A serine sensor for multicellularity in a bacterium

  1. Arvind R Subramaniam
  2. Aaron DeLoughery
  3. Niels Bradshaw
  4. Yun Chen
  5. Erin O’Shea
  6. Richard Losick  Is a corresponding author
  7. Yunrong Chai  Is a corresponding author
  1. Harvard University, United States
  2. Howard Hughes Medical Institute, Harvard University, United States
  3. Northeastern University, United States
5 figures, 1 table and 1 additional file

Figures

Figure 1 with 2 supplements
Switching synonymous serine codons in sinR affects biofilm formation.

(A) Regulatory circuit controlling biofilm formation in B. subtilis. (B) Top: Serine codon usage in the sinR coding sequence. Number within parenthesis indicates the frequency of the corresponding codon in sinR. Bottom: Average serine codon usage across 4153 protein-coding sequences in the B. subtilis genome. Number within parenthesis indicates the relative frequency of each codon in the genome. (C) Colony morphology for the wild-type strain and the indicated sinR synonymous variants grown on solid biofilm-inducing medium. Three TCA codons in the wild-type sequence of sinR were switched to each of the other five serine codons. The wild-type (WT) sinR sequence was replaced by the sinR synonymous mutant at the native sinR locus of the strain 3610. (D) SinR protein level during entry into biofilm formation (OD600 = 2) measured using an anti-SinR antibody that also cross-reacts with SlrR, a protein that is 85% identical to SinR. Western blot against the RNA polymerase subunit SigA was used as the loading control. Whole cell lysates were loaded at different dilutions (indicated as X, X/2, and X/3). (E) Densitometry of SinR bands in (D) after normalization by SigA. (F) Top panel: Western blot against SinR and SlrR using anti-SinR antibody. Bottom panel: Densitometry ratio of the SlrR and SinR bands in the top panel. Error bars represent standard error over three replicate Western blots. The SlrR/SinR ratio for each blot was normalized such that the wild-type strain had a ratio of 1. (G) Matrix gene expression monitored using a PepsAlacZ transcriptional reporter inserted at the chromosomal amyE locus. β-galactosidase activity was measured at OD600 = 2 in liquid biofilm-inducing medium. Error bars represent standard error of three measurements.

https://doi.org/10.7554/eLife.01501.003
Figure 1—figure supplement 1
sinR coding sequence.

The three TCA codons (switched in Figure 1) are highlighted in red. The three TCC codons and the two AGC/AGT codons (switched in Figure 1—figure supplement 2) are highlighted in green and blue respectively. The remaining serine codons are shown in yellow.

https://doi.org/10.7554/eLife.01501.004
Figure 1—figure supplement 2
Effect of TCC and AGC/AGT synonymous substitutions in the sinR gene on colony morphology and biofilm reporter activity.

(A) Colony morphology for the indicated sinR synonymous variants grown on solid biofilm-inducing medium. Either three TCC codons or two AGC/AGT (AGY) codons in the wild-type sequence of sinR were switched to remaining serine synonymous codons. The wild-type (WT) sinR sequence was replaced by the sinR synonymous variant at the native sinR locus of the strain 3610. Colony morphology of the wild-type strain is shown in Figure 1. (B and C) Matrix gene expression monitored using a PepsAlacZ transcriptional reporter inserted at the chromosomal amyE locus. Strains were grown in liquid biofilm-inducing medium and β-galactosidase activity was measured at an OD600 = 2. Error bars represent standard error of three measurements. The synonymous variants highlighted in red do not follow the hierarchy between TCN and AGC/AGT codons seen for the six TCA synonymous variants in Figure 1.

https://doi.org/10.7554/eLife.01501.005
Figure 2 with 3 supplements
Entry into biofilm formation is accompanied by codon-specific increase in ribosome density.

Genome-wide median ribosome density and total mRNA density at 61 sense codons (excluding start and stop codons) (A) during exponential phase growth (OD600 = 0.6), and (B) during stationary phase when biofilm formation is induced (OD600 = 1.4). The six serine (red) and two cysteine (green) codons are highlighted. Genome-wide ribosome density and total mRNA density were measured by deep-sequencing of ribosome-protected mRNA fragments and total mRNA fragments respectively, of a B. subtilis 3610 derivative (ΔepsH) grown in liquid biofilm-inducing medium.

https://doi.org/10.7554/eLife.01501.006
Figure 2—figure supplement 1
Computational workflow for deep-sequencing data analysis.

All steps outlined here were performed in Bash and Python programming languages. For further details on individual steps, see ‘Materials and methods’.

https://doi.org/10.7554/eLife.01501.007
Figure 2—figure supplement 2
Increase in ribosome density downstream of Shine-Dalgarno-like trinucleotide sequences.

Median ribosome density across all protein coding sequences was computed for the 60 nt region around each of six Shine-Dalgarno-like trinucleotide sequences (Li et al., 2012) for the exponential phase sample (left-hand panel) and the biofilm entry sample (right-hand panel).

https://doi.org/10.7554/eLife.01501.008
Figure 2—figure supplement 3
Context independence of ribosome and mRNA densities during biofilm formation.

Each gene was conceptually divided into two equal halves and the ribosome density and mRNA density was computed separately for codons located either in the first half (left-hand panel) or in the second half (right-hand panel) of each gene. All other analysis steps were identical to those in Figure 2.

https://doi.org/10.7554/eLife.01501.009
Figure 3 with 1 supplement
Serine starvation reduces translation speed and inhibits SinR synthesis in a codon-specific manner.

(A and B) Three sinR synonymous variants were synthesized with 10 serine codons switched to AGC, TCA or TCG. The variants were expressed as SinR-YFP fusions from the amyE locus under the control of a lac promoter in a 3610-ΔserA serine auxotroph strain growing in serine-limited medium. Black arrow around 300 min indicates the onset of serine starvation caused by depletion of exogenously-added serine in the growth medium. Cell density (A) and the corresponding SinR-YFP protein level (B) were monitored using a 96-well spectrophotometer. (C) Genome-wide median ribosome density for 61 sense codons (excluding start and stop codons) during serine starvation (vertical axis) and serine-rich growth (horizontal axis) of a serine auxotrophic strain. (D) Fold-change in average ribosome density for individual genes upon biofilm entry (vertical axis) or serine starvation (horizontal axis). Genes that were preferentially up-regulated at least 10-fold upon biofilm entry in comparison to serine starvation are highlighted in red (68 genes, Table 1). Only genes with a minimum of 100 ribosome profiling reads in at least one of the samples were included in this analysis (1724 genes) and the reported log2 fold-changes are median-subtracted values across this gene set.

https://doi.org/10.7554/eLife.01501.010
Figure 3—figure supplement 1
Addition of excess serine or cysteine blocks pellicle formation by B. subtilis.

Single amino acids were added at 300 µg ml−1 to liquid MSgg medium. Biofilm formation of 3610 was assayed visually by pellicle formation at the air-liquid interface 48 hr after inoculation. Serine and cysteine were found to block pellicle formation out of all 20 amino acids tested.

https://doi.org/10.7554/eLife.01501.011
Serine codon bias of biofilm-regulated genes reflects their expression under serine starvation.

(A) Relative serine codon fraction in genes for nucleotide biosynthesis (pyrAA, purB), lactate dehydrogenase (ldh) and a sporulation regulator (spo0A). Numbers in parentheses indicate the number of serine codons in each gene. Relative fraction of serine codons across the B. subtilis genome is shown for comparison. (B) Fold-change (expressed in log2 units) in average ribosome density upon biofilm entry for the four genes shown in A. (C) Colony morphology of a wild-type strain and two spo0A synonymous variants grown on solid biofilm-inducing medium. Seven AGC/AGT codons in wild-type spo0A were replaced by either 7 TCC codons or 3 TCC and 4 TCG codons and inserted at the chromosomal spo0A locus. Both the wild-type spo0A and the synonymous spo0A variants were inserted with a downstream selection marker. (D) Left: Codon Adaptation Index (CAI) for the four genes shown in A. Right: Distribution of CAI values for 4153 protein-coding sequences of B. subtilis.

https://doi.org/10.7554/eLife.01501.013
Author response image 1

Sequencing read density around six serine codons, calculated as a median over the 3 genes with the highest ribosome density (left) or the 10 genes with the highest ribosome density (right). The sequencing read density for each gene was normalized by the total number of sequencing reads for that gene (same procedure as in Figure 2).

Tables

Table 1

B. subtilis genes that have greater than 10-fold difference in expression ratio between biofilm formation and serine starvation

https://doi.org/10.7554/eLife.01501.012
GeneABCDEFFunction
albA9.35−0.693210129175144antilisterial bacteriocin (subtilosin) production protein
alsD6.380.9534138453135alpha-acetolactate decarboxylase
alsS6.961.7976461386396acetolactate synthase
cah2.08−2.74253152541486295S-deacylase
ctc3.78−0.73327220632225750S ribosomal protein L25
cydA5.36−3.18234472040860cytochrome bd ubiquinol oxidase subunit I
cydB9.6−1.889337412946cytochrome bd ubiquinol oxidase subunit II
cysK1.52−2.21157602212089912580cysteine synthase
gcvPA0.97−2.38150014401001255glycine dehydrogenase subunit 1
gcvPB1.05−2.28202920571224335glycine dehydrogenase subunit 2
gspA4.54−0.751111261135106glycosyl transferase (general stress protein)
iseA2.24−1.141117257925211516inhibitor of cell-separation enzymes
lctP6.62−2.962300924744L-lactate permease
ldh3.47−3.88281015,2541732157L-lactate dehydrogenase
maeN4.08−1.611481226347151Na+/malate symporter
mccA3.23−2.72759349438678cystathionine beta-synthase
metE2.72−2.1333,02710647332,23197605-methyltetrahydropteroyltriglutamate/homocysteine S-methyltransferase
mgsR4.72−1.53163209920293stress transcriptional regulator
mtlA3.92−0.22126934104119PTS system mannitol-specific transporter subunit IICB
mtnA1.84−2.79252244253501671methylthioribose-1-phosphate isomerase
mtnD1.72−1.634095660031821361acireductone dioxygenase
mtnK2.54−2.64429012,19658661250methylthioribose kinase
nasD6.21−0.652177872635536assimilatory nitrite reductase subunit
nasE6.050.3834110663109assimilatory nitrite reductase subunit
rbfK3.74−2.571329868143497RNA-binding riboflavin kinase
sboA7.770.6412112,895256530subtilosin A
sboX8.260.1921312876115bacteriocin-like product
ssuA2.09−2.336917682877236aliphatic sulfonate ABC transporter binding lipoprotein
ssuB2.57−2.0924167004779242aliphatic sulfonate ABC transporter ATP-binding protein
ssuC2.17−2.1637848362692206aliphatic sulfonate ABC transporter permease
ssuD2.2−2.1912,96129,1352812817alkanesulfonate monooxygenase
tcyJ3.25−3.261046485642759sulfur-containing amino acid ABC transporter binding lipoprotein
tcyK3.79−3.55281519,0871095124sulfur-containing amino acid ABC transporter binding lipoprotein
tcyL3.33−3.05855422338762sulfur-containing amino acid ABC transporter permease
tcyM3.54−2.97185910,55658199sulfur-containing amino acid ABC transporter permease
tcyN3.38−2.74336317,1491119223sulfur-containing amino acid ABC transporter ATP-binding protein
ureA4.360.78124125277176urease subunit gamma
ycgL0.99−3.0151250027846hypothetical protein
ycgM2.46−1.353910517893proline oxidase
ycgN2.45−1.6928247512675561-pyrroline-5-carboxylate dehydrogenase
ycnJ0.75−2.6316813812126copper import protein
ydaG4.140.497060480150general stress protein
ydbL3.07−0.312941210130139hypothetical protein
yeaA1.34−2.5911213813931hypothetical protein
yezD2.52−4.1714540634025hypothetical protein
yitJ3.01−2.43278510,99046871153bifunctional homocysteine S-methyltransferase/5,10-methylenetetrahydrofolate reductase
yjbC3.67−0.57104647248222thiol oxidation management factor; acetyltransferase
yjnA1.5−2.819721342790149hypothetical protein
yoaB2.28−2.26220652442862792negatively charged metabolite transporter
yoaC2.92−1.89120044591436513hydroxylated metabolite kinase
yrhB2.92−2.93380614,0961904333cystathionine beta-lyase
yrrT2.97−3.12546208944368AdoMet-dependent methyltransferase
ytlI1.66−3.220631814120LysR family transcriptional regulator
ytmI3.34−3.27345217,1731640226N-acetyltransferase
ytmO3.4−2.88386620,0071179213monooxygenase
ytnI3−2.6352213,770867189redoxin
ytnJ3.14−2.8610,64545,9972495456monooxygenase
ytnL3.56−2.451281737135486aminohydrolase
ytnM3.5−2.6454225,2021264277transporter
yuaF1.89−1.748715715863membrane integrity integral inner membrane protein
yvzB0.95−2.4812511816840flagellin
yxaL3.54−0.4110696077442442membrane associated protein kinase
yxbB3.7−0.01108685118155S-adenosylmethionine-dependent methyltransferase
yxeK0.86−2.72270224061073216monooxygenase
yxeL1.29−2.9443752520235acetyltransferase
yxeM0.87−2.57300326921047233ABC transporter binding lipoprotein
yxeP1.75−2.2425774246736207amidohydrolase
yxjH2.02−1.824162824338111432methyl-tetrahydrofolate methyltransferase
  1. A—median-subtracted log2 fold-change: biofilm/exponential-phase, B—median-subtracted log2 fold-change: serine starvation/serine rich, C—raw counts: biofilm entry, D—raw counts: exponential phase, E—raw counts: serine rich, F—raw counts: serine starvation.

Additional files

Supplementary file 1

Lists of strains, plasmids, and primers.

https://doi.org/10.7554/eLife.01501.014

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Open citations (links to open the citations from this article in various online reference manager services)

Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)

  1. Arvind R Subramaniam
  2. Aaron DeLoughery
  3. Niels Bradshaw
  4. Yun Chen
  5. Erin O’Shea
  6. Richard Losick
  7. Yunrong Chai
(2013)
A serine sensor for multicellularity in a bacterium
eLife 2:e01501.
https://doi.org/10.7554/eLife.01501