proved the examined seeds. In the third “experiment” he did detect PNRSV in three seedlings of new rose cultivars (planted in the Royal National Rose Society’s trial grounds). One infected plant was of unknown origin, one was from France, and one was from the USA (via France). Since the hybridizing and growing conditions were not controlled (for example, were virused rootstocks used?), I do not consider that the results from this third “experiment” are scientifically of value relative to the present question.
This paragraph and the next two paragraphs discuss the literature evidence for the model that proposes that the infected plants were most likely cross-pollinated by non-infected pollen in the above virus-seed transfer experiments. In 1986 P. Cole and B. Melton published a paper which investigated the ability of rose pollen to fertilize flowers on the same bush. See footnote (12). The diploid species were all highly self-sterile. None of the 23 diploid specimens exhibited over 4 % fertility and 18 of the 23 produced no self-set seed. They also studied the fertility with pollen from another plant of the same species and found that the diploid group was 50 times more cross-compatible than self-compatible. For roses of higher polyploid level 12 of the 16 studied were no more self-compatible than the diploid group. Unfortunately, only two of the species studied by Thomas were also studied by Cole and Melton, one was R. rugosa which was 100 % self-sterile, and the other was R. rubiginosa Linnaeus (R. eglanteria) which was 83.4 % self-sterile (one sample).
A very recent paper by Han YounYol and Yu SunNam, see footnote (13), did confirm that R. Rugosa was self-sterile but reported that R. multiflora did self-fertilize. I posted a request to members of the Rose Hybridizing Association to check to see if they had information on the question of whether R. multiflora is self-sterile. David Zlesak stated that: " roses have a gametophytic self incompatibility system which can be easier to break down than a sporophytic system…" The reports from the RHS members were varied. They ranged from definitely sterile, a few hips, to reports of many hips. These results appear consistent with David Zlesak’s comments.
The self-sterility factor is not the only factor which favors the cross-pollination model. In the 1963 American Rose Society Rose Annual, Dr. Eileen W. Erlanson Macfarlane, see footnote (14), makes the following statement concerning the self-pollination mechanism in roses (not an exact quote, the sentence structure has been changed to include only the part pertinent to this discussion): “It was found that the pollen cells mature before the embryo sac and egg cells. Most of the pollen is shed when the petals first open and insects will carry some of it to flowers opening for the second day, which may then be ready for cross-fertilization. However, a few pollen grains remain on the shriveled anthers on the third day; and the stamens should still be receptive.”
Thus, even if the infected plant is not self-sterile; the pollen from the nearest neighbor non-infected plant of the same species would have the first opportunity to pollinate the infected bush (assuming that bees were around, which should be a reasonable assumption since the studies were done in an open field).
A possible reason that field spread may be slow through seed transmission is that the PNRSV appears to be on the surface of the pollen and in the part of the pollen that forms the pollen tube but not in the actual “sperm” (the study was on PNRSV in nectarine pollen). See footnote (15). The authors of that study comment: “However, the possibility that sperm cells are infected during or after mitosis, before fertilization, cannot be ruled out. In addition, it is also possible that PNRSV could be transmitted by the vegetative cytoplasm of the pollen, which contains high amounts of the virus … Recent evidence indicates that during fertilization some male cytoplasm can be transferred into the egg cell …”.
I would like to suggest what I consider to be a definitive experiment regarding PNRSV transfer through seed: hand pollinate a fertile non-infected isolated rose (that has had its own pollen removed) with diseased pollen from an infected rose (or several infected varieties) that is (are) known to cross with it. The hand pollinated flowers would then be covered to prevent stray pollen from contaminating the chosen flowers (the seedlings raised from the seeds resulting from this cross would have to be kept in an insect free environment, and sanitary procedures utilized to prevent “after germination” contamination in order for the experiment to be considered truly definitive). Hopefully, visual symptoms would appear on at least some of the seedlings. If no visual symptoms appear, then laboratory tests would be required.
Footnote (1) http://image.fs.uidaho.edu/vide/descr658.htm , Brunt, A.A., Crabtree, K., Dallwitz, M.J., Gibbs, A.J., Watson, L. and Zurcher, E.J. (eds.) (1996 onwards). `Plant Viruses Online: Descriptions and Lists from the VIDE Database. Version: 20th August 1996.’ URL http://biology.anu.edu.au/Groups/MES/vide/ Dallwitz (1980) and Dallwitz, Paine and Zurcher (1993) should also be cited.
Footnote (2) Barbara, D., J., Acta Phytopathologica Academiae Scientiarum Hungaricae, volumn 15, pages 329-332, (1980). Reprinted in Acta Horticulturae, volumn 94, published March 1, 1981.
Footnote (3) Sweet, J., B., Acta Phytopathologica Academiae Scientiarum Hungaricae, volumn 15, pages 231-238, (1980). Reprinted in Acta Horticulturae, volumn 94, pages 231-238, (1981).
Http://www.actahort.org/books/94/94_31.htm
He published a similar (but not just a duplicate) paper in: Journal of Horticultural Science, volumn 55, pages 103-111, (1980).
Footnote (4) Thomas, B. J., Annals of Applied Biology, volumn 94, pages 91-102, (1980).
Footnote (5) Thomas, B.J., Annals of Applied Biology. volumn 98, pages 419-429 (1981).
Footnote (6) Baumgartnerova, H., Acta Horticulture, volumn 377, pages 357-359, (1994). ROSA SP. - RESERVOIR OF THE SOUR CHERRY NECROTIC RINGSPOT VIRUS
Footnote (7) Horst, R., K., book “Compendium of Rose Diseases”, published by The American Phytopathological Society, St. Paul, Minnesota. pages. 26-27, (1983) .
Footnote (8) Baldo Villegas web page, http://members.tripod.com/buggyrose/ipm/83rosemosaic.html
Footnote (9) Cambra, M., Martinez-Torres, J.L., Benaches, M.J., Camarasa,E., Gorris, M.T., Acta Horticulturae, volumn 246, pages 309-312, (1989).
Footnote (10) Adams, G., C, “Rose Mosaic Virus in the Nursery”, http://extension.bpp.msu.edu/rosemosaic/ , last revised 9/7/2002.
Footnote (11) Thomas, B., J., Annals of Applied Biology. volumn 105, pages 213-222 (1984).
Footnote (12) Cole, P., Melton, B., J. Amer. Soc. Hort. Sci., volumn 111, pages 122-125, (1986).
Footnote (13) YounYol, H. and SunNam, Y., Journal of the Korean Society for Horticultural Science, volumn 43, pages 326-332, (2002).
Footnote (14) Erlanson Macfarlane, E. W., American Rose Society Rose Annual, volumn 48 , pages 188-193 , (1963).
Footnote (15) Aparicio, F., et.at., European Journal of Plant Pathology, volumn 105, pages 623-627, (1999).