Supplementary materials for De novo piRNA cluster formation in the Drosophila germline triggered by transgenes

containing a transcribed transposon fragment

Ivan Olovnikov, Sergei Ryazansky, Sergey Shpiz, Sergey Lavrov, Yuri Abramov, Chantal Vaury, Silke Jensen and Alla Kalmykova

This PDF file includes: Materials and Methods Results and Discussion References Figures and legends S1 to S11 Tables ST1-ST5

Materials and Methods RT-qPCR

Total ovarian RNA was reverse-transcribed using Superscript II (Invitrogen). cDNA was

synthesized using random primers, oligo(dT) or gene-specific primers. rp49 PCR served as a

loading control. Universal adapter (5’ GCCTGCCCCAACCTCC 3’) was attached to some gene-

specific primers to avoid non-specific priming.

The following primers were used for RT-PCR (sequence of the adapter is in bold):

rp49 5’-ATGACCATCCGCCCAGCATAC-3’ and 5’-

CTGCATGAGCAGGACCTCCAG-3’

h3_S 5’-CAGCGTCTGGTGCGTCAAATCG-3’

h3_AS 5’-TGGATGTCTTTGGGCATTATGGTGAC-3’

hsp RT 5’-GCCTGCCCCAACCTCCAAAGTAACCAGCAACCAAG-3’

I-TG AS 5’-CACACATCCATACAGGGTTCCAT-3’

I-TG S 5’-CACAGGCAAAGTCCACCGATT-3’

hsp 5’-AAAGTAACCAGCAACCAAGTAAATCAA-3’

5’Ps 5’-AGAGGAAAGGTTGTGTGCGGAC-3’

5’Pas 5’-CTGCGAATCATTAAAGTGGGTATCA-3’

42_4R 5’-CACTGACTACGGTGCCTACAGCTATG-3’

42_4F 5’-TGTTTACCCAGAATGATGTTGAAATATAAGATG-3’

1.9_S 5’-ATTTGAGTTTCTGTCCCACTGTGG-3’

1.9_AS 5’-TGTTGGGAGAGAGAATAGAGATAAGGC-3’

2.1_S 5’-GACCGCCGAACTGGAGGATT-3’

2.1_AS 5’-CGAAGCGAGTGGCAAAGAAA-3’

3.6_S 5’-TCCCCAGAGAGAAAACCCACG-3’

1

3.6_AS 5’-GCTCGTTCGGCTCTCGGC-3’

6252_S 5’-AAATTACGCACGCACCGTTACT-3’

6252_AS 5’-CCATGAAGCAAAGCTCCGC-3’

6721_S 5’-GAGCGATATAAAGAGAGTGGCAAAGA-3’

6721_AS 5’- TTGTGGGTGCTCTCCAATGC-3’

hsp70B_ORF_S 5’-ACAGACTCGGAACGCCTCATTG-3’

hsp70B_ORF_AS 5’-TGATGCTCTCGCCCAGATACG-3’

2.1_TIRANT_S 5’-ATTGGAAATCTTCACACTACAACTATCTGC-3’

2.1_TIRANT_AS 5’-AAACACCTCTCTCCCTCTCTCCAC-3’

Primers I-TG AS, I-TG S, hsp, 5’Ps, 5’Pas, 1.9_S, 1.9_AS, 2.1_S, 2.1_AS, 3.6_S, 3.6_AS,

6252_S, 6252_AS, 6721_S, 6721_AS were used in ChIP analysis.

Northern analysis

Total ovarian RNA samples (20 μg per lane) were separated on formaldehyde/agarose gels

and transferred to Hybond-XL membrane (Amersham). P32-labeled riboprobes corresponding to

sense or antisense strand of I-element (nucleotides 2109-2481 of the GenBank sequence

M14954) were synthesized. As a loading control, hybridization with an rp49 probe was used.

Hybridization was performed at 60°C overnight in 0.5 M NaCl, 0.1 M sodium phosphate (pH

7.5), 25 mM EDTA, 1% SDS, 50 μg/ml denatured salmon sperm DNA, 5x Denhardt’s solution,

50% formamide. The filters were washed three times at 60°C with a solution containing 10 mM

sodium phosphate and 0.2% SDS and visualized by phosphor imager Storm-840 (Amersham).

Bioinformatic analysis

Small RNA libraries were analyzed as described previously. Briefly, after clipping the

Illumina 3’-adapter sequence (TGGAATTCTCGGGTGCCAAGGAACTCCAGTCAC), the

reads that passed quality control and minimal length filter (>18nt) were mapped allowing 0 or 1-

3 mismatches to the Drosophila melanogaster genome (Apr. 2006, BDGP assembly R5/dm3) or

transgenes by bowtie and NucBase (1,2). The plotting of size distributions and read coverage,

measuring nucleotide biases and ping-pong signatures were performed by custom scripts

(available upon request). The ping-pong signature is defined here as normalized occurrence

frequencies of overlap lengths between the 5’ ends of small RNAs mapping to the opposite

strands.

2

Results and Discussion Expression analysis of I-containing transgenes

Small RNAs are produced from both strands of the I-sense and I-antisense transgene

sequences, which implies bidirectional transcription of entire loci. Bidirectional transcription of

transgenes in different transgenic strains was demonstrated using RT-PCR with strand-specific

primers (Figure S8A,B). In strain 2.1, the transgene is inserted in a gene exon in the opposite

direction relative to transcription of the gene. In this case, the antisense transgene transcript may

be a part of a read-through transcript driven by a target gene promoter. Using strand-specific RT-

PCR, we confirmed effective antisense expression of transgenic I-element in strain 2.1 (Figure

S8). However, we do not see correlation between the amounts of small RNAs produced by the

transgene and abundance of sense/antisense transgenic transcripts, at least at the region of primer

location. When transgenes are inserted in the intergenic spacers, antisense transcription may be

activated by transgenic regulatory elements or to be a consequence of a genome-wide pervasive

transcription. Northern analysis of the ovarian RNA from transgenic flies with I strand-specific

probes reveals 3.5 and ~8 kb sense transcripts (Figure S8C). The 3.5 kb transcript corresponds to

the I-TG transcript starting in the hsp70 promoter and terminating with the poly-A signal of the

actin. 8 kb transcripts are most likely to be internal to the transgene, as such transcripts were

observed for transgenic strains with different insertion sites. However, we failed to detect

antisense transgenic transcripts by Northern analysis in ovaries (not shown). Probably, antisense

transcripts are difficult to detect, as they are largely degraded to small RNAs.

Expression analysis of flanking regions in the sites of transgene insertions

The current model of the primary piRNA production suggests that piRNA clusters produce

long transcripts as a result of conventional transcriptional signal suppression (3,4). These

transcripts are also somehow recognized as targets for piRNA production. Distinct chromatin

structure of piRNA clusters provides their efficient transcription (5-7). Given the fact that

transgene insertions promote small RNA production from the adjacent regions, we decided to

compare expression of the genomic regions in the absence and in the presence of transgenes. In

1.9 and 3.6, RT-PCR analysis shows appearance of considerable amounts of transcripts from the

intergenic region upstream of transgene in the transgenic strains as compared to wK (Figure

S9A,B). Moreover, the effect of transcription activation in these strains is masked by the fact that

these transcripts are at least partly processed into piRNAs. It is possible that transgene regulatory

elements could promote transcription from transcriptionally silent regions with subsequent

3

involvement of these transcripts in the production of small RNA. According to another scenario,

formation of the piRNA cluster might modify RNA polymerase activity, resulting in the

appearance of long read-through transcripts, which are processed into small RNAs. Transgene

insertion in the noncoding exon of CG32486 gene (strain 2.1) leads to a decrease in transcripts

abundance, which may be caused by the premature termination of gene transcription within the

transgene (Supplementary Fig. 9C). Using strand-specific RT-PCR, we revealed that transgene

insertion promotes antisense transcription of CG32486 (Figure S9D).

To address the question about factors that determine the abundance of small RNAs

generated by different transgenes, we analyzed the genomic positions of transgenes. Strains

studied in this work may be divided into several groups according to the number of small RNAs

produced from the transgenic sequences. Strains 2.1 and 3.1 show the most pronounced

enrichment in such small RNAs. In strain 3.1, the transgene is inserted into the pre-existing dual-

strand piRNA cluster located in the 3R TAS region, which leads to the generation of abundant

transgene-associated piRNAs. In strain 2.1, where the abundance of transgene-derived piRNAs

is similar to that in 3.1, the insertion has occurred within the 5’UTR of the gene, CG32486, in

the opposite direction relative to gene transcription. In wK, no piRNAs map to this gene. Our data

suggest that active antisense read-through transcription of transgenes, as in the case of strain 2.1,

promotes generation of small RNAs. Transgenes inserted in the introns (strains 2.6 and 3.9)

produce the lowest amount of small RNAs irrespective of the orientation of insertion relative to

the direction of gene transcription. An intermediate position is taken by transgenes inserted

upstream (1.9, 3.6, 2.10) of genes or within the genes in the same orientation (2.3, 3.10). We

suggest that the silencing capacity of I-containing transgenes depends on the transcriptional and

chromatin status of the target region.

References

1. Dufourt, J., Pouchin, P., Peyret, P., Brasset, E. and Vaury, C. (2013) NucBase, an easy to

use read mapper for small RNAs. Mob DNA, 4, 1. 2. Langmead, B., Trapnell, C., Pop, M. and Salzberg, S.L. (2009) Ultrafast and memory-

efficient alignment of short DNA sequences to the human genome. Genome Biol, 10, R25.

3. Brennecke, J., Aravin, A.A., Stark, A., Dus, M., Kellis, M., Sachidanandam, R. and Hannon, G.J. (2007) Discrete small RNA-generating loci as master regulators of transposon activity in Drosophila. Cell, 128, 1089-1103.

4. Muerdter, F., Olovnikov, I., Molaro, A., Rozhkov, N.V., Czech, B., Gordon, A., Hannon, G.J. and Aravin, A.A. (2012) Production of artificial piRNAs in flies and mice. RNA, 18, 42-52.

5. Klattenhoff, C., Xi, H., Li, C., Lee, S., Xu, J., Khurana, J.S., Zhang, F., Schultz, N., Koppetsch, B.S., Nowosielska, A. et al. (2009) The Drosophila HP1 homolog Rhino is

4

required for transposon silencing and piRNA production by dual-strand clusters. Cell, 138, 1137-1149.

6. Pane, A., Jiang, P., Zhao, D.Y., Singh, M. and Schupbach, T. (2011) The Cutoff protein regulates piRNA cluster expression and piRNA production in the Drosophila germline. Embo J, 30, 4601-4615.

7. Rangan, P., Malone, C.D., Navarro, C., Newbold, S.P., Hayes, P.S., Sachidanandam, R., Hannon, G.J. and Lehmann, R. (2011) piRNA Production Requires Heterochromatin Formation in Drosophila. Curr Biol, 21, 1373-1379.

8. Li, C., Vagin, V.V., Lee, S., Xu, J., Ma, S., Xi, H., Seitz, H., Horwich, M.D., Syrzycka, M., Honda, B.M. et al. (2009) Collapse of germline piRNAs in the absence of Argonaute3 reveals somatic piRNAs in flies. Cell, 137, 509-521.

9. Malone, C.D., Brennecke, J., Dus, M., Stark, A., McCombie, W.R., Sachidanandam, R. and Hannon, G.J. (2009) Specialized piRNA Pathways Act in Germline and Somatic Tissues of the Drosophila Ovary. Cell, 137, 522-535.

5

Figure S1. The contents of small RNAs in ovaries of transgenic strains and wK. Annotation was done according to the Drosophila melanogaster genome (FlyBase, r5.43).

1.9

2.1

2.3

2.6

2.10 3.1

3.6

3.9

3.10

62.5

.2

67.2

.1 wK

0

5

10

15tRNAsn/snoRNAExonsSattelitesrRNARepeatsmiRNAOthers

mio

rea

ds

Supplementary Figures

6

P I-TG Pwhite

P

I-TG

Pwhite

P white PI-TG

P white P

X

Figure S2. Description of transgenic constructs used in this study. (A) Designation and schematicrepresenation of constructs. (B) Schematic maps of transgenic insertion sites.

3.962.5.2

2L

3.6

2R

3.10

3L

2.1

2.3

3R

1.967.2.1

2.102.6

3.1

Repeat density

hsp[i1-2Δ/S]pA

hsp[i1-2Δ/AS]pA

pA'[i1-2Δ/S]pA

pW8-hsp-pA

Structure:

prom(hsp70)pA(Act)

pA(hsp70)

Centromere

+ -

High Low

A

B

Transgenes: Strains:

1.9, 2.1, 2.3, 2.6, 2.10

3.1, 3.6, 3.9, 3.10

67.2.1

62.5.2

7

58%10A - 31%1U -

1U - 45%10A - 69%

Figure S3. I-element specific small RNAs in transgenic strains (addition to Figure 1A,B). (A) Normalized small RNA density in 30 bp window along canonical I-element (shown above) in the transgenic fly ovaries (readsper million, rpm; no mismatches allowed). Reads mapped to the sense strand are shown in blue; antisense in brown. The I-element fragment present in the transgenic constructs is marked by dashed line (I-TG). Note that I-TG is inverted relative to hsp70 promoter in strains 3.6 and 3.9. (B) Length distribution of small RNA mapped toI-element. Percentages of reads having 1U and 10A are indicated for each strand (only 24-29 nt reads were considered).

1U - 46%10A - 58%

1U - 48%10A - 60%

1U - 47%10A - 57%

1U - 57%10A - 53%

1U - 46%10A - 67%

Fre

quen

cyF

requ

ency

Fre

quen

cyF

requ

ency

Fre

quen

cyF

requ

ency

Fre

quen

cy

Size(nt)

I-TG

1.9

2.3

2.6

2.10

3.6

3.9

ORF1 ORF2

25

0

25

25

0

25

25

0

25

25

0

25

25

0

25

25

0

25

25

0

25

62.5.2

1 kb

RP

MR

PM

RP

MR

PM

RP

MR

PM

RP

M

I-elementA B

18 20 22 24 26 28

8

0.2

0.1

0

0.1

0.2

0.2

0.1

0

0.1

0.2

0.2

0.1

0

0.1

0.2

0.2

0.1

0

0.1

0.2

0.2

0.1

0

0.1

0.2

0.2

0.1

0

0.1

0.2

0.2

0.1

0

0.1

0.2

Size(nt)18 20 22 24 26 28

1U - 95%10A - 29%

1U - 87%10A - 30%

1U - 91%10A - 40%

1U - 91%10A - 40%

1U - 90%10A - 37%

1U - 88%10A - 27%

92%10A - 20%1U -

0 5 10 15 20 25

0.0

0.1

0.2

0.3

0.4

0.5

0.6

Freq

uenc

y

0 5 10 15 20 25

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0 5 10 15 20 25

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0 5 10 15 20 25

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0 5 10 15 20 25

0.0

0.1

0.2

0.3

0.4

0.5

0.6

Freq

uenc

y

0 5 10 15 20 25

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0 5 10 15 20 25

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0 5 10 15 20 25

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0 5 10 15 20 25

0.0

0.1

0.2

0.3

0.4

0.5

0.6

Freq

uenc

y

0 5 10 15 20 25

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0 5 10 15 20 25

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0 5 10 15 20 25

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0 5 10 15 20 25

0.0

0.1

0.2

0.3

0.4

0.5

0.6

Freq

uenc

y

0 5 10 15 20 25

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0 5 10 15 20 25

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0 5 10 15 20 25

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0 5 10 15 20 25

0.0

0.1

0.2

0.3

0.4

0.5

0.6

Freq

uenc

y

0 5 10 15 20 25

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0 5 10 15 20 25

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0 5 10 15 20 25

0.0

0.1

0.2

0.3

0.4

0.5

0.6

Freq

uenc

y

0 5 10 15 20 25

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0 5 10 15 20 25

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0 5 10 15 20 25

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0 5 10 15 20 25

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0 5 10 15 20 25

0.0

0.1

0.2

0.3

0.4

0.5

0.6

Overlap (nt)0 5 10 15 20 25

0.0

0.1

0.2

0.3

0.4

0.5

0.6

Overlap (nt)0 5 10 15 20 25

0.0

0.1

0.2

0.3

0.4

0.5

0.6

Overlap (nt)0 5 10 15 20 25

0.0

0.1

0.2

0.3

0.4

0.5

0.6

Overlap (nt)

1.9

2.1

2.3

2.6

3.1

3.6

wK

24-29 nt 21 nt 21S/24-29AS 24-29S/21AS

Freq

uenc

y

Figure S4. Signature of the ping-pong amplification cycle. The relative frequencies of 5’ overlapfor 24-29-nt piRNAs and 21-nt RNAs from the I-TG fragment. Small RNAs were separated into sense or antisense relative to canonical I-element. We suppose that low amounts of 21-nt I-sense reads lead to theabsence of the peak at position 10 when 21S/21AS or 21S/24-29AS pairs are analyzed (see Figure 1B).

9

1.0

0.0

42AB/rp49

0.5

wK 1.9 2.1 3.1 3.6 3.9

Figure S5. Expression of 42AB in wK and transgenic strains. RT-qPCR analysis of the amount of 42ABtranscript in the ovaries of transgenic strains. Histogram bars represent a normalized ratio of 42AB transcript abundance in the ovaries of transgenic strains to that of wK. An active processing of long cluster-derived transcripts into small RNAs may be responsible for the observed slight decrease in the 42AB expression level in the transgenic lines in comparison with wK.

10

antisense4 kb

BAprom(hsp70) pA(Act)

3.9

1U - 71%10A - 30% 2.10

sense

1U - 76%10A - 34%

pA(hsp70)pA(Act)

prom(hsp70) pA(Act)

Figure S6. Generation of small RNAs by transgenes containing fragment of I-element (addition to Figure 2). (A) Normalized numbers of small RNAs mapped to transgenic constructs in 100 bp window (blue – sense; brown – antisense; no mismatches allowed). I-sense and I-antisense transgenes are shown above. (B) Length distributionof small RNAs mapping to all transgene sequences except for I-TG and hsp70 promoters. Percentages of reads having 1U and 10A are indicated for each strand (only 24-29-nt reads were considered). (C) Same as (A) for I-promoterless construct (67.2.1, left) and control construct lacking I-TG region (62.5.2, right). (D) Same as (B) for strains shown on panel (C).

prom(hsp70)C

62.5.267.2.11U - 64%

10A - 30%

1U - 68%10A - 24%

D

P PwhiteI-TG

50

0

50

P Pwhite

I-TG

50

0

50

50

0

50

P PwhiteI-TG

75

0

75

P Pwhite

RP

MR

PM

RP

MR

PM

2.10

3.9

67.2.1 62.5.2

Fre

quen

cy

Fre

quen

cy

Fre

quen

cy

Fre

quen

cy

Fre

quen

cy

prom(hsp70) pA(Act)

100

0

100

RP

M

2.6 2.6

11

0.2

0.1

0

0.1

0.2

0.2

0.1

0

0.1

0.2

0.2

0.1

0

0.1

0.2

0.2

0.1

0

0.1

0.2

0.2

0.1

0

0.1

0.2

1U - 64%10A - 27%

1U - 71%10A - 29%

1U - 75%10A - 27%

1U - 62%10A - 23%

1U - 100%10A - 0%

1U - 50%10A - 50%

Size(nt)18 20 22 24 26 28

Size(nt)18 20 22 24 26 28

Size(nt)18 20 22 24 26 28

Size(nt)18 20 22 24 26 28

Piwi Aub Ago3 SpnE Zuc

0

2

4

6

8

fold

cha

nge,

log

2

A

0.2

0.1

0

0.1

0.2

0.2

0.1

0

0.1

0.2

0.2

0.1

0

0.1

0.2

BAub-IP Piwi-IP

antisensesense

Ago3-IP

Figure S7. hsp70 genomic loci produce piRNAs in the germline. (A) Abundance of hsp70B small RNAs in the ovaries of piwi, aub, ago3, spn-E, zuc heterozygous females relative to their amounts in the homozygous ovaries (log2 scale). (B) Length distribution of hsp70 small RNAs bound to Aub, Piwi and Ago3. Analysis ofdata (Li et al. 2009; Malone et al. 2009); accession numbers are SRP000458 (the NCBI trace archives) and GSE15186 (GEO), respectively.

Size(nt)18 20 22 24 26 28 18

Size(nt)20 22 24 26 28

Size(nt)18 20 22 24 26 28

1.0

0.0

sens

e tr

ansc

riptio

n

0.5

1.9 2.1 2.3 3.1 3.6 3.92.6 2.9 3.10

A

1.0

0.0

antis

ense

tran

scrip

tion

0.5

1.9 2.1 2.3 3.1 3.6 3.92.6 2.9 3.10

B

C

w11

18

Spn-E/+

Spn-E/Spn-E

1.9

rp49

1

2

3

456

9 kb

13

Figure S8. Expression analysis of I-containing transgenes. (A) RT-qPCR analysis of the sense transgenic transcript amounts. Reverse transcription was done using oligo(dT) primer. (B) RT-qPCR analysis of the antisense transgenic transcript amounts. Reverse transcription was performed using strand-specific primer,hsp_fw RT. Bars of histograms represent the normalized ratio of sense (A) or antisense (B) transgenic transcriptabundance in ovaries of different transgenic strains to that in 1.9. (C) Northern analysis of I-element in the ovaries of transgenic flies. Total ovarian RNA from w1118, spn-E1/TM3 (+/-), spn-E1/spn-Ehls3987 (-/-) and 1.9 flies was analyzed. Hybridization with antisense I-specific riboprobe was performed to reveal sense transcripts. The lower panels demonstrate hybridization with the rp49 probe as a loading control. Sizes of RNA Millennium Markers (Ambion) are indicated in kb. Accumulation of full-length I-element transcripts is detected in the ovaries of the piRNA pathway component spn-E mutants. 3.5 and 8 kb I-element transcripts are revealed in the ovaries of transgenic strain 1.9.

1

1.9wK

2

0

5' fl

ank/rp49

A B

1

10

05'

flan

k/rp49

5

wK 3.6

C

1.0

0.0

5' fl

ank/rp49

0.5

wK 2.1

D

2

0

AS

5' f

lank

/rp49

1

wK 2.1

Figure S9. Expression analysis of genomic sequences flanking transgenes. (A) RT-qPCR analysis of the transcript amount of the transgene flanking regions in the ovaries of transgenic strains 1.9 (A) and 3.6 (B). (C). RT-qPCR analysis of the sense transcript amount of the gene CG32486 in the ovaries of transgenic strains2.1. Random primers were used for reverse transcription (A, B, C). (D) Strand-specific RT-PCR of the antisenseexpression of the gene CG32486 in which transgene insertion occurred in the strain 2.1. Primers are indicated in Figure 3. Bars of histograms represent normalized ratio of corresponding transcript abundance in the ovaries of transgenic strains to that in wK.

14

5'...GCCACTCGCTTCGTACACAAACTCCCACACCATGATGAAATAACATAAGGTGGTCCCGTC...3'

genome transgene

Figure S10. Transgene insertions induce read-through transcription. Small RNAs overlapping the border between transgene in the 2.1 strain and neighboring genomic sequence are represented.

15

1.9

2.1

2.3

3.1

I-TGTransgeneexcl. I-TG & HSP

Figure S11. Transgenes produce small RNAs of different length with 1U bias. Length distribution of only 1U small RNAs mapping to I-TG and all transgene sequences except for I-TG and hsp70 promoters in transgenic strains. The size distribution of this population formed two peaks, at 21 nucleotides and a broader peak corresponding to genuine 24-29 nt-length piRNAs.

Fre

quen

cyF

requ

ency

Fre

quen

cyF

requ

ency

16

0.2

0.1

0

0.1

0.2

0.2

0.1

0

0.1

0.2

0.2

0.1

0

0.1

0.2

0.2

0.1

0

0.1

0.2

0.2

0.1

0

0.1

0.2

0.2

0.1

0

0.1

0.2

0.2

0.1

0

0.1

0.2

0.2

0.1

0

0.1

0.2

Size(nt)18 20 22 24 26 28

Size(nt)18 20 22 24 26 28

Size(nt)18 20 22 24 26 28

Size(nt)18 20 22 24 26 28

17

18

tg*/wK

wK

19

wK

wK

tg*/wK

20

wK

wK

21

wK