• microRNA;
  • Populus trichocarpa;
  • abiotic stress;
  • cold stress;
  • trees


  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

MicroRNAs (miRNAs), a group of small non-coding RNAs, have recently become the subject of intense study. They are a class of post-transcriptional negative regulators playing vital roles in plant development and growth. However, little is known about their regulatory roles in the responses of trees to the stressful environments incurred over their long-term growth. Here, we report the cloning of small RNAs from abiotic stressed tissues of Populus trichocarpa (Ptc) and the identification of 68 putative miRNA sequences that can be classified into 27 families based on sequence homology. Among them, nine families are novel, increasing the number of the known Ptc-miRNA families from 33 to 42. A total of 346 targets was predicted for the cloned Ptc-miRNAs using penalty scores of ≤2.5 for mismatched patterns in the miRNA:mRNA duplexes as the criterion. Six of the selected targets were validated experimentally. The expression of a majority of the novel miRNAs was altered in response to cold, heat, salt, dehydration, and mechanical stresses. Microarray analysis of known Ptc-miRNAs identified 19 additional cold stress-responsive Ptc-miRNAs from 14 miRNA gene families. Interestingly, we found that individual miRNAs of a family responded differentially to stress, which suggests that the members of a family may have different functions. These results reveal possible roles for miRNAs in the regulatory networks associated with the long-term growth of tree species and provide useful information for developing trees with a greater level of stress resistance.


  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Environmental factors, including water, salinity, temperature, etc., are critical to the long-term growth of tree species. In addition, some environmental conditions, such as wind and gravity, are significant stresses to the tremendous crown structure of trees. In order to adapt to changing environmental factors and to respond to a variety of stressful conditions, trees must express a variety of genes to enhance their tolerance at biochemical and physiological levels. For instance, trees constantly develop specialized woody tissues to correct inclined branch and stem growth caused by abiotic stresses such as wind or snow (Barnett, 1981; Li et al., 2006; Rennenberg et al., 2006; Sinnott, 1952; Street et al., 2006; Timell, 1986). To date, several hundred drought responsive genes have been identified in pine (Dubos and Plomion, 2003; Dubos et al., 2003; Gonzalez-Martinez et al., 2006; Lorenz et al., 2006; Watkinson et al., 2003) and Populus (Nanjo et al., 2004; Street et al., 2006). A much smaller number of cold- and mechanical stress-related genes have also been isolated from various tree species (Benedict et al., 2006; Lu et al., 2008). However, the regulatory networks governing these genes and the overall response of trees to stress are poorly understood.

Recently, a class of negative post-transcriptional regulators, called microRNAs (miRNAs), has been intensively studied (Chen, 2005; Jones-Rhoades et al., 2006; Mallory and Vaucheret, 2006). These small, non-coding RNAs are produced from precursors with unique stem–loop structures (Bartel, 2004); thus far a total of 178 miRNA families, representing 959 founding members, have been found in 10 plant species including Arabidopsis thaliana, Brassica napus, Glycine max, Medicago truncatula, Oryza sativa, Physcomitrella patens, Populus trichocarpa, Saccharum officinarum, Sorghum bicolor, and Zea mays according to release 9.2 of miRBase (Griffiths-Jones et al., 2006). Of the 178 miRNA families, 21 have been authenticated in mutants or functionally characterized in Arabidopsis. It has been demonstrated through post-transcriptional gene silencing (PTGS) that plant miRNAs are involved in various developmental processes, such as organ boundary formation, organ polarity/radial patterning, and in the development of root, stem, leaf, and flower organs (Chen, 2005; Jones-Rhoades et al., 2006; Mallory and Vaucheret, 2006; Sunkar et al., 2007).

Increasing evidence indicates that miRNAs play important roles in the response of plants to biotic and abiotic stresses (Sunkar et al., 2007). The levels of miRNA expression were altered in plants infected with virus (Bazzini et al., 2007) or in plants expressing suppressors of PTGS (Chapman et al., 2004; Chen et al., 2004; Kasschau et al., 2003; Mlotshwa et al., 2005). Expression of 10 of the 11 analyzed miRNA families was significantly repressed in the galled loblolly pine (Pinus taeda) stem infected with the fungus Cronartium quercuum f.sp. fusiforme (Lu et al., 2007). In addition, Arabidopsis (Ath)-miR393 mediates antibacterial resistance by repressing auxin signaling (Navarro et al., 2006). Under abiotic stresses, such as cold, drought, salinity (Sunkar and Zhu, 2004; Zhao et al., 2007), UV-B radiation (Zhou et al., 2007), phosphate or sulfate starvation (Chiou et al., 2006; Fujii et al., 2005; Jones-Rhoades and Bartel, 2004), oxidative stress (Sunkar et al., 2006), or mechanical strain (Lu et al., 2005), the expression of specific plant miRNAs was altered. For instance, the expression of At-miR395 depends on sulfate concentration (Jones-Rhoades and Bartel, 2004). No miR395 transcript was detected in Arabidopsis plants grown in soil or on media containing 2 mm SO42−; however, the expression levels of this transcript increased as the concentration of SO42− was reduced to 0.2 or 0.02 mm. The expression of an Ath-miR395 target ATP sulfurylase gene (APS1), that encodes an enzyme catalyzing the first step of inorganic sulfate assimilation, decreases with the reduced sulfate concentration. Thus, Ath-miR395 is clearly a miRNA involved in sulfate starvation. Similarly, miRNAs are also involved in the adaptability of plants to environmental changes such as the availability of inorganic phosphate (Chiou et al., 2006; Fujii et al., 2005). Upon inorganic phosphate starvation, an Arabidopsis miRNA (Ath-miR399) is induced, whereas the expression of its target, a ubiquitin-conjugating E2 enzyme (UBC) gene, is reduced. Transgenic plants expressing a UBC gene without an Ath-miR399 target site have an altered response to inorganic phosphate starvation (Fujii et al., 2005). In another example, transgenic plants over-expressing Ath-miR399 were observed to accumulate inorganic phosphate in their shoots and display symptoms of phosphate toxicity (Chiou et al., 2006). Another Arabidopsis miRNA that plays an important regulatory role in stress response is Ath-miR398 (Sunkar et al., 2006). It targets two superoxide dismutases that convert superoxide to hydrogen peroxide and molecular oxygen, an important response to superoxide radical formation (Fridovich, 1995). In Arabidopsis plants subjected to oxidative stress inducers, such as high light, heavy metal, and methyl viologen treatments, Ath-miR398 is downregulated, while the expression of superoxide dismutase genes is induced. Over-expression of an Ath-miR398-resistant form of a superoxide dismutase gene leads to a great improvement in the resistance of the plant to oxidative stresses (Sunkar et al., 2006).

To date, a total of 33 miRNA families, representing 215 founding members, have been found in the tree species P. trichocarpa (miRBase Release 9.2; Griffiths-Jones et al., 2006). Of these families, 11 were computationally predicted based on their homology with miRNA sequences identified in other plant species and are poorly characterized. The founding members of the remaining 21 families were cloned from developing xylem tissue of P. trichocarpa (Lu et al., 2005). The expression of a majority of these cloned miRNAs was altered in woody tissues that were subjected to mechanical stress. This indicates the presence of abiotic stress-responsive miRNA regulation in P. trichocarpa similar to that in Arabidopsis (Sunkar and Zhu, 2004). Here, we report the cloning and characterization of 68 miRNAs from P. trichocarpa tissues subjected to cold, heat, dehydration, salinity, flood, or mechanical stress. These miRNAs were predicted to cleave 346 genes, of which six have been experimentally validated. Furthermore, we identified a set of abiotic stress-responsive P. trichocarpa miRNAs using northern blots and miRNA microarrays.


  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Identification of 68 miRNAs from stress-treated tissues of Populus trichocarpa

Populus trichocarpa (Ptc) plants, grown in vitro for approximately 1.5 months, were exposed to cold, heat, dehydration, salinity, or hydration stress treatments. The treated and control (no treatment) plants were pooled and used for constructing an abiotic stressed small-RNA library (AL). The developing xylem tissues from mechanically stressed and normal stems of 1-year-old greenhouse-grown P. trichocarpa plants were pooled to prepare a mechanically stressed small-RNA library (ML). We obtained 3686 (2648 distinct) sequences with sizes of 13 to 25 nucleotides (nt) from the AL, of which 497 with distinct sequences of 20–24 nt. From the ML we obtained 1796 (1179 distinct) sequences with sizes of 13–31 nt, including 175 distinct sequences with 20–24 nt. We focused on the 20- to 24-nt small RNAs for further analysis. Blast analyses against the GenBank ( showed that 423 of the 497 AL and 128 of the 175 ML sequences, which were not further analyzed, correspond to known non-coding rRNAs, tRNAs, small nuclear RNAs, and to those associated with retrotransposons or transposons. Among the remaining 74 AL and 47 ML sequences, eight (Ptc-9, Ptc-141, Ptc-167, Ptc-272, Ptc-273, Ptc-277, Ptc-279, Ptc-280) are present in both libraries, yielding a total of 113 distinct sequences for additional characterization (Figure 1, Tables 1 and 2, Table S1).


Figure 1.  Flowchart for the isolation and prediction of miRNAs in Populus trichocarpa (ptc) plants treated with various abiotic stresses.

Download figure to PowerPoint

Table 1.   Conserved microRNAs (miRNAs) between Populus trichocarpa and Arabidopsis thaliana
miRNA genemiRNA sequence (5′[RIGHTWARDS ARROW]3′)Cloning frequencyLength (nt)miRNA location
  1. nt, nucleotides.

  2. The cloned small RNA IDs and sequences are shown in bold.

  3. *The sequence is a homolog of the cloned sequence. The cloning frequency of small RNA from the abiotic stressed small RNA library (AL) and the mechanically stressed small RNA library (ML) are shown. The genome locations of miRNA are presented as ‘genome ID ( miR start site-stop site with the ID’.

  4. aThe miRNA locus has also been identified in our previous paper (Lu et al., 2005).

Ptc-MIR159aUUUGGAUUGAAGGGAGCUCU  20LG_XII:4201697-4201678a
Ptc-MIR159bUUUGGAUUGAAGGGAGCUCU  20LG_XV:3478615-3478634a
Ptc-MIR159cUUUGGAUUGAAGGGAGCUCU  20scaffold_1140:8367-8386a
Ptc-MIR159dCUUGGAUUGAAGGGAGCUCC*  20LG_XIV:3089086-3089105
Ptc-MIR319aUUGGACUGAAGGGAGCUCCCU  21LG_I:7044536-7044556a
Ptc-MIR319bUUGGACUGAAGGGAGCUCCCU  21LG_III:12635218-12635238a
Ptc-MIR319cUUGGACUGAAGGGAGCUCCCU  21LG_XIII:12992460-12992480a
Ptc-MIR319dUUGGACUGAAGGGAGCUCCCU  21LG_XIX:10852752-10852732a
Ptc-MIR319eUUGGACUGAAGGGAGCUCCUG*  21LG_XII:5108591-5108571a
Ptc-MIR319gUUGGACUGAAGGGAGCUCCUU*  21scaffold_125:778467-778447a
Ptc-MIR319hUUGGACUGAAGGGAGCUCCUG*  21scaffold_82:697146-697126a
Ptc-MIR319iUUGGGCUGAAGGGAGCUCCCA*  21LG_VI:17044727-17044747a
Ptc-MIR160aUGCCUGGCUCCCUGUAUGCCA  21LG_I:32813190-32813210a
Ptc-MIR160cUGCCUGGCUCCCUGUAUGCCA  21LG_X:18982841-18982861a
Ptc-MIR160dUGCCUGGCUCCCUGUAUGCCA  21LG_XI:14495661-14495641a
Ptc-MIR160fUGCCUGGCUCCCUGAAUGCCA*  21scaffold_137:187305-187285a
Ptc-MIR160gUGCCUGGCUCCCUGGAUGCCA*  21LG_XIV:9829763-9829743a
Ptc-MIR160hUGCCUGGCUCCCUGCAUGCCA*  21LG_XVI:7014398-7014378a
Ptc-MIR164aUGGAGAAGCAGGGCACGUGCA  21LG_II:15132582-15132562a
Ptc-MIR164bUGGAGAAGCAGGGCACGUGCA  21LG_VI:7431066-7431086a
Ptc-MIR164eUGGAGAAGCAGGGCACGUGCA  21LG_XVI:12388887-12388907a
Ptc-MIR164fUGGAGAAGCAGGGCACAUGCU*  21LG_III:6492409-6492429a
Ptc-MIR166aUCGGACCAGGCUUCAUUCCC  20LG_I:6097079-6097060
Ptc-MIR166bUCGGACCAGGCUUCAUUCCC  20LG_II:2737702-2737721
Ptc-MIR166cUCGGACCAGGCUUCAUUCCC  20LG_II:13668251-13668270
Ptc-MIR166eUCGGACCAGGCUUCAUUCCC  20LG_V:15276210-15276191
Ptc-MIR166gUCGGACCAGGCUUCAUUCCC  20LG_VIII:11091614-11091595
Ptc-MIR166hUCGGACCAGGCUUCAUUCCC  20LG_X:8492342-8492361
Ptc-MIR166mUCGGACCAGGCUUCAUUCCC  20scaffold_122:445122-445141
Ptc-MIR166nUCGGACCAGGCUUCAUUCCU*  20LG_XIX:5391587-5391568
Ptc-MIR166pUCGGACCAGGCUCCAUUCCU*  20LG_V:824242-824223
Ptc-MIR166qUCGGACCAGGCUUCAUUCCU*  20LG_VII:3395015-3394996
Ptc-MIR167cUGAAGCUGCCAGCAUGAUCUA  21LG_V:14949534-14949554
Ptc-MIR167dUGAAGCUGCCAGCAUGAUCUA  21scaffold_14274:285-265
Ptc-MIR167eUGAAGCUGCCAGCAUGAUCUG  21scaffold_70:484760-484740
Ptc-MIR167fUGAAGCUGCCAGCAUGAUCUU  21LG_X:9042289-9042269
Ptc-MIR168aUCGCUUGGUGCAGGUCGGGAA  21scaffold_86:1165699-1165679a
Ptc-MIR169abUAGCCAAGGACGACUUGCCCAC  22LG_V:2516893-2516872
Ptc-MIR169acUAGCCAAGGACGACUUGCCCAC  22LG_V:2520781-2520802
Ptc-MIR169aeUAGCCAAGGACGACUUGCCCAC  22LG_X:6859529-6859508
Ptc-MIR169afUAGCCAAGGACGACUUGCCCAC  22LG_X:6867156-6867135
Ptc-MIR169wUAGCCAAGGAUGACUUGCCCAC*  22LG_I:20794125-20794146
Ptc-MIR171cAGAUUGAGCCGCGCCAAUAUC*  21LG_VI:4125868-4125848a
Ptc-MIR171dAGAUUGAGCCGCGCCAAUAUC*  21LG_XVIII:12295416-12295396a
Ptc-MIR171eUGAUUGAGCCGUGCCAAUAUC  21LG_I:24172982-24173002a
Ptc-MIR171gUGAUUGAGCCGUGCCAAUAUC  21LG_XII:11612120-11612140a
Ptc-MIR171hUGAUUGAGCCGUGCCAAUAUC  21LG_XV:7305205-7305225a
Ptc-MIR171iUGAUUGAGCCGUGCCAAUAUC  21scaffold_57:112845-112865a
Ptc-MIR171kGGAUUGAGCCGCGCCAAUAUC*  21LG_XV:4259968-4259948
Ptc-MIR390aAAGCUCAGGAGGGAUAGCGC  20LG_I:20745322-20745341
Ptc-MIR390bAAGCUCAGGAGGGAUAGCGC  20LG_VI:7266292-7266311
Ptc-MIR390cAAGCUCAGGAGGGAUAGCGC  20LG_IX:5407928-5407909
Ptc-MIR390dAAGCUCAGGAGGGAUAGCGC  20LG_XVI:11828749-11828768
Ptc-MIR394bCUGUUGGUCUCUCUUUGUAA  20LG_V:17178819-17178838
Ptc-MIR396aUUCCACAGCUUUCUUGAACU  20LG_VI:4075654-4075673
Ptc-MIR396cUUCCACAGCUUUCUUGAACU  20LG_III:14406868-14406887
Ptc-MIR396dUUCCACAGCUUUCUUGAACU  20LG_VI:4082543-4082524
Ptc-MIR396fUUCCACGGCUUUCUUGAACU*  20LG_VI:17347052-17347033
Ptc-MIR396gUUCCACGGCUUUCUUGAACU*  20LG_VII:6490267-6490248
Ptc-MIR397aUCAUUGAGUGCAGCGUUGAU  20LG_XI:3295628-3295609
Ptc-MIR397bCCAUUGAGUGCAGCGUUGAU*  20LG_IV:1782232-1782251
Ptc-MIR398aUGUGUUCUCAGGUCACCCCUU*  21LG_X:14819174-14819194
Ptc-MIR398bUGUGUUCUCAGGUCGCCCCUG  21LG_I:19895257-19895237
Ptc-MIR408AUGCACUGCCUCUUCCCUGGC  21LG_II:16909463-16909443a
Ptc-MIR472aUUUUCCCUACUCCACCCAUCC  21scaffold_131:138364-138384a
Table 2.   Non-conserved miRNAs between Populus trichocarpa and Arabidopsis thaliana
miRNA genemiRNA sequence (5′[RIGHTWARDS ARROW]3′)Cloning frequencyLength (nt)miRNA location
  1. nt, nucleotides.

  2. The cloned small RNA IDs and sequences are shown in bold.

  3. *The sequence is a homolog of the cloned sequence. The cloning frequency of small RNA from the abiotic stressed small RNA library (AL) and the mechanically stressed small RNA library (ML) are shown. The genome locations of miRNA are presented as ‘genome ID ( miR start site-stop site with the ID’.

  4. aThe miRNA locus has also been identified in our previous paper (Lu et al., 2005).

Ptc-MIR475aUUACAGUGCCCAUUGAUUAAG  21LG_II:20949518-20949538a
Ptc-MIR475bUUACAGUGCCCAUUGAUUAAG  21LG_VIII:13883578-13883558a
Ptc-MIR475cUUACAAUGUCCAUUGAUUAAG*  21LG_X:12045864-12045884a
Ptc-MIR476aUAGUAAUCCUUCUUUGCAAA  20LG_XIX:11339587-11339606a
Ptc-MIR476bUAGUAAUUCUUCUUUGCAAA*  20scaffold_2286:2819-2838a
Ptc-MIR476cUAGUAAUUCUUCUUUGCAAA*  20LG_III:15321637-15321618a
Ptc-MIR530aUGCAUUUGCACCUGCACCUU  20LG_IX:11430295-11430314
Ptc-MIR530bUGCAUUUGCACCUGCAUCUU*  20scaffold_163:61415-61434
Ptc-MIR1444bUUCACAUUCGGUCAACGUUC*  20LG_X:10569467-10569486
Ptc-MIR1444cUUCACAUUCGGUCAACGUUC*  20scaffold_8959:1131-1112
Ptc-MIR1445UCCCUUGUAGACUAGAAAAA*  20scaffold_203:157719-157738
Ptc-MIR1446aUUCUGAACUCUCUCCCUCAA  20LG_XII:12249427-12249408
Ptc-MIR1446bUUCUGAACUCUCUCCCUCAA  20LG_XII:12244761-12244742
Ptc-MIR1446cUUCUGAACUCUCUCCCUCAA  20LG_XII:12359326-12359345
Ptc-MIR1446dUUCUGAACUCUCUCCCUCAA  20LG_XV:7922488-7922469
Ptc-MIR1446eUUCUGAACUCUCUCCCUCAA  20LG_XV:8090324-8090343
Ptc-MIR1449UGAGGUGCACGUAAGAUAACUC  22LG_II:23221770-23221791
Ptc-MIR1450UUCAAUGGCUCGGUCAGGUUAC*  22scaffold_163:118894-118873

Aligning the 113 cloned sequences or their homologs (one or two mismatches) with the Populus draft genome assembly (Tuskan et al., 2006) using PatScan (Dsouza et al., 1997), we identified 63 sequences having perfectly matched and 31 near-perfectly matched genome sequences (one or two mismatches) (Figure 1). The remaining 19 have no matched genome sequences (more than two mismatches), which could be due to incomplete coverage of the sequenced genome. It could also be due to nuclear editing (Bass, 2002; Luciano et al., 2004) or some unknown cloning artifacts that caused sequence modifications. Genome sequences (about 600 nt) surrounding the small RNA sequences or their homologs were then used to predict the secondary structure using the mfold program (Zuker, 2003). Hairpin structures were identified for 43 cloned sequences and 25 of their homologs, suggesting that these 68 small RNAs are potential Ptc-miRNAs (the mature miRNA sequences are shown in Tables 1 and 2; the newly identified hairpin structures and sequences are shown in Figure 2 and Table S2). The remaining cloned small RNA sequences without confirmed hairpin structures are considered putative small interfering RNAs (siRNAs; Table S1), of which none were found to be phased siRNAs (data not shown).


Figure 2.  Hairpin structures of the precursors containing newly identified microRNA (miRNA) sequences (red) in Populus trichocarpa (ptc), rice (osa), and loblolly pine (pta). The sequences of ptc-miR479 (Lu et al., 2005) in the ptc-MIR171l precursor are shown in green. The newly identified ptc-miR394a.2 (red) and the ptc-miR394a.1 (green) predicted computationally (Tuskan et al., 2006) are shown in the ptc-MIR394a precursor. The newly identified ptc-miR394b.2 (red) and the ptc-miR394a.1 (green) described previously (Tuskan et al., 2006) are shown in the ptc-MIR394b precursor. Both the newly cloned ptc-miR482.2 (red) and the sequences of ptc-miR481.1 (italic), which were previously cloned by Lu et al. (2005), are shown in the ptc-MIR482 precursor. The newly identified osa-miR530-5p (red) and the previously reported osa-miR530-3p (green) (Liu et al., 2005) are shown in the osa-MIR530 precursor.

Download figure to PowerPoint

Based on sequence similarity, the 68 Ptc-miRNAs were classified into 27 miRNA families (Tables 1 and 2), of which 18 have been previously identified in P. trichocarpa (Lu et al., 2005; Tuskan et al., 2006). It should be stressed that the sequences of eight (Ptc-MIR166, Ptc-MIR167, Ptc-MIR169, Ptc-MIR390, Ptc-MIR394, Ptc-MIR396, Ptc-MIR397, and Ptc-MIR398) of these 18 were based on computational predictions. We report here the experimental validation of those predictions. The remaining nine families, including Ptc-MIR530, Ptc-MIR827, and Ptc-MIR1444 to Ptc-MIR1450, are novel.

Of the 27 identified miRNA families, 15 previously known families and one (Ptc-MIR827) of the newly identified families are conserved between P. trichocarpa and Arabidopsis (Table 1). The remaining three known families (Ptc-MIR475, Ptc-MIR476, and Ptc-MIR482) together with the other eight novel families are absent from the Arabidopsis genome (Table 2).

The 27 Ptc-miRNA families were mapped to a total of 120 loci in the genome of P. trichocarpa (Tuskan et al., 2006) (Tables 1 and 2), of which 16 encode 11 founding members of the nine novel Ptc-miRNA families. Moreover, we identified two new hairpin structures (Ptc-MIR171m and Ptc-MIR171n), and a precursor (Ptc-MIR171l) that produces both Ptc-miR479 from its 5′ arm (Lu et al., 2005) and a new mature miRNA (Ptc-miR171l) from its 3′ arm, for the family Ptc-MIR171 (Figure 2). The only precursor of Ptc-MIR482 family produces both the Ptc-miR482.1 identified by Lu et al. (2005) and a newly cloned sequence Ptc-miR482.2 (Figure 2). The precursors of the Ptc-MIR394 family produce both Ptc-miR394a.1 and Ptc-miR394b.1 (Tuskan et al., 2006) as well as the newly cloned sequences Ptc-miR394a.2 and Ptc-miR394b.2 (Figure 2).

The 43 cloned Ptc-miRNAs and their homologs were also aligned with the Arabidopsis ( and rice (Oryza sativa; Osa) genomes (, and with the DFCI Pine Gene Index, release 6.0 ( using PatScan (Dsouza et al., 1997) to identify their homologs and surrounding sequences. Analyzing the sequences by the mfold program, we identified 40, 62, and 8 loci in the genomes of Arabidopsis, rice, and pine, respectively (Table S3). All of the 40 Arabidopsis loci had been previously reported ( Among the 62 loci in the rice genome, Osa-MIR530 produces the newly identified Osa-miR530-5p and the previously reported Osa-miR530-3p (Liu et al., 2005), while Osa-MIR827 is novel (Figure 2). The eight loci for five miRNA families in the pine genome include six that have been previously reported (Lu et al., 2007) and two newly identified ones (Pta-MIR482c and Pta-MIR482d) (Figure 2, Table S3).

Prediction of targets for the identified Ptc-miRNAs

Using the previously established procedures (Jones-Rhoades and Bartel, 2004; Lu et al., 2005), we predicted, for the 27 Ptc-miRNA families, a total of 346 targets from the 45,555 gene models in the annotation v1.1 JamboreeModels of the P. trichocarpa genome (Tuskan et al., 2006). These targets have penalty scores of ≤2.5 for mismatched patterns in the miRNA:mRNA duplexes (Table 3, Table S4). Included are those of previous computationally-predicted Ptc-miRNAs whose regulatory targets were unknown. Among the predicted targets, 72 are for the Ptc-MIR394, Ptc-MIR482 and the nine novel Ptc-miRNA families (Table 3). Twenty of the 346 targets were predicted to be cleaved by Ptc-miRNA members in two or three families. Five disease resistance protein-encoding targets, eugene3.00440220, eugene3.00180517, grail3.0140004801, gw1.8759.5.1, and gw1.VI.1923.1, were predicted to be cleaved by both Ptc-MIR482 and Ptc-MIR472, and one, eugene3.00190077, was predicted to be cleaved by three miRNAs, Ptc-MIR472, Ptc-MIR482, and Ptc-MIR1448. Fourteen genes that encode pentatricopeptide repeat-containing proteins (PPRs) are the predicted targets of both Ptc-miR475 and Ptc-miR476 (Table S4). In addition, each of the six predicted targets of Ptc-miR476, eugene3.24000001, eugene3.00190210, eugene3.00062011, eugene3.01250075, eugene3.00061748, and fgenesh4_pg.C_scaffold_29000235, has two or three complementary sites of Ptc-miR476 (Table S4).

Table 3.   Potential targets for the Ptc-MIR394, Ptc-MIR482 and the eight newly identified miRNA families in Populus trichocarpa
MiRNA familyTarget functionTargeta
  1. aAll predicted targets with the lowest penalty scores (shown in parentheses) of ≤2.5 are listed.

  2. b5′-RACE validated targets.

  3. cThe gene model is found in Populus trichocarpa v1.0, but not in Populus trichocarpa v1.1.

Ptc-MIR482Disease resistance proteineugene3.00102261(1.5)b, eugene3.00190017(2.5), gw1.8759.5.1(2.5), gw1.VI.1923.1(2.5), eugene3.00190077(2.5), eugene3.00180517(2.5), eugene3.00440220(2.5), grail3.0085005401(2.5), eugene3.01170064(2.5), fgenesh4_pg.C_LG_XIX000056(2.5), grail3.0140004801(2.5), fgenesh4_pg.C_scaffold_7992000001(2.5)
Ptc-MIR530Zinc knuckle (CCHC-type) family proteineugene3.00090148(2)
Homeobox transcription factor KN3estExt_Genewise1_v1.C_LG_II1820(2.5)
Ribosomal protein L1 family proteingw1.XIV.2998.1(2.5)
Unknowngw1.VI.774.1(2.5), estExt_Genewise1_v1.C_LG_XVIII3096(2.5), eugene3.00030970(2.5), fgenesh4_pg.C_LG_VIII001448(2.5), grail3.0100010901(2.5)
Ptc-MIR827sec14 cytosolic factor family proteingw1.III.2050.1(1.5), gw1.29.149.1(2), fgenesh4_pg.C_scaffold_1262000002(2)
Ptc-MIR1444Polyphenol oxidasegw1.182.27.1(1)b, eugene3.00110271(1.5)b, gw1.XI.3507.1(2), gw1.XI.3509.1(2), gw1.178.3.1(2), gw1.178.38.1(2), gw1.178.49.1(2), gw1.182.19.1(2), gw1.8119.4.1(2), eugene3.01780010(2), fgenesh4_pg.C_scaffold_14069000001(2), grail3.0182000101(2), fgenesh4_pg.C_scaffold_178000012(2)
SET domain proteingw1.XIII.2167.1(2.5)
KH domain proteineugene3.00150124(2.5), fgenesh4_pm.C_LG_XII000094(2.5)
Ptc-MIR1445DihydropyrimidinaseestExt_Genewise1_v1.C_LG_IX3131(2.5), gw1.I.1992.1(2.5)
Unknowngw1.XII.661.1(2), eugene3.00021625(2.5)
Ptc-MIR1446GCN5-related N-acetyltransferase (GNAT) family proteineugene3.00060403(2)b
Gibberellin response modulator-like proteinfgenesh4_pg.C_LG_XII000915(2)b
Replication factor C-likefgenesh4_pg.C_LG_XIV001182(2)
Homeodomain transcription factorgw1.I.9208.1(2.5)
Ptc-MIR1447Ankyrin repeat family proteinfgenesh4_pg.C_scaffold_120000019(1.5), eugene3.17360001(1.5), fgenesh4_pg.C_scaffold_120000014(2), fgenesh4_pg.C_scaffold_120000035(2.5), fgenesh4_pg.C_scaffold_120000037(2.5), fgenesh4_pg.C_scaffold_120000026(2.5),
Leucine-rich repeat transmembrane protein kinasefgenesh4_pg.C_LG_XV000379(2)
Disease resistance proteineugene3.00121285(2.5)
Translationally controlled tumor protein homolog (TCTP)estExt_Genewise1_v1.C_LG_X3955(2.5)
Unknowngrail3.0092004201(0.5), gw1.XIII.2552.1(2.5), gw1.X.3990.1(2.5)
Ptc-MIR1448Disease resistance proteineugene3.01310091(2.5)b,c, eugene3.00190077(2.5)
Glutathione S-conjugate ABC transporter (MRP2)gw1.1700.5.1(2.5)
ATP-binding cassette transport proteingw1.IV.2236.1(2.5)
Ptc-MIR1450Leucine-rich repeat transmembrane protein kinaseeugene3.00141443(2.5)
Unknownfgenesh4_pg.C_LG_V001134(2.5), fgenesh4_pg.C_LG_XII001210(2.5)

Among the 72 targets of Ptc-MIR394, Ptc-MIR482, and the nine novel miRNA families, five encode transcription factors, including three targets of Ptc-MIR1446 that encode a gibberellin response modulator-like protein, a homeodomain transcription factor, and a replication factor C-like protein, and two targets of Ptc-MIR530 that code for a zinc knuckle (CCHC-type) family protein and a homeobox transcription factor KN3 (Table 3). Many of the others are stress-related genes, such as the 13 disease resistance protein genes targeted by Ptc-MIR482, Ptc-MIR1447, and Ptc-MIR1448 and the 13 polyphenol oxidase (ppo) genes targeted by Ptc-MIR1444 (Table 3). Polyphenol oxidases (PPOs) are copper-containing enzymes that oxidize mono- or dihydroxy phenols to quinines. The ppo genes are usually expressed in a tissue-specific manner and are involved in plant resistance to biotic and abiotic stresses (for a review see Mayer, 2006). Interestingly, except for the 13 ppo genes involved in secondary metabolism and in response to stress (for a review see Mayer, 2006), the novel miRNA family, Ptc-MIR1444, is also predicted to target a SET domain protein and two KH domain proteins (Table 3). The SET domain protein has 58% identity with a member (At3g03750) of the Arabidopsis histone-lysine N-methyltransferase family that functions in histone methylation and gene silencing (Jackson et al., 2002). Both of the two KH domain proteins have 62% identity with an Arabidopsis protein (At5g53060) that contains three K homology RNA-binding domains (Lorković and Barta, 2002). The KH domain proteins function in RNA metabolism and are critical to vegetative and reproductive development (Cheng et al., 2003; Mockler et al., 2004; Ripoll et al., 2006). They also function in the determinacy of organogenesis in stem cells by interacting with other proteins (Phelps-Durr et al., 2005). Thus, Ptc-MIR1444 may have an important role in plant development, metabolism, and defense.

Because Ptc-MIR530 and Ptc-MIR827, two novel miRNA families, are conserved between P. trichocarpa and rice (Table S3), we searched the TIGR Rice Genome Pseudomolecules and Annotation Release 4 (, as previously performed (Lu et al., 2005) using penalty scores of ≤2.5 for mismatched patterns in Osa-miR530-5p:mRNA or Osa-miR827:mRNA duplexes as the criterion. A total of nine genes were predicted to be targets of Osa-miR530-5p and Osa-miR827 (Table 4). They encode a kinase, a plus-3 domain protein, two 2,3-diketo-5-methylthio-1-phosphopentane phosphatases, two transposable element proteins, and three functionally unknown proteins. These proteins appear to have different functions from the predicted targets of the P. trichocarpa miRNAs Ptc-miR530a, Ptc-miR530b, and Ptc-miR827 (Table 3), suggesting that the founding members of these miRNAs may function differently in these two plant species although their sequences are conserved.

Table 4.   Potential targets for the newly identified miRNAs in rice (Oryza sativa)
miRNATarget functionTargeta
  1. aAll predicted targets with the lowest penalty scores (shown in parentheses) of ≤2.5 are listed.

Osa-miR530-5pPlus-3 domain proteinLOC_Os01g56780(1.5)
2,3-diketo-5-metylthio-1-phosphopentane phosphataseLOC_Os01g01120(2.5), LOC_Os01g01120(2.5)
Transposable element proteinLOC_Os06g16600(2.5), LOC_Os10g29450(2.5)
UnknownLOC_Os02g43314(2), LOC_Os12g36820(2.5)

Validation of six targets for novel miRNA families

To validate miRNA-mediated cleavage of target transcripts, we carried out rapid amplification of 5′ complementary DNA ends (5′-RACE) on mRNAs isolated from P. trichocarpa tissues from which the cloned miRNAs originated. We selected a few predicted targets of the members of Ptc-MIR482 and the novel miRNA families having penalty scores of ≤2.5 in the miRNA:target mismatch patterns (Lu et al., 2005). Six predicted targets, gw1.182.27.1, eugene3.00110271, eugene3.00102261, eugene3.00060403, fgenesh4_pg.C_LG_XII000915, and eugene3.01310091, were cleaved in vivo by members of the novel miRNA families, Ptc-MIR482, Ptc-MIR1444, Ptc-MIR1446, or Ptc-MIR1448 (Figure 3). Examination of the cleavage sites and the sequence complementarity between the miRNAs and their targets revealed that four of the six validated miRNA/target pairs could pass the miRNA:target recognition parameters deduced by Schwab et al. (2005); while two of them (Ptc-miR1446a-e:eugene3.00060403 and Ptc-miR1448:eugene3.01310091) could not. These data are consistent with those for the other validated miRNA/target pairs (Jones-Rhoades and Bartel, 2004; Lu et al., 2005, 2007; Palatnik et al., 2003; Vaucheret et al., 2004), suggesting that eugene3.00060403 and eugene3.01310091 are authentic targets of miRNAs although they could not pass the miRNA:target recognition parameters (Schwab et al., 2005).


Figure 3.  Validation of the predicted mRNA targets. The mRNA cleavage sites were determined by modified 5′ RNA ligase-mediated rapid amplification of 5′ complementary DNA ends (5′-RLM-RACE). The mRNA sequence of each complementary site and its 5′ and 3′ flanking sequences (10 nt) from 5′ to 3′ and the cloned microRNA (miRNA) sequence from 3′ to 5′ are shown. Watson–Crick pairing (vertical dashes) and G:U wobble pairing (circles) are indicated. Vertical arrows indicate the 5′ termini of miRNA-guided cleavage products, as identified by 5′-RACE, with the frequency of clones shown.

Download figure to PowerPoint

gw1.182.27.1 and eugene3.00110271 are two of the 13 ppo genes predicted to be targets of Ptc-miR1444a. Both the validated target of Ptc-miR482.2, eugene3.00102261, and the validated target of Ptc-miR1448, eugene3.01310091, encode putative disease resistance proteins. Eugene3.01310091 was a predicted gene model in the annotation v1.0 JamboreeModels of the P. trichocarpa genome (Tuskan et al., 2006); however, it does not exist in the current annotation (v1.1), indicating that some gene models in the annotation v1.1 are not accurate. These results suggest that Ptc-miR482.2, Ptc-miR1444, and Ptc-miR1448 may be involved in the resistance of plants to biotic and abiotic stresses through the cleavage of ppo genes and disease resistance protein genes.

One of the two validated targets of Ptc-miR1446a-e, eugene3.00060403, encodes a GCN5-related N-acetyltransferase (GNAT) family protein (Figure 3). This protein contains an acetyltransf_1 domain with N-acetyltransferase activities at the N-terminal and a C-terminal BRCT domain that has been found within many DNA damage repair and cell cycle checkpoint proteins (Marchler-Bauer et al., 2005). The functions of the plant GNAT family protein have not yet been fully characterized.

Interestingly, the other validated Ptc-miR1446 target, fgenesh4_pg.C_LG_XII000915 (Figure 3), is also cleaved by two miRNAs, Ptc-miR473 and Ptc-miR477 (Lu et al., 2005). This target gene, sharing a 66% amino acid identity with a rice gibberellin (GA) response modulator-like protein (LOC_Os01g67650,, is a GA-related GRAS transcription factor that could be a repressor of gibberellin signaling (Peng et al., 1997). Because over-expression of the GA-related GRAS protein diminished stem elongation and induced a dwarf phenotype (Itoh et al., 2005), Ptc-miR473, Ptc-miR477, and Ptc-miR1446 may play important roles in the growth of trees by cleaving the fgenesh4_pg.C_LG_XII000915 transcripts. In addition, GRAS proteins also function in the response of plants to mechanical stress (Mayrose et al., 2006), indicating that Ptc-miR473, Ptc-miR477, and Ptc-miR1446 may be involved in the formation of specialized woody tissue in trees.

Northern blot analysis of developmental and stress-responsive expression of Populus miRNAs

In order to predict possible roles for the novel Ptc-miRNAs (Figure 2), we analyzed their expression levels in leaf, phloem, and developing xylem of P. trichocarpa by northern blot analysis using probes with complementary sequences (Figure 4). Northern detection represents a useful criterion for authenticating miRNAs (Ambros et al., 2003). Transcripts of the novel members of the known Ptc-MIR171 and Ptc-MIR482 families, Ptc-miR171l-n and Ptc-miR482.2, and the nine novel Ptc-miRNA families were detected in all the tested tissues, with some exhibiting apparent tissue specificity. The expression of Ptc-miR530a, Ptc-miR1444a, and Ptc-miR1447 are more leaf specific, whereas Ptc-miR171l-n, Ptc-miR1445, Ptc-miR1446a-e, and Ptc-miR1449 transcripts are more abundant in tissues associated with cambium differentiation in the stem. During the formation of tension wood (TW) and opposite wood (OW) in the P. trichocarpa stem suffering from mechanical stress, the expression levels of Ptc-miR530a, Ptc-miR827, Ptc-miR1444a, Ptc-miR1446a-e, Ptc-miR1447, and Ptc-miR1450, which are members of six of the nine novel miRNA families, were altered (up or down) by more than 1.5-fold (Figure 4). In particular, the expression of Ptc-miR1444a was strongly up-regulated in both TW and OW, whereas the transcripts of Ptc-miR530a, Ptc-miR827, and Ptc-miR1450 were significantly reduced in OW.


Figure 4.  Developmental and mechanical stress-responsive expression of Populus microRNAs (miRNAs). Total RNA isolated from leaves (L), phloem (P), developing xylem (X), developing xylem of the tension wood (Xtw), and developing xylem of the opposite wood (Xow) of Populus trichocarpa were analyzed by RNA gel blots with end-labeled antisense oligonucleotides to the Ptc-miR171l-n, Ptc-miR482.2, Ptc-miR530a, Ptc-miR827, Ptc-miR1444a, Ptc-miR1445, Ptc-miR1446a-e, Ptc-miR1447, Ptc-miR1448, Ptc-miR1449, and Ptc-miR1450 sequences. Blots were stripped and rehybridized with a 5.8S rRNA probe. The relative accumulation levels of miRNA to 5.8S rRNA are shown in histograms. The normalized miRNA levels in developing xylem were set arbitrarily to 1.

Download figure to PowerPoint

In addition to mechanical stress, other abiotic stresses, such as cold, heat, salt, and dehydration, also altered the expression of Ptc-miRNAs (Figure 5). Overall, expression levels of the analyzed Ptc-miRNAs were affected, with Ptc-miR171l-n, Ptc-miR1445, Ptc-miR1446a-e, and Ptc-miR1447 being more responsive. While the expression of most of these miRNAs was clearly affected by a particular stress, Ptc-miR1446a-e expression was reduced by all stress treatments.


Figure 5.  Expression of microRNAs (miRNAs) in trees with or without abiotic stress treatments. Total RNA isolated from whole plantlets without (control, Ct) or with treatment [cold (Cd), heat (Ht), salt (St), or dehydration (Dh)] were analyzed by RNA gel blots with end-labeled antisense oligonucleotides to the Ptc-miR171l-n, Ptc-miR482.2, Ptc-miR530a, Ptc-miR827, Ptc-miR1444a, Ptc-miR1445, Ptc-miR1446a-e, Ptc-miR1447, Ptc-miR1448, and Ptc-miR1450 sequences. Blots were stripped and rehybridized with a 5.8S rRNA probe. The relative accumulation levels of miRNA to 5.8S rRNA are shown in histograms. The normalized miRNA levels in plantlets without treatment were set arbitrarily to 1.

Download figure to PowerPoint

Genome-wide profiling of Populus trichocarpa miRNA abundance under cold stress

Cold stress restricts the geographic distribution of trees and other plant species. It also significantly suppresses plant growth and development (Chinnusamy et al., 2007). Elucidating the molecular mechanism of the response of plants to cold stress will provide useful information for the genetic improvement of cold tolerance, thus enabling the growth of previously cold-sensitive species in northern areas and enhancing the ability of native plants against chilling damage (Tzfira et al., 1998). Numerous cold stress-regulated genes have been intensively studied in Arabidopsis (Chinnusamy et al., 2007), and several of these have also been identified in tree species (Benedict et al., 2006). However, no report has been published on the role of small RNAs in the response of trees to cold stress. Microarrays have been used for profiling the abundance of miRNAs in animals and plants (Axtell and Bartel, 2005; Barad et al., 2004; Baskerville and Bartel, 2005; Lim et al., 2005; Nelson et al., 2004, 2006; Thomson et al., 2004). Here, we used miRNA microarrays to examine the abundance of mature miRNAs in cold-stressed plants that were maintained at 4°C for up to 24 h. A total of 168 probes (Table S5) were designed for 203 miRNA genes deposited in the miRBase (release 9.2, Since some miRNA genes produce identical mature miRNAs, these probes allow us to investigate the abundance of 90 mature miRNAs from 33 families. The complete miRNA microarray results are listed in Table S6.

Among the 90 mature miRNAs examined by the microarray, 42 may arise from multiple genes. We designed gene-specific probes for 38 of the 42 miRNAs based on sequence variation in the mature miRNA flanking regions. We were not able to design gene-specific probes for the other four miRNAs since the sequences of the mature miRNA flanking regions are identical. Analysis of variance (anova) was conducted to test for abundance profiles of each gene-specific probe corresponding to the 38 mature miRNAs. Gene-specific probes for 31 of the 38 miRNAs were not able to differentiate the originating gene. However, the results indicated that seven mature miRNAs (Ptc-miR156a-f, Ptc-miR160a-d, Ptc-miR166b-m, Ptc-miR169d-h, Ptc-miR172a-f, Ptc-miR393a-d, Ptc-miR395b-j) displayed differential patterns for the gene-specific probes (data not shown). These probes can thus be used to differentiate the origins of these seven mature miRNAs.

Based on the microarray results, we identified 19 cold stress-responsive Ptc-miRNAs that originated from 14 miRNA families (Figure 6). Among them, 15 were up-regulated while four were downregulated in response to the cold stress treatment. For the majority of these miRNAs the change in regulation occurred within 12 h of the cold stress treatment. However the mature miRNAs of Ptc-MIR156a-f, Ptc-MIR476a, and Ptc-MIR477a,b genes showed a somewhat slower response. Interestingly, we found many examples of members of families that did not respond identically to cold stress, suggesting different functions for miRNAs of the same family.


Figure 6.  The cold-responsive Populus trichocarpa microRNAs (Ptc-miRNAs) identified by microRNA (miRNA) microarrays. Plants were treated at 4°C for 0 (CS0, control), 4 (CS4), 8 (CS8), 12 (CS12), 16 (CS16), 20 (CS20), and 24 (CS24) h, respectively. Clustering was carried out with log treatment/control ratio and grouped by each miRNA family. The intensities of the color represent the relative magnitude of fold changes in log values: −3.0 (green) indicates that the miRNA is highly suppressed by cold treatment; while +3.0 (red) indicate that it is highly induced. Cells marked with grey rectangle indicate highly significant (P values less than 0.01) cold stress responses when compared with CS0.

Download figure to PowerPoint

To confirm the microarray results, the abundance of several miRNAs was further analyzed by quantitative real-time PCR, a method which can readily discriminate the expression of miRNAs having sequence differences of as little as one nucleotide (Shi and Chiang, 2005). Comparing the abundance profiles of the microarray (Figure 6) and the PCR analysis (Figure 7), we found that they share similar trends. Both microarray and PCR revealed that Ptc-miR156g-j, Ptc-miR475a,b, and Ptc-miR476a were downregulated while Ptc-miR168a,b and Ptc-miR477a,b were up-regulated in the cold-stressed plants. We also found discrepancies in the magnitude of response, for some time points, between the microarray and PCR results, which could be due to cross-hybridization between the probe and other highly homologous miRNA family members. Another reason for the discrepancy might be data normalization of the two methods. The quantitative (q)PCR data were normalized to the abundance of 5.8S rRNA; while the microarray data were normalized to the global abundance of all miRNAs detected by the microarray; because it is technically unfeasible to use 5.8S rRNA as the normalization standard for microarray data.


Figure 7.  Real-time PCR validation of the cold-responsive Populus trichocarpa microRNAs (Ptc-miRNAs). The levels of miRNA were quantified in total RNA isolated from the plants treated at 4°C for 0 (CS0, control), 4 (CS4), 8 (CS8), 12 (CS12), 16 (CS16), 20 (CS20), and 24 (CS24) h by quantitative real-time RT-PCR and normalized to the level of 5.8S rRNA in the sample. Error bars represent the standard deviations of three PCR replicates of a single reverse transcription reaction. The normalized miRNA levels in CS0 were arbitrarily set to 1.

Download figure to PowerPoint

To infer the roles of these 19 Ptc-miRNAs in cold stress responses, their putative mRNA targets were computationally predicted from the current release (version 1.1) of the P. trichocarpa gene models (Tuskan et al., 2006). Using the previously reported method (Jones-Rhoades and Bartel, 2004; Lu et al., 2005), a total of 156 protein-coding genes with penalty scores of ≤2.5 were identified as targets of these cold-responsive Ptc-miRNAs (Table S7). The functions of these predicted genes are diverse, indicating that miRNAs may play important regulatory roles in various aspects of the response of plants to cold stress.


  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

A subset of miRNAs are involved in plant responses to abiotic stress

In this study, we cloned small RNAs from abiotic stressed tissues of P. trichocarpa and identified 68 putative miRNA sequences. Based on sequence homology, these putative miRNA sequences were classified into 27 families, of which nine are novel. This raises the number of known Ptc-miRNA families from 33 to 42. During revision of this manuscript for resubmission, Barakat et al. (2007) reported the isolation of 48 novel Populus miRNA families. Sequence comparison showed that some of the novel miRNA families identified in this study, including Ptc-MIR530, Ptc-MIR1444, Ptc-MIR1446, Ptc-MIR1448, and Ptc-MIR1450, were also reported by Barakat et al. (2007), confirming our identification. Interestingly, based on the results from this study and from previous research (Chiou et al., 2006; Fujii et al., 2005; Jones-Rhoades and Bartel, 2004; Lu et al., 2005, 2007; Sunkar and Zhu, 2004; Sunkar et al., 2006; Zhao et al., 2007; Zhou et al., 2007), we conclude that a subset of miRNAs are involved in the responses of plants to abiotic stresses. In P. trichocarpa plants subjected to cold stress, the expression of at least 16 miRNA families, including two novel families, Ptc-MIR1445 and Ptc-MIR1446 (Figure 5), as well as the 14 families identified by miRNA microarrays (Figure 6), was altered. These miRNAs have predicted functions to cleave the transcripts of plant development-related or stress-responsive genes. It is significant that, in many cases, members of the same miRNA family were differentially regulated in response to cold stress. This is consistent with the results obtained from drought-stressed rice (Zhao et al., 2007) and UV-B-treated Arabidopsis (Zhou et al., 2007). In addition, cold-responsive members of a miRNA family may be further differentiated by their temporal response to cold treatment. Thus, the functions of plant miRNAs can be dissimilar even if they share a high degree of sequence similarity and belong to the same family.

Some of the stress-responsive miRNA families are deeply conserved among various plant species, such as Arabidopsis, rice, and Populus. The others are specific to Populus or particular trees. Consistent with our previous results (Lu et al., 2005, 2007), we found that the deeply conserved miRNAs might target functionally different genes for cleavage in a species-dependent manner. These conserved miRNAs, which display Populus-specific or tree-specific functions, may be the result of adaptation to long-term growth and survival in stressful environments. The identification of a set of stress-responsive miRNAs for P. trichocarpa provides the first line of information necessary for the development of transgenic trees with greater resistance to stress.

Cold-responsive miRNAs may also be regulated in response to biotic stress

Interestingly, several cold-responsive miRNAs are also involved in the response of plants to biotic stress. For example, miR156 and miR160 were significantly repressed in the galled loblolly pine stem infected with the fungus C. quercuum f.sp. fusiforme (Lu et al., 2007). The expression levels of miR156, miR160, miR164, and miR169 were increased significantly in tobacco infected with plant viruses (Bazzini et al., 2007). Infection with the TMV-Cg virus caused the accumulation of seven cold-responsive miRNAs (miR156, miR160, miR164, miR168, miR169, miR390, and miR396) in Arabidopsis (Tagami et al., 2007). In Arabidopsis infected with plant virus TYMV p69, miR156 and miR164 were induced (Chen et al., 2004). Similarly, miR156, miR160 and miR164 were also induced in transgenic Arabidopsis plants expressing the viral silencing suppressor P1/HC-Pro (Kasschau et al., 2003). In another study, transgenic Arabidopsis over-expressing miR393a displayed a higher degree of resistance to the bacterium Pseudomonas syringae (Navarro et al., 2006). Taken together, these results indicate that cross-talk exists between miRNA pathways for biotic and abiotic stress responses. These complex regulatory networks of miRNAs contribute to the ability of plants to survive an ever-changing and often stressful environment.

Cold-responsive miRNAs target functionally diverse genes

A total of 165 genes were predicted to be the targets of the 16 cold-responsive miRNA families. They include 156 genes targeted by the 14 cold-responsive miRNA families that were identified by the microarray analysis (Table S7), four targets of Ptc-MIR1445 and five of Ptc-MIR1446 (Table 3). The functions of these predicted target genes are diverse. Several are associated with signaling pathways, such as homologs of the nine leucine-rich repeat protein kinases targeted by Ptc-miR390a-d (Osakabe et al., 2005) as well as the four Ptc-miR476a targets encoding glutamate receptor proteins that are putative ligand-gated channels involved in signaling of environmental stimuli (Meyerhoff et al., 2005). Ptc-miR393 targets four genes that encode homologs of the Arabidopsis IAA receptors involved in the degradation of Aux/IAA proteins and auxin-regulated transcription (Dharmasiri et al., 2005; Kepinski and Leyser, 2005).

Transcription factor genes, such as NAC domain proteins, auxin response factors (ARFs), squamosa promoter-binding proteins and GRAS transcription factors, known to regulate the development and growth of plants (Gandikota et al., 2007; Guo et al., 2005; Laufs et al., 2004; Mallory et al., 2004, 2005; Wen and Chang, 2002; Wu and Poethig, 2006), are another group of predicted targets of these cold stress-responsive Ptc-miRNAs. MiR168 is known to control the Argonaute 1 (AGO1) gene in Arabidopsis and P. trichocarpa (Kidner and Martienssen, 2004; Lu et al., 2005; Vaucheret et al., 2004). Argonaute 1 is involved in the miRNA regulatory pathway (Baumberger and Baulcombe, 2005) and is required for plant development including stem cell function, organ polarity, and adventitious root formation (Kidner and Martienssen, 2005; Sorin et al., 2005). The increased expression of Ptc-miR168 in cold-stressed plants suggests that AGO1 is also important in cold tolerance.

A total of 36 PPRs were predicted to be the targets of Ptc-miR474b, Ptc-miR475a,b, and Ptc-miR476a (Table S7). Several of these targets have been experimentally confirmed (Lu et al., 2005). The PPRs are a large family in plants with more than 450 members in Arabidopsis and about 760 members in P. trichocarpa (Tuskan et al., 2006). The functions of PPRs are largely unknown, although some of them are known to be involved in the circadian-regulated organelle gene expression of plant cells, RNA processing and editing, chloroplast biogenesis, and fertility restoration (Gothandam et al., 2005; Kocabek et al., 2006; Oguchi et al., 2004; Schmitz-Linneweber et al., 2005; Wang et al., 2006). The fact that Ptc-miR475a,b and Ptc-miR476a were downregulated while Ptc-miR474b was up-regulated suggests the diversity of PPR function in response to cold response. It also indicates that the cold stress response involves altered organelle gene expression, mediated by PPRs.

Experimental procedures

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Isolation of small RNAs

Total RNAs were isolated from stress-treated 1.5-month-old P. trichocarpa (clone Nisqually-1) plantlets grown on agar media (0.5× MS macro- and micronutrients, 1× MS vitamins, 100 mg l−1 myoinositol, 20 g l−1 sucrose) at 25°C under 16-h light/8-h dark per day and from developing xylem tissues of mechanical stressed and normal stems of 1-year-old, greenhouse-grown P. trichocarpa (clone Nisqually-1) trees. The plantlets grown in vitro were untreated or treated with cold (4°C for 24 h), heat (37°C for 24 h), dehydration (drought for 14 h in a covered tissue-culture box), salinity (300 mm NaCl for 14 h), or water (cover the plants with water for 14 h). Whole plantlets with or without treatment were pooled and used for construction of the abiotic stressed small-RNA library (AL). Mechanical stress was induced by bending the tree stems into an arch for 4 days and developing secondary xylem tissues from the upper (TW) and lower (OW) portions for the bent segment were collected directly into liquid nitrogen. Tissues from TW and OW were mixed and used for construction of the mechanically stressed small-RNA library (ML). Small RNAs with sizes about 21 nucleotides were isolated as described by Lu et al. (2005).

Blast analyses against GenBank and prediction of stem–loop structures

The Blast analyses against GenBank for the cloned small RNAs corresponding to known non-coding RNAs and to those associated with retrotransposons or transposons were conducted with the Nucleotide Blast program ( using the default parameters. The cloned small RNAs or their related sequences (one to two mismatches) were aligned with the Populus draft genome assembly (Tuskan et al., 2006;, A. thaliana genome annotation version 5.0 (, TIGR Rice Genome Pseudomolecules and Annotation Release 4 (, and DFCI Pine Gene Index, release 6.0 ( using PatScan (Dsouza et al., 1997). Secondary structures were predicted by the mfold program ( using the default parameters (Zuker, 2003) as described (Lu et al., 2005). Small RNA sequences were folded with flanking sequences in five contexts: (i) 300-bp upstream and 20-bp downstream, (ii) 150-bp upstream and 20-bp downstream, (iii) 150-bp upstream and 150-bp downstream, (iv) 20-bp upstream and 150-bp downstream, and (v) 20-bp upstream and 300-bp downstream. In each case, only the lowest-energy structure was selected for manual inspection and the criteria developed by Jones-Rhoades et al. (2006) were applied. These criteria include no more than seven unpaired nucleotides in the 25 nucleotides centered on the miRNA, of which no more than three are consecutive and no more than two are without a corresponding unpaired nucleotide in the near complementary sequence (miRNA*) in the hairpin structure.

Computational prediction and 5′ RNA ligase-mediated-RACE validation of miRNA targets

The MiRNA targets were predicted as described (Jones-Rhoades and Bartel, 2004; Lu et al., 2005). This method includes a penalty scoring system for mismatched patterns in the miRNA:mRNA duplexes within a 20-base sequence window, with 0 points being assigned to each complementary pair, 0.5 points to each G:U wobble, one point to each non-G:U wobble mismatch, and two points to each bulged nucleotide in either RNA strand. Mapping of the mRNA cleavage sites using modified 5′ RNA ligase-mediated (RLM)-RACE was carried out as described by Lu et al. (2005). The 5′-RACE analysis of eugene3.00102261, eugene3.00060403, and fgenesh4_pg.C_LG_XII000915 was carried out on mRNAs isolated from whole P. trichocarpa plantlets untreated or treated with cold (4°C for 24 h), heat (37°C for 24 h), dehydration (drought for 14 h in a covered tissue culture box), salinity (300 mm NaCl for 14 h), or water (plants covered with water for 14 h). The cleavage sites of eugene3.00110271, gw1.182.27.1, and eugene3.01310091 were mapped using mRNAs isolated from the developing secondary xylem and phloem of P. trichocarpa. The nesting and nested primers used are 5′-GCAACAAAGTCGTTCTTTGATCAAATT-3′ and 5′-GAACCATAACTTCCCATTGTTTTGAGA-3′ (eugene3.00102261), 5′-CCATAAGAGTCGAACTGCCAGGAA-3′ and 5′-GGAAAGCATAATGCTCTACGGATGT-3′ (eugene3.00060403), 5′-GATGATTCGGAATCTGTCAACGCAA-3′ and 5′-CCAGTTATTCGAAGGCGACTAGGT-3′ (fgenesh4_pg.C_LG_XII000915), 5′-GGTGTTGGCCTGGATTGTAGCCAA-3′ and 5′-GGTTTCTTGAATCCAGGCAATCTCT-3′ (eugene3.00110271), 5′-GGCACTGTCGGCTTTGGTTTGTT-3′ and 5′-GGTCTAGCATTGATCCAGGGACTA-3′ (gw1.182.27.1), and 5′-CCCTCCTTCAAACCATCCCACTTT-3′ and 5′-CCACACATCATCTAGTACTAGAAGA-3′ (eugene3.01310091).

Northern blot analysis of miRNA expression

Whole plantlets were treated for 20 h with cold (4°C), heat (37°C), salinity (300 mm NaCl), and dehydration (drought in a covered tissue culture box). Untreated plantlets were used as controls. A mixture of five parts leaf, two parts stem, and three parts root (by fresh weight) from each treatment was used for total RNA isolation. Greenhouse-grown trees were bent for 4 days and tissues were colleted as described (Lu et al., 2005). Northern blot analysis of miRNA was performed as described by Lu et al. (2005). A total of 11 DNA oligonucleotides complementary to the cloned miRNA sequences were used for probes. The probes and the detected Ptc-miRNAs are 5′-AGTGATATTGATTCGGCTCG-3′/Ptc-miR171l-n, 5′-AAGGTGCAGGTGCAAATGCA-3′/Ptc-miR530a, 5′-AATGGGAGGAGTAGGCAAGA-3′/Ptc-miR482.2, 5′-TTTTCGTTGATGGTCATCTAA-3′/Ptc-miR827, 5′-GAACATTGACCGAATGTGGA-3′/Ptc-miR1444a, 5′-TTTTTCTAGACTACAACGGA-3′/Ptc-miR1445, 5′-TTGAGGGAGAGAGTTCAGAA-3′/Ptc-miR1446a-e, 5′-AATCAAGGCACTGCAATTCTG-3′/Ptc-miR1447, 5′-GTATGGGAGGCGTTGGAAAG-3′/Ptc-miR1448, 5′-GAGTTATCTTACGTGCACCTCA-3′/Ptc-miR1449, and 5′-TTAACCTGACCGAGCCATTGAA-3′/Ptc-miR1450, respectively. The 5.8S rRNA bands were probed with oligonucleotide (5′-ACGGGATTCTGCAATTCACAC-3′) and served as loading controls. Hybridization signals were imaged and quantified using a Molecular Imager® FX System (Bio-Rad;

Cold stress time course experiment with Populus trichocarpa

The 1.5-month-old, in vitro propagated P. trichocarpa plants were transferred into soil and grown in a greenhouse for 2 months. These plants were subjected to cold stress time course treatments for 0, 4, 8, 12, 16, 20, and 24 h. Cold stress responses were induced at 4°C in the dark. Each treatment was carried out with three 2-month-old tissue culture derived plants and replicated three times. Leaves, stems, and roots were harvested separately at the end of the treatment and stored in liquid nitrogen before extraction of small RNA. Equal amount of leaf, stem, and root tissues harvested from each time point were combined for RNA extraction to minimize the tissue-specific effect. Total RNA was isolated using the protocol of Chang et al. (1993). Integrities of the extracted total RNAs were confirmed by BioAnalyzer (Agilent;

Poplar miRNA microarray

Populus trichocarpa miRNA microarrays were designed based on the poplar miRNA sequences obtained from miRBase (Release 9.2, A 35-nt probe was designed for each of the putative miRNA genes with the mature miRNA complementary sequence on the probes started at the first to fourth base from the 5′ end of the probes, except for six probes that have mature miRNA complementary sequences started at the fifth to seventh base (Table S5). A total of 168 probes were designed for 90 mature miRNAs originated from 33 families. A set of probes based on GFP and GUS genes that do not have more than 7 nt of genome matches were designed as negative controls. The miRNA mircrorrays were manufactured by CombiMatrix ( on the supplier’s 4 × 2000 array format, with each probe replicated at least twice on the array. Sequence information for the probes is listed in Table S5.

miRNA hybridization

Small RNAs from 100 μg of total RNA were isolated using a YM-100 Microcon column (Millipore;, concentrated to 5 μl and dephosphorylated with 2 units of shrimp alkaline phosphatase (Roche Diagnostics; in 50 mm 2-amino-2-(hydroxymethyl)-1,3-propanediol (TRIS) pH 8.5 and 5 mm MgCl2 at 37°C for 1 h. Reactions were terminated by incubating at 65°C for 15 min. Small RNA labeling and hybridization were based on the protocol of Thomson et al. (2004). Dephosphorylated small RNAs were then labeled by incubation with 500 ng 5′-phosphate-cytidyl-uridyl-Cy5-3′ (Dharmacon; and 20 units of T4 RNA ligase in 0.1 mm ATP, 50 mm HEPES pH 7.8, 3.5 mm DTT, 20 mm MgCl2, 10 mg ml−1 BSA, and 10% DMSO at 4°C for 2 h, and cleaned up by ethanol precipitation and incubated overnight at −80°C. The precipitated, labeled small RNAs were resuspended in hybridization buffer containing 12% formamide, 6× SSPE (1×SSPE buffer: 0.15 M Sodium chloride, 0.01 M Sodium hydrogen phosphate, 0.001 M EDTA, pH 7.4) and 0.5% SDS, and applied to microarray chips. Microarray chips were pre-hybridized in buffer containing 400 mm Na2HPO4 pH 7.0, 0.8% BSA, 0.5% SDS, and 12% formamide at 65°C for 10 min. Labeled targets were then applied to microarray chips and hybridized at 37°C for 3 h. Microarray washing was carried out at 23°C once with 2× SSC and 0.25% SDS for 3 min, three times with 0.8× SSC for 3 min each, and twice with 0.4× SSC for 3 min each. Microarrays were scanned with a resolution of 5 μm pixel−1 resolution. The associated array image and signal intensities were analyzed and extracted using the Microarray Imager software provided by CombiMatrix.

Microarray data analysis

Probe signal intensities of the microarray analyzed with the Microarray Imager application by CombiMatrix were imported into SAS for statistical analysis. Data were first filtered for low signal intensities that fall below detection limits and saturated signals that are high beyond the detection limits. Low signal intensities were filtered for spots with the coefficient of variation of the signal intensity larger than 100%. Saturated microarray spots were filtered for those with medium signal intensities larger than 50 000. The remaining microarray data were normalized and analyzed using the mixed model approach by Wolfinger et al. (2001). Log-transformed signal intensities of microarrays were normalized by fitting the model yij = μij + Aij + rij where variable y represents the log-transformed signal intensity and variable A represents the random effect of each array. Residuals of the normalized signal intensity were then analyzed by fitting the model rij = μ + Ti + Rj + Ti(Rj) + rij for each mature miRNA, where i = 1…7 represents seven time points of the cold stress treatment, j = 1…3 represents three biological replications of each cold stress treatment, and rij is the residual of the normalization model of each miRNA at time point i, resp. j. The variable T represents the fixed effect of the cold stress treatment at various time points, and the variable R represents the random effect of the biological replication in the experiment. For each mature miRNA, the P value of the effect T between each time point and the 0-h control was also calculated. The heat maps that represent the ratio between each time point and control were generated using software Mev4 (Saeed et al., 2003). Microarray data have been deposited in the Gene Expression Omnibus database (GEO; under the accession number GSE10582. A P value of <0.01 suggests that the difference of log signal intensities of the probes between cold stress treatment and the control is highly significant, and the miRNA is thus designated as cold stress responsive.

Real-time PCR validation

Real-time PCR was carried out as previously described (Lu et al., 2005; Shi and Chiang, 2005) with the following modification. Poly(A) tails were added to 3′ ends of the total RNA using a Poly(A) Tailing Kit (Ambion; and then reverse-transcribed into single-strand cDNA with the Taqman reverse transcription reagents (Applied Biosystems; and the oligo(dT) 3′-RACE adaptor [5′-GCGAGCACAGAATTAATACGACTCACTATAGG(T)12VN-3′; Ambion]. Real-time PCR was carried out using a 3′-RACE outer primer (5′-GCGAGCACAGAATTAATACGAC-3′; Ambion) as the reverse primer and mature miRNA sequences as the forward primer. These mature RNA sequences are Ptc-miR156g-j (5′-TTGACAGAAGATAGAGAGCAC-3′), Ptc-miR168ab (5′-TCGCTTGGTGCAGGTCGGGAA-3′), Ptc-miR475ab (5′-TTACAGTGCCCATTGATTAAG-3′), Ptc-miR476a (5′-TAGTAATCCTTCTTTGCAAAG-3′), and Ptc-miR477ab (5′-ATCTCCCTCAGAGGCTTCCAA-3′), respectively. The forward primer used for the endogenous reference P. trichocarpa 5.8S rRNA is Ptc5.8-1 (5′-GTCTGCCTGGGTGTCACGCAA-3′).


  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

We would like to thank Judy Jakobek for her suggestions on the manuscript. This work was supported by the North Carolina State University Forest Biotechnology Industrial Research Consortium (grant no. 525768 to SL, Y-HS, and VLC, and no. 158802 to Y-HS, SL, and VLC).


  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Supporting Information

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Table S1. Putative endogenous siRNAs from Populus trichocarpa.

Table S2. Newly identified miRNA genes in Populus, rice and pine.

Table S3. Identification of miRNA genes in Arabidopsis, rice, and loblolly pine.

Table S4. Potential targets for the previously reported miRNA families in Populus trichocarpa.

Table S5. Poplar miRNA microarray probe information.

Table S6. MiRNA microarray data from Populus trichocarpa cold stress treatments.

Table S7. Potential targets for the cold responsive miRNAs in Populus trichocarpa.

Please note: Blackwell Publishing are not responsible for the content or functionality of any supplementary materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.

TPJ_3497_sm_TableS1.xls30KSupporting info item
TPJ_3497_sm_TableS2.xls23KSupporting info item
TPJ_3497_sm_TableS3.xls17KSupporting info item
TPJ_3497_sm_TableS4.xls32KSupporting info item
TPJ_3497_sm_TableS5.xls45KSupporting info item
TPJ_3497_sm_TableS6.xls34KSupporting info item
TPJ_3497_sm_TableS7.xls48KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.