Introduction

In November 2006, Tarpey et al1 identified 22 nystagmus-causing mutations in the FERM domain containing 7 (FRMD7) genes, which resides within the Xq26-q27 interval. Since then, Self et al2 and a number of other groups have confirmed that mutations in this gene are an important cause of congenital idiopathic nystagmus (CIN). At present, the function of this gene is unknown and its expression profile poorly understood. There is some evidence that the mRNA for FRMD7 is present at low levels in most tissues, and possibly the greatest in the kidney and testis (http://symatlas.gnf.org). Using RT-PCR, Tarpey et al1 detected FRMD7 mRNA expression in human adult kidney, liver, pancreas, and at lower levels in heart and brain.1 However, in human foetal tissue, they found the transcript only in kidney. Using in situ hybridisation applied to human embryonic brain at day 56 postovulation , there was FRMD7 expression in the ventricular layer of the forebrain, midbrain, cerebellar primordium, spinal cord, and developing neural retina. At day 37, postovulation embryos, expression was restricted to the mid- and hindbrain, regions known to be involved in motor control of eye movement.1 No temporal profile or further expression study has been published in any species to date.

It has been postulated that the FRMD7 gene is likely to be involved in neuronal development.1, 2, 3 It is also known that gaining an understanding of the spatial and temporal expressions of such genes can provide some insight into complex self-organizing processes, such as mammalian central nervous system development.4 Here, we describe a detailed temporal expression study of Frmd7 in the developing brain, lung, and heart in mice.

Materials and methods

Specific pathogen-free 6-week-old outbred MF-1 mice (obtained from Harlan UK Ltd, Bicester, UK) were time-mated by the detection of a vaginal plug. This day was taken as embryonic day (ED) 0. Pregnant mice between ED10 and 20, newborn and juvenile mice were killed using a schedule 1 method (cervical dislocation). Gravid uteri were removed under sterile conditions, and embryos were killed according to schedule 1 method (neural tube dissection and cervical dislocation). Maternal adult, embryonic and postpartum lungs, hearts, and brains were dissected in a laminar flow hood. Embryos were dissected under a dissecting microscope (LEICA WILD M3Z, Wetzlar, Germany). The dissected lungs, hearts, and brains were then homogenised using a Hybaid RiboLyser Cell Disrupter (Thermo Life Sciences, Hybaid, UK) at speed settings 6.0 s and for 40 s in TRIzol (Invitrogen, Paisley, UK) and RNA extracted according to the TRIzol protocol. Developmental time points were represented by two animals from two mothers. Therefore, two independent samples were obtained from heart, brain, and lung for ED11–19 from adult female mice, and from the 1- and 8-day postdelivery (PD) pups.

RNA was extracted using TRIzol reagent, and samples were treated with DNase (Ambion, Austin, TX, USA) according to the manufacturer's instructions to remove trace contamination by genomic DNA. In all, 1 μg total RNA was reverse transcribed using reverse transcription kits (PrimerDesign Ltd, Southampton, UK) according to the manufacturers' instructions with the addition of both Oligo dT and random nonamer primers to maximise transcription. cDNA was diluted 1 : 10 in RNA/DNAase-free water and stored before use at −20°C.

The ubiquitously expressed gene Eif2s3y, located in a non-recombining region of the Y chromosome, is only expressed in male mouse tissues including the developing brain.5 Quantitative real-time PCR (RT-qPCR) amplification and melt-curve analysis for Eif2s3y were used to determine male and female gender in cDNA samples from mouse brain, heart, and lung tissues.

RT-qPCR protocol

RT-qPCR was performed to assess the relative abundance of murine Frmd7 mRNA compared with two housekeeping genes: Glyceraldehyde-3-phosphate dehydrogenase (Gapdh) and Beta-actin (Actb). Although accuracy of comparative expression analysis can be improved by using greater numbers of housekeeping genes, the greatest increase in efficacy has been shown to be in using greater than one reference gene.6 The choice of these two housekeeping genes was based on their earlier use in heart, lung, and brain tissues,7 and shows that their abundance is high and similar in these tissues.6 Furthermore, these two genes have been employed previously for murine brain studies by RT-PCR, and Gapdh has been shown to have an almost constant expression value from ED10-P0.8

RT-qPCR was carried out using a custom-made RT-PCR assay for use with SYBRgreen chemistry (PrimerDesign Ltd) in a RT thermocycler Rotor-Gene 6000 real time thermo-analyser (Corbett Life Science, Brisbane, Australia). PCR reactions were performed according to the manufacturer's instructions for 45 cycles using the kit reagents, which included a chemical hot-start Taq polymerase. The primer sequences for the Frmd7 gene were sense primer: ATGCAAGGCTTTCTGGAAGAC and an antisense primer: CGGAAACTGGAACCTTTGCTA with an amplification product of 111 nucleotides. For the control genes, commercially available primers were obtained (PrimerDesign Ltd), and these sequences are the intellectual property of PrimerDesign Ltd.

PCR reactions were performed in 100-well Gene-Discs (6001–012, Corbett Life Science) on the Rotor-Gene 6000 real time thermo-analyser (Corbett Life Science). The Rotor Gene run settings were according to the standard protocol suggested for custom qPCR assays by PrimerDesign. Briefly, the two-step amplification conditions were enzyme activation for 10 min at 95°C, followed by 45 cycles of 15 s denaturation at 95°C, 60 s annealing, and extension at 60°C; and fluorescence data collection at the end of the denaturation step. A high-resolution melt-curve analysis was also performed for each run with ramping from 60–95°C with fluorescence data collection at 0.5°C increments. Melt-curve analysis allowed the assessment of specific amplification product at the predicted melting temperature for the primer set described above. Each developmental time point was represented by two samples from separate animals, and all samples were run in duplicate with duplicate water controls (cDNA replaced by water) for each assay.

Data analyses

By monitoring the fluorescence values of each sample, it was possible to determine the quantity of PCR product on a cycle-by-cycle basis. To account for differences in background fluorescence, the Rotorgene software automatically normalised the data. Once normalisation was complete, a threshold at which fluorescence data was analysed was set. This threshold was set at a level where the rate of amplification was greatest during the exponential phase. In this experiment, the same normalised threshold fluorescence value was used in all runs and was set at a value of 0.12. The number of cycles taken for each sample to reach the threshold level is defined as the Ct-value (threshold cycle), and is taken as a measure of the abundance of cDNA target present.

Results were analysed using the ‘comparative Ct method,’ which is also known as the ‘delta-delta Ct’ method.9 In this method, values for the gene of interest are calculated relative to the controls genes expression for each sample, and then different samples are compared with respect to the relative expression of the gene of interest. For the purpose of these experiments, heart, brain, and lung samples were analysed separately and calibrated to the earliest time point sample. This method of analysis relies on a consistently efficient PCR over a range of cDNA concentrations such that each increase of 1 in Ct value is equivalent to a halving of template cDNA. For each assay, we performed a standard curve analysis to asses the PCR efficiencies over a range of template concentrations. For a PCR efficiency that is close to 100%, an efficiency measure of close to 1, and an R2 (a measure of the % data conforming to the given efficiency value) ∼0.99 is expected. Validation of the delta-delta Ct methodology was also provided by Primerdesign Ltd. Each of these assays had been tested for PCR efficiency during production and validation, and all assays were guaranteed to have a PCR efficiency close to 100%. This means that for analysis of relative mRNA expression, 1 Ct value is equivalent to a onefold difference in copy number and the delta-delta Ct method of analysis is appropriate and valid.10

We certify that all applicable institutional and governmental regulations concerning the ethical use of animals were followed during this study

Results

Initial standard curve analysis was performed to asses the validity of the delta-delta Ct method for relative expression analysis. PCR efficiency close to 100% was shown for Gapdh (R2=0.994, efficiency=1.03), Actb (R2=0.9947, efficiency=1.05), and Frmd7 (R2=0.99628, efficiency=0.92).

Frmd7 mRNA was expressed at low levels (up to 14 Ct values later than control genes) in all three tissues. In all runs, Gapdh, Actb, and Frmd7 water controls showed no amplification until later than at least three Ct values behind cDNA samples, and had low broad melts indicative of mispriming resulting in non-specific amplification curves at very late Ct values.

Expression values were calculated relative to the ED11 sample from heart tissue, and are shown in Figure 1.

Figure 1
figure 1

Relative expression of Frmd7 in murine heart lung, and brain tissues during development. Expression data are shown compared with that of ED11 heart samples, and relative to the expression of the Gapdh and Actb housekeeping genes; *=all samples were normalised to this sample. Error bars represent the SEM.

Full size image

RT-qPCR amplification and melt-curve analysis for Eif2s3y were used to determine male and female gender in cDNA samples from mouse brain, heart, and lung tissues. No significant difference of Frmd7 mRNA expression on ED18 between two male and two female brain, heart, and lung samples could be detected suggesting no sex chromosome-specific influence on Frmd7 expression.

Discussion

Our results show that in murine heart and lung, Frmd7 mRNA expression varied very little between ED11 and PD8. Interestingly, for both tissues, expression increased significantly in the adult samples following a similar pattern. In these tissues, overall expression of Frmd7 was low during the developmental stages, and in most samples, the expression was at the limit of resolution for this technique. However, no significant increase in expression of Frmd7 occurred during this critical period in murine development, and that overall expression in these tissues during development is very low and consistent with the previously reported ‘low levels of expression in most tissues’.1 Frmd7 expression in murine brain samples were similarly low at early time points. However, as detailed previously, quantitative comparison between different tissues can be unreliable because of small differences in control gene expression, and so are not presented here. Interestingly, the overall pattern of expression in the brain tissue is noticeably different to that seen in the lung and heart tissues with a marked increase in the expression at ED18.

Genes expressed during murine development exhibit functional clustering in terms of temporal and spatial expression profiles during development.8, 11 In 2005, Matsuki et al8 used an oligonucleotide-based microarray system to study the expression of 12 422 genes in mouse brain between ED11, ED15, ED18, and P0. They found significant differential expression in 1413 of these genes when normalised to the ACTB gene. These genes were then grouped into 15 clusters on the basis of cellular events involved in brain development and known/predicted gene function. Interestingly, only four of these clusters showed an overall temporal expression pattern similar to our Frmd7 findings with low expression until a significant increase at ED18. These four clusters were carbohydrate, lipid, and amino-acid metabolism (CLAM); cell adhesion and recognition; neurotrophin, hormone, growth factor, and cytokine (NHGC); and neurotransmitter and ion channel genes (NT). This is an interesting finding as the CLAM genes supply molecular components required for active neurogenesis and histogenesis; NT supply molecular components for the formation of neuronal connectivity and circuits; and NHGC genes have a role in signalling for neural differentiation and survival.8 All these processes are known to be occurring at this developmental stage, showing the efficacy of this clustering methodology and a possible role for Frmd7 in these processes in murine brain. This group subsequently selected 397 of these differentially expressed genes for which strong functional data were available. These genes were characterised into eight new clusters on the basis of their developmental function. Interestingly, the expression pattern of upregulation at ED18 or P0 was most strongly seen in two clusters. These clusters pertained to genes were involved in synapse formation and function (eg, genes coding GABA receptors and intracellular calcium channels) and axon growth and guidance-related genes (eg, Sema4f and Cdh8). Again, these results suggest a role for Frmd7 in these processes in murine brain development. However, samples were taken from whole brain, and so there is no indication of which brain regions are expressing the gene at this time. Furthermore, earlier brain tissues used in this work (ED11–∼ED14) will have included the developing optic vesicle (developing eye) and all samples of the developing primary visual cortex and connections. Therefore, no inference can be made from this study about which brain/eye structures are responsible for the peak in Frmd7 expression at ED18 or the pathophysiology of CIN. Further expression on this gene would therefore be likely to include the expression analysis of various regions of the adult mouse brain.

It is important to note that in this study, detection of Frmd7 expression was occurring at the limit of resolution for RT-qPCR. In many samples, the Ct difference between housekeeping genes and the Frmd7 gene amplification curves was ∼15 cycles. This corresponds to a difference in expression of 1:65536. At this very low level of expression, primer–dimer formation was occurred in poor samples, which were then necessarily excluded. Possibilities for improving results for such a low copy number gene include gene-specific RT-PCR and the use of fluorescent labelled probes for specific detection of the target transcript. Furthermore, relatively few samples were used per time point, and future work would benefit from a greater N-value, especially in light of the very low levels of expression detected. The final caveat to this study is that expression analysis was performed on murine samples with the possibility of differences to expression profiles in human tissues. For example, mice do not have fovea, and their eyes do not open until ∼P10/11. Therefore, the visual systems of humans and mice are structurally different, and development occurs over differing timescales. However, the role of Frmd7 in mice is likely to direct further study of this gene in humans, especially when a human FRMD7 antibody becomes available.