En relación a la identificación de las debilidades del sistema

Much of the work on detection of specific micro-organisms by PCR has been concerned with presence/absence tests, i.e. non-quantitative PCR. The main constraint in obtaining quantitative data is inherent in the amplification process. Because amplification is (at least initially) an exponential process, small differences in any of the variables that influence the rate of reaction will dramatically affect the yield of PCR product. These variables include just about any chemical or physical parameter that is involved in the reaction. Even with carefully controlled reaction conditions, tube-to-tube variation may be significant (Gilliland et al, 1990b). Additionally, the well documented plateau phase in PCR (Gause and Adamovicz, 1994; Morrison and Gannon, 1994), where the rate of product accumulation decreases with increasing cycle number, prevents direct correlation between the amount of PCR product and the starting number of copies of the target. This can be seen more clearly by looking at the relationship which describes product accumulation during PCR:

log y = log X + n.log (1+E)

where y is the yield, x the starting number of target DNA molecules, n the number of cycles performed and E the efficiency of the reaction. In an ideal PCR, where the number of copies of the amplicon doubles at each stage, the efficiency of amplification is equal to 1. However, in reality efficiency values will typically be in the range 0.7-0.8 (Ferre, 1992). In itself, this does not present a problem in the quantitation of the starting

number of targets since E can be measured. So, if E and y can be measured and n is known then calculation of the starting number of target molecules, x, is possible. A problem arises when the plateau is reached as at this point, E falls dramatically. Hence, the relationship shown above only applies in the portion of the PCR before the onset of the plateau phase.

There are three basic strategies that are used to quantify the starting number of target copies in the PCR: i) measurement of products in the linear (pre-plateau) portion of the reaction; ii) limiting dilution or MPN based PCR; and iii) the use of a competitive internal standard in the PCR (Jansson, 1995; Reischl and Kochanowski, 1995).

Measurement of products in the linear portion o f the PCR is clearly a strategy designed to avoid the problem of variable efficiency caused by the plateau phase. Using this approach, it is necessary to define the linear portion of the reaction and to measure reaction products during that phase. This normally entails the use of highly sensitive detection methods since there may be insufficient product accumulation to allow measurement by ethidium bromide staining (Gause and Adamovicz, 1994). Such methods have been developed and are now marketed specifically for quantitative PCR (see Section 1.5.2.3). However, a further problem that has been identified is the variation between tubes (variable efficiencies of amplication) that might occur when samples are derived from environmental origins and may contain inhibitors of the PCR (Jansson, 1995).

The use of the MPN principle in conjunction with PCR has been illustrated by the work of Picard et al (1992). This group used an MPN approach to estimate the indigenous population of Frankia species in soil. DNA extracted from soil solutions was serially diluted by a factor of three and amplified in triplicate according to a PCR protocol that was capable of amplifying the DNA equivalent of a single cell. The number of amplifiable target DNA sequences corresponding to the number of bacterial cells was determined according to most probable number statistics. A similar approach was used by Sykes and co-workers (1992) to detect and quantitate leukaemic cells within a large population o f normal cells. The major drawback of this method is the necessity for a large number of PCR amplifications per sample.

The third method of quantitation, the use of a competitive internal standard in the PCR, overcomes many of the problems of the other methods (Becker-Andre and Hahlbrock, 1989; Gilliland et al, 1990a; Jansson, 1995). In this approach, a known quantity of an internal standard DNA sequence is co-amplified with the target. The ratio of standard to

Chapter 1. Introduction

target product at the end of the PCR can then be used to deduce the starting ratio and hence the absolute amount of target DNA in the sample. Ideally, an internal standard used in the quantitation of PCR should be a stretch of DNA, or RNA, that is very similar to the target sequence. Most importantly, in order to minimize differences in amplification efficiency, the target and standard DNA sequences should use the same set of primers. There are presently a series of methods for the generation of such internal standards (McCulloch et al, 1995). It is important, of course, that the internal standard and target amplicons can be distinguished after PCR and this can be achieved on the basis of size or minor sequence difference, for example the incorporation of a restriction site (Gilliland et al, 1990a, b). Since this is the method of quantitation used in this thesis, there is further discussion of its application in Section 7.1.1. It is worth noting here that there are now published examples of its use in the measurement of bacteria from environmental samples (Mahon and Lax, 1993; Leser, 1995).