Linking the transcriptome to physiology: response of the proteome of Cupriavidus metallidurans to changing metal availability

Abstract

Cupriavidus metallidurans CH34 is a metal-resistant bacterium. Its metal homeostasis is based on a flow equilibrium of metal ion uptake and efflux reactions, which adapts to changing metal concentrations within an hour. At high metal concentrations, upregulation of the genes for metal efflux systems occurs within minutes. Here, we investigate the changes in the bacterial proteome accompanying these genetic and physiological events after 1.5 cell duplications, which took 3 h. To that end, C. metallidurans CH34 and its plasmid-free derivative, AE104, either were challenged with a toxic metal mix or were cultivated under metal-starvation conditions, followed by bottom-up proteomics. When metal-shocked or -starved cells were compared with their respective controls, 3540 proteins changed in abundance, with 76% appearing in one, but not the other, condition; the remaining 24% were up- or downregulated. Metal-shocked C. metallidurans strains had adjusted their proteomes to combat metal stress. The most prominent polypeptides were the products of the plasmid-encoded metal-resistance determinants in strain CH34, particularly the CzcCBA transenvelope efflux system. Moreover, the influence of antisense transcripts on the proteome was also revealed. In one specific example, the impact of an asRNA on the abundance of gene products could be demonstrated and this yielded new insights into the function of the transmembrane efflux complex ZniCBA under conditions of metal starvation.

Keywords: Cupriavidus metallidurans; metal homeostasis; metal starvation; proteomics; transenvelope efflux systems; zinc.

Graphical Abstract

The transenvelope pump ZniCBA was upregulated under metal-starvation conditions and an antisense RNA was involved in regulation of this process. This suggests a possibly novel function of these efflux systems.

Figure 1.

The abundance of antisense RNAs influenced the protein yield from the associated sense RNA. The abundance of proteins (copy number per cell) was plotted as decadic logarithm against that of the abundance of its transcript (NPKM values as published [58]). Closed circles are CH34 control cells, open circles metal-, squares EDTA-treated cells of strain CH34. Diamonds are AE104 control cells, triangles metal- and inverted triangles EDTA-treated AE104 cells. Data points were grouped according to the abundance (NPKM values as published [58]) of the respective asRNA into six groups: no asRNA (0), NPKM ≤3 (1), 3 < NPKM ≤ 10 (2), 10 < NPKM ≤ 30 (3), 30 < NPKM ≤ 100 (4), and NPKM > 100 (5). A linear curve fit was performed for the six groups (Supplementary Fig. S3) to the function lg₁₀(protein) = lg₁₀(a) + lg₁₀(b) × lg₁₀(senseRNA). Shown here is the slope lg₁₀(b) for the six asRNA groups per condition. The color code has no meaning and was used for contrast. The pale inverted triangle was a slope outside of the remaining data.

Figure 2.

Changes in abundance of proteins correlated with changes in abundance of the associated sense RNA. The 327 proteins with a significantly changed abundance following metal-stressed or starvation in CH34 and AE104 cells were plotted against the changes of the associated sense RNA. Filled symbols represent upregulated protein copy numbers, open symbols the inverse ratio Q of downregulated proteins. Circles are the comparison of metal-shocked CH34 to the control (CM0), squares metal-starved CH34 cells to the control (CE0), diamonds metal-shocked AE104 cells to the control (AM0), and triangles metal-starved AE104 cells (AE0). Colors indicate data points with significant changes (Q ≥2 for up- and 1/Q ≥2 for downregulated RNAs, D >1). CM0 red, CE0 blue, AM0 purple, and the two values for AE0 in green. Gray symbols are the results with significant changes in the proteome but not the transcriptome. The lines are linear curve fittings for the subsequent functions: CM0+, all data: lg10(Qprotein) = 0.487 ± 0.034 + 0.323 ± 0.042 × lg10(QsenseRNA), R² = 74.8% CM0+, Qs ≥2: lg10(Qprotein) = 0.258 ± 0.100 + 0.506 ± 0.077 × lg10(QsenseRNA), R² = 75.5% CM0−, all data: lg10(Qprotein) = 0.406 ± 0.017 + 0.079 ± 0.049 × lg10(QsenseRNA), R² = 19.0% CE0+, all data: lg10(Qprotein) = 0.562 ± 0.042 + 0.371 ± 0.233 × lg10(QsenseRNA), R² = 20.4% CE0−, all data: lg10(Qprotein) = 0.413 ± 0.017 + 0.003 ± 0.246 × lg10(QsenseRNA), R² = 2.2% AM0+, all data: lg10(Qprotein) = 0.459 ± 0.044 + 0.381 ± 0.049 × lg10(QsenseRNA), R² = 85.0% AM0+, Qs ≥ 2: lg10(Qprotein) = 0.358 ± 0.143 + 0.449 ± 0.098 × lg10(QsenseRNA), R² = 86.7% AM0−, all data: lg10(Qprotein) = 0.467 ± 0.032 + 0.002 ± 0.102 × lg10(QsenseRNA), R² = 0% AE0+, all data: lg10(Qprotein) = 0.500 ± 0.079 + 0.261 ± 0.176 × lg10(QsenseRNA), R² = 39.8% Upregulated items in the comparisons are indicated by a ‘+’ and green lines, downregulated (1/Q values) by a ‘−’ and gray lines. For AE0−, R² was <null and this function is not given. Functions in italic letters have regression coefficient <20%. The bold-faced equations are those for the data points with QsenseRNA ≥2 with CM0+ in red and AMO+ in purple.

Figure 3.

Changes in abundance of proteins correlated with changes in abundance of the associated sense RNA. The figure shows the same data points as that in Fig. 1 but uses another symbol and color code to group these points according to the associated changes in the sense-to-antisense ratios QS_AS. Group 1, no QS_AS ratio or QS_AS <0.25 (black circles); group 2, 0.5 < QS_AS < 1.2 (red squares); group 3, 1.2 ≤ QS_AS < 4 (blue diamonds); group 4, 4 ≤ QS_AS < 20 (magenta triangles); and group 5, QS_AS > 50 (green inverted triangles). Dashed lines come from the fitting of all values (small gray dot within the symbols), lg₁₀(Qprotein) = 0.376 ± 0.084 + 0.424 ± 0.069 × lg₁₀(QsenseRNA), R² = 72.1%, or Qprotein = 2.38 × 2.66^QsenseRNA. Data points for group 5 and two zni data points from other groups are labelled with the protein and the respective comparison.

Figure 4.

Model of Zn(II) transport under stress and starvation conditions. Under conditions of zinc (small circles) stress, Zn(II) is imported into the periplasm by porins and either immediately exported by 250 CzcCBA systems to the outside, or imported by 1555 general import systems (GIS) or 254 ZupT into the cytoplasm. Due to the flow equilibrium [25], Zn(II) is exported back into the periplasm by 5330 general export systems (GES) such as ZntA. Under zinc-starvation conditions (right hand), hypothetical zinc-containing particles (gray irregular field) might generate zinc-containing fragments or species (stars), which are also imported into the periplasm. A total of 413 ZniCBA complexes might release the zinc ion from these fragments by export to the outside. Subsequently, Zn(II) is reimported into the periplasm and further on into the cytoplasm. Below, the copy numbers of the transport proteins are shown, bold-faced when measured, otherwise the numbers are generated from the transcript abundance using the functions generated in Supplementary Fig. S3. The arrows indicate transport, lines with crosses possible downregulatory, lines with balls upregulatory effects of transport activities. Lower intensities of the colors of the transport systems indicate a lower abundance of the respective proteins.

Figure 5.

Dependence of the protein copy number on the transcriptional activities. The figure shows the RNASeq results of the chromid region ∼2 012 000 base pairs in metal-challenged (CH34_M), -starved (CH34_E), or control cells (CH34_0), transcripts of the forward strand in red, in reverse direction in blue. Arrows between the abundance plots show the position of the open reading frames and the NPKM values of the gene-specific sense RNA transcripts. Above or below the abundance plots are the position of the annotated asRNAs and their NPKM values. The header on the top gives the Rmet locus number, the gene name, the copy number of the gene product under the three conditions plus a position marker. Flags indicate transcriptional start sites (TSS), white not associated to a sigma factor yet, and red RpoD-dependent. The yellow flag indicates the published TSS zniBp [8], which was, however, no longer found in recent experiments. The TSS signal score is given adjacent to the flags, the no longer found zniBp in italics. The genes zniB, zniC, their 5′ untranslated regions and TSSs are strongly overlapping. Coordinates annotated to position 2 030 000 of the chromid sequence are from the left to the right −165 (zniC-5′UTR), −105 (3′ of zniB), −77 (old zniBp), +10 (zniCp), +16 (previously published zniC-5′UTR), +59 (new found zniBp), +64 (5′ of zniC) and +109 (zniB-5′UTR). This indicates a lively interaction between transcription initiation of zniC and zniB.