Whole genome microarray analysis of C. elegans rrf-3 and eri-1 mutants
Suvi Asikainena,b, Markus Storvikb, Merja Laksoa, Garry Wonga,b,*
aDepartment of Neurobiology, A.I. Virtanen Institute, Kuopio University, Kuopio 70211, Finland
bDepartment of Biosciences, Kuopio University, Kuopio 70211, Finland
Received 10 August 2007; accepted 19 September 2007
Available online 29 September 2007
Edited by Shou-Wei Ding
on L4 stage Caenorhabditis elegans rrf-3(pk1426) and eri-
1(mg366) mutant strains to study the effects caused by loss of
their encoded proteins, which are required for the accumulation
of endogenous secondary siRNAs. Mutant rrf-3 and eri-1 strains
exhibited 72 transcripts that were co-over-expressed and 4 tran-
scripts co-under-expressed (>2-fold, P < 0.05) compared to N2
wild type strain. Ontology analysis indicated these transcripts
were over represented for protein phosphorylation and sperm
function genes. These results provide additional support for the
hypothesis that RRF-3 and ERI-1 act together in the endo-si-
RNA pathway, and furthermore suggests their involvement in
additional biological processes.
? ? 2007 Federation of European Biochemical Societies. Pub-
lished by Elsevier B.V. All rights reserved.
We performed genome wide gene expression analysis
Keywords: RNA interference; RRF-3; ERI-1; Endogenous;
RNAi; Gene expression; C. elegans
Gene silencing through RNA interference (RNAi) pathways
are mediated by small RNAs of ?21–24 nt in length found in
virtually all eukaryotes [1,2]. This evolutionary conserved
mechanism has likely evolved to protect organisms against
invading exogenous nucleic acids but its endogenous roles
are only beginning to be better understood [3–6]. Studies on
Caenorhabditis elegans have revealed a wide variety of micro-
RNAs (miRNAs) and endogenous small interfering RNAs
(endo-siRNAs) with an apparent regulatory role for animal
health and development [2,6–8].
In C. elegans, miRNAs enter their own pathway, while si-
RNAs enter specific pathways depending on the origin and
molecular structure of the initial trigger molecule and specific
downstream processing enzymes. The pathways for exogenous
siRNA (exo-siRNA) and endogenous siRNA (endo-siRNA)
uses in part, separate enzymes while sharing others including
dicer (DCR-1) [7,8]. siRNA amplification has been shown to
use alternative RNA-dependent RNA polymerases (RdRPs)
for synthesis of secondary siRNAs to produce RNAi effects
[7–9]. According to current models, the endo-siRNAs pathway
uses the RdRP homolog RRF-3, while the exo-siRNA path-
way uses RRF-1 [7,8]. Germ cells use a third RdRP homolog,
EGO-1, for exo-siRNAs [7,8,10].
RRF-3 associates, in a large complex, physically with ERI-1,
a conserved protein with DEDDh-like exonuclease and SAP/
SAF-box nucleic acid binding domains in the endo-siRNA
pathway [8,11]. Both of these proteins are required for accu-
mulation of at least some endo-siRNAs [7,8]. Both mutants
are sterile in 25 ?C, exhibit high incidence of males (Him-
phenotype) due to X-chromosome non-disjunction, and are
hypersensitive to exogenous RNAi (enhanced RNAi, ERI-
phenotype) [11,12]. The enhanced sensitivity to exogenous
RNAi has been exploited in a number of whole genome RNAi
screens and has thus led to widespread use of both rrf-3 (RNA-
dependent RNA polymerase-3) and eri-1 (enhanced RNAi-1)
mutants in gene silencing experiments . It has been pro-
posed that the ERI-phenotype is caused by release of limiting
components from the endogenous to the exogenous RNAi
pathway in these mutants [7,8].
Genome wide gene expression analysis was performed on C.
elegans rrf-3 and eri-1 mutant strains to study transcriptional
effects caused by loss of their encoded proteins in the endoge-
nous RNAi pathway and to provide further insight into their
functions. Mutant strains exhibited highly similar expression
patterns and co-regulation of over-expressed transcripts was
observed for genes encoding proteins associated with protein
phosphorylation and sperm function. These results provide
additional support for the model where RRF-3 and ERI-1
act in the same pathway, and suggests that they affect addi-
tional biological functions.
2. Materials and methods
2.1. Sample preparations
rrf-3(pk1426) mutant worms (NL2099), eri-1(mg366) mutant worms
(GR1373), and wild type worms (N2) were grown on culturing plates
containing NGM agar media with OP50 Escherichia coli (E. coli) as a
food source . Animals were synchronized by hatching purified eggs
using 1% hypochlorite solution with 250 mM KOH. Worms were
grown at 20 ?C and harvested at the fourth larval (L4) stage. RNA iso-
lation was made using Ribopure Total RNA isolation-kit (Ambion,
Foster City, CA). Animals were grown separately and RNA isolated
separately for each biological replicate (three for microarrays, three
for quantitative real-time PCR). RNA samples were isolated separately
for microarray, qRT-PCR, and Northern blot experiments. Worm
stains were genotyped by PCR to confirm their identities (data not
2.2. Microarray analysis
Purified total RNA was labeled for chip experiments using One cycle
cDNA synthesis followed by cRNA synthesis according to the manu-
facturer’s protocol (Affymetrix, Palo Alto, CA). cRNA from each
*Corresponding author. Address: Department of Biosciences, Kuopio
University, P.O. Box 1627, Kuopio 70211, Finland. Fax: +358 17
E-mail address: firstname.lastname@example.org (G. Wong).
0014-5793/$32.00 ? 2007 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved.
FEBS Letters 581 (2007) 5050–5054
mutant was hybridized to Gene Chip C. elegans whole genome arrays
containing 22500 transcripts (Affymetrix) as three biological replicates
(3 chips each for N2, rrf-3, eri-1). Microarray hybridization and scan-
ning was performed at the Biomedicum Biochip Core Facility (Hel-
sinki, Finland). GeneSpring-software (Agilent, Palo Alto, CA) was
used to carry out data-analysis. For the probe intensity values gener-
ated by the Affymetrix scanner, Robust Multichip Average (RMA) –
algorithm was used for normalization and statistical processing. Data
were then filtered to remove genes with low expression values (<10). T-
test was performed to sequentially filter out genes with unreliable sig-
nal level between replicates (P < 0.05) and then to collect genes with
significant signal level between samples (P < 0.05). Finally, genes were
filtered for fold change (>2-fold up or <0.5-fold down).
2.3. Functional analysis
Annotations for the regulated genes were obtained from Affymetrix
and supplemented with additional annotations from David 2.1 . In
the case that no known annotations existed, manual annotations were
extracted from Wormbase. Lists of up- and down-regulated genes were
uploaded to David 2.1. Calculations of overrepresented GO biological
process annotations using functional annotation clustering was per-
formed using the GOchart function. Significance cutoff was set at
P < 0.05.
2.4. Quantitative real-time PCR (qRT-PCR) and Northern blot
qRT-PCR was performed using Power SYBR? Green PCR kit (Ap-
plied Biosystems, Foster City, CA) according to the manufacturers
instructions. Amplifications were performed on ABI Prism 7700 (Ap-
plied Biosystems). Act-1 was used as a control to normalize transcript
50tcggtatgggacagaaggac, right 50catcccagttggtgacgata); ssp-16 (left,
50ttaacggaggtgccgataag, right, 50tttgtcctcctttggtgctc); T16G12.7 (left,
50tatggagcaaagggtggaac, right, 50ccagagcagtgtacggcata); C35E7.9 (left,
50gatggattctcgctggatgt; right, 50attctccacaggcggtttc); F25B3.4, (left,
50ctgacacgtcttattccacctg, right, 50gtgagttgacctcaatgagca); clec-69, (left,
50tggtggtgacagttcagagc; right, 50agctggaagattggttgtgc). For the north-
ern blot, an antisense probe for ssp-16 transcript (ssp-16: 50TGGTGC-
GAAATGAACGACAAGTTTGTCCTCCTTTG) and (act-1: 50GG-
with a-[33P]-dATP using terminal deoxytransferase, and hybridized to
Northern blots containing total RNA (10 lg per lane) isolated from
each strain. Northern blots were hybridized overnight at 42 ?C in
5· SSPE,50%formamide,5· Denhardt’ssolution,10%dextransulfate,
1% SDS, and washed thrice at room temperature for 2 min in 4· SSC,
0.1% SDS, and thrice at room temperature for 1–2 min in 0.2· SSC,
0.1% SDS until there was no longer background signal, dried, and
exposed on a phosphor screen overnight.
3.1. Microarray and GO analysis
All microarray data used in this experiment is available in
MIAME format at GEO database (www.ncbi.nih.gov/geo)
(Accession: GSE8659). In the rrf-3 mutant, we observed a total
of 112 over- and 17 under-expressed genes, while in the eri-1
mutant, 199 over- and 66 under-expressed genes were found
(Fig. 1). An overlap of 72 over- and 4 under-expressed genes
were identified. The list of overlapping over-expressed genes
is shown in Table 1. Twenty-five co-over-expressed genes
matched previously cloned siRNAs. Complete gene lists
are shown in supplementary material Table S1 (available
at http://www.uku.fi/aivi/neuro/genomics/rrf-3.shtml). Gene
ontology analysis was performed with DAVID 2.1 using
Affymetrix probe set identifiers to calculate statistically
enriched GO (gene ontology) molecular function annotations
for the overlapping over-expressed genes (Table 2). Complete
GO annotations (molecular function, biological process, and
cellular component) for all regulated gene lists are available
in Table S2 (available at http://www.uku.fi/aivi/neuro/genom-
ics/rrf-3.shtml). From the list of overlapping genes, we ob-
served a statistically significant enrichment of both protein
tyrosine (P < 10?4) and serine/threonine kinases (P < 10?11).
From the list of down regulated genes, significant enrichment
(P < 0.01) of oviposition and reproductive behavior were ob-
served in rrf-3 mutants and sugar/carbohydrate binding
(P < 0.05) in eri-1 mutants.
3.2. Confirmation with quantitative RT-PCR
Five genes were selected for confirmation by quantitative
RT-PCR. These genes were chosen based on magnitude of
change and common over-expression in rrf-3 and eri-1 mu-
tants, and representing sperm specific proteins, phosphatases,
and a lectin type protein. Act-1 was used as a control. The
microarray changes correlated well with the qRT-PCR,
although the magnitudes of the changes were not always iden-
tical (Table 3). We also performed Northern blot analysis on
ssp-16 using act-1 as a loading control and obtained results
similar to those from qRT-PCR (Fig. 2).
Recent studies indicate that RRF-3, an RNA-dependent
RNA polymerase, and ERI-1, a conserved exonuclease are re-
quired for accumulation of endogenous siRNAs. Both proteins
are physically linked in a complex with DCR-1 [7,8,17], and
mutants share identical enhanced RNAi (ERI) phenotypes.
To study gene expression effects caused by loss of these pro-
Fig. 1. Venn-diagram of >2-fold over-expressed (A) and under-
expressed genes (B) in rrf-3(pk1426) and eri-1(mg366) mutants
compared to N2.
S. Asikainen et al. / FEBS Letters 581 (2007) 5050–5054
List of 72 commonly over-expressed genes in rrf-3(pk1426) and eri-1(mg366) mutants vs. N2
Wormbase geneFold change rrf-3 vs. N2 Fold change eri-1 vs. N2 Annotation
Dynein light chain motif
ssp-16, sperm-specific class P protein 16
Protein phosphatase 1
Serine/threonine protein phosphatase
Phosphoglyserate mutase motif
Proton or iron transport? Sideroflexin?
ncr-2, homolog of human NPC1, which when mutated
leads to Niemann-Pick disease
H5 and major sperm protein domain protein
kin-24, protein KINase family member
Tyrosine-protein kinase (FES/FPS subfamily)
Protein-tyrosine kinase, EGF receptor activity
pph-2, Protein PHosphatase family member
Probable 3-hydroxyacyl-CoA dehydrogenase
spt-2, novel protein with high similarity to C. elegans F37A8.1
Homolog of the human gene CBS, which when mutated
leads to homocystinuria
Casein kinase, serine/threonine kinase
PDZ domain (also known as DHR or GLGF)
tag-347, SAND domain
S. Asikainen et al. / FEBS Letters 581 (2007) 5050–5054
teins, we performed microarray experiments on rrf-3 and eri-1
mutant C. elegans strains.
We observed relatively uniform gene expression patterns
with a large amount of shared over-expressed genes in L4 stage
rrf-3 and eri-1 mutants, suggesting a common biological func-
tion. Tyrosine and threonine/serine protein phosphatases and
kinases were shown to be over-represented in genes at least
2-fold over-expressed in the mutants. This suggests a novel role
for posttranslational protein modifications. The targets of
these phosphatases and kinases have yet to be identified and
could be within the endo-siRNA pathway or elsewhere. Sec-
ondary siRNAs are known to be triphosphorylated at the 50
ribose moiety distinguishing them from DCR-1 processed pri-
mary siRNAs with 50monophosphates, indicating the impor-
tance of phosphorylation status in siRNA function [18,19].
Both strains share at least three over-expressed genes that
are sperm specific and is consistent with the sperm defective
high temperature sterility of these mutants [8,12]. Among the
commonly over-expressed transcripts were three genes with
Major Sperm Protein (MSP)-domain (ssp-16, C35E7.9 and
ZC168.6). While the study here was performed on L4 worms
raised at 20 ?C, some germ or sperm cell perturbation may oc-
Table 1 (continued)
Wormbase gene Fold change rrf-3 vs. N2Fold change eri-1 vs. N2Annotation
Protein and zinc ion binding
Values for fold change are the average from 3 independent biological replicates. Annotations were provided by Affymetrix when available and were
supplemented manually by Wormbase and DAVID 2.1 gene ontologies.*, indicates match with cloned siRNA [7,20].
Over-represented Gene Ontology (GO) molecular function terms associated with co-over-expressed gene list from rrf-3 and eri-1 mutant animals
GO term (molecular function)Count%
Protein kinase activity
Phosphotransferase activity, alcohol group as acceptor
Transferase activity, transferring phosphorus-containing groups
Adenyl nucleotide binding
Phosphoprotein phosphatase activity
Purine nucleotide binding
Phosphoric monoester hydrolase activity
Phosphoric ester hydrolase activity
Protein tyrosine phosphatase activity
Protein-tyrosine kinase activity
Protein serine/threonine kinase activity
Hydrolase activity, acting on ester bonds
GO term annotations were assigned and statistically evaluated using David 2.1. The count indicates the number of observations from the input of the
72 genes. Terms are listed in decreasing order of significance (P-value).
Comparison of microarray and quantitative real-time PCR (qRT-PCR) results
Gene Annotation Microarray rrf-3(pk1426) qRT-PCRMicroarray eri-1(mg366) qRT-PCR
Sperm-specific class P protein 16 (ssp-16)
Protein phosphatase 1
Sperm specific domain protein
ser/thr protein phosphatase
c-type lectin family member (clec-69)
12.5 ± 4.5
8.9 ± 1.6
5.7 ± 1.7
4.9 ± 1.3
0.7 ± 0.9
10.3 ± 5.5
5.3 ± 1.5
3.6 ± 0.9
3.1 ± 1.7
0.19 ± 0.07
Values shown are normalized ratios ±S.D. from 3 biological replicates comparing N2 and rrf-3(pk1426) and eri-1(mg366) strains. Independent
biological replicates were used for microarray and qRT-PCR experiments.
Fig. 2. Northern blot was performed to evaluate expression levels of
the ssp-16 gene. Total RNA (10 lg) from the indicated wild type (N2)
or mutant strains (rrf-3(pk1426), eri-1(mg366)) was isolated, and
probes were prepared, hybridized, and exposed as described in Section
2. Act-1 was used as a control. Estimated sizes for the bands were
0.5 kb (ssp-16) and 2.0 kb (act-1).
S. Asikainen et al. / FEBS Letters 581 (2007) 5050–5054
cur despite their superficially wild type appearance. This can be Download full-text
under consideration when using these strains for RNAi screens
Loss of key endo-siRNA pathway proteins causes up-regula-
tion of target mRNAs, likely due to the loss of complementary
siRNA production [7,8,20]. Our results appear to support this
notion, since at least 25 genes that were increased in rrf-3 and
eri-1 vs. N2 have been cloned as endo-siRNAs [7,20]. This
number is likely to be an underestimation and could increase
when more endo-siRNAs are cloned and annotated. Alterna-
tively, over-expression could occur through indirect non-siR-
NA mediated mechanisms.
While this manuscript was in preparation Lee et al. 
reported the expression profiling of mixed stage population
rrf-3 and eri-1 mutants along with several additional RNAi
pathway related proteins. Comparison of 2-fold elevated tran-
scripts revealed only a few matches between their and our
study (rrf-3, 6/112; eri-1, 1/199; data not shown). The small
number of common over-expressed genes from the previous
and current study suggests large differences in gene expression
in these strains that depend upon the developmental stage.
Our results suggest that the majority of gene expression lev-
els are not significantly changed in rrf-3 and eri-1 mutant
strains. Future RNAi screening studies can likely be performed
with these strains without major concern for background
effects. Co-regulation of a large number of genes provides
supporting evidence that RRF-3 and ERI-1 function together,
likely in the endo-siRNA pathway. Finally, our data suggests
their novel involvement in additional biological processes
including protein phosphorylation.
Acknowledgements: The authors thank Petri Pehkonen, Jussi Paana-
nen, and Jarmo Tuimala for help with analysis programs; Suvi Vartiai-
nen, Kaja Reisner and Vuokko Aarnio for help with assays; Outi
Monni and Jaana Saarela from the Biomedicum Biochip Center for
help with microarray experiments, and Jonathan Gent for useful dis-
cussions. Some nematode strains used in this work were provided by
the Caenorhabditis Genetics Center, which is funded by the NIH Na-
tional Center for Research Resources (NCRR). Support was provided
by the Finnish Graduate School of Neuroscience, Saastamoinen Foun-
dation (G.W.), and Sigrid Juselius Foundation (S.A.).
 Fire, A., Xu, S., Montgomery, M.K., Kostas, S.A., Driver, S.E.
and Mello, C.C. (1998) Potent and specific genetic interference by
double-stranded RNA in Caenorhabditis elegans. Nature 391,
 Ambros, V., Lee, R.C., Lavanway, A., Williams, P.T. and Jewell,
D. (2003) MicroRNAs and other tiny endogenous RNAs in C.
elegans. Curr. Biol. 13, 807–818.
 Lu, R., Maduro, M., Li, F., Li, H.W., Broitman-Maduro, G., Li,
W.X. and Ding, S.W. (2005) Animal virus replication and RNAi-
mediated antiviral silencing in Caenorhabditis elegans. Nature 436,
 Wilkins, C., Dishongh, R., Moore, S.C., Whitt, M.A., Chow, M.
and Khaled, M. (2005) RNA interference is an antiviral defence
mechanism in Caenorhabditis elegans. Nature 436, 1044–1047.
 Schott, D.H., Cureton, D.K., Whelan, S.P. and Hunter, C.P.
(2005) An antiviral role for the RNA interference machinery in
Caenorhabditis elegans. PNAS 102, 18420–18424.
 Robert, V.J., Sijen, T., van Wolfswinkel, J. and Plasterk, R.H.
(2005) Chromatin and RNAi factors protect the C. elegans
germline against repetitive sequences. Genes Dev. 19, 782–787.
 Lee, R.C., Hammell, C.M. and Ambros, V. (2006) Interacting
endogenous and exogenous RNAi pathways in Caenorhabditis
elegans. RNA 12, 589–597.
 Duchaine, T.F., Wohlschlegel, J.A., Kennedy, S., Bei, Y., Conte
Jr., D., Pang, K., Brownell, D.R., Harding, S., Mitani, S.,
Ruvkun, G., Yates 3rd, J.R. and Mello, C.C. (2006) Functional
proteomics reveals the biochemical niche of C. elegans DCR-1 in
multiple small-RNA-mediated pathways. Cell 124, 343–354.
 Sijen, T., Fleenor, J., Simmer, F., Thijssen, K.L., Parrish, S.,
Timmons, L., Plasterk, R.H. and Fire, A. (2001) On the role of
RNA amplification in dsRNA-triggered gene silencing. Cell 107,
 Maine, E.M., Hauth, J., Ratliff, T., Vought, V.E., She, X. and
Kelly, W.G. (2005) EGO-1, a putative RNA-dependent RNA
polymerase, is required for heterochromatin assembly on un-
paired dna during C. elegans meiosis. Curr. Biol. 15, 1972–1978.
 Kennedy, S., Wang, D. and Ruvkun, G. (2004) A conserved
siRNA-degrading RNase negatively regulates RNA interference
in C. elegans. Nature 427, 645–649.
 Simmer, F., Tijsterman, M., Parrish, S., Koushika, S.P., Nonet,
M.L., Fire, A., Ahringer, J. and Plasterk, RH. (2002) Loss of the
putative RNA-directed RNA polymerase RRF-3 makes C.
elegans hypersensitive to RNAi. Curr. Biol. 12, 1317–1319.
 Simmer, F., Moorman, C., van der Linden, A.M., Kuijk, E., van
den Berghe, P.V., Kamath, R.S., Fraser, A.G., Ahringer, J. and
Plasterk, R.H. (2003) Genome-wide RNAi of C. elegans using the
hypersensitive rrf-3 strain reveals novel gene functions. PLoS Biol.
 Brenner, S. (1974) The genetics of Caenorhabditis elegans.
Genetics 77, 71–94.
 Glynn Jr., D., Sherman, B.T., Hosack, D.A., Yang, J., Baseler,
W.B., Lane, H.C. and Lempicki, R.A. (2003) DAVID: database
for annotation, visualization, and integrated discovery. Genome
Biol. 4, 3.
 Rozen, S. and Skaletsky, H. (2000) Primer3 on the WWW for
general users and for biologist programmers in: Bioinformatics
Methods and Protocols: Methods in Molecular Biology (Krawetz,
S. and Misener, S., Eds.), pp. 365–386, Humana Press, Totowa,
 Yigit, E., Batista, P.J., Bei, Y., Pang, K.M., Chen, C.C., Tolia,
N.H., Joshua-Tor, L., Mitani, S., Simard, M.J. and Mello, C.C.
(2006) Analysis of the C. elegans Argonaute family reveals that
distinct Argonautes act sequentially during RNAi. Cell 127, 747–
 Pak, J. and Fire, A. (2007) Distinct populations of primary and
secondary effectors during RNAi in C. elegans. Science 315, 241–
 Sijen, T., Steiner, F.A., Thijssen, K.L. and Plasterk, R.H. (2007)
Secondary siRNAs result from unprimed RNA synthesis and
form a distinct class. Science 315, 244–247.
 Ruby, J., Jan, C., Player, C., Axtell, M., Lee, W., Nusbaum, C.,
Ge, H. and Bartel, D. (2006) Large-scale sequencing reveals 21U-
RNAs and additional MicroRNAs and endogenous siRNAs in C.
elegans. Cell 127, 1193–1207.
S. Asikainen et al. / FEBS Letters 581 (2007) 5050–5054