2Environmental Horticulture Department,
1531 Fifield Hall, Box 110670, University of Florida, Gainesville, FL 32611-0670
(352-392-1831, FAX 352-392-3870)
3Agronomy Department, 116 Agriculture Science and Industries Building,
Pennsylvania State University, University Park, PA 16802 (814-863-9805, FAX
814-863-7043)
Improved hybrid bermudagrasses such as Tifway and Tifdwarf are supposed to provide consistent
golf playing surfaces, but they often exhibit a mosaic of patches of different kinds of
bermudagrass. This is especially severe on greens, but Florida golf courses also show the problem
in fairways. To understand the problem, we completed a genetic fingerprinting study,
"Distinguishing off-types in Tifway and Tifdwarf bermudagrasses."
The 2-year, $66,000 project was supported cooperatively by the Florida Golf Course
Superintendents Association and the Florida Turfgrass Research Foundation. Matching funds
were provided by the Golf Course Superintendents Association of America to the Florida Golf
Course Superintendents Association. Using multiple techniques, we identified bermudagrass
patches as genetic off-types, and formulated prevention guidelines.
While greens are most seriously affected by genetic patches, fairways sometimes show the
problem as well, for example, Golf Course X, Palm Beach County, Florida (Fig. 1).
This was an excellent example for the objective of our study: to determine the identity and origin
of unknown bermudagrasses. Golf Course X had been fumigated with methyl bromide and was
planted with what was believed to be Tifway (=T-419) bermudagrass.
Fig. 2. Golf Course X, enhanced
photo.
First, let us analyze the distribution of the patches from a digitally enhanced image of a fairway
from Golf Course X (Fig. 2). On this and other fairways, the
bermudagrass patches occurred as distinctive variants, "Blue green" and "Dark green." There was
a repeating pattern, Blue green here, Dark green there, throughout this and other fairways.
The patches were surrounded by a yellowish matrix grass, which may be described by analogy to
Swiss cheese. The surrounding "matrix" grass looks like the cheese part of Swiss cheese, while
the patches are like the holes.
Our evidence will show that the patches were different grasses, i.e., off-types, thus were not
caused by
fertilization, acid injection, rainy weather, or any other known environmental factor. The
roundish, convex-margined shape of the patches shows they were probably increasing in size from
a point source in the center of the patch. DNA evidence will show that the matrix grass was
Tifway bermudagrass, while the patches were not Tifway, but were contaminants.
Fig. 3. Chromosomes (2n=36) from patch bermudagrass
at Golf Course X.
Fig. 4. Leaf hairs on Tifway 419.
Fig. 5. Morphology profiles of 26
grasses.
Besides DNA fingerprinting, traditional methods were helpful in bermudagrass identification. The
mitotic chromosomes of bermudagrass are extremely small and difficult to distinguish, yet their
number distinguished the patch bermudagrasses from the surrounding matrix grass. The
chromosome number of the patch bermudagrasses was about 36 (Fig. 3), thus they were common bermudagrass, not Tifway. The
chromosome number of the matrix grass was about 27, the same as reported for Tifway.
Morphologic traits also distinguished the patches from Tifway. A replicated grow-out was
performed on 26 grasses in six replicated pots. The patch grasses were very seedy, while Tifway
and matrix grasses produced few seedheads. The patch grasses had normal anthers and shed
pollen, as expected for 36 chromosome common bermuda. Tifway and the matrix grasses were
sterile. Tifway and the matrix grasses had very many (100+), fine, leaf blade hairs (Fig. 4), while the patch grasses that we studied had longer, less
abundant hairs. (Some other bermudagrasses, such as Ormond, have few or no leaf blade hairs.)
A combination of traits, for example, unmown leaf height, inflorescences per pot, and internode
length, revealed a cluster of Tifway-like plants that differed from the Blue green and Dark green
patch grasses (Fig. 5). Bermudagrass samples from different
clusters differed statistically (probability of false differences < 5%) in most cases.
Fig. 6. Trimming bermudagrass leaves for DNA
extraction.
Fig. 7. PCR is done in a thermal
cycler.
DNA fingerprinting is one of the most precise techniques available for comparing biological
samples, such as unknown bermudagrasses. The method requires explanation. We meticulously
extracted DNA from 29 unknown plants plus the Georgia standards for Tifway, Tifdwarf, and
Tifgreen (Fig 6).
We then used a chemical reaction called "PCR" (polymerase chain reaction) to produce millions
of copies of the tiny amount of starter DNA (Fig. 7). PCR
jump-starts the copying by using a "primer" fragment, a chemical which we introduced. There are
many kinds of primers, and each acts as a key to identify one specific kind of recognition site in
the grass DNA. Regions of grass DNA lying between pairs of recognition sites get copied. So the
primer acts like a bookend -- wherever primers bind reasonably closely at two nearby sites on the
grass DNA, the region between the sites is copied. Not all grasses have exactly the same
sequences in the same places, thus copying with PCR yields somewhat different DNA fragments
for different grasses, depending upon the location of the recognition sites. The primers that we
used were "random" -- we had no idea ahead of time which kinds of fragments would show up in
which grasses.
By this method, a DNA fingerprint was based on the varying sizes of DNA fragments copied from
different grasses, determined by our choice of primer and the underlying DNA differences
(presence, absence, and location of recognition sites). Using carefully defined PCR conditions,
two samples having identical DNA must provide the same range of DNA fragment sizes, thus
each sample has the same fingerprint. Grasses with different DNA may show differences in DNA
fragment sizes, for the same reasons. So amplified fragments are the first step to producing a
DNA fingerprint.
Fig. 8. DNA fingerprints (reverse image) of 26
bermudagrasses.
Fig. 9. A labeled series of electrophoresis
lanes..
Because the copied DNA fragments differ from one another in molecular size (measured in
"base-pairs"), they can be spread out (Fig. 8) to produce a
distinctive banding pattern like the UPC (universal price code) bars which are scanned at the
supermarket counter. To create separate bands out of the DNA fragments, the amplified mixture
of each bermudagrass DNA was separated by fragment size by being pulled along a gradient of
electrical charge, a technique known as electrophoresis. Parallel lanes containing DNA from
different samples were run in tandem, for the same length of time. The DNA was then stained, to
reveal separated bands, and photographed using UV fluorescence. Some bands represented large,
slow moving DNA molecular fragments, over a thousand base-pairs long. Other bands
represented small, quick-moving DNA fragments. We found only 3 to 8 useable bands per
primer, but by using a choice of several good primers we found over 50 strong, distinctive bands,
any one of which was clearly present or absent for a particular grass (Fig. 9).
Fig. 10. Digitally-scanned bands are represented by
peaks.
The scanned image of the bands (Fig. 10) revealed distinctive
peaks (each band representing DNA of a particular molecular base-pair size) and valleys (regions
with no DNA). The presence or absence of a peak, or diagnostic fragment, distinguished some
bermudagrasses as different.
How repeatable was DNA fingerprinting? After exhaustively seeking the best "primers," the
procedures were repeated at both Fort Lauderdale and Gainesville. Over 80% of the bands found
at one laboratory were found at the other, while repetitions within the same laboratory were
100% consistent. Small bands were frequently inconsistent, and were not used.
Fairways at Golf Course Y, Palm Beach County, Florida (January 1994) showed the same visual
pattern as Golf Course X: Blue green and Dark green patches in a yellowish matrix (Fig. 11).
Fig. 12. DNA fingerprints of patch grasses matched
exactly across golf courses.
DNA evidence confirmed that the visual impression had a genetic basis -- the Blue green and
Dark green patch grasses on Golf Course Y had the same DNA fingerprint as similar patches on
Golf Course X (Fig. 12). The DNA profile of the matrix
bermudagrass samples from both Golf Course X and Golf Course Y matched one another
precisely, and also matched the Georgia standard of Tifway. About 50 DNA bands matched
precisely across samples from different golf courses, a coincidence that could occur by chance
only once in two billion times. The coincidence was beyond normal chance occurrences.
Fig. 13. DNA fingerprints from Tifway matched the
matrix fairway grass on four golf courses. Dark green off-types matched across four golf
courses.
Did these off-types arise on the golf courses as mutations or seedlings? Neither. Somatic
mutations and seedlings are known in bermudagrass and other plants. They generate a medley of
variation, not the repeating pattern we observed. At some point, the patch grasses may have
arisen as seedlings, but before they were planted on the golf courses. There were other
golf courses involved. Not only was the Blue green grass the same on Golf Course X and Y, but
the Dark green grass was the same grass on Golf Course X, Y, Z, and W. Each type patch must
have had a single clonal origin (barring one-in-two billion probabilities), thus they had to have
been planted (Fig. 13). For example, a 1750 base-pair peak
present on Tifway, and four unknown grasses from different golf courses, was absent in the Dark
green variant (yellow line). In contrast, a 1315 base-pair was present on Dark green
bermudagrasses from four different golf courses, but was absent from Tifway and matrix grasses.
So DNA fingerprinting repeatedly matched unknown matrix grasses to the known Tifway,
distinguished among knowns and unknowns, and matched unknown patch grasses across four golf
courses. The evidence does not say when or how the patch grasses were planted, just that they
were the same grasses.
Thus the first accomplishments of the study were:
We proved for the first time, by the combined evidence of morphology and DNA that some
off-types are contaminants that were planted, not mutations.
It was proven that fairway off-types can be readily distinguished by DNA fingerprinting, with
80%+ repeatability between our separate laboratories, in Fort Lauderdale and Gainesville.
Fig. 14. Patches of off-type bermudagrass on a golf
course green..
Fig. 15. Greens bermudagrasses were classified
morphologically into three groups: Ultradwarf, Tifdwarf, and Tifgreen.
Golf course greens are under considerable stress from close mowing and play, thus off-types are a
greater problem there than on fairways (Fig. 14).
We studied 18 greens bermudagrasses; most represented trade samples, but a few were
off-types from golf course greens. Tifgreen had taller unmown height and a greater number of
seedheads than Tifdwarf (please go back to Fig. 5). Differences
were apparent visually, once the samples were grown out (Fig.
15).
Morphology data showed that the distinction among greens bermudagrasses was highly significant
statistically. There were three distinctive groups, Tifdwarf, Tifgreen, and an ultradwarf group.
Three out of three non-certified trade samples of Tifdwarf were statistically different
morphologically from Tifdwarf, but indistinguishable from Tifgreen, thus another
accomplishment was:
The study proved from morphology that some non-certified material sold as 'Tifdwarf' is not
'Tifdwarf,'
rather it is probably Tifgreen..
Fig. 16. Comparison of the tree of relatedness of
bermudagrasses by DNA-RAPDS and morphology. The deeper the cleft between two branches,
the more distantly related are the grasses.
Could this be verified by DNA fingerprints? No. Early in the study, some DNA preparations
showed apparent differences among greens bermudagrasses, but this was based on weak bands.
Upon reextraction of more DNA from the same plants, purportedly diagnostic bands disappeared.
For example, going back to Fig. 8, notice that 18 of the lanes
have an identical profile-these were all greens samples. What do we make of this? The PCR
procedure for copying DNA is sensitive to many factors, some of which can be reasonably
controlled, but others cannot. In the case of the fairway grasses, there were sufficient strong
bands for diagnostic comparisons; we did not have to resort to weak bands. In the case of greens
bermudagrasses, there were no strong diagnostic bands, and the weak bands we attempted to use
were probably derived from weak chemical reactions.
The "tree of relatedness" of all the 32 grasses, for both DNA and morphology, tells a similar story
(Fig. 16). The known and unknown grasses separated into
distinctive branches based on morphology, representing clusters of similarity. Although DNA
fingerprinting yielded the same branches as morphology for fairway grasses, DNA fingerprinting
could not separate the greens grasses the way that morphology could.
Greens bermudagrasses may represent mutations, a sudden genetic change affecting only a single
piece of DNA. Tifdwarf is believed to have arisen naturally by mutation from Tifgreen.
Bermudagrass has thousands of genes, each associated with thousands of DNA subunits, so a
mutation affecting only one of those subunits within one gene would be a very small part of the
organism's total DNA. Happening onto such a genetic change using PCR amplification would
require great good fortune, like finding a needle in a haystack. Therefore, the chance of copying
DNA from the small region where the mutation had occurred would have been unlikely. And we
conclude,
Greens off-types, while distinguishable by morphology, were not distinguishable by DNA,
therefore they are probably point mutations or limited to a single gene.
Fig. 17. Genetic off-types on the greens often appear to
recur in the same form on both the same and different greens.
Did the probable mutations originate on the greens or before the golf course was planted? Here
there was no direct evidence, only circumstantial evidence. Looking at greens off-types over the
years, one of us (PB) has noticed what appear to be repeating patterns of similar-looking patches,
across individual greens, and across different greens on the same golf course (Fig. 17). A "mutation" which recurs in different places was
probably planted, not a series of mutations, because the latter would be different from one
another.
Fig. 18. Mid-Pacific Golf Club, Kailua, Hawaii, showing
one of 15 original greens which were planted to Tifdwarf in the early 1960's.
Other indirect evidence helps refute the recurring mutation theory as an important source of the
variation on greens. Greens on some courses, and occasionally greens on individual courses, have
remained apparently pure genetically for many years. Examples include the Mid Pacific Golf
Club, Kailua, Hawaii (Fig. 18), and Winter Pines Golf Club, in
the Orlando, Florida area. On both courses, the original greens are almost completely pure.
While off-typing is generally progressive, getting worse over the years, it doesn't always occur,
even on relatively old courses.
In the late 1970's, Boca Greens, a Palm Beach County, Florida golf course, had off-type
variations on all 18 greens, except the practice green near the clubhouse was pure. This was
puzzling at first, because the practice green had been planted at the same time by the same
contractor as the 18 greens on the golf course. On checking the source of plant material, the golf
course superintendent learned that the grassing contractor had run out of sprigs after planting 18
holes, and had to go back for another load to do the practice green. With these anecdotal
observations, we conclude:
Circumstantial evidence suggests that greens off-types are sometimes planted.
Fig. 19. Bermudagrasses can be identified by a
combination of chromosomes, morphology, and DNA fingerprinting.
No, not for routine screening. While DNA testing is a powerful tool for confirming visual clues,
it is our opinion that its best use is as a research tool and for ensuring the integrity of foundation
sprig stock. Common bermudagrass contaminants of Tifway were proven to be contaminants, but
the visual evidence was already overwhelming. DNA typing only proved what we already knew.
DNA testing can help ensure cultivar purity, but only if there is there is diligent field sanitation,
record-keeping, and maintenance of ditches.
Because contaminant off-types in greens appear to arise from mutation, it will be difficult to
identify a diagnostic fingerprint. Each potential new mutation could affect a different part of the
DNA, thus even if there were reliable banding differences, new research would be required for
each new mutation. Any new mutation that was not already detectable visually would have an
extremely rare chance of being sampled. On the other hand if, there were any question,
based on visual evidence, that the wrong plant was growing, it would be cheaper and faster to just
kill it.
DNA testing may also give the buyer of grass a false sense of security that the plant material is
clean, when in fact that could only be determined if every leaf blade in the field were DNA-tested.
Therefore, we conclude:
With present technology, DNA testing of greens grasses was ineffective. Fairways
are a different matter, but even there, visual clues are most powerful in first detecting variants,
and DNA is a secondary confirming tool. An example of the effective use of DNA fingerprinting
would be where a field is visually uniform, but it needs to be determined that it is Tifway.
Off-type bermudagrasses can arise from several sources, but the movement of contaminated
planting sources is the first place to stop it.
The golf course superintendent can seek assurances that the sprig vendor is aware of the potential
problem of contamination and has made efforts to prevent it. While the recurrence of new
mutations is a possible source of contamination, it appears that many instances of off-types are
due to contamination. Therefore, the golf course superintendent can:
Request certified plant material be used when it is available. The Southern Seed Certification
Association (334-821-7400, FAX 334-844-4901, Box 2619, Auburn, AL 36831) currently
certifies golf course bermudagrasses in Florida. The Association inspects fields in Florida.
Request a list of 1-4 year-old plantings from prospective grassing contractors.
Request written documentation on where the source grass originated.
Personally inspect prospective source fields, hopefully having the opportunity to look at areas
where the grass has been mown closely over several months. But don't expect to see too much!
Include appropriate performance specifications in the bid, with a timeline for inspection of
quality and consistency, and an appropriate remedy (e.g., a performance bond). Most users are
very concerned over consistency.
Warn the greens committee that genetic purity does not exist, certainly not beyond the first
few years, but that off-typing can be reduced for a while with considerable effort and expense.
Walk newly sprigged areas on a daily basis as they are growing in, and use nonselective
herbicide (e.g., glyphosate) to kill any questionable patches.
We thank all the golf course superintendents of Florida, who have provided samples,
encouragement, observations, and financial support for this research. We thank Rita Di Bonito,
Diane L. Johnston, and Qinbao Li, who provided technical support for this investigation.
turf@gnv.ifas.ufl.edu