PCR-RFLP–Mediated Detection and Speciation of Bacterial Species Causing Endophthalmitis

Abstract

purpose. To determine the usefulness of polymerase chain reaction–restriction fragment length polymorphism (PCR-RFLP) in the identification and speciation of bacteria causing endophthalmitis.

methods. PCR-RFLP was performed on 53 strains of 14 bacterial species (eight Gram positive and five Gram negative) collected from both keratitis and endophthalmitis patients. Two pairs of oligonucleotide primers based on the 16S rDNA gene were used to PCR-amplify 1.2- and 1.0-kb fragments of bacterial genomic DNA. RFLPs within the PCR product were used to speciate the organisms.

results. The sensitivity of the nested PCR amplification reaction was one organism. All bacteria tested could be identified and speciated using RFLP analysis except for Escherichia coli and Serratia marcescens, which could not be interdifferentiated using RFLP. Molecular analysis of two vitreous samples from two eyes with typical signs of bacterial endophthalmitis confirmed the presence of E. coli in the vitreous from a culture-positive case with E. coli endophthalmitis and revealed the presence of Staphylococcus epidermidis in the vitreous of a culture-negative case.

conclusions. It is expected that this technique will provide a useful laboratory tool for future microbiologic diagnosis of patients presenting with endophthalmitis, especially for those eyes that prove culture negative.

Bacterial endophthalmitis remains a severe sight-threatening disease despite aggressive treatment and results in a poor final visual outcome in the majority of patients. ¹ Early diagnosis and appropriate treatment have been noted to be associated with a better visual outcome, ² but culture techniques can take between 48 hours and 12 days ³ to confirm the presence and identity of the pathogen. Many cultures, however, prove to be negative (21%–63% in the published press ^{1 4 5 6 7 8} ),and although this may reflect the purely inflammatory nature of the disease in some cases, it may also be a reflection of the presence of only a few bacteria in the small sample, sequestration of bacteria on solid surfaces (e.g., IOL, lens remnants and capsule), leading to low numbers in the liquid sample, the use of antibiotics before sampling, and/or the fastidious nature of some of the organisms that cause intraocular infection. ^{1 9 10} These observations suggest that perhaps with a more sensitive and specific detection strategy a microbiologic diagnosis may be obtained in more cases. To this end, this study was designed to develop a highly sensitive, reliable and rapid method for the identification and speciation of bacteria in ocular samples. Primers based on the conserved sequences of the 16S rDNA gene, ubiquitous in bacteria, were used to detect the presence of bacterial DNA in aqueous and vitreous samples. Restriction endonuclease digestion of the polymerase chain reaction (PCR) product enabled species differentiation such that the 14 species of bacteria tested could be identified and speciated, except for Escherichia coli and Serratia marcescens, which could not be interdifferentiated using restriction fragment length polymorphism (RFLP). 

Materials and Methods

Unless otherwise stated all chemicals used were purchased from Sigma Chemical Company (Poole, Dorset, United Kingdom) and were of the highest grade available. 

Samples

All bacteria isolated from clinical ocular samples collected at Moorfields Eye Hospital, whether from cases of keratitis or endophthalmitis, were included for purposes of completion and comparison. A total of 54 strains of 14 bacterial species were tested, of which 43 were clinical isolates and 11 were National Collection of Type Culture (NCTC) strains. Species and strain details appear in Table 1 . 

DNA from clinical isolates of Candida albicans (n = 3), Aspergillus fumigatus (n = 1), and Fusarium solani (n = 1) also were included for specificity testing. 

Procedure for Endophthalmitis Cases.

Intraocular (aqueous and vitreous) sampling was performed on every patient under aseptic conditions. Aqueous sampling was undertaken using a 27-gauge (0.33 mm) needle using topical anesthesia, and 100 to 200μ l was aspirated. Vitreous sampling was undertaken after subconjunctival injection of anesthetic. Vitreous (200–400 μl) was aspirated using a 23-gauge needle that was inserted through the pars plana 3 mm behind the limbus in aphakic eyes and 4 mm behind the limbus in phakic eyes. 

Intraocular samples were examined by Gram’s stain and immediately cultured on solid (blood agar) and liquid media (Robertson’s cooked meat broth and Brain Heart infusion (Difco Laboratories UK, West Molesey, Surrey, United Kingdom) under both aerobic and anaerobic conditions. Aerobic organisms isolated from an intraocular sample as heavy growth from one solid medium or as growth of the same organism in more than one medium (solid and/or liquid) were considered responsible for causing endophthalmitis. All cultures were maintained for up to 14 days. 

Procedure for Keratitis Cases.

Corneal scrapes were performed at the slit lamp under topical anesthesia. Sterile 27-gauge needles were used to scrape off corneal epithelium and anterior stroma in the region of the leading edge of the corneal ulcer. Samples were placed on a slide for Gram and Giemsa staining and also were plated immediately on Blood and Sabauraud’s dextrose agar (Difco Laboratories UK) before transport to the microbiology laboratory. Plates were incubated under aerobic conditions at 30°C and 37°C. 

Identification of Isolates.

Bacterial isolates were identified using standard microbiologic methods. After isolation by culture the API biochemical identification system (API Analytab Products, Division of Sherwood Medical, NY) was used for identification. Organisms were subsequently stored on beads (Mast Diagnostics, Bootle, Merseyside, United Kingdom) at −70°C. 

Collection of Normal and Inflamed Vitreous.

Vitreous was collected by sterile technique at the time of vitrectomy during planned surgical procedures in patients with no evidence of intraocular infection or inflammation or medical history of uveitis and/or diabetes mellitus (“normal” vitreous). Vitreous also was collected from patients with other causes of posterior segment inflammation not associated with bacterial infection (“inflamed” vitreous). Samples of vitreous were aliquoted in a sterile manner and stored at −20°C. 

DNA Extraction

Bacteria.

Method 1 was the full extraction procedure. Previous work in this laboratory has successfully used glass beads and a bead beater apparatus (Stratech Scientific, Bedfordshire, United Kingdom) to effectively release DNA from bacterial cells in suspension. ¹¹ Briefly, a 2-mm colony of bacteria was diluted into 300 μl sterile phosphate-buffered saline solution, with 50 μg/ml proteinase-K and 0.5 g glass beads (0.1-mm size). This mixture was beaten on the bead beater for 10 seconds after which it was incubated at 50°C for 30 minutes. Phenol (500 μl, pH 8.0) was then added, and the samples were vortexed for 30 seconds and centrifuged for 2 minutes at 14,000g. The aqueous phase was removed and extracted twice, the first time using an equal volume of phenol-chloroform-isoamyl alcohol (25:24:1, pH 8.0) and subsequently with an equal volume of chloroform. The DNA was subsequently precipitated with 2.5 volumes of ethanol in the presence of 300 mM sodium acetate (pH 5.2). The DNA was pelleted at 14,000g for 20 minutes, washed with ice cold 80% ethanol, air-dried at 65°C for 10 minutes, and resuspended in 25 μl of sterile TE buffer (10 mM Tris-HCl, pH 8.0, 1 mM EDTA). The DNA was diluted to a working concentration of 10 ng/μl for use in PCR reactions and stored at− 20°C. 

Method 2 was direct PCR. Bacteria were suspended in the distilled water component of the PCR cocktail and used directly in PCR reactions. 

Fungi.

DNA was extracted by a previously described method. ¹² Briefly, DNA extraction tubes were prepared containing a single 1-mm colony of C. albicans or scrapings of A. fumigatus, or F. solani in 100 μl of 0.05 M Tris, pH 7.5, 0.01 M EDTA, 0.028 M β-mercaptoethanol, and 0.3 mg/ml zymolase (ICN Biomedicals, Aurora OH). Samples were incubated for 30 minutes at 37°C, followed by addition of 0.1% SDS and 15 μg/ml proteinase K, and incubation continued for a further 5 minutes. The mixture was subsequently heated to 95°C for 5 minutes and cooled on ice for 15 minutes. Samples were extracted with an equal volume of phenol-chloroform-isoamyl alcohol (25:24:1, pH 8.0) and subsequently with an equal volume of chloroform and precipitated with 2.5 volumes of ethanol in the presence of 300 mM sodium acetate (pH 5.2). The DNA was pelleted at 14,000g for 20 minutes, washed with ice-cold 80% ethanol, air-dried at 65°C for 10 minutes, and resuspended in 25μ l sterile TE buffer (10 mM Tris, 1 mM EDTA, pH 8.0). The DNA was diluted to a working concentration of 10 ng/μl for use in PCR reactions. 

Human Leukocytes.

DNA was extracted from 10 ml human whole blood using the Machery–Nagel Nucleospin-Blood DNA extraction kit (Biogene, Cambridge, United Kingdom) according to the manufacturer’s instructions. The DNA was diluted to a working concentration of 10 ng/μl and stored at− 20°C. 

PCR Amplification of Bacterial 16S rDNA

Design and Optimization of Primers.

Multiple copies of the 16S rDNA genes are present in all bacterial genomes. Details of the four primer pairs used in this study appear in Table 2 . 

Optimal PCR Conditions.

PCR reactions contained 20 mM Tris-HCl (pH 8.3), 100 mM KCl, 0.01% Tween 20, 0.1% NP-40, 100 μM each deoxyribonucleotide (dNTP; Pharmacia Biotechnology, St. Albans, United Kingdom), 0.3 μM of each primer, 1.5 mM magnesium chloride, 3 units Replitherm Taq DNA polymerase (Taq; Cambio Ltd., Cambridge, United Kingdom), in a 25-μl reaction. The cycling parameters for both first-round and nested PCR cycles included 5 minutes at 95°C as an initial step. Subsequent cycling for first-round PCR (primers 16SF + 16SR) was performed by cycling at 94°C for 10 seconds, 56.2°C annealing for 10 seconds, and 72°C for 15 seconds for 30 cycles and for nested PCR (primers NF + NR) by cycling at 94°C for 7 seconds, 64°C annealing for 7 seconds, and 72°C for 10 seconds for 30 cycles. All PCR cycling was performed on a Techne Genius PCR machine (Cambridge, United Kingdom). 

Pretreatment of Taq DNA Polymerase to Remove Contaminating Bacterial DNA.

Taq DNA polymerase is known to be contaminated with low levels of bacterial DNA not originating from either Thermus aquaticus or E. coli and is easily amplified using universal bacterial primers based on ribosomal gene sequences. ^{13 14 15} The Replitherm Taq DNA polymerase (Replitherm Thermostable DNA polymerase; Cambio Ltd.) used in this study is one of a number of commercially available Taq polymerases with known low levels of DNA contamination. Although this level of contamination is insufficient to give a detectable amplification product after just one round of PCR, it is easily detected after nested amplification. Therefore, in this study, before first-round PCR amplification, the Replitherm Taq was treated with AluI restriction endonuclease (ratio of 3:1 units of Taq:AluI; Promega UK Ltd., Southampton, United Kingdom), which recognizes a specific nucleotide sequence AGCT. The criteria for choosing this restriction enzyme were (1) its known high cutting frequency within the 16S rDNA genomic template, which was evident after sequence analysis of the amplified fragments from the 14 bacterial species and (2) the ability to heat-inactivate the enzyme before addition of template DNA. Before PCR amplification, the water, PCR buffer, magnesium, and Taq components were mixed and incubated at 37°C with AluI for 30 minutes. The restriction enzyme was subsequently inactivated by incubation at 95°C for 2 minutes, after which the dNTPs, primers, and template DNA were added, and the PCR cycle commenced. 

DNA Sequencing and Restriction Analysis

Amplified DNAs from PCR reactions were agarose gel purified, excised, and recovered into sterile water (Geneclean II Kit; BIO 101, La Jolla, CA). PCR fragments were directly cycle-sequenced in both directions using an ABI prism automated DNA sequencer (model 377, version 2.1.1; PE Biosystems, Warrington, UK). Sequences were aligned and scanned both manually and using database and software programs available through the HGMP computer center. A search was made for a total of 268 restriction enzyme recognition sites. 

Cloning of PCR Products

Amplified DNA from PCR reactions was purified on agarose/TBE gels, excised, and recovered into solution (Geneclean II Kit; BIO 101). In some cases PCR products were directly cloned into pCR II (Cat. No. K2000–01; Invitrogen BV, Leek, The Netherlands) to aid sequencing. 

Restriction Enzyme Analysis

After PCR amplification, the concentration of PCR product was estimated using ethidium bromide staining of agarose/TBE gels. Restriction enzyme combinations were used that yielded fragments that allowed easy identification of species after separation of digested DNA on polyacrylamide/TBE gels. One restriction enzyme cocktail was developed to directly speciate organisms. All enzymes were purchased from Promega UK Ltd. (Southampton, United Kingdom) except for DraIII and AflIII, which were purchased from New England Biolaboratories Ltd. (Hitchin, Hertfordshire, United Kingdom). 

RFLP Protocol

The restriction enzyme cocktail (5 units each enzyme) was added to approximately 1 μg of DNA/PCR product in PCR buffer that had been adjusted to contain 100 mM NaCl, 1 mM dithiothreitol, and a final concentration of 7 mM MgCl₂. A restriction enzyme cocktail containing the following nine restriction endonucleases was used to achieve speciation: AflIII, BssHII, ClaI, DraI, DraIII, HpaI, NdeI, NsiI, and SalI. Restriction enzyme digests were performed at 37°C for 18 hours. The reaction was halted by freezing, and restriction fragments were analyzed on 10% TBE/polyacrylamide gels. 

Electrophoresis and Imaging

After PCR, amplification products were resolved on a 1% agarose/TBE gel and visualized using ethidium bromide under UV illumination. A molecular weight ladder was included in each run (1.0-kb ladder, Cat. No. 15615-016; Gibco BRL, Paisley, Scotland). 

Restriction enzyme digests were resolved on 10% polyacrylamide/TBE gels and visualized under UV illumination after staining with ethidium bromide. A molecular weight ladder was included in each run (100 bp, Cat no. SLL-100; Advanced Biotechnologies, Leatherhead, Surrey, United Kingdom). 

Results

Specificity

Using primers 16SF + 16SR, a 1.2-kb product was obtained from all strains of all species of bacteria tested (n = 53, Figs. 1a 1b 1c ). When Propionibacterium acnes genomic DNA was used as template, a 600-bp fragment also was obtained (Fig. 1a , lane 8). 

Using primers NF and NR, a single 1.0-kb amplification product was obtained using genomic DNA from all bacteria tested (Fig. 1d) . No amplification product was obtained when genomic DNA from C. albicans, A. fumigatus, F. solani, or human leukocytes was used as template. 

Nested PCR Protocol and Controls

In each case, 1 μl of first-round PCR product was added to the nested amplification reaction as template. Control-negative (no DNA) samples included water and vitreous (normal or inflamed as appropriate). Control negatives were included in each protocol, and the first-round negatives were included as test samples in the nested PCR reaction. Control-negative samples from both rounds of PCR were consistently negative after two rounds of amplification. 

Control-positive samples included extracted genomic DNA (10 ng and 10 fg) or live organisms (1–5 organisms by dilution from fresh overnight culture) in both water and vitreous. Conrtol-positive samples were consistently positive after one round (10 ng extracted DNA only) and two rounds of amplification (10 fg extracted genomic DNA or 1–5 live organisms). 

Sensitivity of PCR Reactions

The sensitivity of the first-round PCR was routinely found to be 10 pg from dilutions of DNA from E. coli, coagulase-negative staphylococci, Staphylococcus aureus and Klebsiella pneumoniae, starting from a concentration of 10 ng/μl (Fig. 2a ). The sensitivity of the reaction was improved to 1 fg after nested PCR (Fig. 2b) . Assuming a total DNA content of 5 fg per organism, ¹⁶ this is approximately equivalent to a sensitivity of one organism. 

To assess the sensitivity of detection with live organisms, serial 10-fold dilutions were made from a small 1-mm colony of bacteria in water (test organisms: coagulase-negative staphylococci and E. coli) (Figs. 3d ). Equal aliquots were immediately cultured and amplified in PCR reactions. Studies were performed to assess the effect of dilutions made in water compared with phosphate-buffered saline on bacterial viability and subsequent detection by culture. The results indicated that if the bacteria were cultured within 20 minutes, no reduction in viability was evident after suspension in water. The sensitivity of detection after one round of PCR was approximately 600 organisms for all species. This sensitivity was improved to one organism after a second round of PCR. 

Vitreous Studies

Normal vitreous was collected from seven patients undergoing elective macular hole surgery according to the criteria already described. Inflamed vitreous was collected from two patients at the time of vitrectomy. One patient had a diagnosis of pars planitis (PP), and the second case was diagnosed with acute toxoplasma retinochoroiditis (TRC). Both normal and inflamed vitreous were tested for their ability to inhibit the PCR reaction. The inhibitory ability of the vitreous from the case with TRC was greater than that from the case with PP, which was greater than normal vitreous and will be discussed in detail elsewhere. These effects were only significant in the first round of PCR. Nested PCR amplification of DNA/organisms in water, normal vitreous, and inflamed vitreous (TRC and PP vitreous) retained a sensitivity of detection of one organism (unpublished observations). The volume of vitreous used in PCR reactions was limited to 20% of the reaction volume for analysis of clinical samples. ¹² Both normal and inflamed vitreous samples were routinely negative for the presence of 16S rDNA after nested PCR amplification. 

DNA Sequencing

The sequencing reaction was performed three times for each PCR product using the same template but different PCR reactions. Comparison of the DNA sequences obtained made with the full and partial sequences of the genes already available in GenBank demonstrated that they were derived from the 16S rDNA gene (sequences 97–100% identical). The sequences obtained were submitted to GenBank and have been assigned the following accession numbers: Streptococcus pyogenes: AFO76028, Streptococcus viridans: AFO76036, Streptococcus faecalis: AFO76027, P. acnes: AFO76032, S. aureus: AFO76030, Streptococcus pneumoniae: AFO76029, Bacillus cereus: AFO76031, E. coli: AFO76037, Serratia marcescens: AFO76038, Haemophilus influenzae: AFO76035, K. pneumoniae: AFO76033, Proteus mirabilis: AFO76034, and Pseudomonas aeruginosa: AFO76039. 

Strain Testing

A set of control patterns was obtained from all bacterial species. Subsequently, PCR-RFLP and sequence analysis was performed on all clinical isolates to ensure the reproducibility of the procedure. All clinical isolates of streptococci, S. aureus, and each of the Enterobacteriacae, yielded results from PCR, RFLP, and sequencing that were in agreement with culture results. RFLP analysis of coagulase-negative staphylococci yielded three patterns. Two of these were still identified as coagulase-negative staphylococci by a masked observer, but a third pattern was found to be unidentifiable. Sequence analysis of clinical isolates of coagulase-negative staphylococci yielded 98% base identity in each case, with several of the following staphylococcal species that belong to this group: S. epidermidis,S. caprae, S. capitis, S. warneri, S. lugdunensis and S. pateurii. These sequences were also found to be 97% identical with that of S. aureus. These new RFLP patterns have been added to the “control patterns” to aid future identification of these species. 

Cloning of PCR Products

Sequencing direct from PCR product did not yield adequate sequencing data for the following bacteria: P. mirabilis, H. influenzae, S. viridans, and P. acnes. PCR fragments were gene-cleaned and cloned into pCRII. Subsequent PCR and sequencing was then performed successfully using primers complementary to Sp6 and T7 sites present in the pCRII vector. 

RFLP Analysis of Amplified PCR Products Differentiates Bacterial Species

The combination of AflIII, BssHII, ClaI, DraI, DraIII, HpaI, NdeI, NsiI, and SalI successfully differentiated between the PCR-amplified products from 13 of 14 bacterial species (Fig. 4) . E. coli and S. marcescens have identical RFLP patterns, reflecting the similarity of the 16S rDNA gene sequences for these two species (96.7% identical). The only stretch of nucleotides that would reliably identify one from the other, however, has a sequence that is not recognized by any of the restriction enzymes commercially available. 

Clinical Samples

PCR amplification and RFLP analysis was successfully applied to clinical samples from two cases of presumed bacterial endophthalmitis (Fig. 5) . Both cases presented clinically with a history and signs typical of acute endophthalmitis. Vitreous from one patient was found to be culture positive for E. coli, but the second case proved to be culture negative. PCR amplification of 5 μl of vitreous, by direct addition to the reaction mix, was successful after two rounds of PCR. RFLP analysis yielded patterns typical of E. coli/S. marcescens in the first case and that of coagulase-negative staphylococci in the second. Sequencing of the PCR product confirmed the identity of the pathogens as E. coli and coagulase-negative staphylococci, respectively. 

Discussion

The methods described in this article combine the detection of specific nucleotide sequences with the easy identification and categorization of the 14 bacterial species of interest into 13 groups. Amplification of 16S rDNA from bacterial genomes has been performed successfully using extracted genomic DNA, live organisms (each spiked into water and vitreous), and clinical samples from patients with presumed bacterial endophthalmitis. 

The choice of primers was affected by the knowledge that 16S rDNA genes are highly conserved in bacterial genomes and may be present in multiple copies. ^{17 18} Bacteria were differentiated by exploiting the variable stretches of DNA sequence in the 16S rRNA gene. The study of sequence variability also can be carried out using other methods, for example, by the synthesis of species-specific oligonucleotide probes, which would identify DNA sequences through hybridization, or by design of PCR primers and amplification protocols, which discriminate for the organism of interest. ^{19 20} However, probes and primers would need to be designed for individual species, and as such these methods were unsuitable for our purposes. 

Standard DNA extraction methods are time-consuming and involve multiple tube transfers wherein organisms or DNA may be lost. The method used simply involves heating to 95°C as part of the PCR-cycling protocol and shows great potential because it reduces the potential loss of organisms or DNA in transfer and is found to be reliable, reproducible, simple, and rapid. A highly sensitive approach was required to confirm the presence of any bacteria present. The ability to detect 10 pg of genomic DNA after one round of PCR may be slightly less than the sensitivity reported by other authors ²¹ and would be improved by reducing the annealing temperature. We have opted for the highest annealing temperature possible, to ensure specificity, which may have reduced the sensitivity of the reaction, but the use of a second-round nested approach improved sensitivity, allowing detection of as little as 1 fg, while maintaining the specificity of PCR detection. 

Thirteen RFLP patterns were obtained from 14 bacterial species and only 12 of the 14 bacterial species had unique RFLP patterns that could identify them from all other species tested. The patterns obtained by gel electrophoresis were found to be reproducible. However, species identification could be made difficult if the minor and fainter bands were not clearly visible on all gels. The identical patterns obtained from E. coli and S. marcescens reflect the similarity in the gene sequence of the 16S rDNA from these organisms. Although the minor sequence difference was not useful in differentiating these two organisms by RFLP analysis, a stretch of nine nucleotides is present, which is more than sufficient to allow the design of oligonucleotide primers that would allow differentiation of these organisms using additional PCR protocols. Because the treatment of these two organisms would be identical (whether endophthalmitis is caused by E. coli or S. marcescens), further identification was not pursued. 

Because this PCR/RFLP protocol is designed to detect the presence of all bacteria, if the organism present is not one of the series studied here, it is conceivable that this specific restriction enzyme digestion may be unable to identify the pathogen. As a precaution, therefore, all amplified PCR products were sequenced to confirm the identity of the bacterium. This technique has been used successfully to confirm the identity of the organism detected by culture in one case and to confirm bacterial involvement in a culture-negative case. No inhibitory effects of vitreous were observed in the analysis of these two samples. To investigate inhibitory effects further, larger numbers of clinical samples (aqueous and vitreous) have been analyzed. Results are reported elsewhere and indicate that the level of PCR inhibition varied from sample to sample and that dilution of the sample was required in the analysis of 8/18 (44%) intraocular samples (5 aqueous and 3 vitreous) before a positive PCR result was obtained (Okhravi N, Adamson P, Carroll N, Dunlop A, Matheson M, Towler HMA, Lightman S, unpublished results). 

The detection of bacterial DNA by PCR-based methodologies, in body sites that are considered sterile, has been used to improve the rate of microbiologic diagnosis for cerebral-spinal fluid, ^{19 22 23} synovial fluid, ²⁴ and vitreous (Okhravi N, Adamson P, Carroll N, Dunlop A, Matheson M, Towler HMA, Lightman S, unpublished results). ¹¹ Results demonstrate that samples from cases clinically suspected as harboring infection contain bacterial DNA, whereas samples from clinically noninfective cases do not (Okhravi N, Adamson P, Carroll N, Dunlop A, Matheson M, Towler HMA, Lightman S, unpublished results). ^{22 23 25}  

At this center, 63% of cultures from cases of postsurgical, presumed bacterial endophthalmitis proved to be negative. ¹ This is a reflection of the large number of cases referred to this tertiary referral center having had prior antibiotic treatment (topical, systemic, and intraocular) for endophthalmitis. The use of antibiotics may affect the culture-positive rate but should not affect the ability to PCR-amplify DNA in the short term. However, because of the lack of quantification, it is not possible to judge the number of organisms present and therefore to differentiate contamination from clear evidence of infection. Although PCR-based methods are able to provide the reliability and sensitivity required, data regarding antibiotic drug sensitivities currently can only be obtained after positive culture. In culture-negative cases, however, information regarding etiology can only be obtained using PCR technology. The presence of multiple organisms in the sample (either as true causative agents of infection or after contamination) is always a possibility, especially in samples collected from the cornea or from cases with endophthalmitis after penetrating injury. The inability of PCR-RFLP techniques to deal effectively with the presence of multiple organisms is a potential drawback of this methodology. A PCR-based study of postoperative endophthalmitis, however, has revealed the presence of multiple organisms in the anterior chamber much more frequently than in the vitreous cavity, allowing unambiguous species identification from the vitreous sample in these cases (Okhravi N, Adamson P, Carroll N, Dunlop A, Matheson M, Towler HMA, Lightman S, unpublished results). It is expected that PCR-based technology will prove to be a useful adjunct to microbiologic culture techniques, especially in culture-negative cases, and provide a useful addition to the diagnostic tools available to the clinician. 

 Supported by a Wellcome Vision Research Fellowship 045203 (NO) and locally organized research funds from Moorfields Eye Hospital Grants 221 and 271.

 Submitted for publication May 14, 1999; revised September 15 and November 2, 1999; accepted November 16, 1999.

 Commercial relationships policy: N.

 

Table 1.

 

Details of Organisms Used in this Study

Organism	Clinical Isolates*	PHLS, †	Total Number of Strains Tested
Gram-positive organisms
Coagulase-negative staphylococci	13	NCTC 11047, ‡	14
Staphylococcus aureus	10	NCTC 08532	10
Streptococcus pneumoniae	3	NCTC 07465	4
Streptococcus viridans	3	0	3
Streptococcus pyogenes	0	NCTC 08198	1
Streptococcus faecalis	1	0	1
Bacillus cereus	0	NCTC 02599	1
Propionibacterium acnes	4	NCTC 00737	5
Gram-negative organisms
Escherichia coli	2	0	2
Serratia marcescens	1	NCTC 10211	2
Haemophilus influenzae	1	NCTC 08143	2
Klebsiella pneumoniae	1	NCTC 09633	2
Proteus mirabilis	2	NCTC 00060	3
Pseudomonas aeruginosa	2	NCTC 10332	3
Total	43	11	54

*  Department of Pathology, The Institute of Ophthalmology, Bath St., London, United Kingdom.

†  Public Health Laboratory Service, National Collection of Type Culture, Colindale, London, United Kingdom.

‡   Staphylococcus epidermidis.

Table 2.

 

Specifications of the Four Primers Used for PCR Amplification of the 16S rDNA Gene of Bacteria

Primer Name	Primer Sequence	Position on the Gene Sequence of E. coli
16SF	5′ TTGGAGAGTTTGATCCTGGCTC 3′	4–25
16SR	5′ ACGTCATCCCCACCTTCCTC 3′	1174–1194
NF	5′ GGCGGCAKGCCTAAYACATGCAAGT 3′	42–66
NR	5′ GACGACAGCCATGCASCACCTGT 3′	1044–1067

 F, forward primer; R, reverse primer; N, nested; K, G/T; Y, C/T; S, C/G.

Figure 1.

 

(a) Using primers 16SF + 16SR complimentary to 16S rDNA amplified a 1.2-kb product from all species of bacteria. Lane 1, molecular weight ladder (Promega); lane 2, S. pneumoniae; lane 3, S. viridans; lane 4, S. pyogenes; lane 5, S. faecalis; lane 6, B. cereus; lane 7, P. aeruginosa; lane 8, P. acnes; lane 9, E. coli; lane 10, S. marcescens; lane 11, P. mirabilis; lane 12, K. pneumoniae; lane 13, H. influenzae; lane 14, control negative (no added DNA). (b) Using primers 16SF + 16SR complimentary to 16S rDNA amplified a 1.2-kb product from all isolates of coagulase-negative staphylococci. Lane 1, molecular weight ladder (Gibco BRL); lane 2, S. epidermidis NCTC 11047; lanes 3 through 14, clinical isolates from cases of culture-positive endophthalmitis reported as secondary to coagulase-negative staphylococci; lane 15, control negative (sterile water with no added DNA); lane 16, control negative (normal vitreous with no added DNA). (c) Using primers 16SF + 16SR complimentary to 16S rDNA amplified a 1.2-kb product from all isolates of S. aureus. Lane 1, molecular weight ladder (Gibco BRL); lanes 2 through 11, clinical isolates from cases of culture-positive endophthalmitis reported as secondary to S. aureus; lane 12, S. aureus NCTC 08532; lane 13, control positive (10 ng of genomic DNA from S. pneumoniae); lane 14, control negative. (d) Using primers NF and NR complimentary to 16S rDNA amplified a single 1.0-kb product from all species of bacteria. Lane 1, molecular weight ladder (Gibco BRL); lane 2, E. coli; lane 3, S. marcescens; lane 4, P. aeruginosa; lane 5, H. influenzae; lane 6, P. mirabilis; lane 7, K. pneumoniae; lane 8, S. pneumoniae; lane 9, S. viridans; lane 10, S. pyogenes; lane 11, S. faecalis; lane 12, coagulase-negative staphylococci; lane 13, S. aureus; lane 14, P. acnes; lane 15, B. cereus; lane 16, control negative from first round after two rounds of PCR; lane 17, control negative for nested PCR; lane 18, control positive from first round after two rounds of PCR; lane 19, control positive for nested PCR.

Figure 2.

 

(a) First-round PCR amplification of 16S rDNA detects 10 pg of genomic E. coli DNA. Lane 1, molecular weight ladder (Gibco BRL); lane 2, 10 ng DNA; lane 3, 1 ng DNA; lane 4, 100 pg DNA; lane 5, 10 pg DNA; lane 6, 1 pg DNA; lane 7, 100 fg DNA; lane 8, 50 fg DNA; lane 9, 10 fg DNA; lane 10, 5 fg DNA; lane 11, 1 fg DNA; lane 12, blank lane; lane 13, control negative (ddH₂O); lane 14, blank lane; lane 15, control positive (10 ng genomic DNA from coagulase-negative staphylococci). (b) Nested PCR amplification of 16S rDNA detects 1 fg of genomic E. coli DNA. Lane 1, molecular weight ladder (Gibco BRL); lane 2, 10 ng DNA; lane 3, 1 ng DNA; lane 4, 100 pg DNA; lane 5, 10 pg DNA; lane 6, 1 pg DNA; lane 7, 100 fg DNA; lane 8, 50 fg DNA; lane 9, 10 fg DNA; lane 10, 5 fg DNA; lane 11, 1 fg DNA; lane 12, control positive from first-round after two rounds of PCR; lane 13, control positive for nested PCR (10 ng genomic DNA from coagulase-negative staphylococci); lane 14, control negative from first-round PCR (ddH₂O); lane 15, control negative for nested PCR (ddH₂O). (c) First-round PCR amplification of 16S rDNA detects 10 pg of genomic DNA from coagulase-negative staphylococci. Lane 1, molecular weight ladder (Gibco BRL); lane 2, 10 ng DNA; lane 3, 1 ng DNA; lane 4, 100 pg DNA; lane 5, 10 pg DNA; lane 6, 1 pg DNA; lane 7, 100 fg DNA; lane 8, 50 fg DNA; lane 9, 10 fg DNA; lane 10, 1 fg DNA; lane 11, control negative (ddH₂O). (d) Nested PCR amplification of 16S rDNA detects 1 fg of genomic DNA from coagulase-negative staphylococci. Lane 1, molecular weight ladder (Gibco BRL); lane 2, 10 ng DNA; lane 3, 1 ng DNA; lane 4, 100 pg DNA; lane 5, 10 pg DNA; lane 6, 1 pg DNA; lane 7, 100 fg DNA; lane 8, 50 fg DNA; lane 9, 10 fg DNA; lane 10, 5 fg; lane 11, 1 fg DNA; lane 12, control positive for nested PCR (10 ng genomic DNA from E. coli); lane 13, control negative from first-round PCR (ddH₂O); lane 14, control negative for nested PCR (ddH₂O).

Figure 3.

 

(a) Comparison of first-round PCR amplification of 16S rDNA and culture of serial dilutions of coagulase-negative staphylococci. Ten microliters of PCR product was loaded in each lane of a 1% agarose/TBE gel. Lane 1, molecular weight ladder (Gibco BRL); lanes 2 through 13, serial 10-fold dilutions; lane 14, control negative; lane 15, control positive (10 ng of E. coli genomic DNA in water). Number of organisms cultured from a typical experiment, from 5 μl of each dilution, appears above each lane. Overnight cultures were used for each experiment and all dilutions were made in water and vortexed. Individual 5-μl aliquots were immediately PCR-amplified and cultured on blood agar. Colonies were counted after overnight aerobic culture at 37°C. c, confluent growth; sc, semi-confluent growth. (b) Comparison of nested PCR of 16S rDNA and culture of serial dilutions of coagulase-negative staphylococci demonstrates a detection sensitivity of one organism by PCR. Five microliters of PCR product was loaded in each lane of a 1% agarose/TBE gel. Lane 1, molecular weight ladder (Gibco BRL), lanes 2 through 13, 1 μl of first-round PCR product in (a); lane 14, control negative from first-round PCR (ddH₂O); lane 15, control negative for nested PCR (ddH₂O). The number of organisms cultured from a typical experiment, from 5 μl of each dilution, appears above each lane of (a). Overnight cultures were used for each experiment, and all dilutions were made in water and vortexed. Individual 5-μl aliquots were immediately PCR-amplified and cultured on blood agar. Colonies were counted after overnight aerobic culture at 37°C. c, confluent growth; sc, semi-confluent growth. (c) Comparison of first-round PCR amplification of 16S rDNA and culture of serial dilutions of E. coli. Ten microliters of PCR product was loaded in each lane of a 1% agarose/TBE gel. Lane 1, molecular weight ladder (Gibco BRL); lanes 2 through 13, serial 10-fold dilutions; lane 14, control negative; lane 15, control positive (10 ng of genomic DNA from coagulase-negative staphylococci in water). The number of organisms cultured from a typical experiment, from 5 μl of each dilution, appears above each lane. Overnight cultures were used for each experiment, and all dilutions were made in water and vortexed. Individual 5-μl aliquots were immediately PCR-amplified and cultured on blood agar. Colonies were counted after overnight aerobic culture at 37°C. c, confluent growth; sc, semi-confluent growth. (d) Comparison of nested PCR of 16S rDNA and culture of serial dilutions of E. coli demonstrates a detection sensitivity of one organism by PCR. Five microliters of PCR product was loaded in each lane of a 1% agarose/TBE gel. Lane 1, molecular weight ladder (Gibco BRL); lanes 2 through 13, 1μ l of first-round PCR product in (c); lane 14, blank lane; lane 15, control negative from first-round PCR (ddH₂O); lane 16, control positive from first-round PCR after two rounds of PCR; lane 17, blank lane; lane 18, control negative for nested PCR (ddH₂O); lane 19, control positive for nested PCR (10 ng of genomic DNA from coagulase-negative staphylococci in water). The number of organisms cultured from a typical experiment, from 5 μl of each dilution, appears above each lane on (c). Overnight cultures were used for each experiment, and all dilutions were made in water and vortexed. Individual 5-μl aliquots were immediately PCR-amplified and cultured on blood agar. Colonies were counted after overnight aerobic culture at 37°C. c, confluent growth; sc, semi-confluent growth.

Figure 4.

 

Restriction analysis of nested PCR fragments with AflIII, BssHII, ClaI, DraI, DraIII, HpaI, NdeI, NsiI, and SalI successfully differentiated between the PCR-amplified products from 13 of 14 bacterial species. Digests were resolved on 10% polyacrylamide/TBE gels and visualized after ethidium bromide staining under UV illumination. Lane 1, 100-bp molecular weight ladder (Advanced Biotechnologies); lane 2, E. coli; lane 3, S. marcescens; lane 4, P. aeruginosa; lane 5, H. influenzae; lane 6, P. mirabilis; lane 7, K. pneumoniae; lane 8, S. pneumoniae; lane 9, S. viridans; lane 10, S. pyogenes; lane 11, S. faecalis; lane 12, coagulase-negative staphylococci; lane 13, S. aureus; lane 14, P. acnes; lane 15, B. cereus; lane 16, 100-bp molecular weight ladder (Advanced Biotechnologies).

Figure 5.

 

(a) Nested PCR amplification of 16S rDNA from vitreous samples from two cases of presumed bacterial endophthalmitis confirmed the presence of 16S rDNA sequences. Results of first-round PCR are not shown because only the control positive was positive. Lane 1, molecular weight ladder (Gibco BRL); lane 2, control positive (10 ng E. coli DNA) after two rounds of PCR; lane 3, direct PCR of a 5-μl aliquot of vitreous from a clinical case of acute bacterial endophthalmitis, positive for E. coli as determined by culture; lane 4, direct PCR of 5-μl aliquot of vitreous from a clinical case of acute bacterial endophthalmitis, culture negative; lane 5, control negative from first-round PCR (ddH₂O); lane 6, control negative for nested PCR (ddH₂O). (b) RFLP analysis of amplified 16S rDNA PCR products from two cases of presumed bacterial endophthalmitis specifically identifies each pathogen. Lane 1, 100-bp molecular weight ladder (Advanced Biotechnologies); lane 2, control positive (10 ng E. coli DNA) after two rounds of PCR; lane 3, RFLP analysis of amplified PCR product from the vitreous sample of a clinical case of acute bacterial endophthalmitis, positive for E. coli as determined by culture, demonstrates the pattern obtained for E. coli/S. marcescens by restriction enzyme digestion; lane 4, RFLP analysis of amplified PCR product from vitreous of a clinical case of acute bacterial endophthalmitis, which proved culture negative, demonstrates the pattern obtained for coagulase-negative staphylococci by restriction enzyme digestion.

The authors thank the Endophthalmitis Study Group at Moorfields Eye Hospital (John Dart, Linda Ficker, Susan Lightman, and Steve Tuft) and Phillip Luthert, Department of Pathology, The Institute of Ophthalmology, for their help and support, which is gratefully appreciated.