Supplemental text

H. somnus 129Pt and H. influenzae Rd both had genes involved in D-xylose and xylitol degradation (Table S5). The genes for D-xylose degradation to D-xylulose 5-phosphate, xylA (HS_0587, HI1112) and xylB (HS_0588, HI1113), were not present in H. ducreyi 35000HP. H. somnus 129Pt and H. influenzae Rd also had most of the components of the xylitol degradation pathway, which converts xylitol to xylulose 5-phosphate. The first step in this pathway, the conversion of xylitol to L-xylulose, is catalyzed by a xylitol 4-dehydrogenase/L-xylulose reductase, which exists in Erwinia uredovora (10) and many eukaryotes. However, there were no gene or protein sequences available from E. uredovora to use in blast searches. Blastx of the L-xylulose reductase gene sequence from Trichoderma reesei (Hypocrea jecorina) hit reductase/dehydrogenase proteins in H. somnus 129Pt (HS_0167), H. ducreyi 35000HP (HD0708) and H. influenzae Rd (HI0155) at 29% – 30% amino acid identity. So, the xylitol 4-dehydrogenase activity may be present in these organisms. The next reaction, conversion of L-xylulose to L-xylulose 5-phosphate is catalyzed by L-xylulokinase (9), which was present in H. somnus 129Pt (HS_0770) and H. influenzae Rd (HI1027), but not in H. ducreyi 35000HP. However, H. somnus 129Pt, H. ducreyi 35000HP and H. influenzae Rd all had rpe (HS_0175, HS_0057, HD1929, HI0566), encoding ribulose phosphate 3-epimerase, which converts D- and L-xylulose 5-phosphate to D- or L-ribulose 5-phosphate (42). There was a cluster of genes in H. somnus 129Pt (HS_0763 to HS_0773) involved in ribose and xylitol metabolism. Although H. influenzae Rd had all of these genes, they were not organized on the chromosome in the same way. As mentioned above, H. influenzae Rd had only 1 copy of the rbs operon (HI0501 – HI0506), which was not adjacent to the sgbK/lyx, sgbH, sgbU and araD genes (HI1024 – HI1027), as one set of rbs genes was in H. somnus 129Pt.

Like H. influenzae Rd, H. somnus 129Pt had the fuc operon (HS_1446 – HS_1451; HI0610 – HI0615), so it can probably metabolize fucose. H. ducreyi 35000HP did not have these genes. Both H. influenzae Rd and H. somnus 129Pt had an extra copy of the fucA gene (HS_0014, HI1012), which encodes fucose-1-P aldolase. Both H. somnus 129Pt and H. ducreyi 35000HP had the mannose utilization genes manAZYX (HS_0605 – HS_0609; HD0765 – HD0768), while H. influenzae Rd did not. However, the genes that flanked the mannose utilization genes in H. somnus 129Pt and H. ducreyi 35000HP were different. H. ducreyi 35000HP had a complete set of genes encoding the mannose PTS and mannose 6-phosphate isomerase (HD0765-HD0768).

H. somnus 129Pt had the galactitol utilization operon (HS_1140 – HS_1146), but H. influenzae Rd and H. ducreyi 35000HP did not. H. influenzae Rd and H. ducreyi 35000HP did have the pflA and pflB genes that flanked the galactitol operon in H. somnus 129Pt, although they were on the opposite strand in reverse order. H. somnus 129Pt had the glucitol/sorbitol utilization operon (HS_0675 – HS_0679) flanked by genes purE (opposite strand HS_0672) and pepP (HS_0682), H. ducreyi 35000HP had the flanking genes on opposite strands with other genes in-between (HD1419 – HD1423; aspC, hypothetical, purK, hypothetical, purK). H. influenzae Rd had purE, purK, aspC in a row (HI1615 – HI1617), but pepP (HI0816) was located approximately 800 genes away.

H. somnus 129Pt had a mannitol utilization operon consisting of mtlADR (HS_1250 – HS_1252). This was like the mtl operon in E. coli, which consists of the mtlA, mtlR, and mtlD genes that encode the mannitol transporter (enzyme IICBAmtl), a transcriptional regulator, and mannitol-1-phosphate dehydrogenase (46). H. ducreyi 35000HP and H.influenzae Rd did not have these genes. HD1859 may be a mannitol/fructose specific IIA component of a PTS, but H. somnus 129Pt did not have it.

H. somnus 129Pt had a gene that was similar to E. coli celB (HS_0437), which encodes a cellobiose-specific IIC component, but did not have genes encoding the complete cellobiose PTS. H. ducreyi 35000HP and H. influenzae Rd did not have any cellobiose PTS genes.

Because H. somnus in culture has been reported to use trehalose and maltose (16), we looked for genes involved in trehalose and maltose uptake and degradation. The enzyme II of the trehalose PTS (encoded by treB) can function with the rest of the glucose PTS (crr, HPr, EI) (1, 34). However, H. somnus 129Pt, H. ducreyi 35000HP and H. influenzae Rd did not have treB. None of the organisms had treC, enoding the trehalose 6-P hydrolase or treA, encoding the periplasmic trehalase. So trehalose probably does not enter the cell via the PTS or follow the trehalose I degradation pathway. The genomes of H. somnus 129Pt, H. ducreyi 35000HP and H. influenzae Rd also didn’t have a gene encoding trehalose-6-phosphate phosphorylase (trePP) or treF, encoding cytoplasmic trehalase, which is part of the trehalose degradation II pathway. Trehalose can also enter cells via a permease (34), followed by trehalose phosphorylase conversion of trehalose to glucose-1-P. However, none of these organisms had genes encoding trehalose phosphorylase. With regard to maltose, H. influenzae Rd, H. somnus 129Pt and H. ducreyi 35000HP were missing all of the key E. coli genes involved in maltose uptake and degradation (malT, malS, malE, malF, malG, malK, malP, malZ) (5). H. somnus 129Pt and H. influenzae Rd did have malQ, which is necessary for maltose metabolism in E. coli (5).