Identification of biofilm proteins in non-typeable Haemophilus Influenzae

Abstract

Background: Non-typeable Haemophilus influenzae biofilm formation is implicated in a number of chronic infections including otitis media, sinusitis and bronchitis. Biofilm structure includes cells and secreted extracellular matrix that is "slimy" and believed to contribute to the antibiotic resistant properties of biofilm bacteria. Components of biofilm extracellular matrix are largely unknown. In order to identify such biofilm proteins an ex-vivo biofilm of a non-typeable Haemophilus influenzae isolate, originally from an otitis media patent, was produced by on-filter growth. Extracellular matrix fraction was subjected to proteomic analysis via LC-MS/MS to identify proteins.

Results: 265 proteins were identified in the extracellular matrix sample. The identified proteins were analyzed for COG grouping and predicted cellular location via the TMHMM and SignalP predictive algorithms. The most over-represented COG groups identified compared to their frequency in the Haemophilus influenzae genome were cell motility and secretion (group N) followed by ribosomal proteins of group J. A number of hypothetical or un-characterized proteins were observed, as well as proteins previously implicated in biofilm function.

Conclusion: This study represents an initial approach to identifying and cataloguing numerous proteins associated with biofilm structure. The approach can be applied to biofilms of other bacteria to look for commonalities of expression and obtained information on biofilm protein expression can be used in multidisciplinary approaches to further understand biofilm structure and function.

Figure 1

Nontypeable Haemophilus influenzae biofilm imaged via scanning electron microscopy. Scanning electron micrographs of NTHi biofilms formed under different growth conditions. A and B) Sterile glass coverslips were covered with a suspension of NTHi in BHI broth. After 24 hr, the coverslips were prepared for SEM examination. (A) Large flat mats of bacteria embedded in an amorphous extracellular matrix were found attached to the glass surface. Scale bar = 2 μm. (B) The individual NTHi are covered in an amorphous layer that conceals the bacterial surface. Scale bar = 1 μm. C and D) Suspensions of NTHi in BHI broth were placed onto sterile Anopore insert filters that were mounted on chocolate agar. Once the NTHi biofilms had formed, after 24 hr incubation, on the upper surface of the filters at the air/liquid interface, the inserts were placed in culture dishes containing sufficient sterile culture medium to exert a positive upward pressure on the bottom of the biofilm, and left for a subsequent 24 hr. (C) The surface of the insert filter is covered with a flat mat consisting of NTHi closely attached to each other. Channels and pockets freee of bacteria have formed within the mat of bacteria. Scale bar = 2 μm. (D) In some orientations, it is possible to see the channels running between the aggregates of bacteria and through the mat. Scale bar = 2 μm. E and F) NTHi biofilms grown on Millipore filters. Sterile Millipore filters were placed onto chocolate agar plates and inoculated with sufficient NTHi in BHI broth to cover the surface at a density of 0.3 bacteria per 10 μmP^2P. The filters were incubated for 24 hr with the upper surface exposed to air, and prepared for SEM examination. (E) The NTHi formed thick biofilms with the base firmly attached to the filter substrate. Scale bar = 2 μm. (F) The top surface of the NTHi biofilm, that had been exposed to air, was covered with a thin film of extracellular matrix. In some instances, the matrix formed a film over regions that resembled bacteria-free pockets. Scale bar = 2 μm.

Figure 2

Correlation of identified protein molecular weights with SDS-PAGE gel slices. The x-axis indicates the molecular weight in Daltons of identified proteins in each gel slice. y represents the slice number of the gel from bottom of the gel (lower molecular weight) to the top of the gel (higher molecular weight).

Figure 3

Strain specific protein identifications in ECM biofilm sample. All indicates no strain specific peptides were observed where the peptide sequence is present in homologous proteins of each of the four strains used for Mascot and Sequest searches. Strain key: 1: Rd aka KW20; 2: R2846; 3: R2866; 4: 86-028NP. U indicates one protein, a thioredoxin, with a peptide not found in any of the four strains searched. See reference 13 for strain information.

Figure 4

COG Category distribution of identified biofilm ECM proteins. The percent distribution of identified proteins in terms of their assigned COG categories. The chart is color-coded as per COG colors at the NCBI COG functional annotation siteP . COG category groupings are as follows: JKL – Information storage and processing; DOMNPT – Cellular processes; CGEFHIQ – Metabolism; RS – Poorly characterized. Categorization presented here reflects original COG categorization. In updated categorization P is included in Metabolism. X indicates a protein with no affiliated COG category.

Figure 5

Relative distributions of identified COG protein compared to genomic distributions. Graphed from low to high are the % COG distribution in our identified sample compared to % COG distribution on bacterial genomes. The y-axis is the log of the ratio of COG % in our sample vs. COG % in a given genome. Diamond shows the distribution ratio in the 86-028NP genome; square is vs. Rd genome; triangle is all Haemophilus influenzae genomes; X is compared to all gammaproteobacteria genomes and the asterix is compared to all bacterial genomes. COG groups are labeled by number and color-coded as per the NCBI COG web page (as also in figure 4).