The Campylobacter jejuni helical to coccoid transition involves changes to peptidoglycan and the ability to elicit an immune response

Summary Campylobacter jejuni is a prevalent enteric pathogen that changes morphology from helical to coccoid under unfavorable conditions. Bacterial peptidoglycan maintains cell shape. As C. jejuni transformed from helical to coccoid, peptidoglycan dipeptides increased and tri‐ and tetrapeptides decreased. The DL‐carboxypeptidase Pgp1 important for C. jejuni helical morphology and putative N‐acetylmuramoyl‐L‐alanyl amidase AmiA were both involved in the coccoid transition. Mutants in pgp1 and amiA showed reduced coccoid formation, with ∆pgp1∆amiA producing minimal coccoids. Both ∆amiA and ∆amiA∆pgp1 lacked flagella and formed unseparated chains of cells consistent with a role for AmiA in cell separation. All strains accumulated peptidoglycan dipeptides over time, but only strains capable of becoming coccoid displayed tripeptide changes. C. jejuni helical shape and corresponding peptidoglycan structure are important for pathogenesis‐related attributes. Concomitantly, changing to a coccoid morphology resulted in differences in pathogenic properties; coccoid C. jejuni were non‐motile and non‐infectious, with minimal adherence and invasion of epithelial cells and an inability to stimulate IL‐8. Coccoid peptidoglycan exhibited reduced activation of innate immune receptors Nod1 and Nod2 versus helical peptidoglycan. C. jejuni also transitioned to coccoid within epithelial cells, so the inability of the immune system to detect coccoid C. jejuni may be significant in its pathogenesis.


List of Contents Supplemental Experimental Procedures
Strain construction (including table of bacterial strains and plasmids and table of primers used in this study) Whole genome sequencing Supplemental Tables: Table S1   Table S2   Table S3   Table S4A  Table S4B   Table S5   Table S6 Supplemental Figures: Figure S1. The C. jejuni transition of ∆pgp2 and ∆pgp1∆pgp2 mutant strains from a helical to coccoid form was similar to wild type.  In order to delete the amiA gene, a portion of the gene was replaced by the Cm resistance cassette (cat). Initial attempts to replace a portion of the C. jejuni amiA gene with the non-polar aphA3 kanamycin (Km) resistance cassette from pUC18-Km (Menard et al., 1993) were unsuccessful. The amiA gene was PCR amplified with iProof (Biorad) from C. jejuni 81-176 genomic DNA using primers amiA-1 and amiA-2 (3060 bp). A polyA tag was added to the PCR product and it was ligated to a commercially available pGEM-T vector (Promega). The resulting construct pEF18 was verified by PCR analysis and sequencing. Inverse PCR was performed on pEF18 with primers amiA-3 and amiA-5 (deleting nucleotides 127-1664 of amiA) and the product was ligated to a cat cassette digested out of pRY109 (Yao et al., 1993) with SmaI to form plasmid pEF18Cm. Correct insertion of the cassette in the same orientation as the amiA gene was verified by PCR with cat-2 and amiA-2 primers, as well as restriction enzyme analysis. C. jejuni 81-176 was naturally transformed with pEF18Cm and mutants were selected on MH-TV containing Cm. Mutants (designated ΔamiA) were very slow growing and appeared after 5-7 days of growth as very small colonies. They were verified by PCR analysis with amiA-6 and amiA-7.
To verify that phenotypes seen with ΔamiA were not due to polar effects on the downstream gene 1284 (CJJ81176_1284; mnmC), the downstream gene was deleted. Gene 1284 was PCR amplified from 81-176 genomic DNA with primers 1284-1 and 1284-2 (2711 bp), a polyA tail was added to the PCR product and it was cloned into pGEM-T, forming construct pEF74. The pEF74 plasmid was verified by PCR analysis and sequencing. Inverse PCR was performed on pEF74 with primers 1284-3 and 1284-4 to delete nucleotides 37-1692 of the 1842 bp gene. The inverse PCR product was digested with DpnI (to remove any native pEF74) and then ligated to the SmaI-digested cat cassette from pRY109 (Yao et al., 1993), resulting in pEF74Cm. Correct insertion of the cassette in the same orientation as the 1284 gene was verified by PCR with cat-2 and 1284-2 primers, as well as restriction enzyme analysis. C. jejuni 81-176 was naturally transformed with pEF74Cm and mutants were selected on MH-TV containing Cm after 5-7 days of growth. Mutants (designated Δ1284) were verified by PCR analysis with 1284-1 and 1284-2.
For complementation of amiA, the amiA gene was PCR amplified with amiA-8 (XbaI) and amiA-9 (XbaI) from C. jejuni 81-176 genomic DNA, digested with XbaI and cloned into the similarly digested pRRK integration vector (Karlyshev & Wren, 2005). The resulting construct (pEF49F) was verified for orientation by PCR with the amiA gene in the same orientation as the antibiotic resistance cassette encoded by the vector, and sequenced. Plasmids were inserted into C. jejuni wild type 81-176 and ΔamiA by natural transformation and transformants selected on the appropriate antibiotics. Single insertions into the rRNA spacer region were verified by PCR with primers ak233, ak234, ak235 (Karlyshev & Wren, 2005) and aphA3-2 for pRRK. Wild type phenotypes were only restored when the amiA gene was deleted in a wild type strain carrying the second copy of amiA at the rRNA locus and not in an ΔamiA strain into which the amiA gene had been inserted. To knockout amiA in 81-176+amiA, this strain was naturally transformed with both ΔamiA genomic DNA and with pEF18Cm and mutants were selected on MH-TV containing Km and Cm. PCR and restriction enzyme analysis were used to verify deletion of amiA and the presence of amiA at the rRNA locus (this strain was designated ΔamiA-c).
A double mutant in ΔamiA and Δpgp1 was constructed by deleting ΔamiA in the Δpgp1 strain. The Δpgp1 mutant strain was transformed with both ΔamiA genomic DNA and with pEF18Cm and mutants designated Δpgp1 ΔamiA were selected on MH-TV containing Km and Cm. PCR analysis was used to verify deletion of amiA.

Strain or Plasmid
Genotype, serotype or description Reference or Source C. jejuni  Wild type isolated from a diarrheic patient (Korlath et al., 1985) Δpgp1  (Apel et al., 2012) rpoA-QRT-R AGTTCCCACAGGAAAACCTA (Apel et al., 2012) Whole-genome sequencing Genomic DNA for all experiments was harvested via Wizard genomic DNA purification (Promega). Illumina libraries were prepared using the KAPA Low-Throughput Library Preparation Kit with Standard PCR Amplification Module (Kapa Biosystems, Wilmington, MA), following manufacturer's instructions except for the following changes: 750ng DNA was sheared using an M220 instrument (Covaris, Woburn, MA) in 50ul screwcap microtubes at 50 peak power, 20 duty factor, 20 o C, 200 cycles per burst and 25 seconds duration. Adapter ligated fragments were size selected to 700-800bp following Illumina protocols. Standard desalting TruSeq LT and PCR Primers were ordered from Integrated DNA Technologies (Coralville, IA) and used at 0.375 µM and 0.5 µM final concentrations, respectively. PCR was reduced to 4 cycles. Libraries were quantified using the KAPA Library Quantification Kit (Kapa), except with 10 µl volume and 90 sec annealing/extension PCR, then pooled and normalized to 4 nM. Pooled libraries were re-quantified by ddPCR on a QX200 system (Bio-Rad), using the Illumina TruSeq ddPCR Library Quantification Kit and following manufacturer's protocols, except with an extended 2 min annealing/extension time. The libraries were sequenced 2x250 bp paired end v2 on a MiSeq instrument (Illumina) at 13.5 pM, following manufacturer's protocols. The MiSeq reads were reference assembled to the genome of the background strains C. jejuni 81-76 (NC_008787) or NCTC 11168 (AL11116) using Geneious 9.1 reference assembler (Biomatters, Auckland, NZ). Table S1. The mean percentage and standard error of helical, coccoid and cells transitioning to the coccoid form in C. jejuni wild type 81-176, ∆pgp1 mutant strain, ∆pgp1 complemented strain (∆pgp1-c) and pgp1 overexpressing strain (81-176+pgp1) grown on solid media at 38 °C over 8 days (depicted graphically in Fig. 1B); C. jejuni wild type 81-176 grown in liquid culture at 38 °C at 4 days; and C. jejuni wild type 81-176 on solid media incubated at 4 °C for 29 days after an initial day of growth at 38 °C. At least three separate fields of view of approximately 200 bacteria/field of view were counted for each strain at each timepoint. Representative cells considered to be helical, coccoid or transitioning to the coccoid form are indicated by a, b or c, respectively in the DIC images in Fig. 1A.     Table S4A. Muropeptide composition of the wild-type 81-176, ∆pgp1, 81-176+pgp1 (pgp1 overexpressing strain), and ∆pgp2 strains grown for either 1 or 4 days at 38 °C (strains were grown at 38 °C unless otherwise indicated) or 29 days at 4 °C summarized in Table  1.
1 The strain designation consists of a letter denoting the series in which the sample was analyzed followed by a number denoting the sample within the series. Samples analyzed in the same batch will have identical series letters. 2 The values for the percentage of O-acetylated species do not represent the true level of O-acetylation in these strains, as most of these substitutions are lost in the standard alkaline reduction procedure used in this study to prepare the PG. These values were included to demonstrate the relative difference in O-acetylation between the samples, but actual comparisons between the samples were not made. 3 nd = not determined.  Table S4B. Muropeptide composition of the C. jejuni ∆amiA, 81-176+amiA (amiA overexpressing strain), and ∆amiA∆pgp1 strains grown for either 1 or 4 days at 38 °C summarized in Table 2 (the muropeptide composition of the additional strains shown in Table 2 are reported in Table S2A). The strain designation consists of a letter denoting the series in which the sample was analyzed followed by a number denoting the sample within the series. Samples analyzed in the same batch will have identical series letters. 2 The values for the percentage of O-acetylated species do not represent the true level of Oacetylation in these strains, as most of these substitutions are lost in the standard alkaline reduction procedure used in this study to prepare the PG. These values were included to demonstrate the relative difference in O-acetylation between the samples, but actual comparisons between the samples were not made. 3 nd = not determined. Table S5. Comparison of the muropeptide dipeptides, tripeptides and tetrapeptides (from Tables 1 & 2) of the C. jejuni wild-type 81-176, ∆pgp2, 81-176+pgp1 (pgp1 overexpressing strain), ∆pgp1, ∆amiA, and ∆amiA∆pgp1 strains grown for 1 and 4 days at 38 °C with varying amounts of coccoid cells present in the population. The percent change between Day 1 and Day 4 of the muropeptides of the same strain are shown below the Day 4 values. 1 The strain designation consists of a letter denoting the series in which the sample was analyzed followed by a number denoting the sample within the series. Samples analyzed in the same batch will have identical series letters.    Figure S1. The C. jejuni transition of ∆pgp2 and ∆pgp1∆pgp2 mutant strains from a helical to coccoid form was similar to wild type. A, DIC microscope images taken over 8 days. C. jejuni wild type 81-176, ∆ pgp2 mutant strain, ∆pgp2 complemented strain (∆pgp2-c) and ∆pgp1∆pgp2 double mutant strain grown on solid media at 38 °C to follow the transition to the coccoid form over time. B, the percentage of helical, coccoid and cells transitioning to the coccoid form as determined from DIC images such as those shown in A. At least three separate fields of view of approximately 200 bacteria/field of view were counted for each strain at each timepoint and this was carried out in triplicate. Representative cells considered to be helical, coccoid or transitioning to the coccoid form are indicated by a, b or c, respectively in the DIC images in A. There were no statistically significant differences in coccoid formation between wild type and ∆pgp2, ∆pgp2-c or ∆pgp1∆pgp2 at each timepoint using the unpaired Student's t-test. No differences in growth characteristics were observed for any of the strains by growth curve analysis in liquid and on solid media (data not shown). This domain is suggested to mediate the periplasmic or extracellular targeting of bacterial proteins to the cell envelope. Residues marked in blue indicate the highly conserved Amidase_3 (pfam01520) domain. This domain is found in N-acetylmuramoyl-L-alanine amidases and is involved in cleaving the amide bond between N-acetylmuramoyl and L-amino acids in bacterial PG. Residues highlighted in red are the active site residues, with the first 3 of the 4 site residues predicted to be involved in metal binding. The active site residues are highly conserved in C. jejuni and H. pylori homologs. Conserved domains and active site residues were identified with NCBI conserved domain searches.