Abstract

	A protocol to semi-automatically pipette the components for 384 DNA 
sequencing reactions using a reduced volume of AmpliTaq DNA polymerase 
FS into a 384 well thermocycler compatible microtiter format plate has been 
developed on the Robbins Scientific, Inc. Hydra96 modified with a plate 
positioner for the Cycleplate384.  After the cycle sequencing reactions are 
complete, the Hydra96 is used to apply the reactions to the top of Sephadex G-
50 filled microtiter format filter plates for the removal of unincorporated 
terminators.  By reducing the amount of AmpliTaq DNA polymerase FS 
reaction mix required for each reaction and semi-automating both the tedious 
and error prone reaction component and cleanup pipetting steps, the cost per 
reaction is reduced, the resulting DNA sequencing data is more reproducible, 
and the throughput of each thermocycler has been quadrupled.

Introduction

	We have developed a cycle sequencing reaction protocol using the 
AmpliTaq DNA polymerase FS terminator reaction kit from Applied 
Biosystems, Inc. (ABI, 402079).  This protocol, as shown in Figure 1, reduces 
the cost of reaction mix by one fourth and primer by one half, and uses one 
fourth the template DNA compared to the ABI protocol (1) resulting in a 
four-fold reduction in reaction cost while improving the quality and read 
length of the resulting DNA sequencing gel readings.

	In a continuing effort to automate, streamline, and improve DNA 
sequencing, we and others (2-7) have taken advantage of new technologies 
when possible.  Recently, Robbins Scientific, Inc. (RSI) has released a 96 well 
cycle plate with dividers molded into each well, the Cycleplate384, that 
provides a total of 384 wells in a 96 well microtiter plate format compatible 
with existing thermocyclers (Fig. 2, plate).  RSI now has merged the 384 well 
technology with their Hydra96 by commercially offering a 384 well plate 
positioner for the Hydra96.  Using this equipment, we now have 
implemented a semi-automated procedure that takes advantage of this 
commercially available instrument, increasing the throughput of each 
thermocycler four-fold while removing much of the error prone manual 
pipetting tedium and reducing the cost of plasticware involved in the process.

Materials and Methods

	A protocol has been developed that allows the sequencing reactions, 
using non-fluorescent labelled primers and terminator chemistry on double 
stranded DNA sequencing templates, to be semi-automatically pipetted on the 
Hydra96 workstation in a Cycleplate384 (RSI #1038-40-0) which has 96 wells 
divided into four quadrants by molded dividers (Fig.1).  This procedure 
entails programming the Hydra96 to pipette a combined mixture of 6.5 
pmoles primer and two microliters of the AmpliTaq DNA polymerase FS 
reaction mix (ABI, 402079) that includes dNTPs, dye terminators (ABI), buffer, 
and thermostable enzyme, and water for a final reaction volume of five 
microliters to the bottom of each well quadrant of a Cycleplate384 in the 384 
well plate positioner (RSI #1029-41-0). This is accomplished by adjusting the 
dispense height (DH) setting on the Hydra96 to approximately 2100, and 
manually rotating the wheel on the plate positioner to align the tips with the 
first set of 96 well quadrants.  After dispensing the mix into the first set of 96 
wells, the wheel on the positioner is rotated to align the tips with the next set 
of 96 wells, the mix is dispensed, and this process is repeated until reaction 
mix has been dispensed into all 384 wells.  The Hydra96 syringes then are 
flushed with sterile water using the wash function of the Hydra.

	Because the plate positioner does not rotate between the lower left and 
lower right quadrants, it is advisable that the well quadrants of the 
Cycleplate384 be loaded in one of two orders.  One possible loading order is:  
first, lower left; second, upper left; third, upper right; fourth, lower right.  The 
other recommended loading order is the reverse, i.e., starting in the lower 
right quadrant and finishing in the lower left quadrant.

	Once the reaction components are dispensed to all 384 wells, the first 
set of 96 DNA sequencing templates, that are stored in a 96 well format, are 
placed on the Hydra96 and a volume of DNA equivalent to three times the 
volume needed for each reaction is pulled into the syringes.  The template 
containing block then is replaced by the plate positioner holding the 384 well 
plate containing the mixes.  The volume required to dispense 100 ngrams of 
DNA, typically two microliters, is dispensed to the side of the first quadrant of 
each well using a height setting of 1500 to prevent contamination of the 
syringe tip with mix.  Setting the height parameter such that the tips do not 
contact the already dispensed mixes is necessary because the remaining DNA 
in the syringe then is added back to the block from which it came.  When very 
low DNA volumes, i.e.. one microliter, are used, a height setting of 1800 is 
required to prevent the liquid from remaining on the end of the syringe tip.  
The syringes then are flushed with sterile water using the wash function of 
the Hydra96, and the second set of 96 DNA samples is placed on the Hydra96 
and pipetted to the next quadrant as above.  These DNA pipettings are 
repeated until the DNA samples from four 96-well storage blocks have been 
pipetted to all four quadrants of each well.  Then the plate briefly is 
centrifuged to insure mixing of the DNA and the reaction mix, and it is 
placed in a thermocycler. Reactions are incubated under the cycle sequencing 
conditions recommended by ABI (1).  

	After the cycle sequencing reactions have been incubated, the reaction 
products are separated from the unincorporated dye-terminators by 
centrifugation through a Sephadex G-50 column in a 96 well filter plate 
(Millipore  MAHVN4550).  Additional information about this procedure also    
is available at the Millipore web site.

        This procedure, that was developed at the Washington University Genome 
Center with minor modifications by us, entails first preparing the columns 
by manually pipetting 250 ul of hydrated Sephadex G-50 medium column 
matrix to each well of four 96 well Millipore filter plates.   
After taping the matrix containing filter plate over a microtiter plate and
centrifugation for 2 minutes at 1500 RPM in a Beckman GS-6R to pack the columns 
and to remove excess buffer.  This step is repeated once more so that there
is a total of 500 ul of column matrix per well.

Recently however, we have been using the 45 µl Column Loader 
(Millipore cat. #: MACL 096 45) to fill the 96 well filter plates as follows:

Add Sephadex G-50 to the Column Loader. 
       Remove the excess of resin from the top of the Column Loader with the scraper supplied. 
       Place MultiScreen HV Plate upside-down on top of the Column Loader. 
       Invert both MultiScreen HV Plate and Column Loader. 
       Tap on top of the Millipore Column Loader to release the resin. 
       Using a multi-channel pipettor, add 300 ul of ddH2O to each well to swell the resin.
       Incubate at room temperature for 3 hours.
       Once the minicolumns are swollen in MultiScreen plates, they can be sealed 
       with saran wrap and stored in the refrigerator at 4 deg C for several weeks.
       A batch of plates also can be stored in the refrigerator at 4 deg C
       for several weeks in a sealed plastic container with a damp towel
       to assure the plates are kept moist.

	When needed, the matrix containing filter plate is taped over a microtiter
plate and centrifuged for 2 minutes at 1500 RPM in a Beckman GS-6R to pack the
colums and to remove any access buffer.

	When the reaction incubation has been finished, the Cycleplate384 
containing the completed cycle sequencing reactions is placed in the 384 well 
plate positioner.  To insure recovery of as much sample as possible and to 
compensate for any drying of the Sephadex G-50 columns, ten microliters of 
sterile water is added to the first set of 96 completed cycle sequencing reactions 
by drawing ten microliters of sterile water into the syringes of the Hydra96, 
placing the 384 well plate positioner containing the Cycleplate384 with the 
completed cycle sequencing reactions on the Hydra96, and dispensing the 
sterile water to the bottom of the first set of reactions with a dispense height 
(DH) setting of 2180.  The diluted samples are drawn into the syringes of the 
Hydra96 with the fill height (FH) setting of the Hydra96 at 2180.  The plate 
positioner containing the  384 well plate is removed from the Hydra96 and 
the first Sephadex G-50 containing filter plate, taped on top of a clean V-
bottom microtiter plate, is placed on the Hydra96.  The samples then are 
dispensed to the top of the columns using a height (DH) setting of 2050, and 
the syringes are flushed with sterile distilled water using the wash function of 
the Hydra96.  These steps are repeated for the remaining three sets of samples 
after manually rotating the wheel of the plate positioner so the Hydra96 adds 
water to and aspirates the desired sample set.  Each V-bottom microtiter plate 
taped to a Sephadex G-50 containing filter plate and the cycle-sequencing 
reactions then is centrifuged for 2 minutes at 1500 RPM in a Beckman GS-6R 
to collect the filtrate in the lower V-bottom microtiter plate.  After loosely 
covering each V-bottom microtiter plate with a Kim wipe, drying the samples 
under vacuum, and sealing with plate sealers (Dynatech 001-010-3501), the 
reaction products can be stored frozen at -20C for several weeks without 
noticeable degradation.

	It is imperative that the height of the columns be consistent, otherwise 
two problems can occur.  If a column is shorter than the dispense height 
setting, a drop of reaction product will be left on the needle of the syringe after 
the rest of the samples have been applied to the top of the columns.  This can 
be rectified by manually picking up the filter plate containing the Sephadex G-
50 columns and touching the top of the column to the drop.  A column that is 
taller than the dispense height setting may result in that column being picked 
up by the needle of the syringe when the reaction products are applied to the 
columns.  If the column remains on the needle, the filter plate containing the 
Sephadex G-50 columns must be manually picked up such that the column is 
returned to the correct well of the filter plate by gently teasing it until it falls 
back in place.  If the column should fall off the needle, the column must be 
manually placed back in the correct well of the filter plate with a clean 
spatula.

	It should be noted that is advisable to create three files on the Hydra96 
to perform the above manipulations, the first is to pipette reaction mixes and 
sterile water to the bottom of the wells of the Cycleplate384, the second is to 
pipette DNA to the side of the wells of the Cycleplate384, and the third is to 
transfer the cycle-sequencing reactions from the Cycleplate384 to the G-50 
columns in the filter plates after incubation for removal of unincorporated 
terminators.  The three different files are necessary because of the different fill 
height (FH) and dispense height (DH) settings for each of these three different 
steps.

Conclusions

	The procedure reported here for semi-automated low volume DNA 
sequencing reaction pipetting in a 384 well format and subsequent semi-
automated sequencing reaction cleanup has several distinct advantages over 
the previously reported manual methods (1-7).  The first and most obvious 
advantage of automation is to reduce manipulation time.  Preparing 384 
reactions using the Hydra96 in conjunction with the Cycleplate384 reduces by 
a factor of three the time necessary to manually prepare the same number of 
reactions, and halves the preparation time for 384 reactions using the Hydra96 
with individual 96 well cycle plates.

	Secondly, as with any manual method, the opportunity for human 
errors exists because of the repetition of tedious manual manipulations. 
However, the Hydra96 removes the tedium of manually preparing these 
reactions, thereby decreasing the opportunity for introducing manual 
pipetting errors.

	Lastly, the Cycleplate384 used in conjunction with the Hydra96 results 
in a significant cost savings.  First, the throughput of each thermocycler is 
increased by a factor of four, and the total cost of cycle plates is reduced to less 
than a third of that for the 96 well plates.  Second, the Hydra96 easily can be 
adjusted to accurately and reproducibly pipette microliter volumes thereby 
further reducing the cost of each reaction kit by a factor of four.  Finally and 
most importantly, since implementing this Hydra96 based semi-automated 
protocol, the resulting DNA sequence data quality has increased in 
consistency, and the number of failed gel readings due to manual pipetting 
errors has been dramatically reduced.

Acknowledgments

	We thank Elaine Mardis at Washington University at St. Louis for 
originally developing the concept of the 384 well plate positioner for the 
Nunc plates, and C. Adonis Reece in our lab for testing the protocol.  This 
research was supported by grants from the NIH/NCHGR, NIH-AI,  and NSF-
EPSCoR.

References

1. ABI PRISM Dye Terminator Cycle Sequencing Ready Reaction Kit with 
AmpliTaq DNA Polymerase, FS.  Protocol P/N 402078 1995.
2. Mardis, E. R., and B. A. Roe.  1989.  Automated Methods for Single-
Stranded DNA Isolation and Dideoxynucleotide DNA Sequencing Reactions 
on a Robotic Workstation.  BioTechniques 7 : 840-850.
3. Smith, V., C. M. Brown, A. T. Bankier, and B. G. Barrell.  1990.  
Semiautomated Preparation of DNA Templates for Large-Scale Sequencing 
Projects.  DNA Sequence 1 : 73 -78.
4. Ansorge, W. , ed.  1996.  DNA Sequencing Strategies.  J Wiley and Sons, 
New York, NY. 
5. Chissoe, S. L., and Y. F. Wang, S. W. Clifton, N. Ma, H. J. Sun, J. S. 
Lobsinger, S. M. Kenton, J. D. White, and B. A. Roe.  1991.  Methods:  
Companion to Methods in Enzymology.  3 :  55-65.
6. Bodenteich, A., S. Chissoe, Y. F. Wang, and B. A. Roe.  1993.  Shotgun 
Cloning as the Strategy of Choice to Generate Templates for High-
Throughput Dideoxynucleotide Sequencing, in Automated DNA Sequencing 
and Analysis Techniques.  Venter, J. C., ed.  Academic Press, London.
7. Roe, B., J. Crabtree, and A. Khan.  1996.  Essential Techniques Series.  
Rickwood, D., ed.  John Wiley and Sons, Chichester, UK.