| Sign In to gain access to subscriptions and/or personal tools. |
Proteome Analysis of Oral PathogensInstitute of Dental Research, Westmead Centre for Oral Health, PO Box 533, Wentworthville, NSW 2145, Australia; Correspondence: * corresponding author, njacques{at}dental.wsahs.nsw.gov.au
The oral environment contains diverse communities of micro-organisms including bacteria, fungi, protozoa, and viruses. Studies of oral ecology have led to an appreciation of the complexity of the interactions that oral micro-organisms have with the host in both health and disease. Despite this, diseases such as dental caries and periodontal diseases are still worldwide human ailments, resulting in a high level of morbidity and an economic burden to society. Proteomics offers a new approach to the understanding of holistic changes occurring as oral micro-organisms adapt to environmental change within their habitats in the mouth.
Key Words: proteomics • two-dimensional electrophoresis • MUDPIT • mass spectrometry • oral micro-organisms
The word "proteome" originated in 1994 when Marc Wilkins coined the term to describe the set of all PROTEins expressed by a genOME (Wilkins et al., 1996). While genomics and transcriptomics provide basic information on DNA sequences, regulatory elements, and gene expression, proteomics provides quantitative information on the total protein profile of a cell, tissue, or organism. Proteomics allows the level of protein expression to be evaluated and can be used to determine the presence of protein isoforms and post-translational modifications or to examine protein-protein interactions.
Several different methods are available for evaluation of a cell’s proteome. The reader is directed to a recent compilation of reviews for specific technical details, since these are beyond the scope of this review (Aebersold and Cravatt, 2002). As far as oral micro-organisms are concerned, only one method of proteome analysis, two-dimensional gel electrophoresis (2-DE), has been used to any extent. 2-DE allows for the visualization of hundreds of proteins at a time (Rabilloud, 2002). It relies on isoelectric focusing (IEF) in the 1st dimension to separate proteins according to their isoelectric point (pI), followed by SDS-PAGE in the 2nd dimension to separate proteins according to their molecular weight (Mr). By using mass spectrometry to identify the proteins, one can produce reference maps that document the functional proteome, similar to the one we have recently published for Streptococcus mutans (Len et al., 2003). Although 95% of a bacterial genome is expressed as protein products (Humphery-Smith, 1999), the theoretical resolving power of 2-DE is still estimated to be approximately 75% of the proteome (Cordwell et al., 2000). Proteins with extreme pI values, especially very basic proteins, as well as very low and high Mr proteins, are not readily resolved by current 2-DE technology (Harry et al., 2000). In practice, many other proteins are poorly represented, including low-abundance proteins and hydrophobic proteins, particularly intrinsic cytoplasmic membrane proteins, which, although representing 15-30% of the proteome of a bacterium, are either not detected or represent less than 1.0% of the proteins displayed on 2-DE gels (Len et al., 2003). Membrane-associated proteins, lipoproteins, and, in the case of Gram-negative bacteria, some outer membrane proteins can, however, be readily detected by 2-DE analysis. The reason for this apparent anomaly is that the denatured outer membrane proteins of Gram-negative bacteria are generally hydrophilic, while integral cytoplasmic membrane proteins remain hydrophobic and are rarely soluble in aqueous environments without the use of strong surfactants, such as SDS, which are incompatible with the 1st-dimension IEF process. Despite recent advances in protein solubilization, separation of hydrophobic intrinsic membrane proteins by 2-DE remains a problem for all cell types (Eucarya, Bacteria, Archea) and will require a significant breakthrough if 2-DE displays are to achieve their full potential (Santoni et al., 2000).
Quantitative visualization of proteins on 2-DE gels is paramount for differential or comparative proteome analyses (Fig. 1
A second method of proteome analysis is now also commonly used. This method, MUlti-Dimensional Protein Identification Technology (MUDPIT), uses liquid chromatography to separate peptides derived from proteins rather than first separating the proteins themselves (Aebersold and Cravatt, 2002; Rabilloud, 2002). Peptides derived from the digestion of the total protein mixture are loaded onto a strong cation exchange column and then eluted with a step-gradient onto a reverse-phase column. The reverse-phase column is eluted with a solvent gradient and the eluate analyzed by mass spectrometry. The drawback of this technique is that it gives only a raw list of proteins present in the sample, without accurate quantification. Lack of quantification, however, can be overcome if the peptides are labeled prior to analysis. One approach is the use of an Isotope Coded Affinity Tag (ICAT) composed of a biotin tag, a linker, and an iodoacetamide handle that binds covalently to cysteine-containing peptides (Rabilloud, 2002). The linker can be made in a light or isotopically labeled heavy version and is therefore chemically indistinguishable from the light version, but possesses a different mass. When comparative analysis of two protein extracts is to be made, one is labeled with the light and the other with the heavy probe. The two extracts are pooled and digested with proteinase prior to being loaded onto an avidin affinity column that retains the labeled cysteine-containing peptides. After elution of the retained peptides onto a reverse-phase column, the peptides are detected by mass spectrometry in a manner similar to that of the MUDPIT technique, except that in this instance it is possible to determine the mass signal intensity ratio of the two forms of the co-eluted peptide and hence the ratio between the two peptides in the pooled samples and thus the proteins from which they were derived. The drawback to this technique is that the cysteine content differs widely from one protein to another, so that some proteins will be assigned by several peptides while others will rely on a single peptide. Since several proteins are cysteine-free, these will not be analyzed by this technique. Application of MUDPIT and ICAT techniques does, however, enable many intrinsic membrane proteins to be analyzed (Han et al., 2001). The drawback to these technologies is that they do not allow for the detection of post-translational modifications, as does 2-DE.
Mass spectrometry is by far the most common technique used in proteome analysis for the identification of an unknown protein, although alternatives such as N-terminal amino acid analysis or Western blotting can also be used. Several different mass spectrometers are commercially available with various levels of sensitivity (Aebersold and Cravatt, 2002). Matrix-assisted laser desorption-ionization (MALDI)-time of flight (TOF) instruments measure peptide mass only. These instruments are robust and relatively cheap and, while not highly sensitive, can generally analyze excised proteins from Coomassie-Blue-stained 2-DE gels. In the MALDI-TOF process, proteins are first digested with proteinases such as trypsin and the peptides mixed with a large excess of UV-absorbing matrix (usually a cinnamic acid derivative) and allowed to dry to a small spot on a MALDI plate (Fig. 2B
With MALDI-TOF, the identity of the unknown protein is determined from a table of the experimentally derived peptide masses, which together are known as the peptide-mass-fingerprint (PMF). The PMF is a signature of a protein and can be compared with the PMF theoretically generated with the same proteinase for each and every protein present in a non-redundant database (Fig. 2  ). In practice, a comprehensive proteome analysis requires the PMF to be compared with an annotated protein database constructed from the genome of the organism under investigation. If the genome has not been sequenced, the PMF can be compared with very closely related species from the same genera. For example, PMFs from 2-DE protein spots obtained from Streptococcus oralis and S. mutans have been used to search the annotated genomic database of Streptococcus pneumoniae and Streptococcus pyogenes (Wilkins et al., 2001, 2002). Unfortunately, successful protein identification relies on a high degree of peptide sequence identity between species. Any alteration in peptide sequences and peptide masses will result in a failure of the protein to be identified by MALDI-TOF techniques. As noted, MALDI-TOF is not that sensitive, and hence many proteins detected on 2-DE gels with the use of fluorescent stains or radiolabeling are of insufficient concentration for analysis by this technique (Len et al., 2003). Where peptides already exist, such as those eluted by MUDPIT, they can be analyzed online by mass spectrometers equipped with electrospray ionization (ESI) injectors rather than MALDI plates. In these instruments, a voltage is applied to a fine needle containing a dilute solution of the peptides. This results in a spray of droplets typically containing just a few molecules. Repeated break-up of the droplets due to evaporation eventually leads to release of intact peptide ions into the gas phase, whence they are sampled into the mass spectrometer and analyzed. Since these instruments work online and at atmospheric pressure, continuous uninterrupted analysis of peptides is possible.
While the characterization of proteins observed under a range of in vitro growth conditions is essential for complete definition of the proteome, differential or comparative proteomics is more frequently used to establish changes in protein expression when cells are grown under different physiological conditions. Recently reported studies with oral micro-organisms invariably use this approach. In the following sections, we have summarized several of these studies.
Dental Caries In a third 2-DE study, 18 proteins were up-regulated and 12 down-regulated when S. mutans was grown at pH 5.2 compared with cells grown at pH 7.0. These proteins were involved in energy metabolism, cell division, translation, and transport (Wilkins et al., 2002). Although the general conclusions of previous studies were confirmed, anomalies were observed. For instance, the protein DnaK was down-regulated when S. mutans was grown at low pH, while transcriptional studies had shown that dnaK gene expression was up-regulated under similar conditions (Jayaraman et al., 1997). The failure of other stress-related proteins to be detected emphasizes an important aspect of current proteomics involving incomplete 2-DE displays, since any undetected changes may be of equal importance in an understanding of the true nature of phenotypic change. The role of stress-related proteins, however, has been examined in S. mutans lacking the Clp ATPase, ClpP. S. mutans lacking ClpP was impaired in its ability to grow at low pH and had a reduced capacity to form biofilms (Lemos and Burne, 2002). Comparison of silver-stained 2-DE gels revealed at least 28 proteins with altered levels of expression in the clpP mutant compared with the parent. In particular, evidence was presented that the molecular chaperones DnaK, GroEL, and GroES were elevated in the mutant. The loss of ClpP appeared to induce a stress response in the cells, possibly due to the accumulation of denatured proteins that are normally targeted by the ClpP proteinase. Following a similar mutagenic approach, a two-component regulator system, defined by the genes hk11 and rr11, was also shown to be involved in biofilm formation and acid resistance in S. mutans (Li et al., 2002). Deletion of the hk11 gene resulted in a mutant that formed biofilms with reduced biomass and possessed greatly reduced resistance to low pH. Autoradiograms of 2-DE gels revealed 594 cellular proteins, 19 of which were acid-regulated in the parent, but only 15 in the mutant. Two of the 4 missing proteins were putatively identified as an exopolyphosphatase previously linked to biofilm formation, and the histidine kinase HK11 itself. The presence of fluoride in the environment represents another stress factor for many cariogenic bacteria. Fluoride not only protects tooth enamel from bacterial acids by forming fluorapatite in the outer layers of enamel, it also directly alters the bacterial phenotype. For example, 2-DE analysis of S. sobrinus grown in the presence of fluoride revealed that several proteins were influenced by the presence of fluoride, including the loss of a glucan-binding lectin activity (Cox et al., 1999).
Infective Endocarditis Antibiotic resistance is a major problem for the control of infective endocarditis, requiring an understanding of the changes in the proteome of the resistant phenotype. To this end, the mechanism of penicillin tolerance has been examined in S. gordonii by 2-DE (Caldelari et al., 2000). A tolerant mutant contained two proteins with increased intensity. Comparison of the sequences of their N-terminal amino acids with known proteins showed that they were homologous to the N-termini of arginine deiminase and ornithine carbamoyl transferase of the arginine deiminase (arc) operon. Although the penicillin-tolerance mutation mapped at a physically distinct location on the chromosome from the arc operon, genetic transformation of tolerance always conferred arc deregulation. It was concluded that the tolerance mutant affected a global regulatory mechanism that was important for survival in the presence of penicillin. The correlation between tolerance and arc deregulation has allowed a reporter system to be developed that will aid in determining the central mechanism of penicillin tolerance underlying this clinically important phenotype.
Periodontal Disease
Oral Candidiasis Several other studies of C. albicans have combined 2-DE with Western blotting to identify antigens that react to antibodies present in human sera. This method has successfully identified several new antigens that may be useful in developing diagnostic strategies aimed at treating systemic candidiasis (Barea et al., 1999; Pardo et al., 2000). Current treatment of Candida infections is impeded by the limited number of antifungal drugs and is complicated by the emergence of azole-resistant strains associated with changes in azole efflux (White et al., 1998). Azole-susceptible and -resistant isolates of Candida glabrata have been compared by 2-DE. Twenty-five proteins were up-regulated and 76 down-regulated in the resistant isolate (Marichal et al., 1997). Since these proteins were not identified, further research is needed to determine their involvement in drug resistance.
Other Pathogens
Unlike protozoan proteomes, viral proteomes are relatively small, and many have already been predicted from their genomic sequence (Blattner et al., 1997). Viral proteome studies tend to center on the effects of a virus on its host. For instance, a recent study based on the 2-DE differential display of HeLa cell proteins has shown that the 2A proteinase of human Rhinovirus and Coxsachie virus cleaves cytokeratin 8 in the early stages of infection (Seipelt et al., 2000). In a similar study, a combination of 2-DE and Western blotting allowed the receptors on HeLa and human lung carcinoma cells that bind Parechovirus 1 and Echovirus 1 to be identified as integrins
The research covered by this review is aimed at approaching old but unresolved questions by shedding new light on microbial responses to the onset and progression of oral diseases. While many of the described studies are clearly in their infancy, they offer a tantalizing insight into the hitherto-ignored complexity of protein regulation and expression, and the resulting change in the phenotype of a micro-organism. Since proteomics is a relatively new science, several technical problems still exist that prevent the complete proteome of a cell being readily identified. Although our recently published map of S. mutans contains 416 identified proteins and could be considered "comprehensive" by current 2-DE standards, our results clearly show that significantly more proteins could be visualized with SYPRO Ruby than could be analyzed by MALDI-TOF mass spectrometry (Len et al., 2003). Pre-fractionation of the proteome—either by cellular compartment (e.g., separation of cytoplasmic, membrane, and wall fractions in bacteria), Mr, or pI, or preferably all three—along with the application of alternative proteome techniques including MUDPIT and ICAT is allowing us to identify many more low-abundance and membrane proteins. In an ideal world, more advanced mass spectrometry, such as the use of highly sensitive Fourier transform ion cyclotron resonance instruments, would be available to all, allowing for a more rapid and comprehensive analysis of oral microbial proteomes in a manner similar to that reported for Deinococcus radiodurans. Using a variation of MUDPIT with this instrumentation, investigators have identified more than 61% of the predicted proteome with high confidence, including many low-abundance and membrane proteins (Lipton et al., 2002).
Proteomics is a technique that will revolutionize the study of oral micro-organisms (as it will other health-related fields) by revealing processes involved in disease onset. Public access databases are already emerging on the Internet containing 2-DE images, databases of sequenced proteins, and software for proteome analysis. In the dental area, "ToothPrint" (http://toothprint.otago.ac.nz) is a perfect example (Hubbard et al., 2001). "ToothPrint" is based on developing rat enamel and provides functionally relevant data and 2-DE protein identification maps. As microbial proteome databases similar to this appear and expand, they will provide an important bioinformatic resource for dental research. Eventually, proteome data of specific phenotypes will become available that will reveal alterations in protein expression and hence cellular function, enabling researchers to identify and focus on pathways or proteins of interest without the need to re-identify each and every protein. This will allow appropriate molecular, biochemical, and physiological studies to be undertaken with the view to developing new diagnostic and therapeutic strategies to combat oral microbial disease.
The authors’ research cited in this review was supported by Grant No. 5 R01 DE 013234 from the National Institute of Dental and Craniofacial Research (NIDCR) as was a graduate scholarship awarded to DJM. This research was facilitated by access to the Australian Proteome Analysis Facility established under the Australian government’s Major National Research Facility program. Received for publication January 3, 2003. Revision received June 2, 2003. Accepted for publication August 6, 2003.
Journal of Dental Research, Vol. 82, No. 11,
870-876 (2003)
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
|
|
 |
|
Sign In to gain access to subscriptions and/or personal tools.
TOP
ABSTRACT
INTRODUCTION
vβ3 and 
a-Toli
L, Anderson GA, Anderson DJ, Auberry DL, Battista JR, et al. (2002). Global analysis of the Deinococcus radiodurans proteome by using accurate mass tags. Proc Natl Acad Sci USA 99:11049–11054.