The nematode homologue of Mediator complex subunit 28, F28F8.5, is a critical regulator of C. elegans development

Sequence analysis

The UniProtKB (http://www.uniprot.org) and NCBI (https://www.ncbi.nlm.nih.gov) databases were searched with BLAST, PSI-BLAST (Altschul et al., 1997), HHblits (Remmert et al., 2011), and HHpred (Söding, Biegert & Lupas, 2005) programs. The protein sequences were identified with their UniProtKB identifiers and the nucleotide sequences with their NCBI ones. The sequences were aligned with T-coffee (Notredame, Higgins & Heringa, 2000; Di Tommaso et al., 2011) and PROMALS (Pei & Grishin, 2007; Pei et al., 2007; Pei, Kim & Grishin, 2008). The secondary structure predictions were performed with PSIPRED (Jones, 1999; Cuff & Barton, 2000; McGuffin, Bryson & Jones, 2000). Multiple sequence alignments were displayed and analyzed with Jalview (Clamp et al., 2004).

RNA isolation and cDNA synthesis

RNA and cDNA were prepared as described (Zima et al., 2015) with modifications. Cultured nematodes were collected in water and pelleted by centrifugation for 5 min at 200 × g and 4 °C. The excess of water was removed and the pellet was frozen at −80 °C. For the isolation of RNA, the pellet was quickly melted and dissolved in 300 μl of resuspension buffer (10 mM Tris–HCl; 10 mM EDTA, 5% 2-mercaptoethanol; 0.5% SDS; pH 7.5). After adding 8 μl of proteinase K (20 mg/ml), the sample was mixed and incubated 1 h at 55 °C. RNA was isolated by phenol–chloroform extraction and ethanol precipitation. The obtained RNA was incubated with RQ1 DNase (Promega, Fitchburg, WI, USA) and purified again by phenol–chloroform extraction and ethanol precipitation. Complementary DNA (cDNA) was prepared with SuperScript III (Invitrogen, Carlsbad, CA, USA) using random hexamers.

Strains, transgenic lines and genome editing

The Caenorhabditis elegans Bristol N2 strain was used whenever not specifically stated and maintained as described (Brenner, 1974).

KV3: (8418)—heterozygous animals carrying one edited disrupted allele of F28F8.5 (P_F28F8.5 (V:15573749)::gfp::let858(stop)::SEC::F28F8.5—edited F28F8.5 disrupted by gfp and self-excising cassette (SEC)) and one WT allele of F28F8.5. This line segregates mutant animals.

KV4: (8419)—edited F28F8.5 carrying gfp::F28F8.5 in its normal genomic position (P_F28F8.5(V:15573749)::gfp::F28F8.5 on both alleles.

Preparation of P_{F28F8.5(400 bp)}::F28F8.5::gfp

For preparation of transgenic lines encoding F28F8.5::GFP from extrachromosomal arrays under regulation of endogenous promoter, we used the PCR fusion-based technique (Hobert, 2002). Primers 7886 and 7888 were used for amplification of the genomic sequence of F28F8.5 (consisting of approximately 400 bp of the predicted promoter region preceding the coding region of F28F8.5). The gene encoding GFP was amplified from the pPD95.75 vector with primers 6232 and 6233. The complete construct was amplified with primers 7887 and 6234. The resulting fusion construct contains the 3′ UTR from pPD95.75 (originally from the unc-54 gene). The PCR mixture was injected into the gonads of young adult hermaphrodite animals together with marker plasmid pRF4. The sequences of all primers used in the paper are in Supplemental Information.

Genome editing

Lines with edited genomes were prepared from wild type N2 animals using the CRISPR/Cas9 system as described (Dickinson et al., 2013, 2015; Ward, 2015; Dickinson & Goldstein, 2016). Using this strategy, the F28F8.5 gene was edited by insertion of a construct including the coding sequence of GFP and a SEC containing the sqt-1(d) gene (a visible selection marker leading to a Rol phenotype), hs::Cre (heat shock inducible Cre recombinase) and hygR (hygromycin resistance) genes. The sgRNA sequence was targeted near the start of the coding sequence for the F28F8.5 gene using a modified pJW1219 plasmid (Addgene, Cambridge, MA, USA) as the Cas9 vector (pMA007); it was prepared by PCR with primers 8403A and 8333 and used in a concentration of 50 ng/μl for microinjections. The plasmid pMA007 was co-injected with the rescue repair template plasmid based upon modified pDD282 vector (pMA006) in a concentration of 10 ng/μl and with three markers (see below). The repair template plasmid pMA006 was prepared in two steps. First the plasmid pMA005 was prepared from gDNA of F28F8.5 (containing both repair arms) and amplified by PCR with primers 8404 and 8405 and cloned into pCU19 backbone. The plasmid pMA005 was subsequently modified—the FP-SEC segment was added and the CRISPR/Cas9 site was altered to protect against Cas9 attack. The linear PCR product of pMA005 was prepared using primers 8406 and 8407 with overlapping regions for Gibson assembly (New England BioLabs, Ipswich, MA, USA). The primer 8406 was prepared with alternate codons for protection against CRISPR/Cas9 site. Linear insert of FP-SEC was prepared by PCR from pDD282 plasmid (Addgene, Cambridge, MA, USA) with primers 8408 and 8409. Primers were prepared with overlapping parts for cloning into linear pMA005 plasmid by Gibson assembly and the final rescue plasmid pMA006 was prepared. Plasmids pGH8 (10 ng/μl), pCFJ104 (5 ng/μl), and pCFJ90 (2.5 ng/μl) (Addgene, Cambridge, MA, USA) were used as fluorescent co-injection markers. After microinjections the population of nematodes were grown for three days at 25 °C and hygromycin (Invitrogen) was added in a final concentration of 250 μg/ml. After three days integrated nematodes were selected according to the rolling phenotype and loss of extrachromosomal arrays.

Using this strategy, we obtained a heterozygous line (KV3) with a disrupted F28F8.5 gene with an inserted gfp regulated by the endogenous promoter of F28F8.5 in one allele and one WT allele. This line segregated homozygous animals for P_F28F8.5:: F28F8.5::gfp (edited F28F8.5 with SEC—P_F28F8.5 (V:15573749)::gfp::let858(stop)::SEC::F28F8.5) with disrupted F28F8.5 on both alleles and expressing GFP under the regulation of the endogenous promoter. Animals of this line were clearly distinguishable by their developmental phenotypes, weak expression of GFP in the cytoplasm and the presence of rol marker. These animals were sterile and had severe developmental defects (see Results). The genotypes were confirmed by single worm PCR of representative animals after their microscopic analysis (with primers 8398 and 8414).

The excision of the SEC was achieved by a 4 h heat shock at 34 °C. The line KV4 was obtained: animals with both alleles carrying the edited F28F8.5 gene in the form of gfp::F28F8.5 in its normal genomic position (edited F28F8.5 with gfp tagged to the N—terminus—P_F28F8.5(V:15573749)::gfp::F28F8.5).

The presence of the knock-in of gfp was confirmed by single nematode PCR with primers 7887 and 8454, 8398 and 8454. The PCR products were purified and sequenced with primers 8455 and 8456. PCR was done by REDTaq ReadyMix PCR reaction (Sigma-Aldrich, St. Louis, Missouri, USA) or by Phusion High-Fidelity DNA Polymerase (New England Biolabs, Ipswich, MA, USA). During the maintenance of the heterozygous line KV3, animals with one edited disrupted F28F8.5 allele and one allele with edited F28F8.5 after self-excision of SEC were also generated (recognizable by the Rol phenotype, expression of gfp in nuclei and lack of developmental phenotypes). Schemes for genome editing are accessible in Files S4–S7.

Downregulation of gene expression by RNA interference

For RNAi done by microinjections, F28F8.5 cDNA was prepared from total cDNA using primers 7889 and 7890. The plasmid pPCRII(Topo) (Invitrogen, Carlsbad, CA, USA) containing F28F8.5b cDNA was linearized using restriction enzymes NotI/SacI. The dsRNA was prepared by in vitro transcription using SP6/T7 Riboprobe® in vitro Transcription Systems (Promega, Madison, WI, USA) from opposing promoters synthesizing complementary single stranded RNA (ssRNA) for both strands of F28F8.5 cDNA and its complementary strand. After in vitro transcription (∼2 h) equal volumes of sense and antisense RNA were mixed, incubated at 75 °C for 10 min and slowly cooled to room temperature during 30 min. Control RNAi was prepared from the promoter region of nhr-60 as previously described (Šimečková et al., 2007) and repeated with dsRNA prepared using the vector L4440. Vectors used for preparation of dsRNA were linearized and transcribed using T7 RNA polymerase. The dsRNA concentration was measured using a UV spectrophotometer and diluted to the concentration of ∼2 μg/μl that was used for injections (Tabara et al., 1999; Timmons, Court & Fire, 2001; Vohanka et al., 2010).

Microinjections

Microinjections of plasmids, DNA amplicons or dsRNA into gonads of young adult hermaphrodites were done using an Olympus IX70 microscope equipped with a Narishige microinjection system (Olympus, Tokyo, Japan). The plasmids were injected into the gonads of young adult hermaphrodites as described (Tabara et al., 1999; Timmons, Court & Fire, 2001; Vohanka et al., 2010).

Microscopy

Fluorescence microscopy and Nomarski optics microscopy were done using an Olympus BX60 microscope equipped with DP30BW CD camera (Olympus, Tokyo, Japan). Animals were analyzed on microscopic glass slides with a thin layer of 2% agarose and immobilized by 1 mM levamisole (Sigma-Aldrich, St. Louis, MO, USA). Confocal microscopy of live homozygous animals with edited F28F8.5 expressing gfp::F28F8.5 was performed using an inverted Leica SP8 TCS SMD FLIM system equipped with a 63 × 1.2 NA water immersion objective, a pulsed white light laser (470–670 nm), AOBS and two internal hybrid single photon counting detectors, and operated by Leica Application Suite X program (Leica Microsystems, Wetzlar, Germany). The GFP fluorescence was excited at a wavelength of 488 nm and the emitted light was simultaneously recorded in two spectral ranges (Channel 1—495 nm to 525 nm, Channel 2—525 nm to 580 nm; the two channel setup was used to help resolve between spectrally different autofluorescence and GFP fluorescence signals).

Fluorescence-lifetime imaging microscopy

For FLIM acquisitions the single photon counting signal from the internal hybrid detectors, acquired during confocal acquisitions, was simultaneously processed by HydraHarp400 TCSPC electronics (PicoQuant, Berlin, Germany) and information about the arrival times of all photons was stored to a hard-drive in TTTR data format. TTTR is freely accessible at https://www.picoquant.com/images/uploads/page/files/14528/technote_tttr.pdf. Data structure, program description and user instructions are also freely accessible at https://github.com/PicoQuant/PicoQuant-Time-Tagged-File-Format-Demos/blob/master/PTU/Matlab/Read_PTU.m. The signal from both time synchronized channels was added up. The false color scale (1–3 ns) is based on the average photon arrival time, with blue color representing short lifetime and red color long lifetime fluorescence.

Single nematode PCR

Single animal PCR was used for verification of all transgenic lines. Following the microscopy examination, selected animals were removed from microscopic slides and transferred into caps of PCR tubes with 4 μl of solution of Proteinase K (20 mg/ml) diluted 1:333 in Barstead Buffer (resulting in Barstead Lysis Buffer which consists of 50 mM KCl, 10 mM Tris pH 8.3, 2.5 mM MgCl₂, 0.45% (v/v) NP40 (Nonidet P-40), 0.45% (v/v) Tween-20, 0.01% (w/v)). Proteinase K was diluted immediately before use as a 20 mg/ml stock solution which was kept on ice and diluted to final working solution at a concentration of 60 μg/ml. The tube was sealed in bottom-up position and the sample transferred to the bottom of the tube by centrifugation. The tube was frozen for 10 min at −70 °C. Next, the tube was heated for 1 h at 65 °C and additional 15 min at 95 °C. The resulting sample was used immediately for amplification of DNA by PCR or stored at −80 °C before further analysis. Similarly, genomic DNA was prepared from selected nematode culture plates and used for further screening by PCR and sequencing.

The resulting precipitated DNA was dissolved in 10 μl of deionized water and used for amplification by PCR using primers outside the edited genomic regions. Specificity of amplification was confirmed by DNA sequencing.

Similarly, homozygous animals with edited F28F8.5 (with gfp inserted in front of the F28F8.5 START codon) were analyzed by single worm PCR with primers 7887 (sense primer) and 7890 or 8454 (antisense primers).

Quantitative RT-PCR

For quantitative RT-PCR, the technique described by Ly, Reid & Snell (2015) was used with modifications. For assessment of the level of expression of F28F8.5 from homozygous animals with the edited disrupted gene, five adult homozygous mutant animals recognized by the phenotype and the same number of young WT hermaphrodites with minimum number of formed embryos were manually harvested and collected in separate Eppendorf tubes. Reverse transcription was done using the Maxima H Minus cDNA synthesis kit (Thermo Fischer, Waltham, MA, USA) as recommended by manufacturer. Universal probe library and primers designed with the help of ProbeFinder Assay Design Software were used and qPCR was run on LightCycler 2.0 purchased from Roche (Roche, s.r.o. Prague, Czech Republic). An average of three sample cDNAs and three control cDNAs were analyzed (twice in duplicates and one time as single experiments), all containing the same amount of RNA for RT for each experiment. The expression of F28F8.5 was normalized to ama-1 and the values obtained in homozygous mutant animals with disrupted F28F8.5 gene were compared to values obtained in control WT N2 animals.

Binding studies

Binding studies were done as described (Kostrouch et al., 2014) with modifications. The coding region of mdt-6 was amplified using primers 8292 and 8293 from cDNA prepared from mixed stages C. elegans cultures and cloned into pTNT vector (Promega, Madison, WI, USA, amplified with primers 8277 and 8278) using the Quick Ligation Kit (New England Biolabs, Ipswich, MA, USA) and expressed in the rabbit reticulocyte TNT-system (Promega, Madison, WI, USA). The in vitro transcribed protein was labeled using ³⁵S Methionine (Institute of Isotopes, Budapest, Hungary). F28F8.5 coding sequence (amplified using 8255 and 8256 primers with 15 bp overhangs for insertion into the vector) was cloned into pGEX-2T vector ((Amersham Pharmacia Biotech, Amsterdam, UK), amplified with primers 8253 and 8254) using the GeneArt Seamless PLUS Cloning and Assembly Kit (Thermo Fisher Scientific, Waltham, MA, USA), transformed into BL21 Escherichia coli cells and the production of protein was induced by isopropyl β-d-1-thiogalactopyranoside (IPTG) (Sigma-Aldrich, St. Louis, MO, USA). The Mediator subunit MDT-30 was amplified from mixed stages C. elegans cDNA with sense primer 8302 and reverse primer 8527 (containing FLAG sequence), cloned into pET28a(+) vector ((Addgene, Cambridge, MA, USA), amplified with primers 8519 and 8520) using the Quick Ligation Kit (New England Biolabs), transformed into BL21 E. coli cells and induced by IPTG. The lysate from bacteria producing His₆-MDT-30-FLAG was used directly or purified on HiTrap Chelating HP column (GE Healthcare, Chicago, IL, USA). Proteins produced by the TNT system or bacterial lysates of bacteria transformed with FLAG labeled Mediator subunits were incubated with glutathione–agarose (Sigma-Aldrich, St. Louis, MO, USA) adsorbed with equal amounts of GST or GST-F28F8.5. Radioactively labeled proteins were detected using TRI-CARB 1600TR, Liquid Scintillation Analyzer (Packard, Meriden, CT, USA).

The resulting samples (labeled proteins bound to GST- or GST-F28F8.5) were separated by polyacrylamide gel electrophoresis. ³⁵S-MDT-6 was visualized by autoradiography and subsequently, the gel containing radioactively labeled protein was localized using superimposed autoradiograms, excised and the radioactivity determined in the scintillation detector. FLAG-labeled MDT-30 was determined by Western blot using an anti-FLAG antibody (monoclonal anti-FLAG, M2 (Sigma-Aldrich)) and quantified densitometrically by ImageJ computer program (https://imagej.nih.gov/ij/download.html) (Schneider, Rasband & Eliceiri, 2012).

Identification of the closest homologue of vertebrate Mediator complex subunit 28 in C. elegans

To identify the C. elegans homologue of MED28, we queried protein databases with curated SwissProt sequences from UniProtKB. They comprised several mammalian and insect proteins (e.g., human MED28_HUMAN and D. melanogaster MED28_DROME). The more sensitive profile-to-profile HHblitz and HHpred algorithms provided hits to a C. elegans annotated protein F28F8.5a and b with highly significant E-values. According to Wormbase (WS248), two protein isoforms are produced from the F28F8.5 gene, isoform a with the length of 200 amino acids and isoform b that has a two amino acid insertion at position 20 of the N-terminal evolutionarily non-conserved region. The best results were obtained when pre-aligned vertebrate and insect MED28 homologues were used as query in three iterations (E < 10⁻⁴⁸ and the probability of true positive >99.99%). When the pre-aligned nematode sequences homologues to F28F8.5 were used to query profiles of human or Drosophila sequences in reciprocal searches, MED28 proteins were obtained with equally significant scores. BLAST and PSI-BLAST searches in their standard settings were not able to reveal a significant hit (E < 10⁻³); the only nematode hit was a Trichinella spiralis protein (E5RZQ1). However, when the searches in protein databases were limited to sequences from Ecdysozoa with Insecta excluded (conservative inclusion threshold E < 10⁻⁶) in the first two iterations and then continued in the complete database of sequences from all species in the subsequent iterations, the final hits of F28F8.5 included human and Drosophila MED28. PSI-BLAST with T. spiralis query sequence in database limited to Ecdysozoa in the first two iterations provided both human and Drosophila MED28 and F28F8.5 in one run (E < 10⁻⁸). We concluded from these searches that F28F8.5 is a homologue of MED28 and very likely its previously unrecognized orthologue.

All PSI-BLAST MED28 homologues possess variable N- and C-termini of 3–80 amino acids showing no conservation across Metazoa. This conservation is loose even just within Drosophilae or Caenorhabditae sequences. Only the central core of about 110 amino acids is preserved in metazoan evolution. Figure 1 shows a sequence alignment of this conserved core of selected MED28 homologues. All sequences are predicted to fold into three helices forming a putative coiled coil fold (UniProt annotation). Submitting the alignment shown in Fig. 1 to HHpred for 3D structure recognition reveals a structural fold of yeast MED21 (PDB identifier 1ykh_B). It is indeed a three-helix coiled coil forming a heterodimer with MED7. It can be expected that MED28 forms a very similar fold interacting with a yet to be determined subunit of the MED complex.

F28F8.5 is a nuclear as well as a cytoplasmic protein

Information available in WormBase suggests that the F28F8.5 gene can be expressed as both an individual and multigene transcript, located as the last gene in a four gene operon that is both SL-1 and SL-2 trans-spliced. To determine the intracellular localization of F28F8.5, we edited the F28F8.5 gene using CRISPR/Cas9 technology. We inserted the gene coding for GFP directly in front of the first codon. The arrangement used in our experiment (based on Dickinson et al. (2013, 2015), Ward (2015) and Dickinson & Goldstein (2016)) employed a SEC that was added after gfp. This strategy initially created a disrupted F28F8.5 gene and putative null allele that can be detected by expression of GFP alone regulated by the endogenous promoter elements of F28F8.5. We found that only heterozygous animals could be propagated due to the sterility of homozygotes tagged in this manner. Assuming this tag is not deleterious to the expression of other genes in the operon, this result suggests that F28F8.5 is an essential gene.

After removal of the SEC from this edited F28F8.5 gene induced by heat shock (visualized by continuous expression of GFP::F28F8.5 fusion protein and loss of the Rol phenotypic marker), the endogenous locus had an N-terminus GFP-tagged F28F8.5 gene that we maintained as homozygous animals, demonstrating this edited allele is fully functional. Note that both known protein isoforms of F28F8.5 (a and b) would be tagged on their N-terminus with GFP by this method.

The GFP::F28F8.5 pattern was ubiquitous, both nuclear and cytoplasmic from embryos to adults (Fig. 2). Prominent nuclear localization was found in oocytes, zygotes, larvae, and adults. Cells with clear nuclear accumulation of GFP::F28F8.5 included epidermal, intestinal, pharyngeal, uterine and vulval muscle cells (Fig. 2). The gonad expressed gfp::F28F8.5 and mitotic as well as meiotic nuclei accumulated GFP::F28F8.5 protein (Fig. 2).

Selected animals were analyzed by confocal microscopy for determination of subcellular distribution of GFP::F28F8.5. Scanning through several focal planes revealed signal in the GFP excitation/emission range in nuclei as well as in the cytoplasm of embryos, all larval stages and adults (Fig. 3). Structures resembling gut granules were also strongly positive in the GFP recording mode. In order to distinguish between GFP-specific fluorescence and autofluorescence, we applied FLIM with an expectation that autofluorescence (such as that from gut granules) is likely to produce a signal with a short fluorescence lifetime opposed to GFP-specific fluorescence. Structures such as gut granules were clearly detected (Fig. 3, panels O, Q, S, T and U, blue color) while fluorescence with a longer lifetime expected for GFP::F28F8.5 was detected in the germline, in oocytes and embryos and in most somatic nuclei of larvae as well as adult animals (Fig. 3, panels O, Q, S, T and U, red and yellow colors).

We also generated transgenic lines encoding F28F8.5::GFP from extrachromosomal arrays consisting of an endogenous internal F28F8.5 promoter regulating a fusion gene with gfp attached to F28F8.5 on its C-terminal end. As with the N-terminally tagged F28F8.5, F28F8.5::GFP showed both nuclear and cytoplasmic localization. As expected for an extrachromosomal transgene, the expression of F28F8.5::gfp was not detected in the germline. This reporter was expressed in embryos starting at the twofold stage and continued throughout development (Fig. S1). We did notice that F28F8.5::gfp was expressed in the excretory canal cell (Figs. S1M, S1N, S1P and S1Q), a pattern not observed with the endogenously edited GFP-tagged gene.

F28F8.5 regulates development

To achieve loss-of-function, RNAi was used to downregulate F28F8.5 expression. Analysis of 2,567 progeny of 17 young adult hermaphrodites inhibited for F28F8.5 function by microinjection of dsRNA into the syncytial gonad revealed that F28F8.5 is essential for proper development (Fig. 4). From the total progeny, 1,127 animals were affected (44%) exhibiting embryonic and larval arrest and a range of less severe phenotypes, including defective molting, protruding vulvae that often burst, male tail ray defects (Fig. 4), and uncoordinated (Unc) movement. In contrast with this, control young adult N2 hermaphrodites injected with control dsRNA showed embryonic arrest in less than 2% of progeny (seven hermaphrodites injected, total progeny observed 1,066, embryonic arrest found in 19 embryos).

Complete loss of F28F8.5 that occurred in homozygous animals with both edited disrupted alleles of the F28F8.5 gene (that are found among the progeny of heterozygous animals carrying one edited disrupted allele and one WT allele) resulted in defective development that was most pronounced in late larval stages. The phenotypes included a dumpy phenotype (Dpy) (Fig. 5C), irregular gut, severely defective growth of the gonad with signs of defect in directional growth (Figs. 5E and 5M) and Pvul phenotype (Figs. 5K and 5L). Most animals had darker gut cells than controls of the same age. The gonad did not develop fully in most animals (Figs. 5C–5L) and often contained empty spaces that were prevalent in some animals, leading to the formation of larvae with optically thin, empty-like tissue in the position of the gonad. The gonads contained foci of irregular tissue with an uncharacteristic appearance. Tissue defects were also visible in extragonadal locations, especially in the place of excretory canals. Body defects were also observed in the position of the uterus that was not properly formed and the spermatheca that was not identifiable in a large proportion of animals.

Estimation of the level of F28F8.5 expression in homozygous mutant animals (originating from maternal load or from SEC self-excision (Dickinson et al., 2015)) from three experiments indicated that mutants with disrupted F28F8.5 had the level of expression about 17 times lower compared to the levels found in WT controls (File S8).

Heterozygous hermaphrodites carrying one edited disrupted allele of F28F8.5 and one WT allele were grossly normal and produced viable embryos. Unlike in homozygous animals carrying the excised SEC allele, the GFP fluorescence was mostly cytoplasmic and most nuclei were not showing accumulation of GFP. In some embryos, however, the nuclei accumulated GFP indicating probable spontaneous SEC self-excision (Fig. 6).

Analysis of progeny of the heterozygous strain KV3 revealed differences compared to the expected Mendelian segregation of phenotypes. Animals with one edited disrupted F28F8.5 allele and one edited F28F8.5 allele with excised SEC were detected. They were recognizable by the Rol phenotype, expression of GFP in nuclei and lack of developmental phenotypes. This genotype was supported by PCR amplification of genomic regions from single nematodes and obtained pattern of amplified DNA fragments. These lines were not stable and were not preserved.

F28F8.5 interacts with Mediator complex subunits

To determine if F28F8.5 could be part of the Mediator complex in C. elegans, we explored its ability to interact with previously identified Mediator subunits. We expressed ³⁵S-labeled MDT-6, part of the “head” module where MED28 is located, in rabbit reticulocyte lysate and assayed its binding to bacterially expressed GST-F28F8.5 or to GST only. As shown in Figs. 7A and 7B, a strong interaction (∼7.7-fold enrichment) was detected between MDT-6 to F28F8.5 that exceeded that seen with GST alone. We also assayed for interaction between GST-F28F8.5 and MDT-30, but we were unable to obtain a satisfactory ³⁵S-Methionine labeled protein in the rabbit reticulocyte system. Therefore, we expressed MDT-30 containing a FLAG sequence inserted at the C-terminus and a His₆ sequence positioned at the N-terminus. After expression in bacteria and purification on a nickel column, we found that the MDT-30-FLAG bound F28F8.5 preferentially (∼2.5-fold enrichment) in comparison to GST alone, as revealed by Western blot using an anti-FLAG antibody (Figs. 7C and 7D).