Effective Placement of. Raymond J. Peterson, Ph.D.

Transcription

1 Effective Placement of LNA into Q-PCR Q Probes Raymond J. Peterson, Ph.D.

2 Why Use LNA in Q-PCR? Q Proper design enables stable & specific hybridization for primers & probes Goals: Ensure sufficient signal by achieving desired T M Avoid self-structure structure Avoid primer dimers and other mis-hybridizations Use to target poor sequence and for difficult designs AT- rich gene region, avoid repeats, other bad context Gene features: splice sites, splice site variants Unique sequence detection: pseudogene,, pathogens Multiplex

3 Where Should LNA Be Placed? 5,, 3, 3, interior, multiple? Depends on function of oligonucleotide: PCR or other amplification primer Hybridization probe: TaqMan, MB, Scorpions SNP probe Gene silencing: sirna, antisense T M is usually integral to decision Therefore, need ability to predict T M

4 DNA & RNA Nearest-Neighbor Neighbor Example Example: GCTATGAT CGATACTA ΔH = Σ ΔH ( GC )+ΔH ( CT )+...+ΔH ( AT ) CG GA TA ΔS = Σ ΔS ( GC )+ΔS ( CT )+...+ΔS ( AT ) CG GA TA T M = +ΔS (initiation)+δs (symmetry) ΔH ΔS +Rln(C T )

5 How Can the Nearest-Neighbor Neighbor Model Be Extended to LNA? For example, single internal incorporation: DNA & RNA have 10 unique dinucleotides LNA has 32 And there will be different parameter values: ΔS and ΔH may both be altered due to conformational entropy, base stacking, & other effects

6 LNA Thermodynamic Databases Systematic approach Internal single match & mismatches: PM, MM Termini and penultimate with and without mismatches: 5TP, 5PP, 3TP, 3TM, 3PP, 3PM All-LNA with and without mismatches: AL, AM Multiple internal incorporations (spacing) ~ 1,600 melt curves

7 Experimental Steps Determine model Design duplexes Melt DNA:DNA and LNA:DNA duplexes Fit Van t Hoff equation to estimate ΔH & ΔS Assemble ΔΔ design matrices & result vectors Perform regression analysis (SVD) to estimate model parameters and errors

8 UV Absorbance Melts "Melting Curve": Increase in Absorbance with Temperature ΔH = -50 kcal/mole, ΔS = -124 e.u. T M = 56 C ΔH = -75 kcal/mole, ΔS = -200 e.u Temperature (C) Derivative Plot of Same Data Temperature (C)

9 Tabulate Thermodynamic Values Training Data Set Oligonucleotides Results from Absorbance Melting Curves Changes due to LNA substitution Name Sequence (5 to 3 ) Length, ΔH ΔG ΔS (e.u.) 37 T M ΔT M ΔΔH ΔΔS ΔΔG 37 Trinuc. (kcal/ mol) (kcal/mol) ( C) ( C) (kcal/mol) (e.u.) (kcal/mol) D1 GTCGAACAGC ± ± ± L1a GTCG L AACAGC CG L A -74.2± ± ± ± ± ±0.16 L1b GTCGA L ACAGC GA L A -76.2± ± ± ± ± ±0.12 L1c GTCGAA L CAGC AA L C -77.2± ± ± ± ± ±0.11 L1d GTCGAAC L AGC AC L A -75.5± ± ± ± ± ±0.11 D2 ATCTATCCGGC ± ± ± L2a ATCT L ATCCGGC CT L A -79.3± ± ± ± ± ±0.18 L2b ATCTA L TCCGGC TA L T -76.8± ± ± ± ± ±0.13 L2c ATCTAT L CCGGC AT L C -73.9± ± ± ± ± ±0.16 L2d ATCTATC L CGGC TC L C -74.0± ± ± ± ± ±0.10 L2e ATCTATCC L GGC CC L G -76.6± ± ± ± ± ±0.08 D3 CGCTGTTACGC ± ± ± L3a CGCT L GTTACGC CT L G -82.9± ± ± ± ± ±0.13 L3b CGCTG L TTACGC TG L T -81.0± ± ± ± ± ± altogether, to cover all trinucleotides Test Data Set Oligonucleotides Results from Absorbance Melting Curves Changes due to LNA substitution Name Sequence (5 to 3 ) Length, ΔH ΔS ΔG 37 T M ΔT M ΔΔH ΔΔS ΔΔG 37 Trinuc. (kcal/ mol) (e.u.) (kcal/ mol) ( C) ( C) (kcal/mol) (e.u.) (kcal/mole) DT1 GTGGATCTTTA ± ± ± LT1a GTG L GATCTTTA TG L G -69.2± ± ± ± ± ±0.08 LT1b GTGGAT L CTTTA AT L C -63.7± ± ± ± ± ±0.09 LT1c GTGGATCTT L TA TT L T -63.6± ± ± ± ± ±0.12 DT2 ACTGGCATCTG ± ± ± LT2a ACT L GGCATCTG CT L G -77.0± ± ± ± ± ± altogether, to test predictions for all 32 nearest neighbors

10 Regression Analysis (SVD)

11 ΔΔ Values for Primer Design 32 Nearest Neighbor Parameter Set Nearest neighbors H (kcal/mole) S (cal/mole K) G 37 (kcal/mole) 5 ALA ± ± ± ALC ± ± ± TLG ± ± ± TLT ± ± ± AAL ± ± ± ACL ± ± ± TGL ± ± ± TTL ± ± ± Uncertainties are the estimated errors in the value of the parameter from SVD analysis, not a standard deviation over a population.

12 Rules of Thumb: Single Internal ΔT M ranges from -1.0 to +8.0 C LNA-A A confers the least additional stability LNA-T T and LNA-C C provide the most Sequence context is important Accuracy ± 2.0 C C of the actual T M

13 LNA Improves Mismatch Discrimination

14 Mismatch Improvement Is Sequence-Dependent

15 Rules of Thumb: Mismatches LNA generally improves discrimination The properties of DNA:DNA and LNA:DNA mismatches are similar G L :T decreases discrimination (Owczarzy( Owczarzy) A L :G, G L :A or G L :G increase discrimination Otherwise, substantial sequence-dependence

16 Rules of Thumb: Others LNA at termini: 5 end makes little difference 3 stabilizes less than internal LNA Penultimate LNA: Effect depends on inside base Decreased entropic improvement All-LNA molecules are so stable that their self- structure can be a serious problem for probes

17 5 LNA Improves Sequencing Trace

18 LNA 5 5 Helps Q-PCRQ

19 Contribution to NN Followed accepted protocol Determined thermodynamics of representative oligonucleotides Contribution is that LNA has more sequence contexts: 10 DNA PM versus 32 LNA PM We spent time worrying about which model would be best NN gave best results

20 LNA Applications Q-PCR: any method, any primer or probe Pathogen detection (DNA or RNA), homology Sequencing primers PCR primers SNP hybridization probes Microarrays Gene silencing (antisense/sirna) therapeutics

21 Acknowledgements Jason D. Kahn, University of Maryland Meinrado F. Samala Joshua D. Levin Dean E. Fiala, Celadon Laboratories Patricia M. McTigue Satya Medicherla