You are on page 1of 2

The Plant Cell, Vol. 20: 1736–1737, July 2008, www.plantcell.

org ª 2008 American Society of Plant Biologists

LETTER TO THE EDITOR

Eleven Golden Rules of Quantitative RT-PCR

Reverse transcription followed by quanti- ploy sequence-specific fluorescent probes, primers for a reference gene that are ;1 kb
tative polymerase chain reaction analysis, such as TaqMan probes, and to absolute apart. Typically, the Ct value for the primer
or qRT-PCR, is an extremely sensitive, quantification methods (http://www.gene- pair at the 5#-end of a cDNA will be higher
cost-effective method for quantifying gene quantification.info/). than the Ct value of the primer pair at the
transcripts from plant cells. The availabil- (1) Harvest material from at least three 3#-end, as reverse transcription begins at
ity of nonspecific double-stranded DNA biological replicates to facilitate statistical the 3# [poly(A)] end of the template mRNA
(dsDNA) binding fluorophors, such as analysis of data, freeze immediately in and does not always extend to the 5#-end of
SYBR Green, and 384-well-plate real-time liquid nitrogen, and store at 280C to the template. Ideally, the Ct value of the
PCR machines that can measure fluores- preserve full-length RNA. 5#-end primer pair should not exceed that of
cence at the end of each PCR cycle make it (2) Use an RNA isolation procedure that the 3#-end pair by more than one cycle
possible to perform qRT-PCR on hundreds produces high-quality total RNA from all number.
of genes or treatments in parallel. This has samples to be analyzed. Check RNA qual- (6) Design gene-specific PCR primers
facilitated the comparative analysis of all ity using an Agilent 2100 Bioanalyzer (RNA using a standard set of design criteria (e.g.,
members of large gene families, such as integrity number, RIN . 7 and ideally . 9) primer Tm ¼ 60 6 1C, length 18 to 25
transcription factor genes (Czechowski or by electrophoretic separation on a high- bases, GC content between 40 and 60%),
et al., 2004). Given the relatively low cost resolution agarose gel (look for sharp which generate a unique, short PCR prod-
of PCR reagents, and the precision, sensi- ethidium bromide–stained rRNA bands) uct (between 60 and 150 bp) of the ex-
tivity, flexibility, and scalability of qRT-PCR, and spectrophotometry (A260/A280 . 1.8 pected length and sequence from a complex
it is little wonder that thousands of research and A260/A230 . 2.0). Quantify RNA using cDNA sample in preliminary tests, to facili-
labs around the world have embraced it as A260 values. tate multiparallel qPCR using a standard
the method of choice for measuring tran- (3) Digest purified RNA with DNase I to PCR program. The 3#-untranslated region is
script levels. However, despite its popular- remove contaminating genomic DNA, a good target for primer design because it
ity, we continue to see systematic errors in which can act as template during PCR is generally more unique than coding se-
the application of methods for qRT-PCR and lead to spurious results. Subsequently, quence and closer to the RT start site.
analysis, which can compromise the inter- perform PCR on the treated RNA, using (7) Reduce technical errors in PCR re-
pretation of results. The letter to the editor gene-specific primers, to confirm absence action setup by standardizing (robotize if
by Gutierrez et al. in this issue highlights of genomic DNA. possible) and minimizing the number of
one of many common sources of error, (4) Perform RT reactions with a robust re- pipetting steps. Mix cDNA with qPCR
namely, the inappropriate choice of refer- verse transcriptase with no RNaseH activity reagents, then aliquot a standard volume
ence genes for normalizing transcript levels (like SuperScriptIII from Invitrogen or Array- of this ‘‘master mix’’ into each reaction well
of test genes prior to comparative analysis Script from Ambion) to maximize cDNA containing a standard volume of specific
of different biological samples. The follow- length and yield. Use ultraclean oligo(dT) primers. Set up reactions in a clean envi-
ing are 11 golden rules of qRT-PCR that, primer of high integrity. qRT-PCR gene ronment free of dust, preferably under a
when observed, should ensure reproduc- expression measurements are comparable positive airflow hood. Routinely check for
ible and accurate measurements of tran- only when the same priming strategy and DNA contamination of primer and reagent
script abundance in plant and other cells. reaction conditions are used in all experi- stocks by performing PCR reactions on no
These rules are for relative quantification of ments and reactions contain the same total template (water) controls.
RNA using two-step RT-PCR (where the amount of RNA (Ståhlberg et al., 2004). (8) For relative quantification of transcript
product of a single RT reaction is used as (5) Test cDNA yield and quality. Perform levels, design and test gene-specific
template in multiple PCR reactions), SYBR qPCR on an aliquot of cDNA from each primers for at least four potential reference
Green to detect gene-specific PCR prod- sample, using primers to one or more ref- genes selected from the literature (e.g.,
ucts, and reference genes for normalizing erence genes that are known to be stably Czechowski et al., 2005) or from your own
transcript levels of test genes before com- expressed in the organ(s)/tissue(s) under experience that are likely to be stably
paring samples. Further details can be the range of experimental conditions tested. expressed throughout all organs and treat-
found elsewhere (Czechowski et al., 2004, Threshold cycle (Ct) values should be within ments to be compared. Validate reference
2005). Most of these rules also apply to the range mean 61 for each reference gene genes in preliminary experiments on the
relative quantification methods that em- across all samples to ensure similar cDNA range of tissues and treatments you wish
yield from each RT reaction. Quality of to compare using a foreign cRNA added to
www.plantcell.org/cgi/doi/10.1105/tpc.108.061143 cDNA can be assessed using two pairs of each RNA sample prior to RT-PCR to
July 2008 1737

normalize data for reference gene tran- reaction. PCR efficiency can be derived REFERENCES
scripts prior to assessment of their expres- conveniently from amplification plots using
sion stability (Czechowski et al., 2005). the program LinRegPCR (Ramakers et al., Czechowski, T., Bari, R.P., Stitt, M., Scheible,
(9) Perform real-time PCR on test and 2003). Estimation via the classical calibra- W.R., and Udvardi, M.K. (2004). Real-time
reference genes in parallel for each sample tion dilution curve and slope calculation is RT-PCR profiling of over 1400 Arabidopsis
transcription factors: Unprecedented sensitiv-
to capture fluorescence data on dsDNA also possible, albeit more complicated
ity reveals novel root- and shoot-specific genes.
after each cycle of amplification. Also, (http://www.gene-quantification.info/).
Plant J. 38: 366–379.
perform dsDNA melting curve analysis at (11) Finally, calculate relative transcript Czechowski, T., Stitt, M., Altmann, T., Udvardi,
the end of the PCR run. When relying on abundance for each gene in each sample M.K., and Scheible, W.R. (2005). Genome-
nonspecific DNA binding fluorophors, such using a formula that incorporates PCR wide identification and testing of superior
as SYBR Green, to quantify relative dsDNA efficiency for the test gene and Ct values reference genes for transcript normalization in
amount, ensure that only a single PCR for both test and reference genes (http:// Arabidopsis. Plant Physiol. 139: 5–17.
amplicon of the expected length and melt- www.gene-quantification.info/). Pfaffl, M.W., Tichopad, A., Prgomet, C., and
ing temperature is produced using gel Neuvians, T.P. (2004). Determination of stable
electrophoresis and PCR amplicon melting housekeeping genes, differentially regulated
target genes and sample integrity: BestKeeper-
curve data, respectively. We typically use a
Michael K. Udvardi Excel-based tool using pair-wise correlations.
commercial mixture of hot-start Taq poly-
Biotechnol. Lett. 26: 509–551.
merase, SYBR Green, and other reagents, The Samuel Roberts Noble Foundation
Ramakers, C., Ruijter, J.M., Deprez, R.H.,
such as Power SYBR Green Master Mix Ardmore, OK 73401
and Moorman, A.F. (2003). Assumption-
from Applied Biosystems, and have ob- mudvardi@noble.org free analysis of quantitative real-time poly-
served significant differences in the effi- merase chain reaction (PCR) data. Neurosci.
cacy (PCR efficiency, specificity, and/or Tomasz Czechowski Lett. 339: 62–66.
yield) of such products from different sup- Department of Biology (Area 7) Ståhlberg, A., Håkansson, J., Xian, X., Semb,
pliers. CNAP Research Laboratories H., and Kubista, M. (2004). Properties of the
(10) Determine which reference gene(s) is University of York reverse transcription reaction in mRNA quan-
Heslington, York YO10 5YW, UK tification. Clin. Chem. 50: 509–515.
best for normalization of test gene tran-
Vandesompele, J., De Preter, K., Pattyn, F.,
script levels amongst all samples (e.g.,
Poppe, B., Van Roy, N., De Paepe, A., and
using geNorm [Vandesompele et al., Wolf-Rüdiger Scheible Speleman, F. (2002). Accurate normalization
2002] or BestKeeper software [Pfaffl et al., Max Planck Institute of Molecular of real-time quantitative RT-PCR data by
2004]), which use as input not only the Ct Plant Physiology geometric averaging of multiple internal con-
value, but also the PCR efficiency for each 14476 Potsdam, Germany trol genes. Genome Biol. 3: RESEARCH0034.