analysis of protocol variations on dna yield

127

GATA 9(5-6): 127-133, 1992

ORIGINAL ARTICLES

Analysis of Protocol Variations on DNA Yield

B A R B A R A A R M S T R O N G and

H A R O L D R. G A R N E R

A modified alkaline lysis method for preparing plasmid DNA from bacterial cells has been developed for automation implementation. The objective of this study was to develop a simplified centrifuge-based protocol that can process samples in a microwell plate. These manual experiments and parametric studies show that the alkaline lysis method can be modified significantly to enable DNA preparations to be done rapidly and reliably by an automated system. The conclusions of this study were (a) centrifugation at <1500 g is sufficient, (b) centrifugation times need not be extended to compensate for the reduced force, (c) reactions can be done at room temperature, (d) reagent volumes can be reduced over those typically used, and (e) certain reagents can be combined to simplify the handling of fluids.

Introduction The alkaline lysis method is one protocol for extract- ing and purifying plasmid DNA from bacterial cells, and several versions of that protocol have been pub- lished [ 1-3]. These protocols cannot be implemented directly on an automation system designed for high- throughput operation, because they are too complex or operate in centrifugation regimens that do not enable the use of microwell plates for sample handling. It is useful to know how each variable in the protocol affects the yield and purity of the plasmid DNA so that reproducible plasmid DNA preparations (preps) can be obtained both manually or when automated.

Recently there has been a need for automating the plasmid DNA preparation procedure to process rapidly hundreds or thousands of bacteria containing DNA plasmids at a time for the human genome project. Centrifuge-based DNA prep systems are avail- able that can extract DNA by the alkaline lysis method; however, these systems cannot be scaled up to meet the sample throughput demands for the human ge-

From the Institute for Development of Advanced Technology, General Atomics, San Diego, California, USA.

Address correspondence to Dr. H.R. Garner, Institute for Development of Advanced Technology, General Atomics, PO Box 85608, San Diego, CA 92186, USA.

Received July 1992; revised and accepted August 1992.

nome project because they process samples in microfuge tubes [4]. To develop a protocol that is sim- ple, rapid, reliable, and possible in a microwell-plate format, it is necessary to understand how variations in each of the plasmid isolation steps will affect the final product.

Here we present the results of a systematic study of the alkaline lysis method. We studied the effects of cell number, reagent volume, reagent temperature, sequence of adding reagents, centrifugation time, and centrifugal force on the yield and purity of plasmid DNA. Also presented is a revised and simplified procedure, based on the relaxed constraints and variables identified, for preparing plasmid DNA from bacterial cells. This new protocol has been implemented on a system designed and built for high-throughput, un- attended sample preparation. This device is described in a companion article ("High-Throughput DNA Prep Systems") [5].

Ma te r i a l s and M e t h o d s

Bacterial Cells The DH5-ot cells were transformed with an amp r derivative of Bluescript plasmid DNA [6, 7]. Cells were grown in an Erlenmeyer flask in T broth (12 g of tryptone, 24 g of yeast extract, and 4 ml of glyc- erol, to which 900 ml of water is added; and 100 ml of K1/2HPO4:0.17 M KH2PO4 and 0.7 M K2HPO4) supplemented with ampicillin at 50 mg/gl. For the purpose of this study, overnight preparations were grown at 37°C in a shaker waterbath and then dis- tributed into the microwell plate. For the purpose of this study, it was desired to guarantee no variations attributable to culture cell densities from well to well. Protocols have been developed to culture 0.5-0.75 ml of plasmid- or cosmid-containing bacteria directly in deep-well microwell plates (Beckman Instru- ments, Fullerton, CA). Basically, the wells are sealed by wide tape (3M, New Zealand), the plates are placed on their sides to maximize the enclosed air interface and grown overnight in a shaker incubator. Typically, the concentrations of viable cells following an overnight preparation was 4 × 109 cells/ml for both the flask- and microwell-grown cultures. This corresponds to a turbidity of 0.25 mm -l measured spectrophotometrically at 500 nm.

Preparation of Plasmid DNA Basic Procedure

We studied variations on the basic miniprep protocol [3], which is defined in Table 1. As each parameter

© 1992 Elsevier Science Publishing Co., Inc., 655 Avenue of the Americas, New York, NY 10010

1050-3862/92/$5.00

128

GATA 9(5-6): 127-133, 1992 B. Armstrong and H.R. Garner

Table 1. Basic P lasmid Miniprep

1. Transfer 1 ml of an overnight culture to a 1.7-ml Epindorf tube.

2. Centrifuge at 10,500 g for 30 s (spin 1). 3. Aspirate media, then resuspend the cell pellet in 100 ~xl of

ice-cold solution I (50 mM glucose, 25 mM Tris-Cl pH 8, and 10 mM EDTA pH 8) by vortexing for 10 s.

4. Add 200 ILl of room temperature solution II (0.2 N NaOH and 1% SDS) and mix by inverting several times.

5. Add 150 I~1 of ice-cold solution Ill (3 M potassium and 5 M acetate) and mix by inverting and shaking for 10 s.

6. Incubate on ice for 3 min. 7. Centrifuge 5 min at 10,500 g at 4°C (spin 2). 8. Transfer liquid to a fresh tube and then precipitate the DNA

by addition of 900 I~1 of 100% EtOH at room temperature. 9. Vortex for 5 s.

10. Incubate at room temperature for 2 min. 11. Centrifuge 5 min at 10,500 g at 4°C (spin 3). 12. Remove EtOH by vacuum aspiration. 13. Wash pellet by adding 1 ml of 70% EtOH. 14. Remove EtOH by vacuum aspiration. 15. Air-dry pellet for 10 min. 16. Resuspend DNA in 50 pA T~oEi pH 8. 17. Incubate at room temperature for 30 rain prior to DNA

concentration measurements.

was studied, all other parameters were held constant. All manual studies were conducted in 1.7-ml Epin- dorf tubes using a microcentrifuge (Beckman Mi- crofuge 12).

DNA Quantization Absorbance readings were made with a Beckman DU 65 Spectrophotometer with 5 to 10 I~1 samples held in capillaries in the Beckman Ultramicrocell [8]. The accuracy of the measurements is - 3 % since no pi- peting is needed in the measurement step. Absorb- ance at 260, 280, and 320 nm was used to determine the yield and purity of the plasmid nucleic acid. A Warburg [9] value was calculated in the following way:

[nucleic acid] = 2 0 ( - 36[A280 - A320] + 62.9 [A260 - A320]) (Ixg/ml)

where A260, A280, and A320 are the optical densities at 260, 280, and 320 nm, respectively. These adjusted values were added together to give the War- burg value for a 1-cm path-length cuvette, and this number was multiplied by 20 (because the path length of the ultramicrocell is 0.5 mm) to give the nucleic acid concentration in micrograms per milliliter (~xg/ml).

The concentration of double-stranded DNA was

measured using fluorescence enhancement of a Hoechst 33258 dye-DNA complex. Measurements were made using a TKO 100 minifluorometer (Hoefer) and fluorescence capillary cell (Helix).

The dsDNA concentration measured fluorometr- ically and total nucleic acids measured spectrophotometrically indicate that there is considerable residual RNA. For some applications, it may be necessary to remove the RNA after executing the basic prep protocol. All samples had a A260-A280 ratio of i> 1.9, indicating that the sample was relatively pro- tein free.

Gel electrophoresis confirmed that plasmid DNA content in the prep dominated the bacteria genome DNA and that there was considerable residual RNA. Within the limitations of quantization by integrating the intensity bands or the gel, the spectrophotometric and fluorometric measurements were verified.

The quality of the resulting preps were monitored by enzyme digestion, followed by confirmation of complete digestion of a gel and by sequencing seg- ments of the DNA. This is further discussed in a companion article [5].

Results and Discussion

Preparation Yield and Quality The basic protocol (Table 1) was followed except for step 1. A 250-ml culture was dispensed in duplicate into 50-ml or 15-ml tubes at the following volumes: 40, 30, 20, 10, 8, 6, 4, 2, 1,0.5, and 0.25 ml. These tubes then were centrifuged in an IEC centrifuge at 960 g to pellet the cells. The excess media was re- moved to leave 1 ml of concentrated culture. The cells were resuspended by mild vortexing and trans- ferred to 1.7-ml Epindorf tubes. Processing these samples followed the basic protocol, starting at step 2.

Figure 1 shows that the alkaline lysis method described can be used to process an overnight culture, yielding an increasing concentration of nucleic acid for every milliliter of cells processed for up to 20 ml of culture. For large volumes of cells (>20 ml), the yield begins to saturate. This indicates that there is insufficient chemistry (solutions I, II, and III) to process >20 ml of culture. Since most experiments process only 1-5 ml of culture, this is not a limitation. Viewed in another way, Figure 2 shows that the efficiency of the procedure drops as the number of cells processed increases. Most minipreps process 1-2 ml of culture where the yield efficiency is the highest. One-milliliter cultures in microwell plates processed by an automation system also would have


Protocol Variations on DNA Yield

129

GATA 9(5-6): 127-133, 1992

30

Figure 1. The double-stranded DNA yield increases linearly as the starting cell volume (total cell number) increases for small culture volumes.

A 20 b=

_.R _=

,<

== ' l l

10

[ ]

[ ]

B

§

[ ]

B

D

10 20 30 40 50

Culture volume (ml)

high efficiencies. This indicates that the volume of reagents used to process a given amount of culture could be reduced without saturating the reagents' ability to process the plasmid DNA.

Comparing Figures 1 and 3, it can be seen that the typical yield for dsDNA was --4 p.g/ml of culture; there was --20 times as much residual RNA re- maining after the prep. This was true for all data presented in this article. RNase is added to reduce this contamination if the RNA must be eliminated for the prep to be of use by follow-on processes-- for example, automated sequencing.

Reagent Volume Sensitivity The basic protocol was followed except for steps 3 - 5, where the amount of solution dispensed in each step was 200%, 100%, 50% 20%, or 10% of that standardly used.

Figure 4 shows that 1 ml of an overnight culture can be processed safely with as little as 1/20 the volume of reagents without a reduction in plasmid DNA yield.

Reagent Combination Sensitivity The basic protocol was followed except for steps 3 - 5, where various combinations of solutions I, II, and III were combined prior to addition to the culture.

Solutions I and II do not have to be added sepa- rately; likewise for solutions II and III. The yield is unaffected by these combinations. If all solutions are combined, the yield is reduced to - 3 3 % of that for sequentially added solutions.

Reagent Temperature Sensitivity The basic protocol was followed, except that half of the tubes were incubated and centrifuged at room temperature with all solutions at room temperature while the other half were incubated on ice and centrifuged at 4°C in a cold room with all solutions at the temperatures stated in the basic protocol.

There is no reduction in plasmid DNA yield when the procedure is performed at room temperature (RT), 25°C. Therefore, the procedure need not be performed on ice or at 4°C (cold) within the centrifuge. Subsequent experiments using I, II, and III stored at room temperature for 24 h to 1 month indicate that it is unnecessary to store these reagents at 4°C. It is surmised that since lysozyme is not a component of solution I of our basic miniprep, the solution is much less sensitive to storage conditions. Therefore, the prep need not be performed in an expensive refrig- erated centrifuge or cold room.

We conclude that two reasons explain the insen- sitivity of the yield to the protocol temperature conditions. First, since solution II was at room


130

GATA 9(5-6): 127-133, 1992 B. Armstrong and H.R. Garner

==

A W

<

B

m m

B B [] [ ] [ ]

B

B

m m

m m

1 0 2'0 30

C u l t u r e v o l u m e (ml)

E

!

40 50

Fisure 2. The efficiency (yield per volume of culture) of the DNA preparation decreases as the starting culture volume is increased.

1000

800

A U

6 0 O

u 400

z

200 |

F 0

B E []

O [ ]

ra

I I B [ ]

B

10 2'0 3'0

C u l t u r e v o l u m e (ml)

!

40 50

Figure 3. Spectrophotometric data indicate that there is a significant amount of non-dsDNA present following the basic protocol.

temperature, when it was added to solution I (ice cold) the resulting combined temperature was measured to be 16°C significantly above the 4°C target temperature. Second, the convective (as it would be

in air in a centrifuge) and conductive (as it would be in contact with ice) thermal transport through the 1.7- ml Epindorf tubes (0.8-mm wall thickness) were measured. The e-folding time for convection was 6



131

GATA 9(5-6): 127-133, 1992

Figure 4. For volumes of reagents, solutions I, II, and III, to process 1 ml of culture, the yield is independent of reagent volume over a range of 10%-200%. The recommended volumes in the basic protocol for solutions, I, I1, and III correspond to 100%.

4 A

2

HI

E

[ ] lid

E []

ID E

E [] m

! !

1 00 2 0 0 3 0 0

Reagent vol (%)

min and for conduction was 3 min. Therefore, es- pecially during centrifugation, since the convective transport time was comparable to the centrifugation time, the temperature of the centrifuge would have little effect on the sample temperature.

Centrifugal Force Sensitivity The basic protocol was followed except for steps 2, 7, and 11. The durations remained the same, but the centrifugal force was varied by adjusting the speed of the rotor.

Figure 5 shows that between 300 g and 3600 g there is only a slight effect on DNA yield. This data is for spin 3 and is representative of data for all three spins. The yield is independent of centrifugal force. The centrifugations can be performed at a slower speed without affecting the nucleic acid yield. Most important is that the centrifugal force required for each spin be ~< 1500 g, the limit for microwell plates and appropriate swinging-bucket rotors. Although the yield was constant down to 300 g, there is a practical reason for using higher forces: specifically, the pellet (cells or DNA) will be better attached to the wall of the microfuge tube and consequently will not be lost during the aspiration of the supernatents. Conversely, using lower forces, the pellet can be resuspended

more easily because it is not highly compacted on the wall of the microfuge tube.

Centrifugation Duration Sensitivity The basic protocol was followed except for steps 2, 7, and 11. The centrifugal force remained the same, but the duration was varied by adjusting the speed of the rotor.

An investigation of the required centrifugation time was studied for each spin and it was determined that the duration need not be increased to compensate for the reduced centrifugal force. Figure 6 shows the samples we processed at 1500 g which is significantly below the 10,500 g that is given in the basic protocol. The data indicate that, for spin 1, DNA yield is independent of centrifugation time. For spins 2 and 3, the yield is also independent of centrifugation time.

The sample processing throughput for an automation system is determined for the number of samples that can be processed in parallel and the time required for each step of the protocol. It is significant that although the force is reduced to approximately one-seventh that of the basic protocol, no compen- sating increase in duration is needed that would se- verely restrict the throughput of an automation system.


132

GATA 9(5-6): 127-133, 1992 B. Arms~ong and H.R. Garner

&

Z 2 q~

B ra

II

II In

0 10~00 'l 2000 3000

F o r c e x g

4000

Figure 5. The dsDNA yield is independent of the centrifugation force of spin 3 over the range investigated. This is also true for spins 1 and 2 (data not shown). To show clearly the dependence at low forces, only data up to 3600 g is shown. At the maximum speed of the centrifuge 10,500 g the yield was also ~ 4 ~g.

3

&

Z 2 . g

B B

B ra

I I I I

0 2 4 6 8

D u r a t i o n o f Spin 1 (min.)

Fisure 6. The dsDNA yield is independent of the centrifugation time of spin 1 over the range investigated. This is also true for spins 2 and 3 (data not shown). The centrifugal force was 1500 g.

10

Conclusions

The data presented show that the alkaline lysis method has considerable flexibility and can be simplified for automation. Most significant for automation is that centrifugation can be done in a range in which mi-

crowell plates can be used. In addition, fluid handling can be greatly simplified for the robotic system since no refrigeration is needed and fluids can be combined, providing considerable savings when con- structing a system.

The results of this parametric study of the alkaline



133

GATA 9(5-6): 127-133, 1992

lysis protocol can also be used to improve , s t ream- line, or debug preps done manua l ly .

The series of exper iments conduc ted to de te rmine the sensi t ivi ty of the D N A yie ld and qual i ty to variations o f procedure and parameters o f the bas ic protocol were examined independent ly , keep ing all o ther parameters fixed. Wi th the knowledge o f the independent sensi t ivi t ies , a new protocol was devised to take advantage of the var ious s impl i f ica t ions to arr ive at a protocol that can be pe r fo rmed in mic rowel l

plates.

We thank D. Burbee for providing DH5ct and Bluescript plasmid and for his comments. We also thank C. Kay and P. Staskus for their comments.

References 1. Birnboim HC, Doly J: Nucleic Acids Res 7:1513-1523, 1979 2. lsh-Horowicz D, Burke JF: Nucleic Acids Res 9:2989-2999,

1981 3. Sambrook J, Fritsch EF, Maniatis T: Molecular Cloning: A

Laboratory Manual, 2nd ed. Cold Spring Harbor, NY, Cold Spring Harbor Laboratory, 1989

4. AutoGen sales literature 5. Garner HR, Armstrong B, Kramarsky DA: Genet Anal Techn

Appl 9:134-139 [this issue] 6. Alting-Meese MA, Short JM: Nucleic Acids Res 17:9494

1989 7. Short JM: Nucleic Acids Res 16:7583-7600, 1988 8. Garner HR, Lewis AC, Thomas MK: Rev Sci Instrum 61:1433-

1435, 1990 9. Warburg O, Christian W: Biochem Z 310:384f, 1942


analysis of protocol variations on dna yield

Documents