Evolution, expression and effectiveness in a cluster of novel bovine β-defensins
The anti-microbial peptides β-defensins constitute a large family of innate immune effector molecules, conserved across a wide species range. In this paper, we describe a systematic search of the sequenced bovine genome to characterise this extensive gene family in Bos taurus, providing an insight into the pattern of conservation of β-defensin genes between species. We have sequenced a sub-set of these newly discovered bovine β-defensin genes and also report expression data for these genes across a range of tissues. We have synthesised the peptide product of one of these genes, bovine β-defensin 123, and found it to be a potent inhibitor of several pathogenic microbes, particularly Escherichia coli and Listeria monocytogenes.
KeywordsAnti-microbial peptide Defensin Bovine Bioinformatic
Defensin genes code for an extensive family of small cationic anti-microbial peptides (AMPs) that constitute an important effector arm of the innate immune system across a wide species range (Ganz 1999) as well as representing a putative link between the innate and adaptive immune responses in higher organisms (Yang et al. 1999). The principle subdivision of the family into α-defensins and β-defensins is characterised by distinctive spacing of cysteine residues in their active peptide regions as well as their tissue expression patterns (Lehrer and Ganz 2002). A third defensin sub-type, θ-defensins, appear to be a recently acquired, primate-specific class of peptides, which are generated by the merging of two α-defensin-like precursors (Tang et al. 1999).
α-Defensins have only been described in mammals, but they are widely distributed in that order (Lynn and Bradley 2007). β-Defensins are the more comprehensively studied sub-class and possess the widest taxonomic distribution, being found in vertebrates and invertebrates as well as plants (Boman 2003), thus indicating an ancient point of origin. Multiple gene duplication events and subsequent sequence diversification in the mammalian lineage has resulted in a large family of proteins with diverse amino acid sequence but virtually identical tertiary structure based on the characteristic disulphide bridging between cysteine residues (Bauer et al. 2001). Recent sequencing of high-quality drafts of many vertebrate genomes has allowed for a comparative genomic approach to be used in characterising the β-defensin repertoire in various species (Schutte et al. 2002; Patil et al. 2005; Radhakrishnan et al. 2005). In species with recent common ancestors, comparable numbers and type of β-defensin have been identified. However, even in closely related species or clades, gain or loss of individual β-defensin genes has been demonstrated—a possible consequence of differing pathogenic insults each individual species faces during the course of its evolution in its own ecological niche (Radhakrishnan et al. 2005).
The bovine lineage was one of the first in which β-defensin-like molecules were discovered (Diamond et al. 1991; Selsted et al. 1993; Tarver et al. 1998). These previous studies resulted in 18 complete and partial bovine β-defensin (BBD) sequences being identified. This number was noticeably smaller than the 35 human, 33 chimp, 38 dog and 45 mouse purportedly functional β-defensin genes subsequently identified by both in silico and laboratory methods (Patil et al. 2005).
In this study, we describe two complementary search strategies for the identification and characterisation of the entire β-defensin repertoire encoded in the bovine genome: the homology-based search methods of the BLAST family of programs (Altschul et al. 1997) and the more sensitive Hidden Markov Models (HMM; Eddy 1998). We have examined the expression of several of these new genes across a wide panel of bovine tissues taken from healthy animals as well as examining the in vitro anti-microbial activity of one new BBD peptide against four common mammalian pathogens, Escherichia coli, Listeria monocytogenes, Salmonella typhimurium and Staphylococcus aureus.
Materials and methods
Bioinformatic identification of bovine AMP orthologs
All publicly available protein sequences corresponding to the known human, mouse and dog β-defensin sequences (Patil et al. 2005) were retrieved from GenBank (http://www.ncbi.nlm.nih.gov/Genbank). The draft 3.1 version of the sequenced and assembled bovine genome was downloaded from Ensembl (http://www.ensembl.org/) and searched using the mammalian β-defensin peptide sequences with the TBLASTN programme (Altschul et al. 1997), which compares the peptide sequence with the deoxyribonucleic acid (DNA) translated in all six reading frames. The subsequent release of the 25,985 Ensembl-predicted bovine gene set was also searched for sequences displaying a high degree of similarity to mammalian β-defensins. To carry out HMM (Eddy 1998) searches of the bovine genome for β-defensin motifs, the entire genome was translated in all six reading frames using a purpose-written Perl script. To generate accurate HMM models representing the β-defensin family, the protein sequences of all β-defensin peptides identified in previous searches of completely sequenced genomes (Patil et al. 2005) were aligned using T-Coffee (Notredame et al. 2000). Homologous sequences including the signature six-cysteine-conserved motif were extracted and used in the construction of the HMM by the hmmbuild program in HMMER 2.1.1 (http://hmer.wustl.edu; Eddy 1998). The generated HMM profile was then searched against the translated genome to identify putative β-defensin-like regions.
Chromosomal location and strand orientation of the identified β-defensins were determined using the BLAST-like Alignment Tool (BLAT) at the University of California—Santa Cruz genome browser (http://genome.ucsc.edu; Kent 2002). Genomic DNA corresponding to putative defensins was retrieved using BLAT and used for prediction of intron/exon boundaries using GenScan (http://genes.mit.edu/GENSCAN.html; Burge and Karlin 1997). The complete repertoire of BBD genes was further analysed by alignment with homologous sequences from other vertebrate species using the T-Coffee multiple sequence alignment programme (Notredame et al. 2000), while neighbour-joining phylogenetic analysis of the proteins was carried out using Mega v.3.1 (Kumar et al. 2001).
Tissue collection, RNA extraction and cDNA synthesis
Tissues were collected at a local abattoir from individual recently euthanised cattle and were immediately flash frozen in liquid nitrogen at the source. Peripheral blood mononuclear cells were extracted from whole blood of uninfected cattle using a Percoll™ gradient (GE Healthcare UK, Buckinghamshire, UK) and previously described methods (Ulmer et al. 1984). Total ribonucleic acid (RNA) was extracted from all tissues using a PRO200 Homogenizer (PRO Scientific) to disrupt cells in the RLT buffer supplied with the RNeasy® mini kit (Qiagen, Crawley, UK) according to the manufacturer’s instructions. Mammary cells were extracted from milk using a protocol as outlined in Sarikaya et al. (2005) and using the RNeasy® lipid tissue minikit. All samples were DNA digested to remove genomic DNA using Qiagen’s on-column DNase and eluted with water. A tissue pool representing RNA pooled from a range of tissues including the liver, mammary gland, lymph node, pituitary gland, hypothalamus, brain, heart, spleen, foetal tissue, ovary and total con A-stimulated leucocytes, was used as a positive control for expression.
RNA was quantified using a NanoDrop® ND-1000 spectrophotometer (Nanodrop Technologies, Wilmington, DE), and purity was assessed on 2% native agarose gels. Single-stranded complementary DNA (cDNA) was synthesised from 1 μg total RNA using oligo dT primers (Promega, Madison, WI) and Omniscript reverse transcriptase (Qiagen). The converted cDNA was then quantified using a NanoDrop® ND-1000 spectrophotometer and diluted to 40 ng/μl working stocks and stored at −20°C for subsequent analyses.
RT-PCR and DNA sequencing
Known and novel AMP gene details
Predicted peptide length
Primer sequences (5′–3′)
Genbank accession number
For gene cDNA sequencing, separate primer sets (SP—Table 1) were designed, and conventional PCR used to amplify the AMP-coding sequences from tissues in which high expression was observed. Each primer pair was optimised using the following protocol. cDNA was amplified in a final volume of 11 μl containing 4.4 pmol of each primer, 200 μM of each deoxyribonucleotide phosphate, 0.5 U of Platinum® Taq polymerase (Invitrogen Life Technologies: invitrogen.com) including 1× reaction buffer (10× buffer consists of 200 mM Tris–HCl pH 8.4, 500 mM KCl). Optimal MgCl2 concentrations were determined by titrating MgCl2 concentration in the 10× PCR mix between 10 and 25 mM in increments of 2.5 mM. PCR amplifications were performed using a PTC-200 thermocycler with a gradient block features (MJ Research: www.mjr.com) with an initial denaturation step for 3 min at 94°C and 40 cycles of 30 s at 93°C, with annealing temperature (Ta) varied across the gradient block from 50 to 65°C for 30 s followed by 72°C for 40 s extension step. The final steps were a 72°C soak for 10 min and a 10°C soak until samples were retrieved. Each sample was then run in a 2% agarose gel using a horizontal gel electrophoresis apparatus (Life Technologies/Whatman-Biometra®: www.biometra.com) in 1× Tris–acetate–ethylenediamine tetraacetic acid buffer with ethidium bromide at 100 V for 30 min. Gels were visualised using an AutoChemi System (Ultra Violet Products). After PCR cleanup using a Qiaquick PCR cleanup kit (Qiagen), samples were sent for commercial sequencing (www.macrogen.com). These sequences were then submitted to Genbank (accession numbers given in Table 1).
Peptide synthesis and functional analysis in vitro
A test panel of common pathogenic bacterial strains was obtained from the National Collection of Type Cultures (NCTC), Central Public Health Laboratory, Colindale, UK: S. aureus NCTC 6571, L. monocytogenes NCTC 11994, E. coli NCTC 9001 and S. typhimurium NCTC 12023. The day before the start of the assay procedure, a starter culture of the required strains were grown up, and on the day of the assay, 2 × 25-ml vials of tryptone soy broth (TSB; Cruinn Diagnostics, Dublin, Ireland) were inoculated with the required number of colonies for 3 h at 37°C and harvested by centrifugation at 1,700 × g for 10 min. The bacterial pellet was resuspended in 20 ml of 10 mM potassium phosphate buffer (pH 7.0) and was serially diluted to the required concentration (~106 CFU/ml) for the assay.
In a 96-well polypropylene plate, 100 μl of potassium phosphate buffer was aliquoted into columns 2, 4, 6, 8, 10 and 12 in rows A and B (duplicates), with negative control wells receiving 100 μl of potassium phosphate buffer with 0.1% TSB. Stock peptide was aliquoted into wells 4A and 4B and serially diluted into columns 6, 8, 10 and 12 using an eight-channel multipipette. Bacterial dilutions were vortexed, poured into sterile Petri dishes and, using an eight-channel pipette, quickly aliquoted to the 96-well plate before the bacteria settled. The plates were incubated for 90 min at 37°C. After incubation, 180 μl of potassium phosphate buffer was aliquoted into rows C to H of columns 2, 4, 6, 8, 10 and 12. The samples were serially diluted (10−1 to 10−3) from rows A/B to G/H. Every other row was left empty to prevent cross-contamination. Each dilution was plated four times using a drop-plate technique on trypticase soy agar (Cruinn Diagnostics). Using an eight-channel pipette, with sterile tips at every other channel, 20 μl of each sample was mixed, removed and spread across the width of the plate. This was repeated three times for each sample, changing the tips for each new plate, resulting in four 20-μl counts for each dilution of each sample. The plates were inverted and incubated overnight at 37°C. The following day, the colonies on each plate were counted. All counts for an individual plate were converted to the same dilution factor and averaged over the four plates. Counts of greater than 100 colonies in any dilution were not included in the calculations. The assay was performed in duplicate from different original bacterial inoculations for each bacterium. The lethal dose (LD50 and LD99) is defined as the minimum concentration of AMP required to kill 50 and 99%, respectively, of viable bacterial cells. Values are expressed in Fig. 4 with error bars indicating standard error of the mean.
Identification and verification of bovine AMP orthologs
Before this study, 18 putative BBDs had been identified through a combination of genomic sequence analysis (Roosen et al. 2004) and direct sequencing of isolated purified proteins from blood neutrophils (Selsted et al. 1993). In the current work, 57 open reading frames (ORFs) bearing a similarity to the characteristic six-cysteine-containing domain of the β-defensin family of genes were identified in a comprehensive bioinformatic search of the sequenced bovine genome (see Supplementary materials). To identify first exons corresponding to each of these second exons sequences, the 5′ upstream genomic region was analysed using the gene prediction software GenScan, as well as homology-based searches using full-length orthologous β-defensin sequences from other species. These methods established possible corresponding first exons for 53 of the 57 β-defensins predicted in the bovine genome. Of the four remaining ORFs, gaps in the sequence of the upstream region or the short nature of the assembled contig are most likely reasons for lack of an associated first exon.
Most of the previously identified BBD genes were found to form a separate cluster from those identified in humans. These data coupled with the higher degree of sequence conservation observed between members of this clade would seem to indicate a relatively recent origin for this β-defensin sub-group. To date, this cluster had only been extensively identified in the bovine lineage, although members of this clade have been reported in other artiodactyls including sheep (Huttner et al. 1998), reindeer (UniProt Q0MR48), water buffalo (UniProt A3RJ36) and goat (UniProt Q0PGY0). Of the 18 previously published bovine defensin sequences, the corresponding genomic loci were identified for only 13 in the published bovine genome. Although subsequent, more complete drafts may resolve this issue, the almost identical amino acid composition of a number of these defensins (BNBD7 and DEFB401 differ by a single amino acid) suggests that these previously identified proteins could represent allelic variants of the same gene. We have also identified a new cysteine-containing domain, which has 92% identity with BT402. This sequence (named BNBD14 here) may represent a previously unknown member of this β-defensin sub-group, which seems to be present in a limited mammalian species range. However, no signal peptide coding first exon sequence was apparent in the current genome draft.
Gene sequences and qualitative analysis of AMP mRNA expression
Previous studies have indicated that loci orthologous to these bovine genes in other species display disparate expression patterns. Primate expression appears to be restricted exclusively to the male reproductive tract (Radhakrishnan et al. 2005), while rodent orthologs of a number of these novel BBDs have been shown to be expressed in a broader range of tissues than in primates (Radhakrishnan et al. 2007). As in these previous studies, the bovine genes appear to be predominantly expressed in the male reproductive tract, although only the two β-defensin 122-like genes in cattle (BBD122 and BBD122A) show expression restricted to the testis. This pair would appear to be a result of a recent gene duplication event occurring later than the last common ancestor of the bovine and canine lineages. These two peptides are 75% identical in their amino acid composition with all the variability limited to the second exon-encoded active peptide. The duplication and subsequent divergence of the bovine 122 locus, coupled with the pseudogenisation of this gene in humans following the human–macaque split, lends further support to view that increased gene diversity in the β-defensin family can be best described by the birth and death model of evolution (Radhakrishnan et al. 2007). Both BBD119 and BBD120 are expressed in the rumen and testis with BBD119 messenger RNA (mRNA) also detected in the uterus. The neighbouring genes BBD123 and BBD124 display the greatest breadth of tissue expression, although they do not share an identical pattern. Overall, the expression range appears to be most similar for genes located close together on the chromosome providing further support for the hypothesis that tandem duplication of single defensin genes rather than duplication of a large chromosomal chunk is responsible for the “birth” of these genes within the cluster (Radhakrishnan et al. 2005).
Functional testing of novel BBD123 peptide against a pathogen panel
The pathogen panel included E. coli, S. typhimurium (Gram-negative), S. aureus and L. monocytogenes (Gram-positive). In addition to BBD123, the well-characterised HBD3 was tested as a standard reference peptide in each of the bactericidal assays performed. HBD3 has previously been shown to be one of the most potent anti-bacterial β-defensins yet discovered with regard to its anti-bacterial activity (Batoni et al. 2006).
Comparison of efficacy of BBD123 and HBD3 against Escherichia coli, Salmonella typhimurium, Staphylococcus aureus and Listeria monocytogenes
The discovery of novel anti-microbials is a welcome development in light of increasing prevalence of antibiotic resistance. AMPs are among the most ancient components of the immune system (Selsted and Ouellette 2005), but their extensive role in mammalian defence has only recently become apparent. We have used the recently released bovine genome sequence to identify novel β-defensin family members in cattle, increasing the count to 57 putative β-defensin genes encoded for in Bos taurus. This is the largest expansion of this gene family identified in any mammalian species. Our phylogenetic tree (Fig. 1) of the novel bovine defensins with the human repertoire indicates the discontinuous nature of the β-defensin catalogue in mammals, where individual defensin genes can be either gained or lost in a particular species, possibly because of selective pressures exerted on that lineage. In the process of domestication, cattle were subjected to much higher population densities than wild bovids. This would have exposed them to a greater abundance and diversity of microbes. This may explain why the bovine genome appears to host the highest number of β-defensins of any mammal. This is also a plausible explanation for the positive selection signatures reported in the bovine specific β-defensins (Luenser and Ludwig 2005). Preliminary analysis (data not shown) indicates that positive selection has influenced the evolution of bovine genes in the syntenic cluster under discussion here.
The PCR-amplified mRNA products were sequenced to verify the accuracy of the bioinformatic predictions. A tissue expression profile for several novel β-defensins have been generated from tissues derived from a healthy animal, indicating that these genes are constitutively expressed under normal conditions. Genes for which no expression data have been determined may yet be found in tissues not examined in our panel, or expression may only be induced after infection by specific pathogens. It seems plausible that the expression of particular β-defensin genes in a tissue would indicate a site of infection for whatever specific pathogens these genes counteract. Therefore, the conservation of these defensins across a wide mammalian species range, coupled with the differing tissue expression ranges in rodents, primates and cattle could indicate that the actual sites of infection for a microbe can vary greatly in different host species. Alternatively, small changes in amino acid composition in the active peptide may change the pathogen specificity of particular orthologous peptides in different species.
It is generally regarded that mammalian cationic AMPs are broad spectrum with anti-microbial activity against a range of human and animal pathogens (Brogden et al. 2003). mRNA expression of two of the known AMPs—lingual anti-microbial peptide (LAP) and tracheal anti-microbial peptide (TAP)—has been recently detected in the mammary glands of healthy and infected cattle (Roosen et al. 2004). Initial studies primarily focused on respiratory tract expression and function (Russell et al. 1996). In our study, we also detect expression of both LAP and TAP in the lung (Fig. 3) but also demonstrate wider expression and possible function of these AMPs across tissues of the reproductive and immune systems. Furthermore, LAP has been demonstrated to be expressed and inducible in tissue epithelial cells in response to mastitis in cattle (Swanson et al. 2004). Our studies confirm the expression of both TAP and LAP in mammary epithelium as well as mammary tissue (Fig. 3). The broad spectrum of anti-microbial properties found for LAP (Schonwetter et al. 1995) correlates with the non-specific pattern of tissue expression we detected (Fig. 3).
Similarly, BNBD4 and BNBD5, originally derived from neutrophils (Selsted et al. 1993), are also expressed across the tissue panel selected, suggesting wider constitutive functions. Induced expression of BNBD5 has been demonstrated in mammary tissue in response to bovine mastitis (Goldammer et al. 2004). Our results show the expression of both BNBD4 and BNBD5 in mammary gland and mammary epithelial cells.
The pathogen panel tested included both Gram-negative and Gram-positive bacteria that are common cattle pathogens representing a significant cost to agriculture as well as posing problems for food safety and human health. E. coli and S. aureus are common causes of mastitis in cattle, while S. typhimurium and L. monocytogenes can cause gastrointestinal problems such as listeriosis and salmonellosis in both animals and man. Importantly, many of these pathogens, especially S. aureus, are developing increased resistance to conventional antibiotics. Certain strains of E. coli and S. typhimurium can exist as commensals in cow (Bettelheim et al. 1974), although it is possible that AMPs play a role in controlling such benign populations as well as the more pathogenic strains (Zasloff 2002). The low LD50 and LD99 of the BBD123 peptide, necessary for killing all four pathogens tested, indicate the potential application of AMPs as novel anti-microbials against both Gram-positive and Gram-negative bacteria. A previous study investigating HBD123 reported a minimal inhibitory concentration (MIC) of 75 and 37.5 μg/ml against S. aureus and E. coli, respectively (Motzkus et al. 2006). Although the reported HBD123 activity cannot be directly compared to our own findings because of the use of different anti-microbial assays, it appears that the anti-microbial activity of these orthologuos peptides is different. The proteins differ at 19 out of 48 amino acid residues in the active region, with these changes possibly contributing to the differences in killing efficiency. Similarly, the use of BBD123 improves on the efficacy previously published using other bovine AMPs against these pathogens. TAP, for example, has been shown to kill E. coli with a MIC of 12–25 μg/ml and S. aureus with a MIC of 25–50 μg/ml (Diamond et al. 1991), as against 0.9 and 15.6 μg/ml for BBD123 against E. coli and S. aureus, respectively (Table 2). It is noted that BBD123 and HBD3 display different potencies against both Gram-positive and Gram-negative bacteria. As a consequence, we speculate that neither the Gram-positive nor Gram-negative status of a bacterial membrane is a major contributing factor in susceptibility of a microbe to β-defensins.
The constitutive expression of these novel genes suggests that circulating levels of AMPs are necessary for optimal immune protection. It is clear that BBDs, as in other species, are very diverse with respect to amino acid sequence, tissue expression pattern and target pathogens. It is possible that some of the variation in these characteristics within the β-defensin family may be related to additional functions distinct from the direct anti-bacterial action (Elsbach 2003). It is probable given the diverse roles that are currently being catalogued for known AMPs that the role of these novel AMPs in the bovine immune response under normal and inflammatory conditions will prove to be pleiotropic.
This work was supported by the Department of Agriculture (Food Institutional Research Measure) funding. The bovine tissue pool RNA was kindly donated by Dr. David MacHugh, Animal Genomics Laboratory, School of Agriculture, Food Science and Veterinary Medicine, UCD.
- Bettelheim KA, Bushrod FM, Chandler ME, Cooke EM, O’Farrell S, Shooter RA (1974) Escherichia coli serotype distribution in man and animals. J Hyg (Lond) 73:467–471Google Scholar
- Tarver AP, Clark DP, Diamond G, Russell JP, Erdjument-Bromage H, Tempst P, Cohen KS, Jones DE, Sweeney RW, Wines M, Hwang S, Bevins CL (1998) Enteric beta-defensin: molecular cloning and characterization of a gene with inducible intestinal epithelial cell expression associated with Cryptosporidium parvum infection. Infect Immun 66:1045–1056PubMedGoogle Scholar