PREVALENCE AND CHARACTERIZATION OF INTEGRONS IN MULTIDRUG- RESISTANT NON-CLINICAL ENTERIC BACTERIAL ISOLATES. A Thesis

PREVALENCE AND CHARACTERIZATION OF INTEGRONS IN MULTIDRUG- RESISTANT NON-CLINICAL ENTERIC BACTERIAL ISOLATES A Thesis Presented to the faculty of the Department of Biological Sciences California State University, Sacramento Submitted in partial satisfaction of the requirements for the degree of MASTER OF SCIENCE in Biological Sciences (Molecular and Cellular Biology) by Aaron Lee Avila SPRING 2013

2013 Aaron Lee Avila ALL RIGHTS RESERVED ii

PREVALENCE AND CHARACTERIZATION OF INTEGRONS IN MULTIDRUG- RESISTANT NON-CLINICAL ENTERIC BACTERIAL ISOLATES A Thesis by Aaron Lee Avila Approved by:, Committee Chair Susanne W. Lindgren, Ph.D, Second Reader Enid T. Gonzalez-Orta, Ph.D, Third Reader Nicholas N. Ewing, Ph.D Date iii

Student: Aaron Lee Avila I certify that this student has met the requirements for format contained in the University format manual, and that this thesis is suitable for shelving in the Library and credit is to be awarded for the thesis., Graduate Coordinator Jamie Kneitel, Ph.D Date Department of Biological Sciences iv

Abstract of PREVALENCE AND CHARACTERIZATION OF INTEGRONS IN MULTIDRUG- RESISTANT NON-CLINICAL ENTERIC BACTERIAL ISOLATES by Aaron Lee Avila Antibiotic resistance in bacteria has been a concern in the medical field for almost as long as antibiotics have been available. The last several decades have seen marked increases in antibiotic resistance, leading to the discovery of multidrug-resistant (MDR) bacteria, which can be resistant to several antibiotics. MDR bacteria are a major problem in the healthcare industry, creating to numerous challenges such as reduced treatment options, increased mortality rates, longer hospital stays, and increased costs. The increasing dissemination of resistance genes is believed to be the result of horizontal gene transfer via mobile genetic elements, including plasmids and transposons. Several studies have also shown that integrons play a significant role in the spread of resistance, acting as v

gene capture and expression mechanisms that are often associated with mobile genetic elements. However, most of the studies investigating the role of integrons in the dissemination of antibiotic resistance utilized bacterial samples from environmental sources or hospitalized patients. Far fewer studies have examined the role of integrons in the propagation of multidrug-resistance in bacteria from the lower intestinal tract of healthy individuals. The purpose of this study was to determine whether or not integrons play a significant role in the proliferation of multidrug-resistance in enteric bacteria isolated from healthy, non-hospitalized adults. Attempts were also made to identify the gene cassettes and organization of cassettes within the identified integrons. Over the course of five years (2005-2009), a total of 92 enteric bacterial samples were collected from students at CSUS via a rectal swab, and isolated on MacConkey agar. These samples were isolated and subjected to a variety of antibiotics and biochemical tests to determine antibiotic resistance profiles and species. PCR amplification of class 1 and class 2 integrase genes (inti1 and inti2) yielded 19 (out of 84 unique samples) class 1 positive isolates, one of which was also found to be class 2 positive. Resistance to trimethoprim/sulfamethoxazole, ampicillin, and piperacillin was found to be significantly greater in class 1 positive isolates compared to class 1 negative isolates (P<0.05). Resistance to two or more classes of antibiotics was also significantly higher in class 1 positive isolates compared to class 1 negative isolates. Resistance to two or more antibiotics, regardless of class was also significantly higher in class 1 positive isolates. PCR amplification of the variable regions of inti1 and inti2 samples yielded seven unique vi

amplicons ranging in size from approximately 250bp to >3kbp. Subsequent sequencing and nucleotide BLAST searches led to the identification of eight different gene cassettes organized in six unique arrays., Committee Chair Susanne W. Lindgren, Ph.D Date vii

ACKNOWLEDGEMENTS I owe the successful completion of this thesis to several people, including those who have contributed directly to the project, as well as those who have supported me along the way. First, I would like to thank Scott Baker for his initial work in collecting samples and gathering raw data that was used in this study. I would also like to thank Windy Miller and Amy Crum for their help in collecting and isolating samples. Thank you also to Myra Rodriguez for her flexibility in always meeting my ever-changing scheduling needs for school. A special thank you must go out to Dr. Susanne Lindgren for inviting me to join her lab when I was much too shy to ask myself. She offered me the opportunity to take over this project and the freedom to choose the direction to take it. I thank her for her support, guidance, and sense of calm when things were not going as planned. I also would like to extend my gratitude to my committee members, Dr. Enid Gonzalez-Orta and Dr. Nicholas Ewing for coming through and meeting deadlines with short notice. I would like to express a warm thank you to my brother and all of my friends for their continued support and encouragement over the years. In times of stress, they were instrumental in helping me to relax, take a break, and enjoy life. Finally, and most importantly, I owe a great deal of gratitude to my parents. Without their unwavering love, steadfast support, and continuous encouragement, I never would have completed this program. Thank you so much for helping me realize my dreams. I love you all, and I am finally finished! viii

TABLE OF CONTENTS Page Acknowledgements... viii List of Tables...x List of Figures... xi INTRODUCTION...1 MATERIALS AND METHODS...8 RESULTS...22 DISCUSSION...49 Appendix...58 Literature Cited...76 ix

LIST OF TABLES Tables Page 1. Primers Used For Detection of Class 1 and Class 2 Integrases and Amplification of Variable Regions... 16 2. Species Identifications... 25 3. Susceptibility Comparisons Between Class 1 Positive and Class 1 Negative Isolates for Each Tested Antibiotic... 26 4. Comparisons Between inti1-positive and inti1-negative Isolates Resistant to Multiple Classes of Antibiotics... 31 5. Comparison Between inti1-positive and inti1-negative Isolates With Intermediate or Resistant Phenotypes for Multiple Classes of Antibiotics... 34 6. Comparison Between inti1-positive and inti1-negative Isolates Resistant to Multiple Antibiotics, Regardless of Class... 34 7. Observed vs. Actual Amplicon and Restriction Fragment Sizes... 40 8. Identified Gene Cassettes... 47 x

LIST OF FIGURES Figures Page 1. Layout of MicroScan Gram Negative Combo Panels... 11 2. Sample Biotype Number Panel Worksheet... 14 3. Relative Primer Locations... 18 4. Class 1 Integron Detection... 27 5. Class 2 Integron Detection... 29 6. Percentage of Cumulative Resistant Integrase-positive vs. Integrase-negative Isolates for Varying Numbers of Antibiotic Classes... 32 7. Number of Resistant Integrase-positive vs. Integrase-negative Isolates for Varying Numbers of Antibiotic Classes... 33 8. Percentage of Cumulative Resistant Integrase-positive vs. Integrase-negative Isolates for Varying Numbers of Antibiotics, Regardless of Class... 35 9. Number of Resistant Integrase-positive vs. Integrase-negative Isolates for Varying Numbers of Antibiotics, Regardless of Class... 36 10. Class 1 Integron Variable Regions... 37 11. Class 2 Integron Variable Regions... 39 12. Class 1 Variable Region Restriction Fragments... 42 13. Variable Region Partial Alignment... 43 14. Gene Cassette Arrangements in Class 1 and Class 2 Integrons... 48 xi

1 INTRODUCTION The discovery of antibiotics in the late 1920s, and their subsequent use in treating and preventing infections beginning in the 1940s, is undoubtedly one of the great medical breakthroughs in the last 100 years (14, 15). In the early years of antibiotic treatment, many scientists and doctors believed that infectious disease had been triumphed once and for all (14). And while it is true that antibiotics have largely nullified several diseases and infections that were once very difficult to treat, there is reason to be concerned that this may not always be the case. Less than a decade after the first antibiotics were introduced in medicine, evidence of bacterial strains resistant to those antibiotics began to surface (14, 15). Shortly thereafter, scientists uncovered evidence that bacteria were not only capable of developing resistance to one antibiotic, but to multiple antibiotics that were also transferable to sensitive strains (14). The rise of multidrug-resistant (MDR) bacteria is a result of unscrupulous antibiotic use in medicine and agriculture over the last several decades (5, 12, 14, 31). Today, MDR bacteria provide numerous challenges and problems for healthcare providers, including increases in hospital-acquired infections, reduced treatment options, higher morbidity and mortality rates, and healthcare cost increases due to longer hospital stays (16, 43). MDR bacteria may be resistant to a couple of antibiotics, several classes of antibiotics, and in some cases every antibiotic (8). Even MDR bacteria that are resistant to only a couple of antibiotics can greatly complicate treatment. Often, such bacteria are resistant to the primary antibiotic preferred for treatment, requiring the use of secondary and tertiary drugs instead, which may be less effective and more toxic to the patient (8).

2 The growing problem of MDR infections is made even more concerning by the fact that new discoveries of antimicrobial agents have been few and far between in recent years (11, 14). Over the last five decades, only two new classes of antibiotics have reached the market, and current information suggests that no new antibiotic classes will be introduced in the near future (11). Without the continuous introduction of new antibiotics, as was seen during the first 20 years of their use, the threat of a return to the pre-antibiotic era is very real (11, 15). Perhaps the most widely publicized strain of MDR bacteria is the much-feared Gram-positive methicillin resistant Staphylococcus aureus (MRSA) (15, 33, 39). However, less well-publicized MDR Gram negative bacteria are also capable of causing serious, difficult to treat infections. The Antimicrobial Availability Task Force, established by the Infectious Diseases Society of America, identified several particularly problematic pathogens, one of which included extended-spectrum beta-lacatamase (ESβL)-producing Enterobacteriaceae (e.g. Escherichia coli and Klebsiella pneumoniae) (46). ESβLs are enzymes produced by bacteria that confer resistance to multiple antibiotic classes, namely cephalosporins, penicillins, monobactams, and beta-lactamases. (46). Over 500 different ESβLs have been identified, the most common belonging to the CTX and CMY gene families (46). Infections caused by ESβL producers usually must be treated with a carbapenem (e.g. imipenem, meropenem). Recently however, ES L- producing Gram-negatives have been identified that are also resistant to the carbapenem class of antibiotics (46, 9). Carbapenem-resistant Enterobacteriaceae (CRE), or superbugs, as the media often refers to them, produce metallo-beta-lactamases (MβLs)

3 that readily cleave most β-lactam substrates (46, 9). As with ESβLs, several MβL enzymes have been discovered, the latest apparently originating from India, identified as NDM-1 (53). Although drug resistance is generally discussed with regard to pathogenic bacteria, not all antibiotic-resistant bacteria are necessarily harmful to their host. Bacteria comprising normal human flora in asymptomatic individuals have certainly been shown to carry resistance to antibiotics (1, 29, 48). E. coli and K. pneumoniae may make up part of a normal intestinal flora, where they cause no problems; however, introduction of these strains to other areas of the body, or to other people, can cause infections such as urinary tract infections, pneumonia, and septicemia (16, 43). These types of infections are most common in people with weakened immune systems and individuals who are hospitalized for other reasons (16, 43). Most often, such infections are acquired in hospitals. The CDC has estimated that as many as 1.7 million hospital-acquired infections result in nearly 100,000 deaths each year in the United States (39). Significant problems arise in the treatment of these infections, especially when they are caused by MDR bacteria. Antibiotic-resistant organisms residing as part of a person s intestinal flora, whether they are pathogenic strains or not, may act as a reservoir for resistance genes that can be transferred to other bacteria (28). Bacteria are able to transfer resistance genes horizontally to one another through various mechanisms. The emergence of MDR bacteria is the result of horizontal gene transfer (7, 14, 15, 40), where genetic information is passed directly from one bacterium to another. Horizontal transfer of antibiotic

4 resistance genes occurs primarily through two different genetic elements: plasmids and transposons (20, 7, 29, 31). Plasmids are small, circular, extrachromosomal DNA molecules that may contain resistance genes (29). Plasmids can be transferred via a pilus from one bacterium to another in a process called conjugation (42). The recipient bacterium acquires all genes present on the plasmid, including resistance genes. Like plasmids, transposons can also carry resistance genes. Transposons are genetic elements that may be inserted into and excised from chromosomes and plasmids (20). Through sharing of DNA via these two mechanisms, bacteria can rapidly acquire new genes that make them immune to various antibiotics. A third group of genetic elements that have been strongly implicated in the emergence of MDR bacteria are called integrons (23). While integrons themselves are not mobile elements, they are frequently associated with transposons and plasmids. Plasmidintegrated transposons carrying antibiotic resistance genes can be transferred to other bacteria through conjugation (20, 7, 29, 31). Integrons are capture-and-expression genetic elements that facilitate site-specific recombination of promoter-less gene cassettes into a site that allows for the transcription of all genetic material contained in the cassettes (7, 18, 31). They consist of three main components located in the 5 conserved region: an integrase gene (inti), a recombination site (atti), and an active promoter (7, 18, 23, 31). The integrase recognizes a conserved, 59-base element (actually varies in length from 57-141 bases), which is found on resistance gene cassettes (7, 18, 31, 45). Upon recognition of this conserved element, the integrase facilitates the integration of the cassette into the integron at the atti site, just downstream of the active promoter (7, 18, 31). Any cassettes

5 that are integrated downstream of the promoter are then free to be transcribed; they may also be rearranged or excised via the integrase, and new promoter-less resistance genes can be integrated (7, 18, 31). Thus, integrons are essentially genetic elements capable of integrating and expressing various rearrangeable antibiotic resistance gene cassettes that can be readily mobilized into neighboring bacteria. At least three classes of integrons have been identified, which are distinguished primarily by the integrase gene. Genes contained within the 3 conserved region also vary between the three classes of integrons. Class 1 and class 2 integrons are the most prevalent and best studied (2, 18). Class 3 integrons appear to be far less common, and therefore less implicated in the spread of multidrug-resistance. Class 3 integrons have been found in Serratia marcescens (3), Klebsiella pneumoniae (13), as well as Delftia species (51). Class 1 integrons, on the other hand, have been found in many Gramnegative Enterobacteria, including species of Escherichia, Klebsiella, Pseudomonas, Enterobacter, Salmonella, Proteus, Serratia, Citrobacter, and Shigella (18, 23). Integrons are known to contain highly conserved regions at the 5 end (which encodes the integrase gene) as well as the 3 end, downstream of integrated gene cassettes. The 3 conserved region of class 1 integrons consists of the qacδ1 and sul1 genes, which confer resistance to quaternary ammonium compounds and sulfonamides, respectively (23, 24). Class 2 integrons appear to be less widespread, although they have been identified in several genera of bacteria, such as Shigella, Salmonella, and Acinetobacter (2, 10, 18, 37, 38), as well as Escherichia, Morganella, and Aeromonas (35). Integrons are believed to a play considerable role in the dissemination of antibiotic resistance genes within Gram-

6 negative bacteria (7, 14, 18, 31). A group of researchers recently created a database, called the Repository of Antibiotic resistance Cassettes (RAC), which contains over 300 different promoter-less gene cassettes (47). Several of these antibiotic resistance gene cassettes are frequently seen integrated into both class 1 and class 2 integrons, including those granting resistance to aminoglycosides, cephalosporins, chloramphenicol, penicillins, and trimethoprim (7). The association between antibiotic resistance and integrons has been well documented. Integrons have been shown to be particularly prevalent in many clinical isolates of Gram-negative enteric bacteria. Integron frequencies in clinical samples as high as 88% (37), and as low as 13% (36) have been found, though more common frequencies fall in the range of 20%-60% (2, 10, 17, 25, 38, 50). There have also been numerous studies investigating the prevalence of integrons in bacteria isolated from sources other than humans. Such sources include wastewater treatment plants (35), irrigation sediments (40), and animals (5, 6, 19, 21, 52). Far fewer studies have been conducted to investigate the prevalence of multidrug-resistance in bacteria obtained from healthy, non-hospitalized individuals. Studies that include commensal bacteria obtained from humans often include clinical isolates (41), or a combination of animal and human derived specimens (12, 32). One study that investigated integrons in a mixed sample set of animal, commensal human, and clinical human isolates did find that MDR was associated with the presence of integrons regardless of origin, indicating that a positive correlation between MDR and commensal human isolates had been established. Another study investigated the transfer of antibiotic resistance genes among nonpathogenic

7 Bacteroides within the human colon, but no attempt was made to identify the presence of integrons or investigate their possible role (42). Through an IRB-approved exemption, a collection of antibiotic-resistant enteric bacteria from healthy CSUS students was accumulated over the course of five years. Multidrug-resistance was observed in several of the enteric isolates. I hypothesized, based on previous research, that the prevalence of class 1 and class 2 integrons would be significantly greater in multidrug-resistant enteric bacteria comprising normal flora of healthy adults than in isolates with low or no resistances. Few studies have attempted to examine the prevalence or role of integrons in the propagation of MDR bacteria that exist as part of the normal human intestinal flora. By determining the prevalence of integrons within the drug-resistant samples collected, some insight may be gained into the role of integrons in the dissemination and maintenance of multidrug-resistance factors in the community.

8 MATERIALS AND METHODS Sample Collection Over the course of five years, through an IRB-approved exemption, enteric bacterial samples were collected from undergraduate microbiology students at CSUS. As part of a voluntary lab exercise, a self-administered sterile rectal swab was used to obtain enteric bacteria from students. Once inoculated, swabs were rubbed over MacConkey agar plates, and four antibiotic diffusion discs were placed on the plate. The antibiotic discs used were ampicillin, erythromycin, tetracycline, and sulfamethoxazole/ trimethoprim. In addition, an antibiotic disc containing ciprofloxacin was also used on one of the agar plates. Plates were then incubated for approximately 24 hours at 37ºC. After the students had finished using the bacteria for their lab exercises, the plates were wrapped with Parafilm and stored at 4ºC for up to three weeks. Bacterial colonies exhibiting growth within zones of inhibition of the antibiotic discs were streaked for isolation onto MacConkey agar plates and incubated for 24 hours at 37ºC. To ensure purity, this process was repeated at least twice for each sample. In some cases, more than one colony was taken from the initial plate (i.e. more than one antibiotic-resistant sample was obtained from the same individual) if there were colonies growing within zones of inhibition of more than one antibiotic. Once isolated, each antibiotic-resistant bacterial isolate was grown overnight in 5 ml of lysogeny broth (LB) in a 37ºC water bath shaking at 50 shakes per minute. Frozen stocks of each isolate were made in duplicate by mixing 0.5 ml of overnight culture with 0.5ml of 80% glycerol in a 1.2 ml freezer vial, vortexing briefly, and placing into a -80ºC freezer. All samples were

9 collected using these methods during the fall of 2005, 2006, 2008, and 2009; no collection was made in 2007. Species Identification and Antibiotic Susceptibility Testing Each antibiotic-resistant enteric isolate was subjected to a variety of biochemical and antibiotic susceptibility tests via Dade Behring MicroScan Dried Overnight Gram Negative Combo Panels (West Sacramento, CA). A total of three different types of panels were used: NC 32, NC 30, and NBPC 30. NC 30 panels were used after the NC 32 panel stock was depleted, and NBPC 30 panels were used after the NC 30 panel stock was depleted. Most samples were tested using only one of the three types of panels. However, some samples were re-tested based on inconclusive results for the species identification. These samples (0806, 0809, 0816B, and 0915) were re-tested on the NBPC 30 panels. All three panels contain identical biochemical tests for species identification. However, each panel does differ in the antibiotics it tests for and/or the concentrations of each antibiotic. Compared to NC 32 panels, NC 30 panels contain two additional antibiotic tests: gatifloxacin and amoxicillin/k clavulanate. NC 30 panels also test additional concentrations of cefotetan, cephalothin, ceftriaxone, cefazolin and piperacillin/tazobactam. NC 30 panels do not contain tests for cefotaxime, ticarcillin/k clavulanate, moxifloxacin, or meropenem, and use fewer concentrations of amikacin and tobramycin. NBPC 30 is a breakpoint panel, containing all of the antibiotics from NC 30 and NC 32 panels, an additional concentration of nitrofurantoin, as well as four additional

10 antibiotics: chloramphenicol, norfloxacin, cefoxitin, and tetracycline. Because it is a breakpoint panel, NBPC 30 panels contain fewer concentrations for many of the antibiotics only the concentrations necessary to determine susceptibility. Figure 1 shows a diagram of each panel used, including the concentrations of all antibiotics. Panels were inoculated using the Turbidity Standard Technique according the Dade Behring MicroScan Dried Gram Negative Procedural Manual (34). After incubation of the panels at 37ºC for 18 hours, the biochemical results of each panel were read manually and interpreted as indicated by the manufacturer s instructions. Based on the results of the biochemical tests, a worksheet was used to generate a biotype number for each isolate (Figure 2). The MicroScan Biotype Lookup Program (44) was used along with the biotype number to identify the species of each isolate as well as a confidence percentage. Minimum inhibitory concentrations (MICs) for each antibiotic were also read manually according to the procedural manual for the panels (34). Once MICs for each antibiotic were recorded, susceptibility was determined based on the Interpretive Breakpoints chart included in the procedural manual (34). Each sample was assigned a ranking of susceptible (S), intermediate (I), or resistant (R) based on their MIC for each antibiotic. Template DNA Preparation Template DNA was prepared using a simple, crude preparation technique, similar to that described by Mazel et al. (32). Frozen bacterial samples were first streaked for

11 C G P4 GLU RAF INO URE LYS TDA CIT TAR OF/G NIT K4 Cl4 SUC RHA ADO H2S ARG ESC MAL ACE OF/B 2/38 T/S Cf8 Fd64 SOR ARA MEL IND ORN VP ONP G CET DCB 4 8 16 32 Ak 4 8 16 Cfz ESBL -a 8 16 Am 8 16 Azt 1 2 4 8 Gm 4 8 16 Crm ESBL -b 8/4 16/8 A/S 1 2 Cp 1 2 4 8 To 8 16 64 P/T LOC 16 64 Pi 2 4 Lvx 4 8 16 32 Cft 16 32 Ctn 8 32 Cax 16 64 Tim 2 4 Mxf 2 4 8 16 Caz 2 4 8 16 Cpe 4 8 Imp 4 8 Mer Figure 1-a. Layout of MicroScan Gram Negative Combo Panels. Negative Combo Panel Type 32. Orange: biochemical tests used in the determination of species; Green: biochemical tests not used; Pink: antibiotic tests, concentrations in μg/ml, abbreviations listed in Appendix B; Yellow: putative screen for ES L production; Blue: locator for automated panel analysis (not used).

12 C G P4 GLU RAF INO URE LYS TDA CIT TAR OF/G NIT K4 Cl4 SUC RHA ADO H2S ARG ESC MAL ACE OF/B LOC 2/38 T/S Fd64 SOR ARA MEL IND ORN VP ONP G CET DCB 2 4 8 16 Cfz 2 4 8 16 Am 16 32 Ak 1 2 Cp 4 8 16 32 Ctn 8 16 32 64 P/T 8/4 16/8 Aug 2 4 Gat 2 4 8 16 Caz 1 2 4 8 Gm 8/4 16/8 A/S 2 4 Lvx 4 8 16 32 Cax 2 4 8 To ESBL -a 16 64 Pi 8 16 Azt 2 4 8 16 Cpe 4 8 16 Crm ESBL -b 8 16 Cf 4 8 Imp Figure 1-b. Layout of MicroScan Gram Negative Combo Panels. Negative Combo Panel Type 30. Orange: biochemical tests used in the determination of species; Green: biochemical tests not used; Pink: antibiotic tests, concentrations in μg/ml, abbreviations listed in Appendix B; Yellow: putative screen for ES L production; Blue: locator for automated panel analysis (not used).

13 C G P4 GLU RAF INO URE LYS TDA CIT TAR OF/G LOC K4 Cl4 SUC RHA ADO H2S ARG ESC MAL ACE OF/B NIT ESBL -a ESBL -b SOR ARA MEL IND ORN VP ONP G CET DCB 8 16 Am 16 64 Pi 16 32 Ak 8 16 Cfz 8 32 Cft 1 2 Cp 8/4 16/8 A/S 16 64 P/T 4 8 Gm 8 16 Cf 8 16 Caz 2 4 Gat 8/4 16/8 Aug 4 8 Te 4 8 To 16 32 Ctn 8 32 Cax 2 4 Lvx 32 64 Fd 16 64 Tim 8 16 Azt 8 16 Cfx 8 16 Cpe 2 4 Mxf 4 8 Imp 4 8 Mer 2/38 T/S 4 8 16 Crm 8 16 C 4 8 Nxn Figure 1-c. Layout of MicroScan Gram Negative Combo Panels. Negative Breakpoint Combo Panel Type 30. Orange: biochemical tests used in the determination of species; Green: biochemical tests not used; Pink: antibiotic tests, concentrations in μg/ml, abbreviations listed in Appendix B; Yellow: putative screen for ES L production; Blue: locator for automated panel analysis (not used).

14 Glucose Fermenter 4 + 2 + 1 GLU RAF INO URE LYS TDA CIT Cl>4 TAR OF/G Cl>4 NIT SUC RHA ADO H 2S ARG ESC MAL Cf>8 ACE P>4 Fd>64 OXI SOR ARA MEL IND ORN VP ONPG OXI CET K>4 To>4 4 + 2 + 1 Glucose Non-Fermenter Identification Figure 2. Sample Biotype Number Panel Worksheet. Only the first eight columns are used to generate biotype numbers for glucose fermenters (100% of samples tested). Positive results in the top row score four points, second row scores two points, and third row scores one point. Points are added in each column to generate an eight-digit biotype number.

15 isolation on LB agar and incubated overnight at 37ºC. An isolated colony from the plate was used to inoculate 1 ml of LB media, which was then grown overnight in a 37ºC water bath shaking at 50 shakes per minute. The overnight culture was transferred to a sterile 1.5 ml Eppendorf tube and centrifuged at 6000 rpm for approximately 1 minute. The supernatant was then discarded, and the pellet of bacteria was re-suspended in 0.5 ml sterile de-ionized water. After briefly vortexing the suspension, the tubes were placed in a 100ºC water bath for 10 minutes to lyse the bacteria. The tubes were then centrifuged again at 6000 rpm for 5 minutes to pellet cell debris. The supernatant was removed and placed into sterile 0.5 ml tubes for use as template DNA. PCR Detection of class 1 and class 2 Integrons Detection of class 1 and class 2 integrons relied on amplifying a section of the integrase gene (inti1 and inti2, respectively) via two separate PCR assays. Successful amplification of either gene indicated the presence of an integron of the corresponding class. Primer sets are listed in Table 1. Positive controls were used for both class 1 and class 2 integron assays. Salmonella enterica serovar Typhimurium strain DT104 was used as the positive control for the class 1 assay, as it is know to carry a class 1 integron (22). For the class 2 positive control, a strain of E. coli (ATCC# 47055) was chosen because it contains a Tn7 transposon, which is known to contain a class 2 integron (6, 12).

Integron Class Class 1 Class 2 Target inti1 (integrase) Class 1 Conserved inti2 (integrase) Class 2 Conserved Primer int1-f int1-r Hep58 Hep59 inti2-f inti2-r hep74 hep51 Sequence (5 to 3 ) GGGTCAAGGATCTGGATTTCG ACATGCGTGTAAATCATCGTCG TCATGGCTTGTTATGACTGT GTAGGGCTTATTATGCACGC TTATTGCTGGGATTAGGC ACGGCTACCCTCTGTTATC CGGGATCCCGGACGGCATGCACGATTTGTA GATGCCATCGCAAGTACGAG Amplicon Size (bp) 484 Variable (~1k-5k+) 233 Variable (~1k-5k+) Annealing Temp (ºC) Table 1. Primers Used For Detection of Class 1 and Class 2 Integrases and Amplification of Variable Regions. 61 56 53 59 Reference 32 49 19, 35 27, 35, 50 16

17 All PCR reactions were performed in 50 μl volumes. The class 1 dectection assay was composed of the following: 1.0 μl of 10mM dntp Mix (datp, dttp, dctp, dgtp- 0.2mM final concentration) (Promega), 5.0 μl 10x Taq Buffer Advanced (5Prime), 0.5 μl Taq Polymerase (5U/μl) (5Prime) added after 4 minutes of denaturation, 10 μl of 2.5 μm inti1-f primer, 10 μl of 2.5 μm inti1-r primer (0.5 μm final) (Sigma), 10.0 μl template DNA, and 13.5 μl dh 2 0. Negative controls were run in all assays with 10.0 μl dh 2 0 in place of template DNA. The cycle parameters were as follows: 7 minutes of predenaturation at 94ºC, followed by 30 cycles of 94ºC for 1 minute, 61ºC for 1 minute, and 72ºC for 1 minute, and a final elongation step of 72ºC for 8 minutes at the end. The class 2 detection assay reaction mixture was identical to the class 1 assay except for the following changes: 2.0 μl of 25mM MgCl2 was added, the dh 2 0 volume was reduced to 11.5 μl, and inti2 forward and reverse primers were used to target the inti2 gene. The cycle parameters were as follows: 7 minutes of pre-denaturation at 94ºC, followed by 32 cycles of 94ºC for 1 minute, 53ºC for 1 minute, and 72ºC for 45 seconds, with a final elongation step of 72ºC for 8 minutes at the end. PCR Amplification of Integron Variable Regions Samples that tested positive for either class 1 or class 2 integrase were used in separate PCR assays designed to amplify the variable region of the integron. Primers were used to target conserved sequences on opposite sides (5 and 3 conserved sequences) of the variable region of the integrons (See Figure 3 for relative primer locations). The class 1 variable region assay was identical to the class 1 detection assay

18 A B Figure 3. Relative Primer Locations. A: Primer locations for amplifying section of inti1 (class 1) and inti2 (class 2) genes in 5 -conserved region of integron; B: Primer locations amplifying variable region between 5 -conserved region and 3 -conserved region

19 except for the hep58 and hep59 primer pair that was used (49). The cycle parameters for the class 1 variable assay were as follows: 5 minutes of pre-denaturation at 94ºC, followed by 33 cycles of 1 minute at 94ºC, 45 seconds at 56ºC, and 5 minutes at 72ºC, and a final elongation step of 5 minutes at 72ºC at the end. The class 2 variable region assay was identical to the class 1 variable assay except for the use of a higher annealing temperature of 59ºC and a different primer pair, hep74 and hep51, which targets conserved regions of class 2 integrons (27, 35, 50). Agarose Gel Electrophoresis of PCR Products All PCR products were visualized by running 20.0 μl of PCR product mixed with 2.0 μl of loading dye on agarose gels. Products from the class 1 and class 2 integrase detection assays were run on 2% gels, as their products were 484bp and 233bp respectively. Products from the class 1 and class 2 variable region assays were run on 1% gels as most of their products ranged from approximately 1kbp to >3kbp, depending on the length of the respective integron. DNA ladders (Sigma-1kbp and Morganville Scientific-100bp) were run on each gel. Gels ran at 90V for approximately 45 minutes and then stained in ethidium bromide before being visualized under UV light. Restriction Digest of Variable Region PCR Products Variable region PCR products that appeared to be similar in size were exposed to a restriction enzyme, AluI (BioLabs), in order to determine if the products were of the same sequence. AluI was chosen because its recognition sequence is only four bases, thus

20 increasing the likelihood of its activity over other enzymes that target six-base sequences. Approximately 30 μl of PCR product was mixed with 1.0 μl of 10U/ml AluI and incubated at 37ºC for 4 hours. The products were then run on a 2% agarose gel and visualized using the same procedures as described for PCR products. Identical sized patterns on the gels indicated the variable regions were likely of the same sequence, while different banding patterns on the gel would indicate different sequences. This step was taken to help reduce the risk of needlessly sequencing multiples of identical variable regions. Gel Extraction, Variable Region Sequencing, and Cassette Identification Based on the results of the restriction digest assay for variable region products of similar size, one sample representing each unique amplicon size was chosen for sequencing. Following PCR amplification and gel electrophoresis of variable region products as described above, DNA bands were extracted using the QIAquick Gel Extraction Kit (Qiagen) following the manufacturers protocol. A total of seven extracts (six class 1 samples and one class 2) were sent to Sequetech in Mountain View, CA for sequencing. Complete sequencing by primer walking was not performed due to cost. Instead, sequencing was performed using single primer extensions from both the 5 and 3 conserved regions in an effort to reduce cost while sequencing as much of the template as possible. For shorter variable regions, this was sufficient to identify all cassettes. However, for longer products, cassettes located in the middle of the variable region could not be identified. Sequencing data was used to conduct nucleotide searches using BLAST in order to identify gene cassettes.

21 Nomenclature of Antibiotic-resistant Enteric Isolates Antibiotic-resistant enteric isolates were assigned identification numbers according to the year in which they were collected. Sample ID numbers beginning with F06 indicate samples that were collected during the fall of 2006, while ID numbers beginning with 08 or 09 indicate samples that were collected during the fall of 2008 and 2009 respectively. Sample ID numbers that were collected in the fall of 2005 begin with either L or M. Arbitrary numbers were also assigned to identify samples that were derived from different individuals during the same collection year. These numbers, found after the number or letter indicating the collection year, were not used to identify specific individuals, nor were they used to track any characteristics about the individuals. In some cases, ID numbers were also labeled with regard to the antibiotic to which they initially exhibited resistance during the sample collection process. There are five different antibiotic-resistance labels: SXT (sulfamethoxazole/trimethoprim), TET (tetracycline), E (erythromycin), AMP (ampicillin), and CIP (ciprofloxacin). Finally, letters A, B, C, D, and E found at the end of the ID number indicate multiple samples that were collected from the same individual. For example, sample 0920A is sample number 20 collected in the fall of 2009, and was one of three isolates collected from the same individual.

22 RESULTS Identification of Samples and Resistance Profiles A total of 92 antibiotic-resistant enteric bacterial samples were collected and isolated from 66 unique healthy human donors. A probable species identification of each of the 92 samples was made by running each sample on a MicroScan Dried Gram Negative Panel to generate a biotype number based on the results of multiple biochemical tests contained on the panels. Each panel consisted of three rows of biochemical tests (top three rows on each panel, see Figure 1), not all of which were necessary for species identification. Only the tests necessary for the identification of glucose fermenters (100% of tested samples, n=92) were used to generate biotype numbers, indicated in Figure 1 by the orange shaded portions, and as shown on the panel worksheets (Figure 2). Each eightdigit biotype number generated a list of probable bacteria with a rough confidence percentage. The most probable species for each isolate was recorded. Species identifications and antibiotic susceptibilities for all samples not labeled as 08xx or 09xx were derived from biochemical results and MIC data from previous work in the Lindgren lab by Baker (4). All 92 antibiotic-resistant enteric isolates were also tested against a wide range of antibiotics of varying ranges of concentrations using the same panels that were used for species identification (Figure 1, shaded in pink). MICs for each antibiotic were generated based on the ability of the organism to grow at various concentrations. The MICs were then used to determine susceptibility of the organism to each antibiotic. Not all antibiotics contained on the panels were useful in determining susceptibility, since some of them

23 were tested only at one concentration to aid in species identification. Kanamycin, cephalothin, penicillin, chloramphenicol, nitrofurantoin, and colistin were tested at only one concentration for most of the samples, so susceptibility data for these antimicrobials was incomplete. Additionally, the use of three different panels resulted in not all of the samples being tested for the same antibiotics. In order to analyze the data appropriately, only those antibiotics that were tested on every sample and were able to generate an MIC were used for tabulation of results and calculations. Antibiotics that were not tested on every sample, and therefore omitted from calculations, include the following: cefotaxime, cefoxitin, tetracycline, ticarcillin/k clavulanate, amoxicillin/k clavulanate, gatifloxacin, norfloxacin, moxifloxacin, and meropenem. These omissions resulted in a reduced total of 18 antibiotics (representing seven classes) that were used to generate resistance profiles for all samples. Complete MIC data for all 18 of these antibiotics, as well as omitted antibiotics described above, is listed in Appendix A. For samples obtained from the same individual, their biochemical results and MIC results were compared to determine uniqueness. Samples derived from the same donor, but with differing results for two or more biochemical tests, differing MICs for more than one antibiotic, or different species identifications were deemed to be unique. The only exceptions to these criteria were for samples 0922A/0922B and M-5-Ea/M-5-Eb because 0922B and M-5-Ea were found to contain an integron, while 0922A and M-5-Eb do not contain an integron. In all, seven samples were determined to not be unique, and therefore were omitted from further analysis. Finally, one more sample (L-3-E) was removed from the analysis of the results due to failure to propagate the sample from

24 frozen storage after it had been tested on the panels, but before it could be tested for the presence of integrons. Therefore the final number of unique antibiotic-resistant enteric isolates that were tested for integrons was 84. Of the 84 unique isolates that were subsequently tested for the presence of integrons, E. coli was the most commonly identified species comprising 76.2% (n=64) of the samples. Other isolates included K. pneumoniae (4.8%, n=4), E. cloacae (4.8%, n=4), K. ascorbata (4.8%, n=4), and several other species at lower frequencies (see Table 2 for full species identification results, Appendix A for biotype numbers and confidence levels). The average number of resistances per sample was 3.33. Resistance profiles varied widely from sample to sample, from 13 resistances to zero resistances out of the 18 antibiotics tested for all samples. The antibiotic resisted most frequently was ampicillin, with 55 isolates out of 84 (65.5%) growing at the highest concentration. Piperacillin, ampicillin/sulbactam, and trimethoprim/sulfamethoxazole were also frequently resisted (50.0%, 35.7%, and 34.5% respectively). Table 3 lists the 18 universally tested antibiotics along with susceptibility numbers. PCR Detection of Class 1 and Class 2 Integrases A total of 91 samples were tested for both class 1 and class 2 integrase genes via PCR amplification. Of these samples, 84 were determined to be unique. A total of 19 (22.6%) isolates tested positive for a class 1 integron based on the amplification of 484bp DNA fragments that matched the amplicon generated from the class 1 positive control (Figure 4). As shown in Table 2, 15 samples were identified as E. coli, one as K.

25 Species # Samples Tested on Panels # Samples Tested for inti # Unique Samples # Unique Samples Tested for inti Number Unique Samples inti + (%) 1 Escherichia coli 68 67 65 64 15 (78.9) Klebsiella pneumoniae 5 5 4 4 1 (1.1) Enterobacter cloacae 5 5 4 4 0 (0) Kluyvera ascorbata 4 4 4 4 1 (1.1) Escherichia fergusonii 2 2 1 1 0 (0) Klebsiella oxytoca 2 2 2 2 1 (1.1) Raoultella ornithinolytica 2 2 2 2 1 (1.1) Salmonella sp. 1 1 1 1 0 (0) Cedecea davisae 1 1 1 1 0 (0) Citrobacter freundii 1 1 1 1 0 (0) Enterobacter aerogenes 1 1 1 1 0 (0) Total 92 91 85 84 19 (22.6) Table 2. Species Identifications. 1 Percentage of positive unique isolates tested.

Antibiotic Amikacin Gentamicin Tobramycin Trimethoprim/ Sulfamethoxazole Ceftazidime Cefazolin Cefuroxime Ceftriaxone Cefepime Cefotetan Piperacillin/Tazobactam Ampicillin Ampicillin/Sulbactam Piperacillin Aztreonam Ciprofloxacin Levofloxacin Imipenem Total (Mean) #S (%) 19 (100) 16 (84.2) 16 (84.2) 5 (26.3) 18 (94.7) 14 (73.7) 14 (73.7) 18 (94.7) 18 (94.7) 19 (100) 17 (89.5) 2 (10.5) 6 (31.6) 4 (21.1) 17 (89.5) 15 (78.9) 16 (84.2) 19 (100) 253 (13.3) inti-positive (n=19) #I (%) 0 (0) 1 (5.3) 1 (5.3) 0 (0) 1 (5.3) 0 (0) 2 (10.5) 0 (0) 1 (5.3) 0 (0) 0 (0) 0 (0) 5 (26.3) 4 (21.1) 0 (0) 0 (0) 0 (0) 0 (0) 15 (0.79) #R (%) 0 (0) 2 (10.5) 2 (10.5) 14 (73.7) 0 (0) 5 (26.3) 3 (15.8) 1 (5.3) 0 (0) 0 (0) 2 (10.5) 17 (89.5) 8 (42.1) 11 (57.9) 2 (10.5) 4 (21.1) 3 (15.8) 0 (0) 74 (3.9) #S (%) 62 (95.4) 53 (81.5) 54 (83.1) 50 (76.9) 59 (90.8) 35 (53.8) 47 (72.3) 58 (89.2) 62 (95.4) 62 (95.4) 49 (75.4) 23 (35.4) 35 (53.8) 30 (46.1) 48 (73.8) 52 (80.0) 55 (84.6) 64 (98.5) 898 (13.8) inti-negative (n=65) #R (%) 1 (1.5) 6 (9.2) 3 (4.6) 15 (23.1) 5 (7.7) 17 (26.2) 6 (9.2) 6 (9.2) 2 (3.1) 3 (4.6) 10 (15.4) 38 (58.5) 22 (33.8) 31 (47.7) 17 (26.2) 13 (20.0) 10 (15.4) 1 (1.5) 206 (3.2) P-value 1 Table 3. Susceptibility Comparisons Between Class 1 Positive and Class 1 Negative Isolates for Each Tested Antibiotic. S = Susceptible; I = Intermediate; R = Resistant according to Dade Behring Procedural Manual (34). 1 N/S = Not Significant (P>0.05). 2 Significantly higher levels of resistance in inti-positive isolates. 3 Significantly lower susceptibility in inti-positive isolates #I (%) 2 (3.1) 6 (9.2) 8 (12.3) 0 (0) 1 (1.5) 3 (4.6) 12 (18.5) 1 (1.5) 1 (1.5) 0 (0) 6 (9.2) 4 (6.2) 8 (12.3) 4 (6.2) 0 (0) 0 (0) 0 (0) 0 (0) 56 (0.86) N/S N/S N/S <0.01 2 N/S N/S N/S N/S N/S N/S N/S <0.04 2 N/S <0.05 3 N/S N/S N/S N/S N/S 26

27 1 2 3 4 5 6 7 500bp Figure 4. Class 1 Integron Detection. Post-EtBr stained agarose gel showing 484bp PCR-amplified fragments of the Class 1 integrase gene inti1. Lane 1: positive control, Salmonella DT104; Lane 2: 100bp DNA ladder; Lane 3: sample L-1-TET; Lane 4: L-4-TET; Lane 5: L-5-TET; Lane 6: M-16-TET; Lane 7: negative control.

28 pneumoniae, one as Kluyvera ascorbata, and one as Raoultella ornithinolytica. Only one sample tested positive for a class 2 integron, determined by the visualization of a 233bp band identical to that produced by the class 2 positive control (Figure 5). Interestingly the sample containing the class 2 integron also was one of the 19 samples that tested positive for a class 1 integron. Statistical Significance of Resistance and Integrons The statistical significance of the relationship between resistance and the presence or absence of integrons was examined in several ways. For all analyses, the Fisher Exact Probability Test or Chi-Square Test was used to calculate P-values, with preference for the Chi-Square Test where applicable. Calculations were made using an online program at vassarstats.net. Based on the nature of the input data, the program determined whether or not the Chi-Square Test could be performed. Significance was deemed to be a P-value of <0.05. For each of the 18 tested antibiotics, the significance of the relationship between susceptibility (resistant, intermediate, or susceptible) and presence or absence of the gene inti1 was calculated. Resistance was significantly more prevalent in inti1- positive isolates than in inti1-negative isolates for two antibiotics: trimethoprim/ sulfamethoxazole and ampicillin (Table 3). In addition, susceptibility to the antibiotic piperacillin was found to be significantly lower in inti1-positive isolates compared to inti1-negative isolates (Table 3). Since there is no clear consensus on the definition of multidrug-resistance in the literature (26), the significance of the relationship between multidrug-resistance and

29 1 2 3 4 5 6 7 8 300bp 200bp Figure 5. Class 2 Integron Detection. Post-EtBr stained agarose gel showing 233bp PCR-amplified fragments of the Class 2 integrase gene inti2. Lane 1: 100bp DNA ladder ; Lane 2: positive control, E. coli ATCC# 47055; Lane 3: sample F06-2-AMP; Lane 4: F06-11-AMP; Lane 5: F06-12-AMP; Lane 6: F06-28-AMP; Lane 7: F06-30-AMP; Lane 8: negative control

30 presence or absence of an integron was calculated in three different ways. Table 4 and Figure 6 show that a significantly greater number of inti1-positive isolates were resistant to at least one antibiotic in two or more antibiotic classes compared to inti1-negative isolates. Comparisons between the actual number of resistant integrase-positive and integrase-negative isolates for each number of antibiotic classes yielded no significance (P>0.05, See Figure 7). Inclusion of intermediate susceptibilities in the definition of resistant (26), also yielded no significance between multidrug-resistance and integron presence (Table 5). Finally, Table 6 and Figure 8 show the significance between integron presence and the number of resistances to individual antibiotics. It was found that resistance to two or more antibiotics, regardless of class was statistically greater in inti1- positive isolates compared to inti1-negative isolates. Figure 9 shows the actual number of resistant integrase-positive and integrase-negative isolates for each number of antibiotics, regardless of class; no significance was found in this analysis. PCR Amplification of Variable Regions All 19 samples that produced positive results for the presence of the class 1 integrase gene, inti1, were further investigated by amplifying the variable region of the integron. Analysis of EtBr-stained agarose gels yielded 6 amplicons of distinctly different sizes: ~250bp (n=3), ~900bp (n=3), ~1100bp (n=2), ~1800bp (n=5), ~2000 (n=3), and >3000bp (n=1). Additionally, two samples failed to produce any noticeable product, most likely due to loss or mutation of the 3 -conserved region. The gel pictured in Figure 10 shows the various sizes (900->3kbp) of class 1 variable region amplicons.

31 Integrase Positive (n=19) Integrase Negative (n=65) # Antibiotic # samples # samples not # samples # samples not Classes Resistant (%) Resistant (%) Resistant (%) Resistant (%) P-value 1 1 18 (94.7) 1 (5.3) 52 (80.0) 13 (20.0) N/S 2 16 (84.2) 3 (15.8) 36 (55.4) 29 (44.6) <0.03 3 7 (36.8) 12 (63.2) 17 (26.2) 48 (73.8) N/S 4 4 (21.1) 15 (78.9) 9 (13.8) 56 (86.2) N/S 5 2 (10.5) 17 (89.5) 2 (3.1) 63 (96.9) N/S Table 4. Comparison Between inti1-positive and inti1-negative Isolates Resistant to Multiple Classes of Antibiotics. 1 N/S = Not Significant (P >0.05)

Figure 6. Percentage of Cumulative Resistant Integrase-positive vs. Integrase-negative Isolates for Varying Numbers of Antibiotic Classes. Resistance to two or more antibiotic classes was found to be significantly higher in inti1-positive isolates compared to inti1-negative isolates (P<0.05). 32

Figure 7. Number of Resistant Integrase-positive vs. Integrase-negative Isolates for Varying Numbers of Antibiotic Classes. No significance found for any number of antibiotic classes (P>0.05). 33

34 Integrase Positive (n=19) Integrase Negative (n=65) # Antibiotic # samples Res. # samples not # samples Res. # samples not Classes or Int. (%) Res. or Int. (%) or Int. (%) Res. or Int. (%) P-value 1 1 18 (94.7) 1 (5.3) 56 (86.2) 9 (13.8) N/S 2 16 (84.2) 3 (15.8) 43 (66.2) 22 (33.8) N/S 3 8 (42.1) 11 (57.9) 24 (36.9) 41 (63.1) N/S 4 4 (21.1) 15 (78.9) 14 (21.5) 51 (78.5) N/S Table 5. Comparison Between inti1-positive and inti1-negative Isolates With Intermediate or Resistant Phenotypes for Multiple Classes of Antibiotics. 1 N/S = Not Significant (P>0.05). Integrase Positive (n=19) Integrase Negative (n=65) # of # samples # samples not # samples # samples not Antibiotics Resistant Resistant Resistant Resistant P-value 1 1 18 1 52 13 N/S 2 18 1 44 21 <0.04 3 14 5 33 32 N/S 4 9 10 23 42 N/S 5 6 13 17 48 N/S 6 4 15 14 51 N/S 7 3 16 10 55 N/S 8 2 17 6 59 N/S 9 1 18 2 63 N/S 10 0 19 2 63 N/S 11 0 19 2 63 N/S 12 0 19 1 64 N/S 13 0 19 0 65 N/S Table 6. Comparison Between inti1-positive and inti1-negative Isolates Resistant to Multiple Antibiotics, Regardless of Class. 1 N/S = Not Significant (P>0.05).

Figure 8. Percentage of Cumulative Resistant Integrase-positive vs. Integrase-negative Isolates for Varying Numbers of Antibiotics, Regardless of Class. Resistance to two or more antibiotics, regardless of class, was found to be significantly higher in inti1-positive isolates compared to inti1-negative isolates (P<0.05). 35

Figure 9. Number of Resistant Integrase-positive vs. Integrase-negative Isolates for Varying Numbers of Antibiotics, Regardless of Class. No significance found for any number of antibiotics (P>0.05). 36

37 1 2 3 4 5 6 7 8 3000bp 1500bp 1000bp Figure 10. Class 1 Integron Variable Regions. Post-EtBr stained agarose gel showing multiple variable region sizes. Lane 1: 100bp DNA ladder; Lane 2: positive control, Salmonella enterica DT104 showing double integron; Lane 3: sample 0901; Lane 4: 0919; Lane 5: L-1-TET; Lane 6: F06-12-AMP; Lane 7: 0812; Lane 8: negative control.

38 The variable region of the only detected class 2 integron (sample F06-2-AMP) was also amplified by PCR. Analysis of the PCR product via gel electrophoresis and subsequent staining with EtBr showed a single DNA band of approximately 2.5kb. Figure 11 shows the stained gel along with the class 2 positive control, which yielded an amplicon of the same size. Restriction Digest of Variable Region Amplicons Class 1 variable region PCR products that appeared similar in size were subjected to restriction enzyme AluI to aid in identification of unique sequences. Analysis of restriction fragments on EtBr-stained gels yielded identical patterns for all samples within respective size groups (250bp, 900bp, 1.1kbp, 1.8kbp, and 2kbp). The 250bp samples produced no observable changes after exposure to AluI. This result is consistent with the assumption that these PCR products were simply the amplified conserved regions without any cassettes, and the fact that the conserved regions between the primers do not contain a target site for AluI. Each of the 900bp products yielded two distinct DNA fragments, while the 1.1kbp products yielded four distinct DNA fragments. The 1.8kbp amplicons yielded at least five distinct bands of DNA, though some smaller fragments of similar size may have been present and indistinguishable from each other. Likewise, the 2kbp amplicons yielded at least seven bands, not all of which are completely distinguishable due to multiple small fragments of similar size (See Table 7 for restriction fragment lengths). Though the 1.8kbp and 2kbp amplicons produced restriction fragments that are not all distinguishable from each other, it is clear that the patterns are identical within

39 1 2 3 4 5 6 7 8 3000bp 1500bp Figure 11. Class 2 Integron Variable Regions. Post-EtBr stained agarose gel showing ~2500bp PCR-amplified variable regions. Lane 1: 100bp DNA ladder; Lane 6: positive control, E. coli ATCC# 47055; Lane 7: sample F06-2-AMP; Lane 8: negative control.

40 Samples Amplicon Length Observed (~bp) Amplicon Length Actual (bp) Restriction Fragment Lengths- Observed (~bp) Restriction Fragment Lengths- Actual (bp) 0904, 0911B 0 0 N/A N/A M-5-Ea, F06-39-AMP, M- 16-TET 250 232 250 232 L-1-TET, 0807, 0904 900 845 400, 500 367, 478 0812, 0916B 1100 1085 120, 280, 350, 400 114, 257, 342, 372 0914, 0919, 52, 57, 64, 60, 225, 0922B, 0922C, 1800 1742 50, 220, 250, 350, 261, 366, 657 0805 650 0901, F06-2- AMP, 0813 2000 1979 1 50, 100, 150 2, 200, 250, 350, 400 9, 10, 25, 26, 76, 99, 113, 114, 119, 145, 147, 166, 248, 330, 362 1 F06-12-AMP >3000 Unknown NT Unknown F06-2-AMP 3 2500 2224 4 NT Unknown Table 7. Observed vs. Actual Amplicon and Restriction Fragment Sizes. Observed amplicon and restriction fragment sizes are based on visualization of bands on Et-Br stained agarose gels compared to a 1kb DNA ladder. Actual amplicon and fragment lengths were calculated based on database sequences that aligned with sample sequences. N/A = not applicable; NT = restriction fragment analysis was not tested-only one sample yielded an amplicon of this size; Unknown = cannot be determined due to incomplete cassette identification; 1 Fragment lengths and actual amplicon size based on possible unverified centrally located orf ; 2 Multiple indistinguishable bands between 100 and 200bp; 3 Only class 2 positive sample; 4 Amplicon length based on probable unverified cassette.

41 both groups and therefore likely the of the same sequence (See Figure 12 for restriction digest gels). Sequencing and Cassette Identification A single representative sample from each variable region size group was sequenced from both the 5 and 3 conserved ends using the same primers that were used in the variable region PCR amplifications. While sequencing results varied and did not reflect full coverage of the entire template, most sequencing results did produce enough data to identify most, if not all inserted cassettes. Figure 13 shows the partial alignment of sample 0812 with a class 1 integron from a Vibrio cholerae strain (GenBank ID: GQ214171.1) containing cassette aada1. Sequencing reactions yielded sufficient data to produce an alignment with part of the gene cassette at both the 5 and 3 ends, with sequence identities of 96% to 100%. Cassette identifications were further corroborated through comparison of the restriction digest results of the amplicons compared to the expected restriction fragment sizes. Each occurrence of the AluI target site, AGCT, was identified in the GenBank sequences to which the sample sequences aligned, and the expected fragment sizes were calculated (Figure 13 shows the expected AluI target sites for the 1.1kbp amplicons). Restriction fragment sizes matched very closely for the 250bp, 900bp, 1.1kbp, and 1.8kbp variable regions, as shown in Table 7. Similarly, the expected size of the complete amplicons were calculated and compared to estimated observed lengths. Table 7 shows that total observed and actual amplicon lengths are also closely matched. Together, the partial cassette alignments, identical restriction fragment patterns,

42 1 2 3 4 5 6 7 8 1 2 3 4 5 A. B. 400bp 500bp 100bp C. 1 2 3 4 5 6 700bp 400bp 100bp Figure 12. Class 1 Variable Region Restriction Fragments. Post EtBr-stained agarose gels showing DNA fragments of class 1 variable regions after exposure to AluI. A.) Lane 1: 100bp DNA ladder; Lane 2: positive control, multiple fragments of double integron; Lanes 3-4: fragments of 1.1kbp amplicons; Lanes 5-7: fragments of 2kbp amplicons; Lane 8: negative control. B.) Lane 1: 100bp DNA ladder; Lane 2: positive control; Lanes 3-5: fragments of 900bp amplicons. C.) Lanes 1, 3-6: fragments of 1.8kbp amplicons; Lane 2: 100bp DNA ladder.