POLYMERASE CHAIN REACTION
Objectives
1) To understand factors governing choice of PCR primers.
2) To understand factors limiting PCR, and ways to get around them.
3) To understand real time PCR.
Basic mechanism of PCR
The polymerase chain reaction can be used to radically amplify a short region of a genome. The
genomic DNA is denatured, primers are annealed bracketing the region, and DNA synthesis is conducted
from the primers doubling the number of molecules in this region. The cycle is repeated many times, until
most of the DNA in the tube is a fragment whose ends are defined by the primers. The process is
automated by a machine that cycles between polymerization temperature and a denaturing temperature,
and the use of a thermostable polymerase (AmpliTaq) that survives the denaturing cycle.
Ref: Saiki, et al., Science 439: 487 (1988).
Note that the ends of the final product are defined by the primers, including any noncomplementary bases
within the primer or added to its 5' end. The most common way to do PCR is to conduct the reactions,
then load a fraction of the product on a gel for visualization. For quantitative purposes, the reaction can
instead be conducted with continuous monitoring. This is called "real time PCR", RTPCR, or qPCR.
1
A typical PCR profile for an older instrument is shown above. Denaturation, annealing, and
extension times are partly determined by the time the particular machine in use requires to stabilize at the
new temperature, and may be quicker for more modern instruments. The extension time is also governed
by the speed at which the polymerase can extend, about 1000-2000 bases/min. The extension
temperature will be the optimum for the thermostable instrument in use, although the polymerase will also
extend at least partially at the annealing temperature in order to stabilize primer template association. The
annealing temperature is usually about 5oC below the TM of the primers (although see below), and will
have to be long enough for an effective kinetic annealing step at the concentration of primers used. Faster
cycling may be associated with smaller volumes and higher primer concentrations. Two step cycles,
where the primers have been designed to anneal at the extension temperature can also be used to achieve
economy in the total cycle time.
A 2 step cycle
An excellent tool for exploring the priming properties of
an oligo with a specific sequence is the program Oligo,
which is on site license at UTHSCSA. To get access,
contact Jeremy Mann (jeremy@biochem.uthscsa.edu).
2
The use of the term TM in PCR hides a somewhat complex situation. TM of a duplex is the
temperature at which it is half duplexed at equilibrium.
The TM of high molecular DNA is predicted by:
The width of the curve is broader for an oligonucleotide (red).
That value is dependent on salt concentration and is usually indexed to 1 M NaCl. The Tm of an oligo in
excess against a template is the temperature at which the template is half duplexed at equilibrium, often
indexed to 1M Na+, 0 mM Mg++, and 100 pM oligo concentration. That value is also dependent on oligo
concentration (proportional to 1/log(C); about 1oC per factor of 2). Basically, the PCR conditions have
been adjusted so that the success of amplification will roughly follow the behavior of the
thermodynamic TM, even though PCR priming is not an equilibrium experiment.
One consequence that you may not anticipate from equilibrium TM calculations are that the efficiency of
amplification can fall off dramatically if the primer concentration is reduced below standard conditions.
For an effective polymerase chain reaction the template should be mostly all duplexed (primed) so that
each cycle will double the amount of product. Therefore there is a kinetic limitation in that the
hybridization must approach completion in the short time period set aside for annealing. The primer
concentration and salt conditions in a PCR reaction have been chosen to achieve nearly complete priming
in 0.5 - 1 min. at a temperature close to the thermodynamic TM of the oligo. PCR recopies typically
3
specify between 0.1 - 1.0 uM for the concentration of the primers. These results from the hybridization
time screen of Oligo illustrate that 0.1 uM oligo concentration may create a condition where there is not
enough time in the cycle for adequate hybridization.
t1/2 = sqrt N * ln 2 / (350000 * C)
Keller, G.H. (1993) DNA Probes: Molecular Hybridization Technology, pp. 1-25; Stockton Press.
The main causes of low primer concentration are hairpin formation or primer dimer formation (see
below).
Primer dimer
If the primers can prime on themselves or each other, then they will be used up making a small
dimeric product. As little as a 3 base perfect complementarity of a 3' end can cause detectable primer
dimer synthesis. Primer dimers are often mistaken for the primers themselves when the reaction product
is examined on a gel. This is particularly problematical when one tries to make small products (<= 200
bp).
An unsatisfactory PCR primer:
Note that the polymerase tends to add an extra untemplated A at the end of each product. This property is
used for an efficient cloning strategy called "TA cloning" discussed elsewhere in the course.
A satisfactory PCR primer subjected to Oligo's "duplex" analysis"
4
Oligo will carry out a primer dimer analysis on each primer interacting with itself, and on the forward and
reverse primers interacting with each other. When doing a multiplex experiment (trying to use more than
one primer pair in the same reaction) be sure to check out all pairwise combinations of primers for primer
dimer problems.
An oligo with a potential hairpin close to the problem range according to Oligo:
Delta G at the priming temperature of 0 would indicate that half the primer is unavailable. One can fight
this problem off by designing the primer to anneal at the extension temperature, and by raising the primer
concentration.
5
Priming specificity
The breath of the melting curve for an oligo is broader than for high molecular weight DNA, allowing
appreciable priming above the TM. For a 20 nt primer, every 5 oC extra stability allows priming with an
extra mismatch. Having two specific PCR primers is very important to achieving discrimination between
amplifying the target and amplifying extraneous products. Remember that once a primer primes at a
partially complementary site, a false-positive product will be generated that is perfectly complementary to
the primer in future cycles. That is, discrimination against priming in wrong places is not applied
recursively at each cycle.
The capability of PCR primers to discriminate against slightly different sequences is surprisingly poor
compared to allele-specific discrimination in a filter hybridization experiment. Imagine trying to
discriminate between two different templates based on a single one base difference in the priming site. A
filter hybridization experiment can make this discrimination. Filter hybridization is not an equilibrium
experiment, but the effect of a mismatch on both the TM and on discrimination during washing of a
hybridized filter depends on the increased dissociation rate. As a rule of thumb, an oligo dissociates about
10x faster for each 4oC above its TM. For example, consider a filter hybridization with an oligo probe,
such that the probe is washed off under pure dissociation conditions at the TM predicted above. The
perfectly matched oligo washes off with a half life of about an hour. The same oligo with a 1 bp
mismatch is less stable by 1oC/percent mismatch or about 5C. It will come off about 10x faster leaving
1/100 to 1/1000 as much signal.
Now imagine the same situation with the oligos used as PCR primers. On the first round, at best the
perfectly matched oligo will be on the template 100% of the time and the mismatched oligo on its
6
template 10% of the time. Even if that resulted in 10x lesser product production, on subsequent cycles
those products will amplify just as well, lagging 3 or fewer cycles behind the perfectly matched signal. In
an non-quantitative experiment, it would be almost impossible to tell the difference between the perfectly
matched and the once mismatched template. In a quantitative (real time) PCR experiment, the
mismatched signal might make it appear as if the template were present in lower concentration. Other
factors degrade the specificity even further. For example, the polymerase will extend transiently bound
primer, possibly yielding nearly complete priming by the end of the annealing period even if priming is
only at the 10% level. The efficiency of this low stringency priming mode tends to be more affected by
mismatches near the 3' end of the primer than by the overall primer stability.
There are several consequences of the relatively poor specificity of sequence discrimination during PCR
priming. 1) There is always a possibility of getting false products. Allowing the annealing temperature to
fall, or amplifying an excessive number of cycles will greatly increase the numbers of false products, 2)
some methods exist that intentionally make use of below TM priming, and 3) PCR methods intended to
discriminate 1 base differences (such as SNP detection) will generally be based on some principle other
than stringency of annealing of the PCR primers.
Note that the likelihood of false priming sites with sequences similar to the targeted priming sites
increases with the complexity of the template. Amplification from human DNA is prone to raise false
products, and primers and conditions must be carefully designed to achieve specificity. On the other
hand, one would have to be very far away from optimal primers and conditions to raise a false product
from a plasmid template. Some experiments are more vulnerable to false products than others. In
experiments where the PCR products are examined by gel electrophoresis, false products that are the
wrong size can often just be ignored. In a real time amplification strategy where the total amount of
duplex DNA in the reaction is followed by fluorescence detection, false products will compromise the
results.
Primer Design (ref: Wetmur, J. (1991) Crit. Rev. in Bioch. and Mol. Biol. 26:227-259.)
The importance of matching the TM of the primer to the annealing temperature.
When a PCR amplification comes up blank, the natural tendency is to think that one of the primers may
not be able to form a stable duplex at the annealing temperature. This is not actually a very common
problem, and can always be diagnosed by just dropping the annealing temperature 5 or 10 degrees. It
is really hard to mis-predict the annealing temperature by more than 10 degrees on the high side. The only
blunder of this type I have personally seen is to include in the TM calculation bases added to the 5' end that
are not intended to match the initial template (such as an added restriction site). Even those primers
usually work since only the first round is affected.
7
Two strategies for amplifying and adding 5' extensions:
Other reasons for a primer to fail to prime:
Usually when a primer fails to prime, it has a more intrinsic flaw, such as primer dimer problem, or a
stable hairpin, which physically removes it from the reaction mix. Then, somewhat counter-intuitively, a
simple solution may be to raise the annealing temperature. One easily overlooked problem of this type
is to misestimate the concentration of the primer stock when assembling the reaction and not add enough
primer in the first place. Primers are usually delivered as a dried down mix labeled with the theoretical
yield. It is common for the theoretical yield to be an exaggeration, and it is also common to simply fail to
dissolve most of the primer when the primer stock is prepared. A sample from the dissolved
concentrated primer stock should always be diluted and the concentration checked
spectrophometrically before making up the working stock. Although they are often 80 - 100% of the
theoretical yield, routinely I find them to be at least 2 fold on the light side, and sometimes as much as 10
fold on the light side. Nowadays it is rare for the synthesis of the primer to actually fail, but when it does
the primer may be truncated. The most common synthesis error is for you to write the sequence down
incorrectly on the order form. A simple way to check that the primer primes as designed is to use it as a
sequencing primer, assuming that a suitable template is available.
When the primer is too stable:
A common mistake is to make the primer too long. After it has enough bases to prime at the annealing
temperature, adding extra bases can degrade its performance by one of several mechanisms: 1) increasing
the stability of duplexing to incorrect sites, 2) increasing the likelihood of hairpins or primer dimer
configurations. In both cases, try raising the temperature right up to the extension temperature.
In a large target (i.e., the human genome), any primer will adventitiously prime at many incorrect
sequences. Specificity is only possible, because it is unlikely for two such false priming sites to be close
enough together to make a PCR product. However, if there is an error that increases the potential for nonspecific priming, the reaction will raise a lot of strange and unrelated PCR products. A lot of trouble can
be saved by designing the primers according to the criteria below. Computer programs are available that
do a good job of identifying primers that will work cleanly.
8
1) Balance the TM's of the two primers. Avoid overly-stable primers.
2) Avoid self-complementarity and complementarity between the two primers.
3) Avoid complementarity with repetitive sequences in the target DNA.
4) Avoid GC rich 3' ends.
Balancing the TM's
The reason for this is obvious. If the primers have radically different TM's, then one of them must
be working at low stringency. The original PCR protocols called for making the final adjustment of
stringency by altering the Mg+2 concentration. (This is simpler than changing the annealing temperature
itself, because tubes with different [Mg+2] can be put in the PCR machine at the same time). However,
that won't make the primers better balanced; it will just find the best compromise stringency. The better
approach is to get the TM's right in the first place.
The effect of different Mg+2 concentrations on TM.
total Mg+2
mM
--------0.8
1.3 (std.)
2.0
2.5
3.0
3.5
4.0
free Mg+2
mM
---------0.175
0.575
1.375
1.875
2.375
2.875
3.375
T elev.
C
-----5.2
8.1
9.9
10.7
11.4
12.0
12.5
% G+C tolerated
in template
--------------70.5
63.4
58.9
56.9
55.2
53.9
52.7
A crude estimate for the TM of an oligonucleotide is: 2oC x (A+T residues) + 4oC x (G+C
residues). This equation is indexed to 1 M NaCl, 0 mM MgCl2, and 100 pM concentration. As it turns
out, the changes to PCR conditions (50 mM NaCl, 1.3 mM MgCl2, 1 uM primer concentration) nearly
cancel out. This equation tends to predict a temperature about 5oC under the true TM and so the
temperature predicted is often used directly as the annealing temp. This estimate is reasonably accurate if
the oligo is about 50% G+C and the sequence is not internally repetitive (like AAAAAAAGGGGGGG, or
AGAGAGAGAGAG).
Mg+2 concentration also affects the TM of the PCR product with respect to denaturation on each
cycle. At %GC in the product of > 65%, the product may not be able to denature at all in the presence of
the reaction buffer. In this case some DMSO (or other denaturants) may be added, [Mg+2] may be reduced,
and/or an alternate thermostable polymerase may be used that survives a few degrees higher denaturing
temperature. This problem is somewhat lessened if the PCR product is short (<500 bp), since that
destabilizes it a degree or two. To realize this benefit, the initial (presumably larger) template will have to
be denatured before the polymerase is added. An ultimate solution would be to substitute a base analogue
for G (McConlogue, L., Brow, M.A.D., and Innis, M.A. (1988) Nucl. Acids Res. 16: 9869.
9
There are computer programs that do a more sophisticated calculation of the oligo's Tm taking into
account the dinucleotide composition, the anticipated ionic conditions, and the anticipated primer
concentration. Most people try for 20 base oligos with about 50% G+C and a TM of 60oC. I recommend
designing for 55 or 60 degree annealing rather than going for 72 degree annealing and a 2 step cycle,
because the latter leaves no option to raise the annealing temperature to rescue the experiment from
certain problems discussed above.
Another issue that requires a good understanding is whether or not to choose primers with G+C rich 3'
ends. The idea behind G+C rich 3' ends is to clamp down the end recognized by the polymerase in the
case where there are some mismatches between the oligo and the template. If the purpose is to prime in
the presence of mismatches, then a G+C rich 3' end is good. On the other hand, if the primer is known to
perfectly match its template, then the G+C rich 3' end is only going to promote false priming at the wrong
sites. In this case make the 3' end of the primer A+T rich. This principle also applies to the design of
DNA sequencing primers.
From the Oligo version 6 manual. Oligo is now vended by Molecular Biology Insights.
On a 20 base primer, it costs about 5oC to tolerate each mismatch to the template. So if nonspecific
products appear, move the annealing temperature up by 5oC. It is unclear if any of the predictive methods
are better than +- 5oC.
10
Software
There are many programs available that profess to design PCR primers. UTHSCSA has a site licensed
copy of Oligo 6.0. This is a full capability program with excellent documentation concerning the
scientific principles. I recommend using this program, or benchmarking other programs against it. To
successfully handle oligos with pathological sequences, a program should use a nearest neighbor
algorithm. A well respected web-based program is primer3 (at MIT, http://frodo.wi.mit.edu/). For jobs
that require batch processing of lots of sequence for automatic PCR primer picking, request the
bioinformatics center to import and install primer3.
Length effect, and the effect of over-amplification.
The PCR product doubles nearly quantitatively through early cycles, but then the accumulation
levels off. One limiting factor is that the polymerase misincorporates bases and then fails to extend the
unpaired 3' end. Another is that as the concentration of product builds up, it begins to anneal with itself
rather than the PCR primers. For reasons that are not totally clear, if the cycling is continued long enough,
the product is converted to a smear of lower gel mobility, and eventually sticks entirely in the slot. I
suspect that this happens because the final PCR products prime illegitimately on one another making
mosaic structures that denature and reanneal to form networks. This most commonly happens when the
template is a plasmid or a previous PCR product for which it is easy to overload the reaction with
template. In this case the reaction plateaus in the first few cycles and then spends the rest of the time
getting into this aberrant mode. If this happens, try using 1/1000 as much template, then 1/1,000,000 as
much.
Larger fragments reanneal more effectively, and therefore are product inhibited earlier in the
reaction. They are also prone to accumulate partially extended products with a mismatched base on the 3'
end. There are practical size limits of several kilobases on the size of the products that can be amplified
without special measures. 75 to 500 bp is optimal.
11
The PCR primers can be expected to prime spuriously at a low level; and eventually to generate a
small spurious product with a primer on both ends. Once such a product is produced, it will amplify at a
normal rate and eventually overtake the expected products. Consequently, it doesn't pay to over-amplify.
Hot start PCR.
The greatest risk of spurious priming is when the reaction is assembled and before it is heated up
for the first cycle. This problem can be suppressed by leaving out one of the components of the reaction
until it reaches the first denaturation cycle. This is sometimes called "hot start" PCR.
Hot start PCR is implemented in several ways. The simplest is to leave out the Mg+2 until the
tubes are hot, then open them and add the Mg+2. Another way is to solidify a wax on the top of the
reaction volume and then add the Mg+2 over the wax. When the wax melts, the Mg+2 mixes into the
reaction. A third way is to add single stranded binding protein to the reaction. It ties up the primers until
the temperature goes up and denatures the protein. Finally, the polymerase can be tied up with antibody
until the denaturing temperature denatures the antibody.
Ref: D'Aquilla, et al. 1991. Maximizing sensitivity and specificity of PCR by pre-amplification heating.
Nucl. Acids Res. 19:3749.
Many procedures call for a longer than usual first round denaturation step. This is thought to be helpful if
the template is high molecular weight, and therefore requires more time for denatured strands to unravel
completing the denaturation process. For supercoiled plasmid templates it may help increase the nicking
of the template to make strands available to act as templates. This extra step is usually not critically
important.
Long-accurate PCR
The maximum length of a PCR product can be increased up to 50 kb by using a mix of a
thermostable polymerase that cannot proofread with a little bit of thermal stable polymerase that can
proofread. Apparently the proofreading enzyme resolves mismatches caused by misincorporation. But a
pure proofreading enzyme does not work for long-accurate PCR. It has been speculated that the
proofreading enzyme during the longer incubation times needed to extend the product is damaging the
primers. Sometimes making longer primers is recommended for long-accurate PCR. This method is a
little finicky, and merits purchasing a specialized kit. Extension times have to be elongated based on the
length of the expected product. Some procedures call for varying the extension times during the
amplification, although I haven't been able to find any sensible explanation as to why do that.
Ref: Wayne M. Barnes. 1994. PCR amplification of up to 35-kb DNA with high fidelity and high yield
from lambda bacteriophage templates. Proc. Natl. Acad. Sci. USA 91: 2216-2220.
Uses of PCR
12
Diagnostics
The simplest use of PCR is the determination of whether the target sequence is present or not in
some sample. PCR is sufficiently sensitive to detect the presence of one copy of HIV genome per million
blood cells. The theoretical amplification is 2 to the power of the number of cycles.
Number of cycles
20
40
Amplification
106
1012
The true amplification can be expected to be lower than the theoretical amplification, but well
designed primers and conditions will usually give over 1.9 fold amplification per cycle. The usual number
of cycles used is 30-40 cycles for amplification of single copy mammalian genes, and up to 60 cycles for
HIV detection.
Contamination problem
Due to the extreme sensitivity of the PCR assay, it is vulnerable to contamination. The amount of
the target DNA fragment in the sample might be 10-6 to 10-13 of the amount of PCR product produced by a
positive assay. Consequently, even a minute contamination with a prior PCR product into a DNA sample
will produce a false positive signal. This is perhaps the most commonly encountered PCR problem. The
amount of contamination required can spread as an aerosol. The following precautions are recommended:
1. Use of a separate set of micropipettes and other implements to prepare templates from those used to
manipulate PCR products. Some labs use an entirely separate template preparation room. Note: a
common oversight is to prepare a control reaction by diluting a control plasmid with the "clean"
pipettes. The control plasmid, having a high concentration of positive DNA, should be kept away
from the "clean" pipettes and template area for the same reason the PCR products are kept away.
2. Amplification of a negative control with all of the components except the template. This control will
be positive if there are any contamination problems. Even experienced PCR practitioners suffer
from contamination problems on a regular basis, so a negative control is necessary every
experiment. To be maximally effective in picking up contamination, a master mix should be
prepared with all components except the template, It should then be distributed into the reaction
tubes with the last aliquot being the negative control.
3. Use of dUTP instead of dTTP in PCR amplification coupled to pretreatment of the reaction with uracilN-glycosylase. The idea is that a major source of contaminants is previously produced PCR
products. If the previously made PCR products were made with dUTP, then they can be destroyed
with the enzyme uracil-N-glycosylase (ung). The uracil-N-glycosylase is mostly destroyed during
the high temperature initial DNA denaturation, and remains inactive above 55oC. However, it can
interfere with the amplification if annealing is done below 55oC. It can also interfere with post
amplification analysis of the PCR product, and will preclude cloning of the PCR product unless the
host strain is ung-.
13
Since contamination is such a common problem, most practitioners of PCR keep PCR reagents
separate from use in other experiments, and keep them in numerous aliquots. Typically, when
experiencing a false positive reaction in the negative control, they would discard the present aliquot
of all reagents.
Mutation Screening.
PCR products can be directly sequenced. Therefore, it is possible to directly amplify and
sequence DNA from any genome without the intermediate cloning step. The preferred sequencing method
is cycle sequencing (because it works well on double stranded linear molecules). If the sequencing
method employs dye terminators, then the products will have to be purified away from the PCR primers.
The PCR product can be run on a gel system for which any base change within the fragment alters
the mobility. To accomplish this, the PCR product is denatured and then run in a native gel system
without reannealing. Each strand then forms hairpins as available from its sequence. Virtually any base
change alters the potential to form some hairpins and causes a mobility shift on the gel. This is called a
Single Strand Conformational Polymorphism (SSCP). In addition to screening exons for novel mutations,
this method can also be used to create a genetic marker where no more convenient method exists. The
fragments analyzed for SSCP have to be relatively small (~100 bases), because larger fragments form
complex patterns by falling into multiple secondary structure patterns.
Ref: Orita, M., Iwahana, H., Hayashi, K., and Sekiya, T. 1989. Detection of polymorphisms of human
DNA by gel electrophoresis as single-strand conformation polymorphisms. PNAS 86: 2766-2770.
To hunt for a specific mutation, PCR products can be hybridized to Allele-Specific Oligonucleotides
(ASOs). Note below that allele-specific probes have been introduced into Real Time PCR. Also, there
are strategies to detect mutations by denaturing, reannealing with wild type DNA, and then cleaving
heteroduplexed sites with one of a variety of enzymes.
Fidelity Problem
The original thermostable polymerase (Taq polymerase) lacked a 3' exonuclease activity (for proof
reading) and therefore has a poor fidelity of copying. This lack of fidelity is compounded by the multiple
rounds of synthesis. The error density can exceed 1/100 bp, but more typically it should be around 1/3000
bp (after 20 cycles).
Several precautions are available to avoid fidelity problems:
1) New thermostable DNA polymerases are now on the market that do proofread. However, their fidelity
turns out to only be about 10 x better than Taq polymerase. Also, they tend to nibble at the primers,
so longer primers may be needed to compensate.
2) The PCR product can be sequenced without cloning. That way the individual errors in the different
molecules average out. This is helpful if the goal is to learn the sequence of the original template,
but it is not a helpful way to learn if clones generated from the product will be unmutated.
14
3) If the PCR product is cloned, then two different clones from different amplifications should be
sequenced (unless there is a previously known sequence to match).
Physical mapping with PCR
Ref. Olson et al., Science 245, 1434 (1989)
Any pair of PCR primers that amplify a unique sequence out of a genomic clone serve to identify that
sequence in any other genomic clone. Such a sequence is called a sequence tagged site (STS). If one
investigator publishes the sequence of a STS from his clone, another investigator can synthesize the
primers and tell if his clone overlaps without exchanging any materials. Maps of how very large clones
overlap are primarily put together by determining the overlap of STS's. A map of overlapping Yeast
Artificial Chromosome clones covering the complete human Y chromosome has been put together in this
way (Foote et al, Science 258:60-66 (1992). A very small sample appears below:
The same STSs were screened against the DNA from people with naturally occurring deletions of
part of the Y chromosome. This allows correlation between the genetic map defined by these deletions
and the physical map defined by the YAC contig. STSs are similarly used to define what is missing in a
deletion mutant.
Recovering DNA next to a known sequence (anchored PCR).
There are a variety of methods to do anchored PCR from double stranded DNA. The object is to
recover DNA of unknown sequence next to known sequence. Most methods involve ligating something
to the end of a fragment to act as the second priming site. The key problem is to avoid end-to-end
amplification of all of the fragments in the mixture.
Bubble PCR
One can ligate on a pair of oligonucleotides with a noncomplementary bubble, which is designed
to support amplification only on fragments where the gene specific internal primer has primed.
15
Inverted PCR
In inverted PCR, the fragments are first circularized and then re-cleaved at a different site to move the
unknown region between two primers.
16
Recombination Mapping
Arbitrary Primed PCR (AP-PCR) also known as Random Amplified Polymorphic DNA (RAPD).
Ref: Welsh, J., & McClelland, M. 1990. Fingerprinting genomes using PCR with arbitrary primers. Nucl.
Acids Res. 18:7213-7218.
When genomic DNA is amplified with a PCR primer of 9 or 10 bases and an arbitrary sequence, a
pattern of products is amplified by chance complementarity. Some of these products are polymorphic. By
working through a collection of such arbitrary primers, a recombination map can be constructed with these
markers. This method is often used to gain a preliminary genetic map of an organism for which there is
no prior characterization. This method is also used to make a diagnostic pattern representing the genomes
of different species of bacteria.
Simple Sequence Length Polymorphisms (SSLPs)
If the PCR product for a STS contains an internal tandem repetitive sequence, the length of the
PCR product will be polymorphic. In this case the PCR product can be used as a genetic marker, and
placed on the genetic map along with polymorphic genes, Restriction Fragment Length Polymorphisms,
and disease loci. These genetic markers are favored by human geneticists because they have a large
number of alleles. They are variously called Simple Sequence Length Polymorphisms (SSLPs),
minisatelite repeats, microsatelite repeats, VNTRs, dinucleotide repeats, & CA repeats. A typical case is
shown below:
17
Stuttering problem
Each individual allele in the figure above appears not as a single size, as you would expect, but as
a small cluster of bands separated by 2 bp in size. This is because the polymerase is able to allow the
polymerizing strand to slip on the template during polymerization through the tandem repeat. This
problem is sometimes referred to as "stuttering". Stuttering is tolerated in the above experiment as long as
one can tell the alleles apart.
If one attempted to sequence a PCR product of the kind illustrated above, the sequence would
become unreadable as one passed through the tandem repeat. Given the option, one should try not to put
simple tandem sequences between the PCR primers (e.g. poly(A) stretches). Alternatively, one could
clone the amplified fragment, and thus purify a homogeneous sequence.
Recovery of homologous sequences
PCR can be used to amplify any sequences straight out of the genome for which enough
information is available to make the primers. This might be done to recover different alleles of the same
gene, different gene family members, or the homologous genes from a different organism.
When the PCR primers may encounter mismatches on the intended priming sites, it is wise to use
a longer primer and to prime at a reduced stringency. This, however, promotes the production of
extraneous products. Consequently, one needs to be prepared to identify the correct product among a
variety of different products amplified. This can be aided by predicting the size of the correct product,
predicting a restriction site found within it, or probing the products with an oligonucleotide probe
designed to hybridize between the two primers.
Quantitative RT (Reverse Transcription) PCR
The intensity of a PCR product can be used to estimate the amount of the original template. One must
measure the amount while it is still in the log-linear portion of the reaction. Originally one measured
incorporation of a radioactive primer in a series of amplifications differing in the number of cycles. The
advent of Real Time fluorescence PCR (see below) has made quantitative PCR much more accessible.
Often this method is used to find the ratio of an inducible RNA to a constitutive RNA in a sample. There
are a variety of strategies to conduct the reverse transcription reaction followed by PCR in a single
reaction tube. One of these is to use an enzyme (rTth) that is both a thermostable reverse transcriptase and
a thermostable DNA polymerase. The most common method of priming the reverse transcription reaction
is with a mixture of DNA oligonucleotides referred to as random hexamers. Random hexamers are
synthesized with a mixture of all four bases at each position. So more exactly, they are a mixture of all
possible hexamers.
Real Time PCR.
In Real Time PCR, the thermocycler is equipped with a fluorescence detector to monitor the buildup of
product. Dept. of Biochemistry has an ABI 7500 Real Time PCR machine that can simultaneously
18
amplify and monitor 96 samples. Training is required before you use this instrument. See Linda Roman
to set up a password on the machine for you and to arrange training.
Detection methods.
SYBR Green
A method that is straight forward, but lacks specificity for the target sequence, is to incorporating a
fluorescent dye that detects double stranded DNA into the PCR reaction. Although some initial real time
experiments employed ethidium bromide for detection, now a more sensitive fluorophore named SYBR
Green is used for this purpose. The dye fluoresces after associating with double stranded DNA, so it
directly detects the buildup of product. Since this detection method is not sequence-specific, one is
vulnerable to being mislead by amplification of spurious products. The instrument can be programmed to
do a denaturation curve on the sample after amplification, which may provide a warning that the amplified
product is very short as might occur for a primer dimer.
Denaturation curves from the ABI manual. The Y axis is actually - d (Rn)/dT, where Rn is the ratio of the
SYBR green signal to a reference dye that does not change fluorescence during the reaction. The peak
therefore appears at the midpoint of the melting transition (the point of maximum slope). The reference
dye (ROX in this case) compensates for inconsistencies in the optical clarities of the different wells.
Sequence-specific detection
19
The available sequence specific detection methods involve an oligo probe with a fluorophore on one end
and a quencher on the other. These probes are added in addition to the two PCR primers, and are
designed to fluoresce if they can hybridize to the accumulating PCR product. In order to prevent the
probe from acting as a primer, they are phosphorylated at the 3' end. There are several variations on two
general formats:
One type of probe fluoresces directly upon hybridization. The molecular beacon web site has extensive
documentation: ( http://www.molecular-beacons.org/).
From http://www.molecular-beacons.org/MB_introduction.html
The oligos with a fluorophore on one end and a quencher on the other has the ends in a hairpin so that
usually the quencher is brought close to the fluorophore and quenches it. If the oligo is hybridized to
another nucleic acid, the quencher is separated from the fluorophore and fluorescence increases.
Hybridization is therefore detected directly by fluorescence without having to remove the unhybridized
oligo. This kind of probe can also be used for quantitative in-solution hybridization or as a substrate on a
microchip or microarray.
Another type of probe is called a hydrolysis probe (TaqMan). These are also called "fluorogenic
probes". The fluorophore is at the 5' end and the quencher at the 3' end. In this case the probe is fully
complementary to the intended product and is short enough that fluorescence is quenched as long as the
probe is intact. This probe hybridizes to the template in the path of the polymerase. Upon colliding with
the probe, the polymerase hydrolyzes the 5' nucleotides releasing the fluorophore which is now
unquenched. When designing hydrolysis probes, one is advised to give it an extra 5oC over the TM of the
primers to account for any destabilizing effect of the dyes. Some of the available fluorophores are
quenched by guanine, so one is advised to not put a G at the 3' end of the probe. Also one is advised to
choose the strand that puts more C's than G's in the probe. A common strategy to avoid detection of
genomic DNA during a reverse-transcription PCR assay is to make the probe cross an intron exon
boundary. Finally, more than one probe with different fluorophores may be used to detect the same
amplicon.
20
From the ABI web site.
Primer design for qPCR
Primers for use with SYBR Green can be made in the usual way. Beware that primer design software that
comes on the ABI synthesizer (Primer Express) predicts TM's substantially below the true TM, and then
their instructions for assembling reactions compensate by instructing you to conduct annealing
substantially above the TM. For use in conjunction with the fluorescent probe systems, primers and the
probe are usually purchased as a set from the company specializing in the particular probe design. They
typically cost about $150 for 2000 reactions. Invitrogen sells an alternative arrangement for about half the
cost. For large numbers of probes, ABI sells preloaded "cards" (384 well microtiter plates) that address
anywhere from 8 RNA samples with 11 probe sets up to one RNA sample with 380 probe sets:
(https://products.appliedbiosystems.com/ab/en/US/adirect/ab?cmd=catNavigate2&catID=601274&tab=De
tailInfo). These require an Applied Biosystems 7900HT Fast Real-Time PCR System.
Mutation detection: In the hydrolysis probe method, the probe oligo can be designed such that a single
nucleotide mismatch will prevent hybridization. Hence the probe could distinguish a one base difference
in a sequence.
21
Multiple fluorophore detection: The instrument simultaneously monitors 5 different wavelengths
corresponding to 5 commonly used fluorophores. In principle, one could have 5 differently colored
probes in each reaction. In practice, there should probably not be more than two differently colored
probes in a reaction.
Quantification: For quantification, typically quadruplicate samples are amplified for each unknown.
Reactions that fail to follow a log-linear accumulation are thrown out, and the rest are averaged to find the
fractional cycle number (Ct) that the reaction crosses a threshold within the log linear part of the
accumulation.
Ct will have a log-linear relationship with the initial concentration of the template in each sample.
Consider a standard curve created by amplifying a series of samples composed by serial dilution from a
stock solution of known concentration. Such a stock can be created by quantifying the product of a prior
PCR amplification. The instrument will plot the standard curve as Ct versus log concentration.
22
This should ideally produce a straight line with a slope of -3.322 (= -1/log10(2)).
In that case, for measuring the ratio of two unknowns:
delta Ct = -3.322 *log10(ratio of the two samples)
Ratio of the two samples = 10^(delta Ct/-3.322)
Amplifying an unknown and finding its initial concentration from this curve would be a very tight way to
calculate the initial concentration. The standard curve would provide a true slope to use if the
amplification was proceeding with less than an 2 fold increase on each cycle. The software also reports
deviation from ideal amplification as the efficiency of amplification. An efficiency of less than 0.9 means
that the assay is significantly suboptimal in some parameter such as annealing temperature, primer selfannealing, or primer dimer formation. In this case the slope would be -3.917. Without standardization, if
you observed delta Ct = -2, you would compute that the induction ratio was 10^(-2/-3.322) = 4. With
standardization you would compute 10^(-2/-3.917) = 3.2.
The major cause of scatter in the standard curve (assuming your pipetting is accurate) is inadequate
calibration for differences in the optical density of the different wells in the 96 well plate. If this is the
case, it will be revealed by a failure of replicates to plateau at the same Rn if cycles are run deep in to the
plateau.
23
By normalizing the plateau values on all the curves for a given primer pair, the scatter in the standard
curve can be essentially eliminated.
When comparing the RNA level between two different preparations, a major risk is that the samples might
not have been prepared comparably. For example, a different number of cells may have been included. A
notorious problem for RNA is that a differential amount of RNA degradation may have occurred. So,
particularly for measuring levels of RNA, it is customary to also measure the Ct for another molecule that
is believed to be expressed the same in both samples. The signal of interest can then be normalized by
this reference signal. To be especially careful, you might do a dilution series on one of the samples and
plot the Ct vs. the dilution factor to reveal if the true slope for either the target or the reference RNA
deviated from the ideal -3.3 and take that into account.
The Real Time PCR machine has software to assist in these analyses and is accompanied by substantial
documentation. Hydrolysis probes are designed similarly to PCR primers with respect to the Tm, but
have the additional constraints that they should not have a C at the 3' end, and should have more C's than
G's.
24
Confirmation of microarray results.
Real time PCR has become the most common method of confirming microarray results. Microarray
hybridization experiments can access the expression of up to 20,000 transcripts at a time, However, the
results have poor quantitative properties, even if subjected to replicative determinations. Because of the
expense of individual microarray experiments, most of the time it is used to nominate a much smaller set
of genes for verification by a different method, usually Real Time PCR.
Some resources for real time PCR:


Tevfik Dorak's Real Time PCR page: http://dorakmt.tripod.com/genetics/realtime.html
Lunge VR, Miller BJ, Livak KJ, Batt CA. 2002. Factors affecting the performance of 5' nuclease PCR
assays for Listeria monocytogenes detection. J. Microbiol. Methods. 51:361-8.
Study Questions
1) For perfectly matched PCR primers, it generally isn't considered of any benefit to make them longer
than about 30 bases. Why?
2) Does the amount of PCR product at the end of the amplification reflect the amount of original template
in the sample? What conditions must be met for the amount of PCR product to reflect the amount
of the original template?
3) Overly stable PCR primers often raise a background of nonspecific products, but overly stable
fluorogenic probes don't cause any particular problem. Why?
4) A real time reverse transcriptase/PCR assay for a message is designed with one gene-specific PCR
primer paired with oligo dT as the other PCR primer, and detection by SYBR Green. What is
wrong with this design?
5) One PCR primer in a particular set is diagnosed as having a hairpin of marginal significance. What
symptoms might you expect if this hairpin causes a problem? How might conditions be adjusted to
overcome the problem?
Last update 4/10/2011 - Steve Hardies
25