IC patients were chosen by following the criteria of National Institute of Arthritis, Diabetes and Kidney Diseases (Gillenwater et. al., 1987). Total DNA, which was extracted from IC patients bladder biopsies, was PCR amplified by nested primers. P5 was the only one clone to be amplified from horse serum. The horse had similar IC symptom. These universal primers could amplify various genera of bacterial 16S rRNA genes. In designing these primers, Dr. Domingue was careful to eliminate significant complementary or secondary structure while matching optimal annealing temperatures. Figure 2 showed the location of the nested primers on 16S RNA secondary structure of E. coli. The blind control, MG, was exception. This control DNA could not be amplified by 16S rRNA universal primers for unknown reason. One specific protein of Mycoplasma was amplified by the designed primers. This part of the work was done by Dr. Gerald J. Domingue, Ms. Liset G. Human and their colleagues in the Department of Urology, Microbiology and Immunology and Pathology, Tulane University.

In collaborative work with Dr. Domingue, we obtained pUC18-IC clones described above. In order to sequence the clones efficiently, PCR with pUC18 biotionylated primers was carried out. The PCR program was set 35 cycles to amplify pUC18-IC clones by one biotinylated primer and another non-labeling primer. Each clone had two sets of PCR products. FB had biotinylated forward primer sequences and RB had biotinylated reverse primer sequences. RB biotinylated PCR products were examined by 1% agarose electrophoresis (Figure 3). FB had the same size (data not shown). This ethidium bromide-stained gel showed PCR products from seven different pUC18-IC clones excess primers. Lane 1 was blind control MG; lane 2 clone P5 was from horse serum; lane 3 was clone 63; lane 4 was clone 68; lane 5 was clone 125; lane 6 was clone 129; lane 7 was clone 131. These numbers were log numbers from different patients samples. Lane 8 was negative control which didn’t contain template DNA but water. It only showed primers. The sizes of the PCR products ranged from 300 to 450 base pairs after we knew sequencs. Figure 4 showed that excessive primers and free nucleotides of each clones were removed when DNA PCR products were purified by QIAquick PCR product purification kit. Lane 4, clone 68, showed atypical primers aggregation during the same procedure of purification. By observing density of ethidium bromide stain agarose gel, the amount of all clones DNA decreased because the DNA was lost and primers were removed during purification.

Magnetic beads separation and sequencing biotinylated single-strand DNA

After purifying biotinylated pUC18-IC PCR products, we used alkaline solution to denature double-strand DNA following binding to the streptavidin-coating magnetic beads. Streptavidin-coating magnetic beads had high affinity to biotinylated single-strand DNA and separated it from non-labeled single-strand DNA. By using the Sequencing Kit, polymerization reaction was undertaken on single-strand DNA. Each clone had two sets of biotinylated single-strand DNA, FB and RB, which were sequenced by non-labeled reverse and forward primer, respectively. We continued to run two sets (FB and RB) of sequencing samples on sequencing gel and gained sequences of each clone from its two end untill sequences were overlapped. Having read the DNA sequences of each clones, we input these sequences into the PC/Gene program (A. Bairoch/ University of Geneva/ Switzerland; IntelliGenetics Inc.) and proofread them twice. The sequences of each clones were printed out on the Sequences Lists. Clone MG, blind control, contained 290 base pairs; Clone P5 had 426 base pairs. Clone 63 and Clone 68 had 473 and 373 base pairs, respectively. Clone 125 and Clone 129 contained 456 and 470 base pairs, respectively. Clone 131 has 476 base pairs. These sequences were analyzed and arranged by PC/Gene program. The clone 63 and 68 had higher G-C contents than the others.

Alignment and Comparison with previously published sequences.

The sequences of each clones were compared to two computer genes databases. One was National Center of Biotechnological Information (NCBI) of NIH. The other was the Ribosomal Database Project (RDP) of University of Illinois, Urbana, Illinois. The analyzed results of each clone were different from two database. Some of results had different species of bacteria. Some of them had same species but different order. I would discuss them separately.

From results of NCBI, figure 5 showed scores and a part of High-Scoring Segment Pairs (HSP) of the MG, blind control. Some of its query sequences were 100% similar to the Mycoplasma gentalium attachment protein that was control DNA from purified culture. I didn’t know that it was positive control until I gained the results from NCBI databases. The fidelity of sequencing technique and reliability of NCBI database could be proved by blind control. Clone P5 had high rRNA similarity sequences of Pseudomonas pickettii ribosomal RNA. The ranges of some queries identity were from 85% to 94% (Fig. 6); Clone 63 was mostly homologous to Caulobacter subvibroidies 16S ribosomal RNA; The identity of some fragments was from 76% to 100% (Fig. 7) ; Clone 68 had high 16S ribosomal RNA similarity to Azospirillum sp. and the ranges of some queries identity were from 80% to 100% (Fig. 8); Clone 125 was also mostly homologous to Azospirillum sp. 16S ribosomal RNA and the identity of some fragments was from 84% to 100% (Fig. 9); Clone 129 had high similarity to Pseudomonas flavescens B26 16S ribosomal RNA. The identity of some fragments was from 84% to 100% (Fig. 10); Clone 131 was mostly homologous to E. coli. rrnH gene, which was the locus of one rRNA gene. The percentage of some fragments identity was 99%. (Fig. 11). Based on NCBI results, the mostly homologous bactreria of each clones were summarized in Table 2.

After we gained these results, we used another retrieve e-mail server of NCBI to search database records and literature of these bacteria. We just reviewed these bacteria’s backgound breifly. The sequences of clone 131 was mostly homologous to E. coli rrnH gene. E. coli was most commonly isolated from all types of UTI and most of them belonged to a restricted range of serogroups. Clone P5 and clone 129 were highly homologous to Pseudomonas pickettii and Pseudomonas flavescens, respectively. Pseudomonas was hardy saprophytes of water, soil and other moist environments. Their ubiquity stems formed their minimal nutritional requirements. They were infrequent commensals of humans but could proliferate in the oropharynx and gastrointestinal tract of elderly, debilitated, or immunocompromised patients as well as in such sites as burns, ulcer and wounds. These sites were the usual portals of entry for organisms causing generalized infections. Clone 63 was highly related to Caulobacter subvibrioides. Caulobacter was a distinctive genus of prosthecate bacteria. It was a gram-negative, aquatic bacterium which was characterized by an unusual cell cycle. They could adhere to surface to surfaces and were found in diverse locales, their role in oligotrophic environments and bacterial biofilm communities was of interest. The phylogenetic relationships of a group of marine and freshwater caulobacter were examined in part to consistent with 16S rRNA sequence divergence. Stahl et. al. (1992) hypothesized C. subvibrioides are ancestrally related. Clones 68, 125 were mostly similar to Azospirillum spp. Azospirillum spp. was gram-negative, microaerobic, nitrogen-fixed bacteria associated with the roots of many economically important crops and grazes. Five species of this genus have been described. A. brasilense and A. lipoferum could perform and regulate nitrogen fixation only in a narrow range of low O² concentration (03. to 0.8kPa). Following a shift to an anaerobic condition, they underwent ADP-ribosylation of dinitrogenase reductase and loss their nitrogenase activity (Zhang et. al., 1993). Caulobacter subvibrioides and Azospirillum spp have not been shown to cause any human infectious diseases.

Based on different specificity, capability and statistical strategy of computer searches, the results showed a few difference from Ribosomal Database Project (RDP) of University of Illinois, Urbana, Illinois. Table 3 showed the top ranks of species of each clone from RDP. MG query sequence was mostly homologous to the Methanospaera stadtamanii. Clone P5 and Clone 63 had high rRNA similarity sequences of Bradyrhizobium japonicum. Clone 68 had most 16S ribosomal RNA similarity sequences to Pseudomonas syringae. Clone 125 also was mostly homologous to Flavobacterium lutescens. Clone 129 had high Pseudomonas aeruginosa similarity. Clone 131 was mostly homologous to E. coli K-12 (Table 3).

When we searched these bacteria’s records from retrieve e-mail server, clone P5 and Clone 63 were highly homologous to Bradyrhizobium japonicum, the soybean symbiont. It was one of legume root and stem nodule bacteria. They could live either free in the soil and in laboratory culture or endosymbiotically in infected host cells of the central nodule tissue, where they were capable of fixing N2. The clone 68 was highly homologous to Pseudomonas syringae. It was a plant pathogen, gram-negative bacterium and typical biotropic parasite that obtained its nutrients from living host cells. Clone 129 had high similarity to P. aeruginosa sequences. Many of the serious infections were caused by P. aeruginosa. Clone 131 had the same species as that of NCBI.

Collecting these differences from two databases, we made summary tables of the top ten species of each clone from two databases. When we expanded our view from first bacteria to the top ten, we could find out the same bacteria appeared on both of databases. Mycoplasma was the top rank of NCBI and the common species of two databases from MG clone, but Methanosphaera stadtmanii was the MG clone first choice from RDP (Table 4). Table 5 showed P5 had highest similarity to Bradyrhizobium japonicum on RDP; Pseudomonas pickettii was common and the first rank of NCBI. Caulobacter subvibroidies was on the top of NCBI list for clone 63 but did not show on RDP; Bradyrhizobium japonicum was the first rank on RDP and also listed on NCBI (Table 6). Table 7 showed clone 68 is Azospirillum sp. on NCBI top list but Pseudomonas syringase was on the top of RDP. Pseudomonas syringae was clone 68’s common species. Clone 125 was Flavobacterium lutescens from RDP and Azospirillum sp. from NCBI (Table 8). Flavobacterium lutescens was their common choice. Table 9 showed the different results of clone 129. Pseudomonas aeruginosa was from RDP and Pseudomonas flavescens was from NCBI; both of them picked Pseudomonas aeruginosa on their list. Clone 131 was the only one to show the same species- Escherichia coli- from RDP and NCBI (Table 10).

We found out a few different results on the lists of each clone, i.e. different species and order of bacteria from two computer databases,. But both of computer database shared the same species of bacteria on their top ten lists of each clone. To sum up, we made a summary table that showed these common bacteria of each clone (Table 11).

We needed to considered which of these detected agents could be IC pathogen. First, we used the multiple sequence alignment in PC/Gene program to compare sequence relation among six clones, the data showed that 92 base pairs were identical (17.7%) and 103 base pairs were similarity (19.8%) among the six sequences. The dendrogram of the alignment showed the relationship among these clones. The clone 131 and 129 were the closest and clone P5, while collecting from horse serum, was distantly related to the others (Fig. 12). Each clone came from different patients who had the same symptom of IC and P5 was from horse which had similar symptom of IC. Before sending sequences to computer database, understanding their sequence relationship, we could gain general possibility if we could collect more samples from different IC patients in future.

Second, we searched the related literature and published cases reported infection of human caused occasionally by the bacteria on the lists. This literature search might give us general backgournd knowledge about these bacteria and some of opportunistic etiologies, which especially were UTI-related diseases, might have higher possibility to cause IC. At the same time, we also could find that some of well-known virulence factors and pathogenesis of these bacteria might be involved in IC symptom.

Eventually, to use the conserved 16S rRNA sequences as universal primers, we did find out some unknown DNA form IC patients’ bladder biopsies in this thesis. When we comparied sequences of each clones to two computer database, we caught some species of bacteia. Pseudomonas , Escherichia coli were likely the etiologic agents of IC. But we could not exclude the possibility of the other bacteria and Pseudomonas and Escherichia coli also could be contaminants. From the different results of two computer databases, we should ponder several questions scrupulously. The 16s rRNA primer might not PCR amplify all organisms from IC patients’ biospies. The etiologic agents might not show on our lists. The fidelity between PCR products directly sequencing and cloning sequencing also should be considered. We also need to pay attention to the reproducibility of sequencing, heterogeneity of genes and computer databases. We would discuss in next section.