722-723
Nucleic Acids Research, 1995, Vol. 23, No. 4
© 1995 Oxford University Press
Pulsed field gel electrophoresis of bacterial DNA
isolated directly from patients' sputa
Stefanie Breitenstein, Burkhard Tiimmler and Ute Romling*
Klinische Forschergruppe and Abteilung Biophysikalische Chemie, OE 4350, Medizinische Hochschule Hannover,
Konstanty Gutschow Strap"e 8, D-30623 Hannover, Germany
Received October 18, 1994; Revised and Accepted December 16, 1994
Direct examination of bacterial chromosomal DNA from environmental and clinical sources in vivo was confined to the
amplification of selected genes by PCR (1). All other procedures
referred to for bacterial genome analysis require the organisms to
be cultured on artificial media. However, we have developed a
pulsed field gel electrophoresis (PFGE) procedure enabling the
direct analysis of bacterial genomes colonizing the respiratory
tract of patients with cystic fibrosis (CF) (2). The method exploits
the differential susceptibility of bacterial and human cells to
hypotonic lysis. Released eukaryotic DNA is removed from
bacterial cells by enzymatic degradation and/or field inversion gel
electrophoresis (FIGE). The fraction of purified bacteria is
encapsulated into agarose and processed for subsequent PFGE
analysis. Appropriately modified, the protocol should be generally applicable to other patients' specimens and samples from
the environment.
CF patients are colonized with microbial pathogens like
Slaphylococcus aureus or Pseudomonas aeruginosa at a concentration of 107— 109 cells/ml sputum, which is sufficient for
PFGE analysis. Copious amounts of human DNA from immune
cells and bronchial epithelium, however, blur the bacterial
genomic pattern in PFGE. The sputum also contains high
molecular weight mucin which together with the free DNA
makes the sputum viscous and renders bacterial cells hard to
harvest. These problems were overcome by applying the
following procedure.
Two to three ml sputum were heavily mixed with 2ck (vol/vol)
(i-mercaptoethanol at a 1:5 (vol/vol) ratio and incubated at room
temperature -2-5 h under shaking until the sample seemed
homogenous. The suspension was centrifuged for 30 min at 3800
$ and the sediment was resuspended in 10 ml deionized water.
The suspension was incubated for 15 min at room temperature in
order to lyse human cells. The latter centrifugation and lysis steps
were repeated once more. The pellet was resuspended in 10 ml
DNase I buffer (50 mM Na-acetate. 10 mM MgCk 2 mM CaCh.
pH 6.5) and 2.5 (ig DNase I (Sigma. Deisenhofen. Germany) was
added. The released free DNA was hydrolysed for 2 h at 37 °C
under repeated mixing. After two washing steps with SE (75 mM
NaCl. 25 mM EDTA. pH 7.5) the mucins were precipitated from
the suspension, which was still viscous, by centrifugation at 100
? for 15 min. The supernatant containing the bacteria was
removed and centrifuged again at 1400 t> for 15 min. The bacterial
pellet was resuspended in the minimum possible volume of SE.
mixed u ith the same \olume ot 2' < ' \ o]/\o| i |ou meltine ntiarove
lo whom correspondence should be .iddrcssed
1 2 3 4 5
6 7 8 9
10
Figure 1. Direct PFGE sputum analysis of bacterial species. (Al Selective
detection of P.ueruginosa and E.coli in one sputum sample Lanes: I: Sputum
derived plug showing Spe\ digest of bacterial chromosomes. The P.aeruxinosu
chromosome was cleaved into fragments of up to 800 kb. and the E. colt
chromosome into fragments of up to 250 kb. 2. 3: Spe\ digest of P.aeruninmu
control strain. Colony isolate from the same patient's sputum was cultured in
viirii and characterized by standard procedures (3). 4: .S'/;pI-digested Exult
control processed as above 13). 6: Sputum derived plug showing Xha\ digest ot
bacterial chromosomes. E.coli DNA was cut into fragments of up to 550 kb and
P.aeruxinom DNA into fragments of up to 230 kb. 7: Xhul digest of E.coli
control. 8. 9: Xhal digest of P.aeruginosci control. 5. 10: K-oligomer standard.
(Bl Comparison of different isolation procedures: Oral fragment pattern of
P.aeruginosa DNA Lanes: 1: Oral digested sputum plug treated with DNase
I. 2: Oral-digested sputum plug (FIGE procedure). 3: Control isolated from in
Htm culture b \ standard procedures. Electrophoretic conditions: Panel (Al
contour clamped homogenous electric field (CHEF). 210V. linear ramps from
5-25 s for 20 h. 5-60 s for I 7 h. 0.5 x TBE; Panel (B): CHEF. 200 V. linear
ramps for 1-15 s for 25 h. I -30 •> for 14 h. 0.5 x TBE.
type VII (Sigma. Deisenhofen. Germany), poured into moulds
and hardened. Cell lysis was carried out according to established
privedurt> ' 3 > AiMrose-embeddi'd bacterial chromosomes were
Nucleic Acids Research, 1995, Vol. 23, No. 4
cleaved with restriction endonucleases that cut infrequently and
were then separated by PFGE (3)
As an alternative to digestion with DNase I, human DNA was
removed by pre-running the plug for 6 h in a FIGE apparatus with
a linear increase of pulse times from 3 to 60 s and with a
forward:reverse ratio of 3:1. After FIGE the plugs were removed
carefully from the slot and agarose-embedded bacterial chromosomes were digested according to the standard protocol (3). This
alternative protocol is more laborious, but does produce stronger
signals and more resolved macrorestriction fragment patterns,
which probably result from the removal of yeast or fungi
chromosomes during the FIGE run. Figure IB compares the
ability of the two procedures to yield clear Dral fragment patterns
of P.aeruginosa DNA from CF sputum.
Figure 1A shows an example of how bacteria can be differentiated by selectiverestrictiondigests. Spel digestion of the sputum
plug yielded large fragments from the P.aeruginosa strain and
smaller fragments from the coexisting Eschenchia coli strain. Xba\
digestion of the same plug revealed larger fragments for E.coli
DNA, and smaller fragments for P.aeruginosa DNA. The intensity
of the ethidium bromide stain indicates that the P.aeruginosa
concentration in the sputum was much higher than the E. coll
723
This is the first report of a protocol for the PFGE analysis of
bacterial genomes that does not require the growth of bacteria in
vitro. This procedure is capable of estimating therelativeamounts
of bacterial species in natural habitats and resolves the presence
and percentage of various clones and subclones in vivo (3). In
contrast to in vivo PCR, the PFGE procedure displays the whole
bacterial chromosome. Estimation of the relative amounts of
bacteria is not biased by the variation of amplification rates of the
species specific genes. With appropriate modifications this
method should be applicable to any high density population of
bacteria, especially pathogens, for example, those found in
urinary tract infections and burn patients.
REFERENCES
1 Persing, D.H, Smith, T F, Tenover, F C , and White, TJ (eds) (1993)
Diagnostic Molecular Microbiology American Society for Microbiology,
Washington DC
2 Boat, T F, Welsh, M J and Beaudet, A.L. (1989) In Scnver, C.R.,
Beaudet, A L., Sly, WS and Valle, D. (eds), The Metabolic Basis of
Inherited Disease McGraw Hill, New York, pp 2649-2480
3 ROmling, U., Heuer, T and Tiimmler, B. (1994) In Chrambach, A , Dunn,
MJ and Radola, BJ. (eds), Advances in Electrophoresis VCH,
Weinheim, Vol 7, pp. 353-406