About two weeks ago I posted a note seeking to gather information about
Arabidopsis transformation using Agrobacterium and vacuum infiltration.
The questions were:
1) What Agro strains have worked well for you and your colleagues? Which
strains have not worked well? Which ecotypes did you try?
2) Which ecotypes have you successfully transformed? Which haven't
worked? Which ecotypes seem especially willing to be transformed? Did the
Agro strain used seem to matter?
3) Have you observed any primary transformants that are or are apparently
homozygous for the insert DNA?
OUT OF ABOUT A DOZEN REPLIES, THE FOLLOWING POINTS EMERGED:
- The Col-0, No-0, and Ws-0 ecotypes have all been transformed using this
method, as have a few others. No one specifically said they had succeeded
using Landsberg genotypes (if anyone has succeeded with La, please post
- Different Agro strains have been successfully used, with a general
preference for C58 strains carrying a disarmed C58 Ti plasmid (such as
- Most transformation events apparently occur in cells formed from the
floral primordium, but earlier transformation has also been observed
(detected as homozygous primary transformants or multiple sibling
transformants with the same insertion).
Also noteworthy was the successful use of LBA4404 by some. Many of the
failures I have heard about have been from people using LBA4404, but
obviously some have had success with this strain.
One thing generally NOT mentioned by those who replied, but which I know
from other communications, is that many people have tried this method and
failed, or have failed in their first attempts. So before listing some of
the replies, I might add four generic comments about the method:
1) It is important to have healthy, happy, well fertilized plants (don't
overwater, be sure they get plenty of light - 150 microEinsteins or more -
they ought to just really look good).
2) Infiltration should be done at a time (after clipping the first bolt
off) that there are lots of litte inflorescences that are just emerging
from the rosette, with lots of young unopened floral buds, so that you see
a few 10cm tall bolts but the pot is full of bolts that are more in the 1-5
cm range. The important thing is to have lots of young bolts and not too
many open flowers and very few if any siliques.
3) The Agro and Arabidopsis genotypes used no doubt has an effect, and if
you aren't constrained by other factors I know that C58 derivatives with a
C58 Ti plasmid have worked well for many people in Col-0.
4) Initially at least, Follow The Directions in your protocol.
HERE ARE EXCERPTS FROM SOME OF THE REPLIES:
1) We only try the C58 strain for the infiltration method and it work well
with pGKB5, pBin19 or pDE1001 binary plasmid.
2) We transform routinely Ws and ColO
3) We are testing if we have some primary homozygous transformants in
our collection. We are
testing the transformants that for 100 seeds of the T2 generation
tested in vitro on kanamycine didn't show kanamycine sensitive plants.
4) We try 2 methods of infiltration, with or without taking the plants
out of the soil, and with the first method (rooting out) we had at least 10
times more transformants.
One observation of importance is that in one experiment where whe
checked the 12 transformants, all were due to the same insert and
were obviously clonal. Therefore, it is apparent that at least some
transformation events arise from transformed sectors.
I initially used LBA4404 and Bin19/kan selection to transform Columbia. I
got 12 transformants from one pot, following your protocol BUT SOUTHERN
BLOT ANALYSIS SHOWED THEM ALL TO BE DERIVED FROM THE SAME TRANSFORMATION
I HAVE SINCE USED GV101 WITH BOTH KAN AND HYG VECTORS TO TRANSFORM COLUMBIA
(AGAIN USING THE PROTOCOL YOU POSTED TO THE NET) AND THIS HAS WORKED VERY
WELL. WE FIND ABOUT 1 TRANSFORMANT PER 1000 SEED PLATED AND SOUTHERN
ANALYSIS SHOWS THEM TO BE DERIVED FROM A SINGLE TRANSFORMATION EVENT.
agro strain: GV3101(pMP90)
vector: pBIN19 derivatives, pOCA18 derivatives
Arabidopsis ecotypes: Col-0, Col-0xNd-0 hybrid, No-0 (or a closely related
ecotype) (All work fine.)
I myself routinely use Col-0 these days, and the typical efficiency I have
been obtaining is 5-20 transformants/4000 seeds. I haven't intensively
looked at insertion sites within a same batch of transformants (i.e., from
a same pot). So I cannot tell whether all those within a same batch are
independent. However, it is clear that there are some, if not all,
independent events within a same batch because plants from a same batch
include ones with different segregation ratios. I have never seen anything
like homozygous for the insertion in primary lines.
1) ASE, in combination with pCGN vectors, works very well. Max. frequency
better than 1:50.
2) No-0, Col, Ws all work well, no obvious differences between these.
3) We have not observed homozygous T1s (based on analysis of >50 lines).
Apparent no. of inserts in all lines between 1 and 2
4) Additional comments:
In contrast to your original suggestion, we have had most luck with watering
plants on the day of infiltration, and not watering them for several days (5-7
days) afterwards. Doing it the other way around killed a lot of plants.
Also, this method has worked very well for a transgene for which I could
previously not get any (expressing) transformants with the Valvekens
5) We haven't checked this yet, but I'm curious how many people have
confirmed that transformants are indpendent. I have limited data based on
variability of staining from a GUS transgene, variability in phenotype of a
dominant transgene, and number of inserts. These indicate that we get
independent transformants from small pools (a dozen plants or so).
We have had very good results
using Columbia with the Agro strain C58C1 (pMP90). So far I've looked at seed
from over 150 individual infiltrated plants and all but a couple have given me
a good number of selection resistant plants (we're using kanamycin). 25 of
the 26 plants tested for GUS activity were GUS positive. The plants aren't at
a stage yet where they can be tested for homozygosity so I have no information
about that. I should mention that we are using a surfactant called Silwet
L-77 in our infiltration buffer, as suggested in a protocol by Takashi Araki,
which I believe may be at least partially responsible for our good results.
[Editor's Note : Use of Silwet L-77 was actually first suggested by Brian
Staskawicz. It is toxic to Arabidopsis at concentrations much above 0.01%,
but when used at 50 ul/liter it can make plants that are stubborn to
infiltrate much more willing. It is available from Union Carbide, but only
in ridiculously large quantities. I and many others generally do not use
L-77, although it has helped in some cases. - Andrew Bent]
GV3101 practically killed the plants after infiltration using the protocol that
Andrew Bent posted last year. LBA4404 is not as lethal and the plants get
larger and you get more seed. (However, I do not yet know the relative
proportion of transformed seed in the GV3101 lots, so a little seed may still be
OK if lots of it is transformed.) I've heard anecdotally that folks don't
actually submerge the rosettes when using GV3101, but just the bolts. Is this
LBA4404 has actually seemed to work OK for me. From the 2 independent
experiments that I did with that strain, at least one pot out of 3-5 pots
infiltrated had a fairly high proportion of Kan-resistant seeds. Southerns
remain to be done, however, so this is a preliminary result.
You can use my name, since I'd be interested in corresponding with others who
are working on this.
University of Wisconsin
I have been using Agro ASE and Columbia with great success.
I have observed a couple of visible mutations (pin-like and curled leaf) in the
primary transformants (out of a couple of hundred lines). These mutations
could of course be dominant. Several other bits of anecdotal evidence lead me
to believe that the transformation event is taking place late, probably during
pollen and ovule development. First, the frequency of transformation is quite
consistent from transformed plant to transformed plant, roughly between 0.5 and
5% of the seeds are KanR. If the transformation event could occur early as
well as late, then you might expect the frequency in some plants to be very
high (>30%) and we don't see that. The second bit of evidence comes from
analysis of enhancer trap lines. If the transformation happened early, then
"independent" lines derived from the same plant would yield the same enhancer
trap GUS staining pattern. We don't see this so it seems that each KanR seed
....there is so much variation in the frequency of
intergration. From <1 to >50 (probably many hundred)transformants per
pot of plant (~10plants per pot).
3 experiments completed all with LBA4404 x WS
variable results: seeds were planted in pots and progeny from each pot
bulked and tested (KanR)
expt. #1 165 plants in 11 pots = 18 transformants from 3 pots, 0
transformants from 8 pots.
expt. #2 150 plants in 10 pots = 4 transformants from 3 pots, 0
transformants from 7 pots.
expt. #3 60 plants in 4 pots = 118 transformants from 4 pots
We did nothing obviously different in the 3rd experiment.
no homozygotes (have been observed).
Thanks to all who replied -
Department of Agronomy
University of Illinois
Urbana, IL 61801