Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Identification and selection of optimal reference genes for qPCR-based gene expression analysis in Fucus distichus under various abiotic stresses

  • Marina Linardić ,

    Roles Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Project administration, Validation, Visualization, Writing – original draft, Writing – review & editing

    marina.linardic@cantab.net (ML); siobhanb@ucla.edu (SAB)

    Affiliations Department of Molecular, Cell and Developmental Biology, University of California Los Angeles, Los Angeles, California, United States of America, Department of Energy Institute of Genomics and Proteomics, University of California Los Angeles, Los Angeles, California, United States of America

  • Siobhan A. Braybrook

    Roles Funding acquisition, Project administration, Supervision, Writing – review & editing

    marina.linardic@cantab.net (ML); siobhanb@ucla.edu (SAB)

    Affiliations Department of Molecular, Cell and Developmental Biology, University of California Los Angeles, Los Angeles, California, United States of America, Department of Energy Institute of Genomics and Proteomics, University of California Los Angeles, Los Angeles, California, United States of America, Molecular Biology Institute, University of Los Angeles, Los Angeles, California, United States of America

Abstract

Quantitative gene expression analysis is an important tool in the scientist’s belt. The identification of evenly expressed reference genes is necessary for accurate quantitative gene expression analysis, whether by traditional RT-PCR (reverse-transcription polymerase chain reaction) or by qRT-PCR (quantitative real-time PCR; qPCR). In the Stramenopiles (the major line of eukaryotes that includes brown algae) there is a noted lack of known reference genes for such studies, largely due to the absence of available molecular tools. Here we present a set of nine reference genes (Elongation Factor 1 alpha (EF1A), Elongation Factor 2 alpha (EF2A), Elongation Factor 1 beta (EF1B), 14-3-3 Protein, Ubiquitin Conjugating Enzyme (UBCE2), Glyceraldehyde-3-phosphate Dehydrogenase (GAPDH), Actin Related Protein Complex (ARP2/3), Ribosomal Protein (40s; S23), and Actin) for the brown alga Fucus distichus. These reference genes were tested on adult sporophytes across six abiotic stress conditions (desiccation, light and temperature modification, hormone addition, pollutant exposure, nutrient addition, and wounding). Suitability of these genes as reference genes was quantitatively evaluated across conditions using standard methods and the majority of the tested genes were evaluated favorably. However, we show that normalization genes should be chosen on a condition-by-condition basis. We provide a recommendation that at least two reference genes be used per experiment, a list of recommended pairs for the conditions tested here, and a procedure for identifying a suitable set for an experimenter’s unique design. With the recent expansion of interest in brown algal biology and accompanied molecular tools development, the variety of experimental conditions tested here makes this study a valuable resource for future work in basic biology and understanding stress responses in the brown algal lineage.

Introduction

Brown algae represent one of the five major lineages that have developed a multicellular body organization independently from other species (with red algae, plants/green algae, fungi, and metazoans as the other four) [1,2]. Among all of the algal groups, the brown algae have the largest diversity in size and morphology, from filamentous to ‘complex’ thalli (bodies) [1]. In addition to their complex morphology, brown algal life cycles are similarly diverse, exhibiting a broad range of variation between the gametophyte and sporophyte generations [3]. As the main inhabitants of the intertidal zone, brown algae have to deal with extreme conditions; tidal cycles expose them to desiccation and osmotic shock, evaporation, and heat shock daily. Furthermore, anthropogenic impact (e.g. pollution) serves as an additional source of abiotic stress in these already challenging environments. How brown algae survive, and thrive, in these environments remains unknown. Taking into account their divergent evolutionary history [4], importance as an ecological and economical resource [5,6], and their astonishing ecology [7], the brown algae represent a unique and important group to explore. However, they have been extremely understudied on a molecular level. It is only within the past decades that scientists have begun to explore their rich molecular biology.

Gene expression analyses are necessary to understand how brown algal genetic networks dynamically change upon stimulus to regulate their responses to stressful life conditions. In recent years, microarrays and next-generation sequencing (NGS) technologies have arisen as the most used methods for quantification of gene expression on a global level. In spite of these advancements, qPCR remains one of the simplest and accessible methods for studies of small gene numbers. It also serves as a confirmational tool for NGS-derived results. qPCR is most widely used for fast and reliable quantification of mRNA steady-state levels because of its high sensitivity, accuracy, specificity, reproducibility, and low cost [8,9]. However, the accuracy and reliability of qPCR experiments are highly affected by several factors such as RNA integrity, reverse transcription efficiency, and primer efficiency [10]. It is therefore of utmost importance to combat methodologically introduced variation by normalizing to stable reference genes as internal controls. An ideal reference gene should have a similar expression in all tissues and experimental conditions. Choosing reference genes with unstable expression could result in misleading results and inappropriate conclusions regarding target gene expression. However, the expression of commonly used reference genes may vary depending on the life stage of the organism, experimental conditions, as well as tissue source [10]. Therefore, it is unlikely that a single reference gene would exist for all experiments and it is necessary to evaluate the best reference genes for each experiment. To quantitatively evaluate the efficiency and stability of reference genes for qPCR experiments, several statistical algorithms have been developed, such as geNorm [8], NormFinder [11], and BestKeeper [12]. These methods have been used to determine the best reference genes in red, green, and brown algae [1316], plants [1722], and metazoans [2327].

In some brown algae such as Ectocarpus siliculosus and Undaria pinnatifida, efforts have been made to identify reference genes, with results indicating variable stability of candidate reference genes depending on experimental treatments [13,14]. The brown alga Fucus has been an important cell biology model. Experiments in Fucus led to several crucial discoveries in cell polarization and asymmetric cell division [2832]. Recent studies in Fucus further show its scientific importance as a cell biology model [3339], but also as a tool to investigate ecological and physiological effects of abiotic stresses in natural habitats [4046]. As new molecular methods become available, the potential for Fucus to serve as a modern model system for exploring biology in extreme environments becomes tractable. qPCR serves as a basis to address questions of gene expression; it is, therefore, necessary to develop a set of normalization genes in Fucus to enable reproducible quantitative gene analysis by qPCR. In Fucus, Elongation Factor alpha, Beta-Actin, Tubulin, and/or a 14-3-3 protein have all been used as reference genes for qPCR [41,42,44,45]. However, there has been no systematic study of the stability and suitability of such reference genes over a wide array of conditions.

In this study, we identified and tested a set of potential normalization genes for Fucus distichus. Nine ‘housekeeping’ genes were selected as candidate reference genes: Elongation Factor 1 alpha (EF1A), Elongation Factor 2 alpha (EF2A), Elongation Factor 1 beta (EF1B), a 14-3-3 gene, Ubiquitin Conjugating Enzyme (UBCE2), Glyceraldehyde-3-phosphate Dehydrogenase (GAPDH), Actin Related Protein Complex (ARP2/3), Ribosomal protein (40s; S23), and Actin (ACT). Their expression was analyzed by qPCR in samples submitted to various stress conditions; salinity, desiccation, pollution, nutrient deprivation, wounding as well as temperature, light, and phytohormone treatment. Three algorithms were used to evaluate the expression stability of the reference genes: geNorm, NormFinder, BestKeeper; a rank aggregation algorithm was employed to reach a consensus between the three. The number of reference genes to use for qPCR-based normalization was also explored. We provide a recommendation of paired reference genes for the conditions tested, alongside a recommended method for suitability assessment in an individual’s experiments. To validate our recommendations, the differential expression of Hsp70 and Hsp90 genes, encoding stress-responsive heat shock proteins, were examined under salinity stress.

Results

Choosing candidate reference genes

To identify the ‘housekeeping’ gene sequences for potential normalization genes, two approaches were taken. First, housekeeping gene sequences previously reported for Ectocarpus siliculosus [13] were aligned against the related species Fucus serratus embryo development transcriptome [36] to identify putative homologs. Furthermore, additional sequences were identified directly from the F. serratus transcriptome dataset, by identifying highly and constantly expressed genes across the four embryo developmental stages in the study. As such, candidate gene sequences were chosen based on three criteria: 1) a high percentage of alignment to E. siliqulosus (for those with Ectocarpus sequence matches), 2) high expression values in the Fucus embryo transcriptome (normalized expression metric ‘transcripts per million transcripts’; ‘TPM’) and 3) a relatively constant expression across samples in the Fucus embryo transcriptome. Nine housekeeping genes were selected that satisfied these criteria: elongation factor 1 alpha (EF1A), elongation factor 2 alpha (EF2A), elongation factor 1 beta (EF1B), a 14-3-3 gene, ubiquitin conjugating enzyme (UBCE2), glyceraldehyde-3-phosphate dehydrogenase (GAPDH), actin related protein complex (ARP2/3), ribosomal protein (40s; S23), and actin (ACT) (Table 1, sequences in S1 Fig). Primer sets were designed for all nine genes and their efficiencies and melting temperatures determined by qPCR (Table 1, S1 Fig; See Materials and Methods). The efficiency of the primer sets varied from 98.5% (GAPDH) to 106.9% (EF1A), and correlation coefficients ranged between 0.987 and 0.999 (Table 1). All efficiencies were considered acceptable for use. We strongly recommend the calculation of primer efficiencies for use qPCR-based differential gene expression analyses.

thumbnail
Table 1. Description statistics of housekeeping gene candidates.

https://doi.org/10.1371/journal.pone.0233249.t001

Expression of reference genes across samples

To determine reference gene expression across a wide set of conditions, the steady-state levels of all nine genes were analyzed in cDNA samples from fifteen different treatments (in biological triplicate) with two time-points each (3 hours and 3 days post-treatment): auxin, gibberellic acid, EtOH control, imidacloprid, CuSO4, hypersaline 2x artificial seawater (ASW), hyposaline 0.5X ASW, desiccation, wounding, high light, low light, high temperature, low temperature, ASW + Provasoli nutrients, and ASW with no additional nutrients as the control (detailed in Materials and Methods; S1 Table). The quantification cycle (Cq) was identified for each gene in the triplicate samples by qPCR. The Cq values of all tested genes lay between 18.63 and 28.21 (except a single replicate outlier in EF2A at Cq = 32.7) (S4 Fig). The most highly expressed gene was EF1A (Cqmean = 20.6), followed by 40S. The lowest expression level was observed for GAPDH (Cqmean = 26.9) (Fig 1). For each gene, the variation across samples did not exceed 3 cycles, which suggests a relatively constant expression in all conditions. All nine candidate genes were determined as suitably ‘consistent’ (qualitatively) across conditions to move forward in our analyses.

thumbnail
Fig 1. Expression level of tested housekeeping genes.

The distribution of Cq (quantification cycle) values of 9 reference genes pooled across 30 samples (15 experimental conditions x 2 time-points) obtained using qPCR. The boxplot marks the median (line) and 25th (lower) and 75th (upper) percentile; x marks the mean; the underlying violin plots show the data distribution for each housekeeping gene. Outliers are depicted as black dots. Data for individual experimental conditions may be found in S4 Fig.

https://doi.org/10.1371/journal.pone.0233249.g001

Reference gene expression stability in all samples

The stability of expression for a given reference gene is a quantitative measure by which the suitability of the reference gene may be assessed. Stability refers to how invariant the expression pattern of a given gene is in an experimental setup. The stability of our nine candidate reference genes was evaluated across all 15 conditions and 2 time-points (Fig 2). Stability was analyzed using three algorithms: geNorm [8], NormFinder [11], and BestKeeper [12]. These algorithms rank the reference genes according to the calculated gene expression stability values and acceptability threshold (geNorm (M<1.5); NormFinder (SV<1.0)) or standard deviation and threshold (SD<1.0, BestKeeper) (detailed in Materials and Methods). Fig 2 shows the calculated stability metric for each gene in all samples combined, by method, ranked from best to worst (left to right).

thumbnail
Fig 2. Individual ranking of housekeeping genes by stability.

Stability values for nine candidate reference genes generated by the following algorithms: A) geNorm, B) NormFinder, C) BestKeeper, and D) a consensus (by rank aggregation). The maxiumum scale for A-C represent the maximum acceptability value for each stability metric.

https://doi.org/10.1371/journal.pone.0233249.g002

The rankings generated by geNorm and NormFinder were similar; EF1B and GAPDH were identified as the most stable genes, followed by ACT, 14-3-3, EF1A, and 40S (Fig 2A and 2B). BestKeeper, on the other hand, ranked 40S and 14-3-3 as the most stable (Fig 2C). The three least stable genes identified by all three algorithms were UBCE2, EF2A, and ARP2/3 (Fig 2A–2C). To create a consensus of the best-to-worst gene ranking across all three programs, a rank aggregation method was performed using a Cross Entropy Monte Carlo algorithm (R-package RankAggreg; See Materials and Methods). This analysis indicated that GAPDH and EF1B were the two most stable genes and EF2A and ARP2/3 the most unstable (Fig 2D). Based on these analyses, GAPDH and EF1B are recommended as general normalization genes for qPCR gene expression studied in Fucus.

Reference gene expression stability in sample groups (conditions)

Numerous reports acknowledge the importance of using an appropriate reference gene for a given experiment since individual reference genes may not maintain normalized expression in all conditions [10,47,48]. To test which reference genes would be most suitable for each of our stress-condition groups, expression stability was analyzed per group using geNorm, NormFider, and BestKeeper. The samples were grouped based on the nature of their stressor as follows: Fucus is an intertidal alga that is challenged by a very harsh natural environment where it undergoes daily desiccation and osmotic shock (group 1 –physiological stress); furthermore, these environments are under the constant pressure of pollutants, such as copper and herbicides/pesticides (group 2—pollution); also, brown algae serve as food for marine organisms such as mollusks, so the effect of grazing (proxied by mechanical wounding) was examined as well (group 3 –wounding); in Stramenopiles, phytohormones have been found to influence developmental processes [4951] and as such we have tested the effect of auxin and gibberellic acid (group 4—hormones); Fucus was also grown with and without nutrients to identify the best housekeeping genes for nutritional studies (group 5—nutrients); lastly, we cultured Fucus under different light and temperature regimes (group 6 –temperature-light). The ranking of reference genes in each treatment group, based on stability, is shown in Fig 3. The Cq values for each reference gene, by treatment group, can be found in S4 Fig. The stability values for each gene by treatment group, per algorithm, can be found in S5 Fig.

thumbnail
Fig 3. Optimal stability ranking of candidate reference genes using geNorm, NormFinder, BestKeeper, and rank aggregation method.

Gene ranks from the three stability algorithms are shown by treatment group. Line colors represent: purple (geNorm), green (NormFinder), yellow (BestKeeper), black (mean rank), and red (consensus rank).

https://doi.org/10.1371/journal.pone.0233249.g003

GAPDH was identified as one of the top three most stable genes across multiple conditions (4 out of 6) and ARP2/3 as one of three least stable genes (4 out of 6), whereas the stability of other candidate reference genes varied depending on the condition examined (Fig 3; S5 Fig). As mentioned earlier, the three algorithms rank reference genes according to their calculated gene expression stability values and acceptability thresholds (geNorm: M<1.5; NormFinder: SV<1.0) or standard deviation (BestKeeper: SD<1.0). Based on these thresholds, in the hormone and the temperature-light groups, ARP2/3 and UBCE2 were rejected as suitable reference genes (S4 and S5 Figs).

geNorm, BestKeeper, and NormFinder showed differences in the ranking of candidate reference genes within treatment groups, to some extent (Fig 3). To achieve a consensus ranking across all three methods that could be used for recommendation of normalization genes by condition, we again performed a rank aggregation analysis (Fig 3, red line and left to right order). For pollution, hormone, nutrient and temperature-light treatments there was less variation in ranking by the three algorithms and the consensus appears to be a reasonable recommendation. However, for physiological stress and wounding, the results of the three algorithms differed more, and the consensus seemed qualitatively less reliable. As such, for these two conditions, we recommend using the ranking produced by geNorm or NormFinder; our recommendation here is based on geNorm (See Discussion).

Identifying the optimal number of reference genes for normalization

Our previous analyses allowed us to provide recommendations for normalization genes to be used in general and specific experimental designs. However, using a single reference gene during qPCR experiments can cause bias and the use of multiple reference genes is suggested as standard practice [8]. To determine the optimal number of reference genes to be used in Fucus qPCR experiments, pairwise variation Vn/n+1 was calculated for the proposed reference genes using the geNorm algorithm [8]. In this method, the reference gene variances were combined additively and successively according to geNorm rank. After each addition the pairwise variance was calculated between that sum (Vn+1) and the variance of the prior sum (Vn); for example, V2/3 was the pairwise variance calculated when using the top two genes was compared with using the top three. In general, the more reference genes that are added, the lower the successive pairwise variance becomes. When the pairwise variance was below 0.15, n number of normalization genes was considered sufficient.

When all conditions were pooled, to simulate a general experiment, V2/3 was below the 0.15 threshold, suggesting that using the top two stable reference genes was sufficient (Fig 4; V2/3 = 0.145; optimal reference gene set GAPDH+EF1B). When the analysis was performed by condition group, V2/3 was below 0.15 for all conditions except pollution (physiological stress = 0.130, hormones = 0.096, nutrients = 0.103, temperature-light = 0.120, wounding = 0.105, pollution = 0.173; Fig 4). For pollution stress, V3/4 was less than the threshold (V3/4 = 0.145) suggesting the need for three reference genes for these experiments (Fig 4). Taken together, two stable reference genes are sufficient for most conditions tested here. By combining the recommended number of reference genes from this analysis with the consensus rank achieved for stability, we generated a table of most recommended and least recommended reference genes (Table 2).

thumbnail
Fig 4. Calculation of optimal number of housekeeping genes.

Pairwise variation (Vn/n+1) was calculated in all tested samples; all (all conditions together), hor (hormones), nutr (nutrients), phy (physiological stress), pol (pollution), t-l (temperature-light), wou (wounding). The Vn/n+1 values below the 0.15 threshold (dotted line) indicate that n normalization genes are sufficient.

https://doi.org/10.1371/journal.pone.0233249.g004

thumbnail
Table 2. List of recommended reference genes for each of the tested conditions.

https://doi.org/10.1371/journal.pone.0233249.t002

Validation of reference genes in conditions of physiological stress

To experimentally validate our reference gene recommendations, we performed qPCR for two heat shock proteins, Hsp70 and Hsp90, in samples from Fucus distichus under physiological stress (salinity and desiccation). Hsp70 and Hsp90 were chosen as potential target genes as they had exhibited differential expression in Fucus under various stress conditions [40,42,52]. Hsp70 and Hsp90 sequences were identified in the F. serratus embryo transcriptome [36] and primers were designed and tested (S6 Fig). We then examined gene expression using the ΔΔCq method with 1) normalization to the two most stable reference genes (GAPDH and EF1A) or 2) with normalization to the least stable genes (EF2A and ARP2/3) for physiological stress (Table 2). Gene expression analysis with the most stable pair (EF1A + GAPDH) showed a significant increase of Hsp70 cDNA levels with 0.5x ASW treatment and a significant decrease in levels with the 2x ASW stress and desiccation treatments (Fig 5A). Conversely, normalizing to the ARP2/3/EF2A (least stable) reference gene set resulted in a loss of significant differential expression in the 0.5X ASW treatment (Fig 5B). Hsp90 cDNA levels did not show significant change with treatment using either normalization set (Fig 5). These data indicate that using the most stable pair of normalization genes allowed for the capture of more information on differential gene expression for Hsp70; however, even the lowest-ranked normalization pair did result in some differential expression detection. Using a suboptimal normalization pair would likely be most detrimental when high Cq variance within technical replicates or between biological replicates is present. As such, using the most recommended set of normalization genes, in general or by condition, is likely to increase experimental sensitivity and reduce false negatives.

thumbnail
Fig 5. Detection of significant change in gene expression depending on the reference gene choice.

Expression of two heat shock protein genes (Hsp70 and Hsp90) normalized by the two most stable (A: EF1A and GAPDH) and two least stable (B: ARP2/3 and EF2A) reference gene pairs. Statistically different gene expression from ASWP is indicated by * (Student’s t-test; normal distribution, unequal variance; p<0.05).

https://doi.org/10.1371/journal.pone.0233249.g005

Discussion

The use of stable and suitable reference genes when analyzing qPCR data is of utmost importance to correctly assess, present, and interpret gene expression in any experimental design. In this study, we analyzed the suitability of nine candidate genes for normalization of gene expression data obtained from qPCR in a brown alga Fucus distichus. This paper reports recommendations for reference gene pairs to be used in Fucus experimental qPCR studies (Table 2). The recommendation is a result of two analyses: stability assessment and calculation of an optimal number of reference genes to be used. Below we discuss the outcomes and comparative merits of these analyses.

Assessing stability

The expression stability of the candidate genes was tested using the geNorm, NormFinder, and BestKeeper algorithms. Our analysis with the three algorithms showed that the stability of most candidate genes strongly depended on the experimental condition used, with GAPDH being the most stable gene in most conditions and ARP2/3 being the least stable in most conditions. We did observe some discrepancies in ranking candidate gene stability within conditions when using these three different algorithms.

The differences in stability values observed between different algorithms were not entirely unexpected, taking into account that there are differences inherent in the three approaches [8,11,12]. geNorm is based on the principle that the most stable housekeeping genes should have a nearly identical variation in expression ratio, or co-expression pattern. This approach, however, may identify co-differentially expressed genes as highly stable in the chosen experimental system [8,11]. Conversely, NormFinder is not affected by the co-differential expression problem and as such should be more robust; it estimates both inter- and intra-group variation and then combines them into a stability value [11]. BestKeeper, on the other hand, uses raw Cp values as input to identify the best among the investigated candidate genes. It uses a Pearson correlation analysis, a parametric method, which is valid for normally distributed data with a homogeneous variance; if the data do not match these dependencies it may lead to a false interpretation of the obtained results.

When the stability of genes in all the treatments was tested together, some discrepancies were detected in the ranking of the most stable candidate reference genes using the three algorithms, but there was an agreement between ranking the least stable genes. GAPDH and EF1B were ranked as the most stable housekeeping genes tested by geNorm and NormFinder, however, BestKeeper only placed them 3rd and 5th, respectively. The least stable genes were identified as ARP2/3, EF2alpha, and UBCE. This is contradictory to results in Ectocarpus siliculosus, where both ARP2/3 and UBCE were the most stable genes and GAPDH was one of the most unstable genes [13]. It may be that patterns of gene expression vary between different brown algal species, which further emphasizes the need to precisely define best reference genes for specific studies.

According to our results, in four out of seven tested conditions GAPDH and EF1B were the most stable reference gene candidates. It is valuable to note that GAPDH and EF1B had different Cqmean values (26.9 and 20.6, respectively) providing candidate normalization genes at different expression level tiers. These genes have been previously identified as stably expressed in other systems such as green algae, brown algae, plants, and animals [14,15,17,18,53,54]. Some studies in Fucus directly selected common reference genes such as EFAs to normalize their target genes [41,42,45]. EFA genes have been one of the most widely used in normalizing gene expression of algal and plant species under stress conditions [13,17,22,54,55]. However, EFAs are not necessarily optimal for all stress conditions in Fucus, as shown here.

Some of the candidate genes were mostly present in the middle of the stability range, such as 14-3-3, 40S, and actin. Actin and ribosomal subunit or RNA genes have previously been reported as variably expressed in algae [13,16] and other organisms [17,22,56,57]. The most unstable genes from our study were ARP2/3, UBCE2, and EF2A. Even though their stability ranking position varied depending on the treatment, for all of the conditions except wounding, at least one of the aforementioned genes was ranked last. ARP2/3 or UBCE have previously been used as normalization genes in the brown alga Ectocarpus siliculosus and were stably expressed in that system [13], consistent with our analysis. In line with our results, some previous reports in algae and plants identified UBCE as a variable and non-suitable reference gene [16,54].

What is perhaps most pertinent to note is that even though ranking differences were observed between the different algorithms, the majority of the calculated stability values were still suitable as the fell under the rejection threshold (geNorm >1.5, BestKeeper>1, NormFinder>1; S5 Fig). This suggests that, generally, all nine of the proposed reference genes we started with should be suitable for normalization. We note that there are specific stress conditions where a handful of candidate genes were rejected for use (hormone and temperature-light; S5 Fig) and so care in specific stress conditions should be taken when selecting the exact pair of normalization genes.

The optimal number of reference genes

It has been shown that the inclusion of more than one reference gene is required to detect subtle changes in gene transcript levels by qPCR analysis [8,11,58]. Vandesompele et al. [8] provide evidence that normalization using a single gene can lead to an inaccurate normalization of up to 25% of cases (with up to 6.4-fold difference). In our experiments and analyses, we could conclude that in all treatment groups, except pollution, using the two most stable genes should be sufficient for normalization. For the pollution treatment, our experiment would require three normalization genes according to analysis. We strongly recommend that experimental designs for Fucus RT-PCR and qPCR include at least two reference normalization genes.

Evaluating normalization pairs in a desiccation and salinity stress study

Overall, our recommendation of a ‘best’ normalization pair for physiological stress condition (GAPDH and EF1A) proved to be more sound when compared to the ‘least’ suitable pair when evaluating the differential gene expression of Hsp70 during salinity stress and desiccation. The statistically significant difference in the expression of Hsp70 in low salinity failed to be detected when normalizing to the two least stable genes. This has previously been shown in other reports as well, where the choice of unstable genes led to misinterpretation of expression data [47,59,60]. It should be noted that we have assumed that the differential expression of Hsp70 in low salinity is true and not a false positive; we believe this to be the most likely case given the higher stability of the ‘best’ pair.

Conclusion

The selection of suitable reference genes to test qPCR based gene expression is a necessary first step in understanding key molecular networks in the brown algal lineage—here we recommend the usage of specific reference gene sets in specific conditions when performing comparable experiments in the brown alga Fucus distichus. In general, at least two reference genes are recommended, and the best ‘general’ pair to use are GAPDH and EF1B. When designing new experiments, we recommend checking the top five general recommended genes from this study (Table 2) to test their stability in one’s experimental system, as conducted here, before deciding on the best normalization gene set for your study.

Materials and methods

Culture conditions and experimental design

Fucus distichus (Fucales, Phaeophyceae) individuals were collected from their natural rockpool habitat at the University of California Kenneth S. Norris Rancho Marino Reserve (Cambria, CA). Multiple apical segments were cut from adult individuals and cultivated in 2L glass flasks in a Percival incubator (Percival Scientific, USA) at 16⁰C in filter-sterilized artificial seawater (ASW; 450 mM NaCl, 10 mM KCl, 9 mM CaCl2, 30 mM MgCL2∙6H2O, 16 mM MgSO4∙7H2O) enriched with Provasoli medium (PES) for acclimation [61]. Samples were cultured under a white fluorescent light at 60 μmol m-2 s-1 with a 12h:12h light: dark cycle. After 7 days, individual apical segments were transferred into Petri dishes with 15 different treatments, one segment per condition (S1 Table). Three biological replicates were obtained for each treatment at two time-points: 3 hours and 3 days after treatment start, resulting in 90 samples in total. A summary of the treatments may be found in S1 Table.

The chemical treatments were: control ASW with nutrients (PES), nutrient-deficient ASW, 0.2 μg/L imidacloprid (Marathon 1%, OHP, Inc., USA), 50 μM indole-3-acetic acid (IAA; Cat#102037, MP Biochemicals, Irvine, CA.), 50 μM gibberellic acid (GA; Cat#G7645, Sigma). An equal volume of absolute ethanol was used as a control for the GA and IAA treatment. In addition, a saline shock was performed using a hypersaline solution (2x ASW) and a hyposaline solution (0.5x ASW). The desiccation treatment was affected by placing the algal segments onto dry Petri dishes, after blotting gently, under the same environmental conditions as in other treatments. A mechanical wounding treatment, to simulate the effect of grazing, was performed by damaging the algal segments with a razor blade in several places along the thallus. Alteration of light and temperature was achieved as follows: a portion of samples was cultured under modified light conditions (120 μmol m-2 s-1 and complete darkness) or two non-standard temperatures (8⁰C and 22⁰C).

RNA extraction and cDNA synthesis

Tissue was flash frozen in liquid nitrogen after treatment (3 hours and 3 days) and immediately ground in liquid nitrogen with a pestle and a mortar. In our hands, extraction of RNA from Fucus tissue is impaired by storage of the intact and/or ground tissue at -80 for longer periods of time. To further refine the tissue homogenate samples were ground in 3 mL Duall glass grinders (Cat# K885451/0021, Smith Scientific, UK) grinder in 1ml of CTAB buffer (2% CTAB, 100 mM Tris-HCl, 1.5 M NaCl, 50 mM EDTA, 50 mM DTT). RNA extraction was adapted from Apt et al. [62] as follows briefly. Samples were shaken on a tilt shaker at room temperature for 20 minutes after which 1v of chloroform was added to each. Solutions were mixed by inverting the tubes several times and additionally incubating for 20 minutes while gently shaking. Samples were then centrifuged at 10,000g for 20 minutes at 4⁰C, after which 0.3v of 100% ethanol was added to remove the polysaccharides. Polysaccharides were further extracted with 1v of chloroform and centrifuged for 20 minutes at 10,000g at 4⁰C. The upper phase was transferred to a new tube and RNA was precipitated overnight at -20⁰C by adding 0.25v LiCl and 1% (v/v) beta-mercaptoethanol. Samples were centrifuged for 30 minutes 13,000g (4⁰C), after which the pellet was re-suspended in 50 μl of DEPC-treated MiliQ water. To remove residual DNA, a DNase treatment was performed (TURBO DNase, ThermoFisher) according to manufacturer’s instructions. The final volume was adjusted to 500 μl with RNAse free water and extraction was performed by adding 1v of phenol: chloroform: isoamylic alcohol (25:24:1 v/v). The samples were centrifuged at 10,000g for 20 minutes (4⁰C), after which another 1V chloroform extraction was carried out and centrifuged again. The upper phase was transferred to a clean RNAse free tube and RNA was precipitated by addition of 0.1v sodium acetate (pH5.5) and 2.5v of 100% ethanol overnight at -20⁰C. After centrifugation (30 minutes at full speed, 4⁰C), the supernatant was carefully aspirated with a pipette and the pellet was washed with 75% ethanol. The tubes were centrifuged for 10 minutes at full speed (4⁰C). The supernatant was carefully removed and the tubes were left to air dry for 5–10 minutes. The pellet was then re-suspended in RNAse free water.

The purity of RNA was assessed by measuring the ratio OD260/OD280 and OD230/OD260 using a NanoDrop 2000 (Thermo Fisher Scientific, USA). RNA integrity was measured using the High Sensitvity RNA ScreenTape Assay in an Agilent TapeStation (Agilent Technologies, Inc.; S7 Fig). The RNA sample (40 ng) was reverse-transcribed to cDNA using RevertAid First Strand cDNA Synthesis Kit with oligo (dT)20 primers (Thermo Fisher Scientific, USA) according to the manufacturer’s instructions. cDNA samples were diluted to 10 ng/μl for qPCR and kept at -20°C until further use.

Quantitative real-time PCR (qPCR)

For each candidate gene, a pair of oligonucleotide primers was designed using Primer3 Input (v. 4.1.0) online tool (Table 1). qPCR was performed using a CFX384 Real Time PCR System (Bio-rad Laboratories Inc., USA). For each test, 3 μl of cDNA (in technical duplicate wells) template was amplified using the SsoAdvanced™ SYBR® Green Supermix (Bio-rad Laboratories Inc., USA) in a final volume of 15 μl to test housekeeping gene expression levels. The cycling was performed as follows: 95⁰C for 5 min followed by 41 cycles of 30s at 95⁰C, 30s at 60⁰C and 30s at 72⁰C and a final step of 95⁰C for 1 min. Each run was finished with heating up the samples from 65⁰C to 95⁰C to obtain a melting curve to test the specificity of amplification. All of the amplicons tested here had single melting peaks indicating a unique amplification product (S2 Fig).

Primer efficiencies were calculated as follows: a pooled cDNA sample of all samples was mixed and a dilution series generated to yield 1x, 0.1x, 0.01x, and 0.001x dilutions. Each primer set was used to amplify from the dilution series using the conditions above. The primer amplification efficiency (PE) and the correlation coefficient (R2) of each primer pair were calculated (Table 1, S3 Fig). All of the primer sets reported here fell within accepted boundaries of PE and R2. Cq for the NTC for each primer set were all above 35 cycles (EF1A = 37.71, EF2A = 39.44, EF1B = undetected, GAPDH = 38.33, UBCE2 = 39.51, ACT = 35.78, 14-3-3 = undetected, 40s = 38.39, ARP2/3 = undetected). A MIQE checklist (https://rdml.org/miqe.html) is provided in the supplement (S2 Table).

Assessing the stability of candidate gene expression

Data stability analysis was performed on all samples (15 conditions x 2-time points x 3 biological replicates) for all of the 9 genes identified, as well as separate groups with respect to the nature of the treatment (physiological stress, hormone addition, pollution stress, temperature-light modification, nutrient modification, and wounding/grazing (S1 Table). Quantification cycle (Cq) values were analyzed with geNorm, NormFinder, and BestKeeper.

GeNorm calculates an average expression stability value (M) for each reference gene. The analysis allows ranking of the genes according to their expression stability based on an iterative stepwise exclusion of the genes with the highest M value (lowest stability). Additionally, it calculates the optimal number of reference genes to be used for normalization, through a pairwise variation (V) test [8]. NormFinder calculates the expression stability value for each gene using ANOVA-based mathematical analysis, taking into account intra- and inter-group variations of the samples. A low SV-value indicates the high expression stability of this gene [11]. BestKeeper is an Excel-based tool that uses the standard deviation (SD) and the coefficient of variance (CV) as evaluation criteria for stably expressed reference genes. Stability values for geNorm were analyzed using R package NormqPCR [63] (NormFinder stability values were calculated using the NormFinder R script, and BestKeeper Excel-based analysis was performed and standard deviation (SD) was used as a stability measure [12]. In addition, a rank aggregation method based on a Monte Carlo cross-entropy algorithm (R-package; RankAggreg 0.6.5; https://CRAN.R-project.org/package=RankAggreg; with default settings and following set parameters: method = "CE", distance =“Spearman”, convIn = 7, seed = 100) was used to combine the gene ranks of the three above algorithms and create a consensus housekeeping gene ranking.

Validation of chosen housekeeping genes

Apical segments were placed the following physiological stress conditions: 2x salinity, 0.5x salinity, and desiccation condition for 6 hours. Untreated samples were transferred to 1X ASW. After 6 hours, RNA was extracted as detailed above and gene expression analyzed. To test the normalization efficiency of reference genes in a specific condition (ΔΔCq method), two of the most stable and two of the most unstable candidates for this group (physiological stress) were used to test the expression level of two heat-shock proteins (Hsp). Hsp gene sequences were identified through a local BLAST of the Ectocarpus Hsp70 and Hsp90 to the Fucus serratus transcriptome [36] and primers were designed using Primer3 Input (v. 4.1.0) online tool. ΔΔCq was calculated by normalization of the Hsp genes with the two housekeeping genes and to the targeted gene expression detected in a separate control sample.

Supporting information

S1 Fig. Sequences of all used reference genes.

Each reference gene was matched to a Fucus transcriptome gene (Trinity gene CDS; Linardić et al., 2020) and PCR amplicons resulting from the PCR amplification with the designed primers in Table 1.

https://doi.org/10.1371/journal.pone.0233249.s001

(PDF)

S2 Fig. RT-qPCR Melt curves of 9 candidate housekeeping genes.

-ΔF/ΔT (change in fluorescence/change in temperature) is plotted against temperature to obtain a clear view of the melting dynamics of each reference gene.

https://doi.org/10.1371/journal.pone.0233249.s002

(PDF)

S3 Fig. PCR amplification efficiency curves.

Amplification efficiency is determined from the slope of the log-linear portion of the calibration curve (y function coefficient). The initial template concentration (the independent variable; log10 of dilution 1x, 0.1x, 0.01x, 0.001x) is plotted on the x axis and corresponding Cq (the dependent variable) is plotted on the y axis.

https://doi.org/10.1371/journal.pone.0233249.s003

(PDF)

S4 Fig. Reference gene expression in individual condition groups for all 9 reference genes, with consistent coloring by gene name.

The boxplot marks the median (line) and 25th (lower) and 75th (upper) percentile; x marks the mean; the underlying violin plots show the data distribution for each housekeeping gene. Outliers are plotted as black dots.

https://doi.org/10.1371/journal.pone.0233249.s004

(PDF)

S5 Fig. Stability analysis values of candidate reference genes in the 6 stress condition groups.

Stability values are are conditionally formatted (shade of color) according to their stability value, from the most stable (darkest shade) to the least stable (lightest shade). Purple corresponds to geNorm, green to NormFinder and yellow to BestKeeper analysis. Red text indicates genes over the threshold of acceptability, by algorithm.

https://doi.org/10.1371/journal.pone.0233249.s005

(PDF)

S6 Fig. Primers and sequences for Hsp70 and Hsp90 used in the study.

https://doi.org/10.1371/journal.pone.0233249.s006

(PDF)

S7 Fig. TapeStation analysis of the 90 RNA samples.

The High Sensitivity RNA ScreenTape assay was used for analyzing and assessing integrity of total RNA in all samples (following the manufacturer’s instructions) on a 2200 TapeStation Bioanalyzer (Agilent Technologies Inc.). RIN = RNA integrity number.

https://doi.org/10.1371/journal.pone.0233249.s007

(PDF)

S1 Table. Table of conditions used for housekeeping gene analysis.

Column 1 –treatment, column 2 –parameters of the treatment, column 3 –length of treatment, column 4 –associated group of treatments.

https://doi.org/10.1371/journal.pone.0233249.s009

(PDF)

Acknowledgments

We thank the developers of geNorm, NormFinder, and BestKeeper for their tools. We also thank Dr. Lauren Dedow for critical reading of the manuscript. We thank the members of the Braybrook and Pellegrini labs for their comments and support during this project.

References

  1. 1. Charrier B, Le Bail A, De Reviers B. Plant Proteus: Brown algal morphological plasticity and underlying developmental mechanisms. Trends Plant Sci. 2012;17(8):468–77. pmid:22513108
  2. 2. Silberfeld T, Leigh JW, Verbruggen H, Cruaud C, de Reviers B, Rousseau F. A multi-locus time-calibrated phylogeny of the brown algae (Heterokonta, Ochrophyta, Phaeophyceae): Investigating the evolutionary nature of the “brown algal crown radiation.” Mol Phylogenet Evol. 2010;56(2):659–74. pmid:20412862
  3. 3. Cock JM, Godfroy O, Macaisne N, Peters AF, Coelho SM. Evolution and regulation of complex life cycles: a brown algal perspective. Curr Opin Plant Biol. 2014;17:1–6. pmid:24507487
  4. 4. Baldauf SL. The deep roots of eukaryotes. Science. 2003;300:1703–6. pmid:12805537
  5. 5. Barbier M, Araújo R, Rebours C, Jacquemin B, Holdt SL, Charrier B. Development and objectives of the PHYCOMORPH European Guidelines for the Sustainable Aquaculture of Seaweeds (PEGASUS). Bot Mar. 2020;63(1):5–16.
  6. 6. FAO. The State of World Fisheries and Aquaculture 2018—Meeting the sustainable development goals. Rome; 2018. 227 p.
  7. 7. Hurd CL, Harrison P, Bischof K, Lobban CS. Seaweed Ecology and Physiology. Cambridge University Press; 2014. 551 p.
  8. 8. Vandesompele J, De Preter K, Pattyn F, Poppe B, Van Roy N, De Paepe A, et al. Accurate normalization of real-time quantitative RT-PCR data by geometric averaging of multiple internal control genes. Genome Biol. 2002;3(7):research0034.1. pmid:12184808
  9. 9. Phillips MA, D’Auria JC, Luck K, Gershenzon J. Evaluation of candidate reference genes for real-time quantitative PCR of plant samples using purified cDNA as template. Plant Mol Biol Report. 2009;27(3):407–16. pmid:24489433
  10. 10. Kozera B, Rapacz M. Reference genes in real-time PCR. J Appl Genet. 2013;54(4):391–406. pmid:24078518
  11. 11. Andersen CL, Jensen JL, Ørntoft TF. Normalization of real-time quantitative reverse transcription-PCR data: A model-based variance estimation approach to identify genes suited for normalization, applied to bladder and colon cancer data sets. Cancer Res. 2004;64(15):5245–50. pmid:15289330
  12. 12. Pfaffl MW, Tichopad A, Prgomet C, Neuvians TP. Determination of stable housekeeping genes, differentially regulated target genes and sample integrity: BestKeeper–Excel-based tool using pair-wise correlations. Biotechnol Lett. 2004;26(6):509–15. pmid:15127793
  13. 13. Le Bail A, Dittami SM, Franco PO, Rousvoal S, Cock MJ, Tonon T, et al. Normalisation genes for expression analyses in the brown alga model Ectocarpus siliculosus. BMC Mol Biol. 2008;9:75. pmid:18710525
  14. 14. Li J, Huang H, Shan T, Pang S. Selection of reference genes for real-time RT-PCR normalization in brown alga Undaria pinnatifida. J Appl Phycol. 2019;31(1):787–93.
  15. 15. Liu C, Wu G, Huang X, Liu S, Cong B. Validation of housekeeping genes for gene expression studies in an ice alga Chlamydomonas during freezing acclimation. Extremophiles. 2012;16(3):419–25. pmid:22527038
  16. 16. Kowalczyk N, Rousvoal S, Hervé C, Boyen C, Collén J. RT-qPCR Normalization Genes in the Red Alga Chondrus crispus. PLoS One. 2014;9(2):e86574. pmid:24498277
  17. 17. Hong S-Y, Seo P, Yang M-S, Xiang F, Park C-M. Exploring valid reference genes for gene expression studies in Brachypodium distachyon by real-time PCR. BMC Plant Biol. 2008;8(1):112. pmid:18992143
  18. 18. Iskandar HM, Simpson RS, Casu RE, Bonnett GD, Maclean DJ, Manners JM. Comparison of reference genes for quantitative real-time polymerase chain reaction analysis of gene expression in sugarcane. Plant Mol Biol Report. 2004;22(4):325–37.
  19. 19. Czechowski T, Stitt M, Altmann T, Udvardi MK, Scheible W-R. Genome-wide identification and testing of superior reference genes for transcript normalization in Arabidopsis. Plant Physiol. 2005;139(1):5–17. pmid:16166256
  20. 20. Dekkers BJW, Willems L, Bassel GW, van Bolderen-Veldkamp RP (Marieke) , Ligterink W, Hilhorst HWM, et al. Identification of Reference Genes for RT–qPCR Expression Analysis in Arabidopsis and Tomato Seeds. Plant Cell Physiol. 2012;53(1):28–37. pmid:21852359
  21. 21. Souček P, Pavlů J, Medveďová Z, Reinöhl V, Brzobohatý B. Stability of housekeeping gene expression in Arabidopsis thaliana seedlings under differing macronutrient and hormonal conditions. J Plant Biochem Biotechnol. 2017;26(4):415–24.
  22. 22. Nicot N, Hausman J-F, Hoffmann L, Evers D. Housekeeping gene selection for real-time RT-PCR normalization in potato during biotic and abiotic stress. J Exp Bot. 2005;56(421):2907–14. pmid:16188960
  23. 23. Gong H, Sun L, Chen B, Han Y, Pang J, Wu W, et al. Evaluation of candidate reference genes for RT-qPCR studies in three metabolism related tissues of mice after caloric restriction. Sci Rep. 2016;6(1):38513. pmid:27922100
  24. 24. Holmgren G, Ghosheh N, Zeng X, Bogestål Y, Sartipy P, Synnergren J. Identification of stable reference genes in differentiating human pluripotent stem cells. Physiol Genomics. 2015;47(6):232–9. pmid:25852171
  25. 25. Zhang Y, Chen D, Smith MA, Zhang B, Pan X. Selection of Reliable Reference Genes in Caenorhabditis elegans for Analysis of Nanotoxicity. PLoS One. 2012;7(3):e31849. pmid:22438870
  26. 26. Pombo-Suarez M, Calaza M, Gomez-Reino JJ, Gonzalez A. Reference genes for normalization of gene expression studies in human osteoarthritic articular cartilage. BMC Mol Biol. 2008 9(1):17. pmid:18226276
  27. 27. McNeill RE, Miller N, Kerin MJ. Evaluation and validation of candidate endogenous control genes for real-time quantitative PCR studies of breast cancer. BMC Mol Biol. 2007;8(1):107. pmid:18042273
  28. 28. Bothwell JHF, Kisielewska J, Genner MJ, McAinsh MR, Brownlee C. Ca2+ signals coordinate zygotic polarization and cell cycle progression in the brown alga Fucus serratus. Development. 2008;135(12):2173–81. pmid:18480164
  29. 29. Bouget FY, Berger F, Brownlee C. Position dependent control of cell fate in the Fucus embryo: role of intercellular communication. Development. 1998;125(11):1999–2008. 5 pmid:9570765
  30. 30. Goodner B, Quatrano RS. Fucus embryogenesis—a model to study the establishment of polarity. Plant Cell. 1993;5(10):1471–81. pmid:12271041
  31. 31. Brownlee C, Bouget FY. Polarity determination in Fucus: from zygote to multicellular embryo. Semin Cell Dev Biol. 1998;9:179–85. pmid:9599413
  32. 32. Kropf DL, Kloareg B, Quatrano RS. Cell wall is required for fixation of the embryonic axis in Fucus zygotes. Science. 1988;239(4836):187–90. pmid:3336780
  33. 33. Torode TA, Marcus SE, Jam M, Tonon T, Blackburn RS, Herve C, et al. Monoclonal antibodies directed to fucoidan preparations from brown algae. PLoS One. 2015;10(2):e0118366. pmid:25692870
  34. 34. Torode TA, Siméon A, Marcus SE, Jam M, Le Moigne MA, Duffieux D, et al. Dynamics of cell wall assembly during early embryogenesis in the brown alga Fucus. J Exp Bot. 2016;67(21):6089–100. pmid:27811078
  35. 35. Hervé C, Siméon A, Jam M, Cassin A, Johnson KL, Salmeán AA, et al. Arabinogalactan proteins have deep roots in eukaryotes: Identification of genes and epitopes in brown algae and their role in Fucus serratus embryo development. New Phytol. 2016;209:1428–41. pmid:26667994
  36. 36. Linardić M, Cokus S, Pellegrini M, Braybrook S. Growth of the Fucus embryo: insights into wall-mediated cell expansion through mechanics and transcriptomics. bioRxiv. 2020;25107.
  37. 37. Farnham G, Strittmatter M, Coelho S, Cock JM, Brownlee C. Gene silencing in Fucus embryos: Developmental consequences of RNAi-mediated cytoskeletal disruption. J Phycol. 2013;49(5):819–29. pmid:27007308
  38. 38. Hable WE. Rac1 signaling in the establishment of the fucoid algal body plan. Front Plant Sci. 2014;5(690). pmid:25540648
  39. 39. Tarakhovskaya E, Lemesheva V, Bilova T, Birkemeyer C. Early embryogenesis of brown alga Fucus vesiculosus L. is characterized by significant changes in carbon and energy metabolism. Molecules. 2017;22(1509): pmid:28891948
  40. 40. Pearson GA, Hoarau G, Lago-Leston A, Coyer JA, Kube M, Reinhardt R, et al. An expressed sequence tag analysis of the intertidal brown seaweeds Fucus serratus (L.) and F. vesiculosus (L.) (Heterokontophyta, Phaeophyceae) in response to abiotic stressors. Mar Biotechnol. 2010;12(2):195–213.
  41. 41. Jueterbock A, Kollias S, Smolina I, Fernandes JMO, Coyer JA, Olsen JL, et al. Thermal stress resistance of the brown alga Fucus serratus along the North-Atlantic coast: Acclimatization potential to climate change. Mar Genomics. 2014;13:27–36. pmid:24393606
  42. 42. Smolina I, Kollias S, Jueterbock A, Coyer JA, Hoarau G. Variation in thermal stress response in two populations of the brown seaweed, Fucus distichus, from the Arctic and subarctic intertidal. R Soc open Sci. 2016;3(1):150429. pmid:26909170
  43. 43. Rugiu L, Panova M, Pereyra RT, Jormalainen V. Gene regulatory response to hyposalinity in the brown seaweed Fucus vesiculosus. BMC Genomics. 2020;21(1):42. pmid:31931708
  44. 44. Mota CF, Engelen AH, Serrao EA, Coelho MAG, Marbà N, Krause-Jensen D, et al. Differentiation in fitness-related traits in response to elevated temperatures between leading and trailing edge populations of marine macrophytes. PLoS One. 2018;13(9):e0203666. pmid:30212558
  45. 45. Lago-Lestón A, Mota C, Kautsky L, Pearson GA. Functional divergence in heat shock response following rapid speciation of Fucus spp. in the Baltic Sea. Mar Biol. 2010;157(3):683–8.
  46. 46. Pearson G, Serrão EA, Leonor Cancela M. Suppression subtractive hybridization for studying gene expression during aerial exposure and desiccation in fucoid algae. Eur J Phycol. 2001;36(4):359–66.
  47. 47. de Boer ME, de Boer TE, Mariën J, Timmermans MJ, Nota B, van Straalen NM, et al. Reference genes for QRT-PCR tested under various stress conditions in Folsomia candida and Orchesella cincta (Insecta, Collembola). BMC Mol Biol. 2009;10(1):54.
  48. 48. Dong M, Zhang X, Chi X, Mou S, Xu J, Xu D, et al. The validity of a reference gene is highly dependent on the experimental conditions in green alga Ulva linza. Curr Genet. 2012;58(1):13–20. pmid:22205301
  49. 49. Basu S, Sun H, Brian L, Quatrano RL, Muday GK. Early embryo development in Fucus distichus is auxin sensitive. Plant Physiol. 2002;130:292–302. pmid:12226509
  50. 50. Le Bail A, Billoud B, Kowalczyk N, Kowalczyk M, Gicquel M, Le Panse S, et al. Auxin metabolism and function in the multicellular brown alga Ectocarpus siliculosus. Plant Physiol. 2010;153:128–44. pmid:20200071
  51. 51. Bogaert KA, Blommaert L, Ljung K, Beeckman T, De Clerck O. Auxin function in the brown alga Dictyota dichotoma. Plant Physiol. 2018;pp.01041.2018. pmid:30420566
  52. 52. Creis E, Delage L, Charton S, Goulitquer S, Leblanc C, Potin P, et al. Constitutive or inducible protective mechanisms against UV-B radiation in the brown alga Fucus vesiculosus? A Study of Gene Expression and Phlorotannin Content Responses. PLoS One. 2015;10(6):e0128003. pmid:26030665
  53. 53. Gu C, Chen S, Liu Z, Shan H, Luo H, Guan Z, et al. Reference gene selection for quantitative real-time PCR in Chrysanthemum subjected to biotic and abiotic stress. Mol Biotechnol. 2011;49(2):192–7. pmid:21416201
  54. 54. Han X, Lu M, Chen Y, Zhan Z, Cui Q, Wang Y. Selection of reliable reference genes for gene expression studies using Real-Time PCR in tung tree during seed development. PLoS One. 2012;7(8):e43084. pmid:22912794
  55. 55. Cuevas JC, López-Cobollo R, Alcázar R, Zarza X, Koncz C, Altabella T, et al. Putrescine is involved in Arabidopsis freezing tolerance and cold acclimation by regulating abscisic acid levels in response to low temperature. Plant Physiol. 2008;148(2):1094–105. pmid:18701673
  56. 56. Jain M, Nijhawan A, Tyagi AK, Khurana JP. Validation of housekeeping genes as internal control for studying gene expression in rice by quantitative real-time PCR. Biochem Biophys Res Commun. 2006 Jun 30;345(2):646–51. pmid:16690022
  57. 57. Hoogewijs D, Houthoofd K, Matthijssens F, Vandesompele J, Vanfleteren JR. Selection and validation of a set of reliable reference genes for quantitative sod gene expression analysis in C. elegans. BMC Mol Biol. 2008;9:9. pmid:18211699
  58. 58. Xiao L, Zhang L, Yang G, Zhu H, He Y. Transcriptome of protoplasts reprogrammed into stem cells in Physcomitrella patens. PLoS One. 2012;7(4). pmid:22545152
  59. 59. Li M-Y, Wang F, Jiang Q, Wang G-L, Tian C, Xiong A-S. Validation and Comparison of Reference Genes for qPCR Normalization of Celery (Apium graveolens) at Different Development Stages. Front Plant Sci. 2016;7:313. pmid:27014330
  60. 60. Pinto F, Pacheco CC, Ferreira D, Moradas-Ferreira P, Tamagnini P. Selection of Suitable Reference Genes for RT-qPCR Analyses in Cyanobacteria. PLoS One. 2012;7(4):e34983. pmid:22496882
  61. 61. Anderson RA. Algal culturing techniques. Vol. 53, Elsevier Academic Press. 2005. 596 p.
  62. 62. Apt KE, Clendennen SK, Powers DA, Grossman AR. The gene family encoding the fucoxanthin chlorophyll proteins from the brown alga Macrocystis pyrifera. Mol Gen Genet. 1995;246(4):455–64. pmid:7891659
  63. 63. Perkins J, Kohl M. NormqPCR: Functions for normalisation of RT-qPCR data. 2013;1–23.