Tropheryma whipplei Twist: a human pathogenic Actinobacteria with a reduced genome

Abstract

The human pathogen Tropheryma whipplei is the only known reduced genome species (<1 Mb) within the Actinobacteria [high G+C Gram-positive bacteria]. We present the sequence of the 927303-bp circular genome of T. whipplei Twist strain, encoding 808 predicted protein-coding genes. Specific genome features include deficiencies in amino acid metabolisms, the lack of clear thioredoxin and thioredoxin reductase homologs, and a mutation in DNA gyrase predicting a resistance to quinolone antibiotics. Moreover, the alignment of the two available T. whipplei genome sequences (Twist vs. TW08/27) revealed a large chromosomal inversion the extremities of which are located within two paralogous genes. These genes belong to a large cell-surface protein family defined by the presence of a common repeat highly conserved at the nucleotide level. The repeats appear to trigger frequent genome rearrangements in T. whipplei, potentially resulting in the expression of different subsets of cell surface proteins. This might represent a new mechanism for evading host defenses. The T. whipplei genome sequence was also compared to other reduced bacterial genomes to examine the generality of previously detected features. The analysis of the genome sequence of this previously largely unknown human pathogen is now guiding the development of molecular diagnostic tools and more convenient culture conditions.

Figure 1

(A) Circular representation of the T. whipplei Twist genome (upper panel), and the alignment of the Twist and TW08/27 genomes (lower panel). The origin of replication was predicted to be near the dnaA gene, based on the conservation of dnaA-dnaN-recF gene cluster and the change of AT-skew signal. One leading strand displays a pronounced excess in T vs. A and a slight excess of G vs. C. The location of the oriC was verified by Southern-blot hybridization with a digoxigenin-labeled oriC probe onto the Not I restriction profile. In the upper panel, the outermost (1st) circle indicates the nucleotide positions. 2nd circle: The two chromosomal segments, one of which exhibits an inversion (see text). 3rd and 4th circles: The ORF locations on the plus and minus strands, respectively. Functional categories are color-coded (see Fig. 1B). 5th and 6th circles: tRNAs. 7th circle: The locations of three rRNAs are indicated by black arrows. 8th circle: The repeat locations for the two largest families. The most internal circle shows the AT skew (A-T/A+T) computed with a sliding window size of 10 kb. The lower panel shows the detailed structure of the chromosomal inversion (red arrows), flanked by repeated sequences coding for WND-domains (green arrows). Black arrows: WiSP protein-coding genes at the extremities of the inverted segments. (B) Linear view of the complete circular genome of T. whipplei. Arrows = genes, with color-coded functional categories; small red flags = TRNAs; open arrows = other RNAs; open boxes = repeats belonging to the two major categories.

Figure 1

(A) Circular representation of the T. whipplei Twist genome (upper panel), and the alignment of the Twist and TW08/27 genomes (lower panel). The origin of replication was predicted to be near the dnaA gene, based on the conservation of dnaA-dnaN-recF gene cluster and the change of AT-skew signal. One leading strand displays a pronounced excess in T vs. A and a slight excess of G vs. C. The location of the oriC was verified by Southern-blot hybridization with a digoxigenin-labeled oriC probe onto the Not I restriction profile. In the upper panel, the outermost (1st) circle indicates the nucleotide positions. 2nd circle: The two chromosomal segments, one of which exhibits an inversion (see text). 3rd and 4th circles: The ORF locations on the plus and minus strands, respectively. Functional categories are color-coded (see Fig. 1B). 5th and 6th circles: tRNAs. 7th circle: The locations of three rRNAs are indicated by black arrows. 8th circle: The repeat locations for the two largest families. The most internal circle shows the AT skew (A-T/A+T) computed with a sliding window size of 10 kb. The lower panel shows the detailed structure of the chromosomal inversion (red arrows), flanked by repeated sequences coding for WND-domains (green arrows). Black arrows: WiSP protein-coding genes at the extremities of the inverted segments. (B) Linear view of the complete circular genome of T. whipplei. Arrows = genes, with color-coded functional categories; small red flags = TRNAs; open arrows = other RNAs; open boxes = repeats belonging to the two major categories.

Figure 2

Comparative analysis of the number of genes present in each functional category as defined by the COG database. For the sake of comparing bacterial genomes of largely different sizes, a percentage graph is provided in Suppl. Fig. S2.

Figure 3

Predicted amino acid metabolisms of T. whipplei based on the M. tuberculosis metabolisms described in KEGG. Metabolic pathways were predicted to be lost for nine amino acids (pink ovals). Additional deficiencies were predicted for other seven amino acids (blue ovals). Metabolic pathways for three amino acids (green ovals) were retained. Enzymes for the alanine (yellow oval) biosynthesis were not identified in both M. tuberculosis and T. whipplei. Enzymatic steps predicted to be present in both M. tuberculosis and T. whipplei are shown in black. Enzymatic steps present only in M. tuberculosis or in T. whipplei are shown in red and green, respectively. Enzymatic steps absent in both M. tuberculosis and T. whipplei are shown with dashed black lines.

Figure 4

Thioredoxin and glutaredoxin systems. Proteins in pink rectangles are predicted to be present in M. tuberculosis, for which only remote homologs are detected in T. whipplei. Glutaredoxin gene is absent in both T. whipplei and M. tuberculosis (see text). 2G-SH and 2G-S represent reduced and oxidized forms of glutathione, respectively.

Figure 5

Typical examples of anomalous phylogenies exhibited by T. whipplei genes. Trees were constructed using CLUSTAL W alignments and the neighbor-joining method (Thompson et al. 1994). Bootstrap values are indicated when larger than 80%. Other examples are provided in Supplemental Figure S3.