Two membrane-bound transcription factors regulate expression of various type-IV-pili surface structures in Sulfolobus acidocaldarius

Sulfolobus acidocaldarius strains, plasmids and growth conditions

All strains used in this study are described in Table S1. Strains were grown essentially as described, using basal Brock medium (pH 3.5) supplemented with 0.1% NZ-amine, 0.2% sucrose and 10 µg/ml Uracil (Wagner et al., 2012). Plasmids and their creation are described in Table S2. Primers are listed in Table S3.

Escherichia coli strains, plasmids and growth conditions

All E. coli strains used in this study are described in Table S1. All strains were grown in LB medium supplemented with the respective antibiotics. Plasmids were generated as described in Table S2 using primers described in Table S3.

SDS-PAGE and Western-blot analysis

Total membranes or purified protein samples were supplemented with 5× SDS-loading buffer (10% (w/v) SDS, 300 mM Tris/HCl pH 6.8, 500 mM DTT, 50% (v/v) Glycerol, 0.04% (w/v) bromphenole blue) and subjected to SDS-PAGE analysis according to the method of Laemmli using 11% gels. For Western-blot analysis, proteins were transferred onto PVDF membranes (Roche diagnostics) using the semi-dry method. His-HRP antibody (Abcam) diluted 1:10,000 in PBST was used to detect ArnR and ArnR1. PBST was prepared by diluting a 10 × PBS stock (composition: 580 mMNa₂HPO₄ × 2H₂O_, 170 mM NaH₂Po₄ × H₂O, 680 mM NaCl, 0.05% Tween20 pH 7.3) to 1 × PBS using distilled H₂O and 0.05% Tween20 was added. Chemiluminescent signals were detected as described (Hoffmann et al., 2016).

Expression and purification of ArnR and ArnR1

Overexpression of ArnR from pSVA2543 and ArnR1 from pSVA2538 was performed essentially as described (Studier, 2014) using the E. coli BL21 (DE3) derivative strain E. coli OverExpress(tm)C43(DE3) (Lucigen). Cells were grown for approximately 48 h to OD₆₀₀ of 12. Membranes were isolated essentially as described (Bischof et al., 2016), using buffer A (20 mM Tris/HCl, pH 8, 300 mM NaCl). solubilization was performed for 2 h at 4 °C, using five mg/ml total membranes in solubilization buffer (buffer A supplemented with 10 mM Imidazole, 2% n-Dodecyl-β-d-maltopyranoside (DDM) (Glycon)). Solubilized protein was isolated from residual membranes using ultracentrifugation (400,000 × g, 4 °C, 1 h) and purified using one ml of His-Select nickel affinity gel resin (Sigma) equilibrated in solubilization buffer. Solubilized protein was applied to the resin and the flow-through was collected. After washing of the column with 5 × 1 ml of wash buffer (buffer A supplemented with 20 mM Imidazol, 0.5% DDM (Glycon)), bound target protein was eluted using 5 × 500 µl buffer A supplemented with 0.03% DDM and 150 mM Imidazole. Eluted ArnR protein was desalted to buffer A supplemented with 0.03% DDM using a PD10 column (GE Healthcare, Chicago, IL, USA), according to the manufacturer’s protocol. Eluted ArnR1 protein was further purified using a Heparin-HP column (GE Healthcare, Chicago, IL, USA) on an Äkta purifier system (GE Healthcare, Chicago, IL, USA). Therefore, fractions containing ArnR1 were diluted to 20 mM NaCl in buffer A supplemented with 0.03% DDM and loaded on a five ml HisTrap HP at one ml/min flow rate. After washing of the column with 20 mM Tris/HCl pH 8, 20 mM NaCl, 0.03% DDM for at least five column volumes, ArnR1 was eluted in two ml fractions using 20 mM Tris/HCl, pH 8, 500 mM NaCl, 0.03% DDM. ArnR as well as ArnR1 was concentrated using ultrafiltration in a 100 kDa cut-off Amicon (Merck Millipore, Burlington, MA, USA). Protein concentration was determined using BCA assay (Serva), according to the manufacturer’s protocol. The Analysis of the oligomeric state of purified ArnR and ArnR1 was performed using Blue-Native PAGE analyses as described (Claeys, Geering & Meyer, 2005) using NativeMark™ unstained protein standards (Life Technologies, Inc., Carlsbad, CA, USA).

Microscale thermophoresis (MST)

For MST measurements, double-stranded AlexaFluor647-labeled DNA was generated by annealing two primers (Table S3). Therefore, primers were mixed in equal amounts (five mM final concentration) in 20 mM Tris/HCl pH 8, 200 mM NaCl and 0.05% (w/v) DDM, heated to 90 °C and then cooled to 20 °C in steps of 2 °C/10 s and annealing was monitored by analyzing primers before and after annealing on 15% TBE Gels. To determine the binding affinity of ArnR, 2.5 nM labeled, double-stranded DNA was titrated with increasing concentrations of ArnR (0.156–18 µM). The measurements were performed at 24 °C and 80% LED power. Samples were incubated for 10 min before measurements and centrifuged at 16,000 rpm in a tabletop centrifuge (Eppendorf, Hamburg, Germany). The laser on and off times were adjusted to 25 s and 5 s, respectively. Measurements were performed on a NanoTemper Monolith NT.115 instrument in hydrophilic‐treated capillaries, and analyzed using NT analysis software version 1.4.27 (NanoTemper Technologies GmbH, Munich, Germany). The data of two independent experiments performed in duplicate were used to calculate binding affinities. Therefore, fluorescence was normalized (ΔF) by subtracting the lowest measured fluorescence value determined within each experiment from all measured Fluorescence values. Obtained curves were fitted using non-linear regression, with one site-specific binding according to (y = F_max*X/(K_D + X)).

Quantitative reverse transcription-polymerase chain reaction (qRT-PCR)

Total RNA samples were isolated from shaking cultures before (0 h) and after 2 h of starvation. Samples were isolated as described (Lassak et al., 2012). TRIzol reagent (Invitrogen, Carlsbad, CA, USA) was used for total RNA isolation following manufacturer’s instructions. Digestion of gDNA and preparation of cDNA was performed using Quantinova Reverse Transcription Kit (Qiagen, Hilden, Germany), according to the manufacturer’s protocol. The gDNA removal was performed for 3 min, and the Reverse Transcription reaction for 7 min. Quantitative reverse transcriptase (qRT-PCR) analysis of aapF and upsX was performed as described (Haurat et al., 2017) using Magnetic Induction Cycler (Bio Molecular Systems, Upper Coomera, QLD, Australia), 2× qPCRBIOSyGreen Mix Lo-ROX (Nippon genetics, Dueren, Germany) according to the manufacturer’s protocol, in combination with primers listed in Table S3. Three biological replicates and two technical replicates were assayed. Gene expression of aapF and upsX was normalized to their expression at time point 0 h (before starvation) and depicted in log2-fold change. Alcohol dehydrogenase (saci_1690) was used as housekeeping gene as transcript levels of this gene are not affected by nutrient limiting conditions (Bischof et al., 2019).

ArnR and ArnR1 form oligomeric structures

To characterize ArnR and ArnR1, both regulators were heterologously expressed in E. coli and purified to homogeneity (Fig. 2). To optimize translation rates for the respective proteins, ArnR and ArnR1 encoding genes were codon optimized for E. coli (Table S4) and synthetically manufactured (Genscript, Piscataway, NJ, USA). Codon optimized arnR and arnR1 were expressed with an N-terminal His10-tag under control of an arabinose inducible promoter (Geertsma & Dutzler, 2011; Valla & Lale, 2014) in E. coli OverExpress C43(DE3) cells using autoinduction medium (Studier, 2014).

ArnR and ArnR1 were solely detected in the E. coli membrane fraction (Fig. 2A), which was used further for protein purification via Nickel-affinity chromatography coupled to desalting of eluted protein samples (Fig. 2A, middle). The identity of ArnR and ArnR1 was confirmed by immunoblot analysis using His-specific antibodies (Fig. 2B). To determine the oligomeric state of purified ArnR and ArnR1, Blue-Native PAGE analysis was performed (Fig. 2C). ArnR and ArnR1 formed diverse multimeric species from monomeric to dodecameric (Fig. 2C). This is in agreement with the observation that transcription factors adopt homo-dimeric or multimeric states in their active conformation (Wolberger, 1999). The formation of various oligomeric species moreover showed that both regulators are stable and correctly folded after purification.

ArnR and ArnR1 are phosphorylated by the Serine/Threonine kinase ArnC

Expression and presumably activity of the archaellum are regulated by reversible protein phosphorylation. The two eukaryotic protein kinases ArnC and ArnD phosphorylate multiple serine and threonine residues of the archaellum repressor ArnB (Hoffmann et al., 2016). ArnC (but not ArnD) further phosphorylates ArnA, which is thought to be essential for the interaction and activity of ArnA and ArnB (Reimann et al., 2012; Hoffmann et al., 2016). Phosphoproteome analysis revealed phosphorylation of Tyr217 and Thr222 in S. acidocaldarius ArnR1 and two neighboring tyrosine residues in the HAMP-domain (Tyr154 and Tyr155) of S. solfataricus ArnR (Esser et al., 2012; Reimann et al., 2013) (Sso_ArnR, Fig. 3A). Therefore, phosphorylation assays were performed next to analyze if ArnR and ArnR1 are phosphorylated in vitro. Given the fact that the protein kinases ArnC and ArnD are involved in archaellum regulation (Hoffmann et al., 2016), they were incubated with ArnR or ArnR1 in the presence of radioactively labeled γ³²P-ATP and phosphotransfer from the kinases to the regulators was monitored. Signals corresponding to auto-phosphorylation of ArnC (Fig. 3B, lane 4) and ArnD (Fig. 3C, lane 4) was obtained. Strikingly, phosphotransfer was observed from ArnC to both ArnR and ArnR1 (Fig. 3B, second to last and last lane), while no phosphotransfer to ArnR or ArnR1 was observed from ArnD (Fig. 3C, second to last and last lane). In addition, two distinct bands were obtained for phosphorylated ArnR and ArnR1 that might correspond to phosphorylated monomeric and dimeric species of the proteins (Fig. 3B, second to last and last lane), indicating that the multimerisation observed on BN-PAGE (Fig. 2B) might have a functional relevance.

ArnR and ArnR1 bind to the flaB promoter

To understand their function as transcriptional activators of the archaellum operon, binding of full-length ArnR and ArnR1 to different flaB promoter sequences was investigated. It was recently described that the flaB promoter (pflaB) harbors two 15 bp cis-regulatory elements (ArnR box-1 and ArnR box-2), which are inverted repeats with the consensus sequence TCGAC-(N)5-GTCGA (Lassak et al., 2013) (Fig. 4A). To characterize the binding of ArnR and ArnR1 to the flaB promoter, three different Alexa Fluor 647 labeled promoter fragments were generated and subjected to MST experiments in the presence of increasing amounts of detergent-purified ArnR or ArnR1. A 41 bp fragment containing both ArnR boxes (pflaB⁴¹) as well as a fragment containing either ArnR box-1 or ArnR box-2 were used (Fig. 4A). With increasing concentrations of ArnR (Fig. 4B, left graph) or ArnR1 (Fig. 4B, right graph), binding to pflaB⁴¹ was detected as was visualized by increasing fluorescence signals (Kd ArnR: 4.6µM, Kd ArnR1: 3.31 µM) (Fig. 4B). In experiments with flaB promoter fragments containing only ArnR box-1 or box-2, an increase in fluorescence signals was not observed (Fig. 4B), demonstrating that for efficient binding both boxes are required.

ArnR and ArnR1 are involved in cross-regulation of various type-IV-pili structures

A cross-regulation between Aap-pili and the archaellum was proposed recently, as a deletion mutant of the central membrane protein-coding gene aapF of the Aap-system lead to hypermotile cells (Henche et al., 2012a). To analyze if apart from the archaellum ArnR and ArnR1 regulate the Aap- or the Ups-system of S. acidocaldarius, fragments of the aapF and upsX promoter sequences were subjected to MST analysis (Fig. 5A). Strikingly, ArnR bound to paapF (Kd 1.5 µM) and pupsX (Kd 1.78 µM) (Fig. 5B, left graph), whereas ArnR1 did not (Fig. 5B, right graph). The promoter fragment of saci_2122 (encoding a putative sugar binding protein) was used as negative control. Neither ArnR nor ArnR1 showed binding to pSaci_2122, supporting the idea that ArnR and ArnR1 exclusively regulate T4P related promoters. Taken together these results suggest that ArnR is involved in regulation of the archaellum, the Aap- and Ups-pilus of S. acidocaldarius while ArnR1 exclusively exerts its function in the archaellum.

Subsequently, we aimed to understand if there is an effect of phosphorylation of ArnR and ArnR1 on binding to pflaB, pflaB ArnR box-1, pflaB ArnR box-2, pupsX and paapF. Therefore, ArnR and ArnR1 were incubated with ArnC and subjected to MST analysis. However, ArnR as well as ArnR1 was aggregating during the experiment, which hindered further analysis.

The deletion of arnR and arnR1 affects transcription of aapF and upsX

It was recently shown, that the deletion of arnR and arnR1 negatively affects the transcription rate of flaB and reduces motility of cells (Lassak et al., 2013). Since apart from binding to pflaB ArnR also bound to paapF and pupsX in our MST analysis, the effect of arnR deletion on aapF and upsX levels was investigated in vivo. Up to now, the only known condition under which transcription of arnR is induced is nutrient depletion (Lassak et al., 2013). Therefore, a deletion mutant of arnR (S. acidocaldarius MW328) was subjected to nutrient depletion for 2 h and levels of aapF and upsX transcription were monitored relative to their transcription before starvation. The deletion mutant of arnR1 was also assayed to analyze whether the deletion of arnR1 affects transcription levels of aapF and upsX even though binding to their respective promoter regions was not detected in vitro. As depicted in Fig. 6, a log2 four-fold induction of aapF transcription after 2 h of nutrient limitation was detected in S. acidocaldarius MW001 (wild type) cells. Remarkably, aapF transcription was log2 four-fold downregulated in the arnR deletion mutant and also significantly (>0.05) reduced in the arnR1 deletion strain (Fig. 6). In S. acidocaldarius MW001 (wild type) a log2 two-fold induction of upsX was detected upon starvation, and the deletion of arnR (S. acidocaldarius MW328) led to a log2 four-fold downregulation of upsX (Fig. 6). No significant (>0.05) change of upsX transcript levels was found in the arnR1 deletion mutant (Fig. 6). This confirms our MST binding assays and shows that ArnR is involved in the regulation of the aap and ups operons. Further, as identified by qRT-PCR ArnR1 is involved in the regulation of the aap operon.