Environmental Toxicology and Chemistry, Vol. 34, No. 6, pp. 1362–1368, 2015 # 2015 SETAC Printed in the USA

TRANSCRIPTIONAL PROFILING OF THE SOIL INVERTEBRATE FOLSOMIA CANDIDA IN PENTACHLOROPHENOL-CONTAMINATED SOIL MIN QIAO,y GUANG-PENG WANG,y CAI ZHANG,yz DICK ROELOFS,x NICO M. VAN STRAALEN,x and YONG-GUAN ZHU*k yState Key Lab of Urban and Regional Ecology, Research Center for Eco-Environmental Sciences, Chinese Academy of Sciences, Beijing, China zCollege of Resources and Environment, Shandong Agricultural University, Tai’an, China xDepartment of Animal Ecology, VU University, Amsterdam, The Netherlands kKey Laboratory of Urban Environment and Health, Institute of Urban Environment, Chinese Academy of Sciences, Xiamen, China (Submitted 4 March 2014; Returned for Revision 24 April 2014; Accepted 4 February 2015) Abstract: Pentachlorophenol (PCP), a widely used pesticide, is considered to be an endocrine disruptor. The molecular effects of chemicals with endocrine-disrupting potential on soil invertebrates are largely unknown. In the present study, the authors explored the transcriptional expression changes of collembola (Folsomia candida) in response to PCP contamination. A total of 92 genes were significantly differentially expressed at all exposure times, and the majority of them were found to be downregulated. In addition to the transcripts encoding cytochrome P450s and transferase enzymes, chitin-binding protein was also identified in the list of common differentially expressed genes. Analyses of gene ontology annotation and enrichment revealed that cell cycle-related transcripts were significantly induced by PCP, indicating that PCP can stimulate cell proliferation in springtail, as has been reported in human breast cancer cells. Enrichment of functional terms related to steroid receptors was observed, particularly in 20 significant differentially expressed genes involved in chitin metabolism in response to PCP exposure. Combined with confirmation by quantitative polymerase chain reaction, the results indicate that the adverse effects on reproduction of springtails after exposure to PCP can be attributed to a chemical-induced delay in the molting cycle and that molting-associated genes may serve as possible biomarkers for assessing toxicological effects. Environ Toxicol Chem 2015;34:1362–1368. # 2015 SETAC Keywords: Pentachlorophenol

Folsomia candida

Endocrine disruptor

Molting

Soil invertebrates play an essential role in maintaining soil function, and they are widely used for the assessment of soil contamination. Collembola (springtails) are among the most abundant and widespread soil invertebrates and have been recommended as reliable biological indicators of soil quality [11,12]. Folsomia candida has been used as a model organism in ecotoxicity, and internationally accepted guidelines have been available since 1999 [13,14]. It is blind and reproduces parthenogenetically. Numerous studies have assessed the toxic effects of pure chemicals or contaminated soils using the F. candida reproduction assay, because it is very sensitive and easy to handle [15]. With the advent of genomic techniques, the springtail F. candida also is becoming a genomic model organism for soil toxicology. The expression profiles of thousands of genes for F. candida exposed to environmental toxicants can be measured in parallel with microarray technology based on the available sequenced expressed sequence tag database [16]. It allows fast and sensitive detection of pollution and clarifies molecular mechanisms behind the toxicology [17]. There has been some exploration of transcriptional effects on F. candida in response to the stress of heavy metals [18,19], polycyclic aromatic hydrocarbons (phenanthrene) [20], and environmental factors [21]. However, the molecular effects of herbicides on collembola, especially those with endocrine-disrupting potential, are largely unknown. Compared with earthworms, collembola are more sensitive to herbicides [22]. Thus, the objectives of the present study were to assess the effects of PCP by transcriptional profiling of F. candida and to discriminate among the specific responses associated with PCP, to indicate the potential mechanisms of action. The results provide insights for the development of biomarkers, which can be used for screening

INTRODUCTION

Pentachlorophenol (PCP) is one of the most frequently detected pesticides in the environment. It is widely used as an herbicide, wood preservative, defoliant, germicide, and fungicide [1]. Although banned in many countries, it is still used in some restricted applications. In China, approximately 104 tonnes of PCP were produced in 1997, which was approximately 20% of the global production. From the 1960s until it was banned in 2000, PCP was used as a molluskacide to kill oncomelania [2,3]. It is persistent in soil or water because of its stable aromatic ring structure and high chlorine content. Many studies have indicated that PCP can disrupt the normal function of the endocrine system [4,5]. Animal studies have shown significant alternations in thyroid hormone levels in several species [6–8]. There is also evidence that pure PCP can affect immune function in rats, and long-term exposure to PCP can decrease fertility [9]. It has been shown to inhibit zebrafish ovulation in vitro (at 0.6 mM) [10], to reduce the number of eggs laid and their subsequent hatching rates, and to induce the formation of testis-ova in Japanese medaka (50–100 mg/L) [4]. At environmentally relevant concentrations, the endocrinedisrupting potential of PCP was found in a recombinant yeast assay, and it can disrupt the steroidogenic pathway in cultured Xenopus oocytes [5]. Although its impacts on vertebrates and in vitro evidence have been widely reported, little is known about the effects of PCP on soil organisms, especially invertebrates. All Supplemental Data may be found in the online version of this article. * Address correspondence to [email protected]. Published online 20 February 2015 in Wiley Online Library (wileyonlinelibrary.com). DOI: 10.1002/etc.2930 1362

Transcriptional profiling of springtail in PCP-polluted soil

new and existing chemicals in terms of their endocrinedisrupting potential. MATERIALS AND METHODS

Test organisms and PCP exposure experiments

The collembolan F. candida, whose expressed sequence tags are available, was used in the present study. Organisms were cultured in a mixture of plaster of Paris and charcoal (9:1) and kept under controlled conditions with a 12:12-h light:dark photoperiod, a temperature of 20  2C, and a humidity of 70  5%. Animals were age synchronized before exposure to PCP-contaminated soil, following standardized methods. Briefly, adult animals were incubated under the same culture conditions to lay eggs and then removed after 2 d. The eggs hatched after approximately 10 d. All exposures were performed in artificial soil (corresponding to International Organization for Standardization standard 11268-1; 70% quartz sand, 20% kaolinite clay, and 10% peat ground, dried and sieved to 0.5 mm; CaCO3 was added to adjust the pH to 6.0  0.5) under the same temperature and light conditions described above. The organisms were exposed to the 50% effective concentration for reproduction (EC50) of PCP, which is a nominal concentration of 87 mg kg1 dry weight soil. Pentachlorophenol (99.5% purity, Sigma-Aldrich) was spiked as an acetone solution into 10% of the soil for the test concentration. After homogeneous mixing by shaking and stirring, the spiked soil was allowed to equilibrate for 24 h in a closed glass container and then opened for 24 h to allow the acetone to evaporate. The remaining 90% of the soil was added before moisturizing to 60% of the water-holding capacity. Nonspiked artificial soil and equal volumes of acetone-spiked soil were prepared as controls. A standard 28-d reproduction assay of collembola in artificial soil was carried out to verify that the EC50 concentration of PCP significantly reduced reproduction. Briefly, 10 synchronized juveniles (10–12 d old) were exposed to 30 g of control or polluted soil in 100-mL glass jars, using 5 replicates for each treatment. The jars were opened for aeration twice per week and fed once per week. After exposure, the offspring were counted by means of floatation. Next, short-term exposures of 1 d, 2 d, or 7 d were performed to examine the functional gene expression in F. candida exposed to PCP-spiked soil. Similarly, synchronized cultures were obtained, and 30 animals (23 d old) were exposed per jar containing 30 g of control or polluted soil with 3 replicates per treatment. Given that PCP was dissolved in acetone, the solvent control was used for comparison. Animals were extracted from soil by floating and deposited in microcentrifuge tubes after surplus water was removed; they were immediately frozen in liquid nitrogen before RNA isolation and microarray hybridization. RNA preparation, labeling, and hybridization

The RNA was extracted from each replicate (pool of 30 animals) using the Trizol extraction method (Invitrogen). Three biological replicates were used in microarray tests for each time point. After further purification, the content of isolated RNA was quantified by spectrophotometric analysis, and integrity was confirmed by denaturing agarose gel electrophoresis. A single-color design was used, and 875-ng input of total RNA was used for amplification and labeling with the Agilent LowInput Fluorescent Linear Amplification Kit, according to the

Environ Toxicol Chem 34, 2015

1363

manufacturer’s guidelines. Labeled and amplified complementary RNA (cRNA) was purified using an RNeasy Mini Kit (Qiagen), and the Agilent Gene Expression Hybridization Kit was used for hybridization. Each biological replicate was hybridized individually on 1 array, for a total of 12 hybridizations. After hybridization, the arrays were washed using the Agilent Gene Expression Wash Buffer Kit and scanned with the default settings on an Agilent DNA microarray scanner. In the present study, 4  44 k format Custom Agilent Microarrays (design ID 030000) were used. As described by Yuan et al. [23], each microarray contained 5920 different probes spotted randomly 7 times, representing 5920 different F. candida gene clusters from Collembase. Microarray data analysis

Scanned images were analyzed with Feature Extraction (10.7) Software (Agilent). Raw data were background-corrected and normalized in the limma package from the R software environment. Linear models and empirical Bayes methods were applied to determine different gene expressions. The results were then corrected with subsequent multiple testing using the Benjamini–Hochberg method. Genes with an adjusted p value < 0.05 were considered to be differentially expressed. Quality control was done by making MA plots and box plots of each array. The assessment of differential gene expression resulted in a mean log2 expression ratio (treated/untreated) and a p value for each probe on the array. For hierarchical clustering, mean normalized log2 expression ratios were used from each array, and were done in TIGR MEV version 4.5.1, using Euclidean distance and average linkage. The microarray data described in the present study were deposited into the US National Center for Biotechnology Information (NCBI) GEO database under accession number GSE52389. Annotation of the transcripts was performed using the BLAST algorithm, and gene ontology term enrichment analysis was done with the gene ontology annotated gene clusters present in Collembase using the topGO package in R. Quantitative real-time transcription polymerase chain reaction

To validate microarray data, RNA samples from the microarray experiment of the 2-d exposure were used for real-time quantitative reverse transcription polymerase chain reaction (qRT-PCR). The 2-d exposure samples were chosen to present qRT-PCR data compliant with previous transcriptional studies using F. candida [19]. Three biological replicates for treatment and control were analyzed. Primer sets were designed for 8 target genes (Fcc00152, Fcc00181, Fcc00494, Fcc00583, Fcc01312, Fcc04212, Fcc04366, and Fcc05668) with the software package Primer Express Ver 2.0 (Applied Biosystems). The Fcc02512 (YWHAZ) gene was used as a reference gene for normalization of input complementary DNA (cDNA). Efficiency of PCR was determined by obtaining standard curves in triplicate for all primer sets with 4-fold dilutions of a standard batch cDNA. Primer sequences and PCR efficiency values are available in the Supplemental Data. Approximately 1 mg input of total RNA per sample was used for reverse transcription using M-MLV reverse transcriptase (Promega) according to the manufacturer’s protocol. The derived cDNA was diluted 1:5, and 1 mL was used in 20-mL PCR reaction volumes containing forward and reverse primers and Maxima SYBR Green PCR Master Mix (Applied Biosystems). The qRT-PCR reactions were performed in triplicate for each sample on a DNA engine, Opticon (MJ Research), using universal cycling conditions (10 min at 958C, 15 s at 958C, 55 s at 608C; 45 cycles). A mean

1364

Environ Toxicol Chem 34, 2015

normalized expression value was calculated from the obtained Ct values with the Q-Gene module. Statistical analysis

The offspring in PCP-contaminated or control soils were counted, and a 2-sample t-test was performed in SPSS to verify that the PCP concentrations used in the present study had a significant effect on springtail reproduction. The correlation between the microarray and qRT-PCR platforms was analyzed with the Spearman’s Rho correlation in SPSS (Ver 11.5), based on the mean log2 transformed normalized expression values. RESULTS

Gene expression analysis

To verify that the PCP concentration in soil reported in the literature had a significant and reproducible effect on the reproduction of F. candida in our experiment, we performed a 28-d exposure toxicity test. A nominal PCP concentration of 87 mg kg1 (EC50), a solvent (acetone) control, and an untreated artificial soil control were tested. There were no significantly different effects on reproduction between solvent control and untreated artificial soil control, but a significant effect was observed for the EC50 concentration of PCP. Reproduction was reduced significantly by 40% compared with the solvent control (Supplemental Data, Figure S1). Microarrays were used to identify transcripts that responded to PCP-polluted soil. Three biological replicates were analyzed for each exposure time point. Significantly differentially expressed transcripts were identified for each time point using linear models and empirical Bayes methods (Benjamini– Hochberg adjusted p value < 0.05). The analysis resulted in 685, 580, and 349 differentially expressed genes at 1 d, 2 d, and 7 d; the respective significantly differentially expressed genes are shown in the Supplemental Data, Tables S1–S3. The alteration in gene expression was visible even after 1 d of exposure, and the number of altered transcripts decreased with exposure time. This may indicate that the organisms experience physiological adaptation to PCP treatment, although fluctuating numbers of differentially expressed genes over exposure time have been reported [24]. As can be seen from the Venn diagrams (Figure 1), 92 genes were significantly differentially expressed

M. Qiao et al.

at all exposure times, which may represent a general stress response to PCP. The expression profile of these 92 common differentially expressed genes was similar over time; 53 of them were found to be downregulated and 18 were upregulated at all time points. Functional categories of differentially expressed genes

All differentially expressed genes were subjected to BLAST annotation and functional classification based on gene ontology analysis. The gene ontology terms for biological process and molecular function were identified and enriched, because these seemed to be the most indicative of any effects on a biological level. The gene ontology terms that had a p value