Do birds differentially distribute antimicrobial proteins within clutches of eggs?

Abstract

Parents can considerably affect the fitness of their offspring through the additive influences of inherited traits and parental effects (Rossiter 1996; Mousseau and Fox 1998). These latter, nongenetic factors may be adaptive phenotypic responses to environmental variation that are invested in the egg at some cost and can increase offspring growth and survival (Schwabl 1993, 1996; Schwabl et al. 1997; Mousseau and Fox 1998; Eising et al. 2001). Among their many functions, eggs must protect embryos from microbial infection because marine and terrestrial environments harbor microbes capable of infecting embryos. Eggs of a variety of animal taxa contain antimicrobial agents (Brogden 2005), but antimicrobial defenses in eggs of wild animals remain poorly understood despite the obvious significance of this function (Giacomello et al. 2006). Avian eggs contain a network of defenses against microbial infection. The shell provides physical protection that is complemented by the strong antimicrobial effects (through unknown mechanisms) of incubation (Cook et al. 2005a). Microbes multiply rapidly on the shells of unincubated eggs, such as those laid early in the sequence before the onset of incubation, and can penetrate to the embryo and cause hatching failure (Cook et al. 2003, 2005b). Further, chicken eggs contain at least 6 antimicrobial proteins in the albumen (Board and Fuller 1974, 1994). One of these antimicrobials, ovotransferrin, chelates Fe3+, Cu2+, and Zn2+, necessary for growth of most microbes, and may also have active antimicrobial properties (Valenti et al. 1983, 1985; Ibrahim 1996). A second antimicrobial, avidin, binds biotin, a vitamin necessary for growth of most microbes (Board and Fuller 1974), and is presumed to slow the growth of a broad spectrum of bacteria. A third antimicrobial, lysozyme, catalyzes the lysis of gram-positive bacteria (Rogers and Perkins 1968) by hydrolyzing the peptidoglycan in their cell walls. Activity of these antimicrobial proteins is higher at incubation temperatures (34–36 °C) than at ambient temperatures (Tranter and Board 1984) and increases with protein concentration (Salton 1957; Yamanishi et al. 2002). Birds could, therefore, increase protection of eggs through deposition of higher concentrations of antimicrobial proteins. In addition, indirect evidence suggests that effects of antimicrobials like lysozyme may be carried over from albumen into the nestling for a short period of time after hatching (Saino et al. 2002). Because earlier-laid eggs are at greater risk for infection because of the rapid growth of microbes on unincubated eggs (Cook et al. 2005a), earlier-laid eggs should have higher levels of these proteins than later-laid eggs. However, this should only be true if production or deposition of these proteins is costly in some way, which was suggested for lysozyme (Saino et al. 2002, 2007). A decline in the concentration of lysozyme with laying order in the eggs of European barn swallows (Hirundo rustica) (Saino et al. 2002) supports this prediction. Whether antimicrobials are allocated in a similar manner in the eggs of other wild birds is unknown. Our goal in this study was to test whether birds nonrandomly allocated antimicrobial proteins to eggs within clutches. We analyzed 444 eggs from 97 complete clutches of known laying sequence for 8 species from 6 families of birds (Table 1). These species inhabit environments ranging from temperate wetlands and grasslands to tropical forests and exhibit preincubation delays ranging from 0 (green-rumped parrotlet, Forpus passerinus) to 8 days (blue-winged teal, Anas discors). We predicted that antimicrobial concentration would decline with laying order in all species except the parrotlet. Our null hypothesis was that antimicrobials would be evenly or randomly distributed across all eggs in clutches of each species. Taxonomy and reproductive characteristics of the bird species used in this study whose clutches were sampled and tested for differential deposition by laying order of the antimicrobials ovotransferrin, avidin, and lysozyme All data (except location of sampling) are from Birds of North America account (http://bna.birds.cornell.edu/BNA) for each species, except the green-rumped parrotlet (Grenier and Beissinger 1999). Common names, typical breeding habitat, location of sampling for this study, nest type, family, mean clutch size, incubation onset, and exposure are listed. Incubation onset refers to the egg upon which incubation is initiated. Exposure is the number of days the first-laid egg sits in the nest prior to the onset of full incubation. Taxonomy and reproductive characteristics of the bird species used in this study whose clutches were sampled and tested for differential deposition by laying order of the antimicrobials ovotransferrin, avidin, and lysozyme All data (except location of sampling) are from Birds of North America account (http://bna.birds.cornell.edu/BNA) for each species, except the green-rumped parrotlet (Grenier and Beissinger 1999). Common names, typical breeding habitat, location of sampling for this study, nest type, family, mean clutch size, incubation onset, and exposure are listed. Incubation onset refers to the egg upon which incubation is initiated. Exposure is the number of days the first-laid egg sits in the nest prior to the onset of full incubation. Study species were chosen to represent a range of preincubation exposure periods (i.e., the number of days the first-laid egg sits in the nest prior to incubation) in birds (see Table 1) and for logistical reasons. Exposure periods ranged from 0 to 8 days, although most were between 1 and 3 days. Infection can occur within 1–3 days in unincubated eggs (Cook et al. 2003), so these exposure periods were sufficiently long to test our hypothesis. Logistically, we needed to find nests before egg laying began and to collect eggs daily. Thus, we chose species that nested conspicuously (primarily in nest boxes), were not sensitive to disturbance and were already under long-term study. We collected complete clutches of eggs from 8 species (see Table 1 and references therein for more details): pearly-eyed thrashers Margarops fuscatus in a tropical montane forest (El Yunque National Forest; 18°19′N, 65°47′W) in Puerto Rico; green-rumped parrotlets in tropical savannahs of Venezuela (8°34′N, 67°35′W); western bluebirds Sialia mexicana and tree and violet-green swallows Tachycineta bicolor and Tachycineta thalassina, respectively, in temperate grasslands in northern coastal California (Hopland Field Station; 38°99′N, 123°07′W); eastern bluebirds Sialia sialis in subtropical grasslands in central Alabama (32°35′N, 82°28′W); red-winged blackbirds Agelaius phoeniceus in temperate marshes in Kansas (Konza Prairie Biological Station; 39°05′N, 96°35′W); and blue-winged teal in temperate marshes in North Dakota (48°41′N, 99°15′W). Details of study sites and species are listed in Table 1. Eggs were collected in May and June 2006, except for parrotlets (July–November 2005). Nests were checked between 0900 and 1300 every 1–3 days until nest cups were fully formed, at which point we checked them daily. We removed eggs on the day they were laid and replaced them with painted clay dummy eggs to facilitate normal laying behavior. Collected eggs were then cracked open in a clean manner using ethanol and flame-sterilized forceps, yolk was separated from albumen using sterile pipettes, and albumen was stored at −20 °C. We collected 13 clutches (41 eggs; clutch size range 2–4, mean 3.19 ± 0.09 standard error) from pearly-eyed thrashers, 10 clutches (44 eggs; range 2–5, mean 4.43 ± 1.13) from tree swallows, 10 clutches (41 eggs; range 3–5, mean 4.17 ± 0.08) from violet-green swallows, 15 clutches (43 eggs; clutch size range 2–5, mean 3.23 ± 0.15) from red-winged blackbirds, 18 clutches (69 eggs; range 2–5, mean 4.09 ± 0.09) from eastern bluebirds, 10 clutches (44 eggs; range 3–5, mean 4.60 ± 0.10) from western bluebirds, 11 clutches (79 eggs; range 5–10, mean 7.61 ± 0.20) from green-rumped parrotlets, and 10 clutches from blue-winged teal (83 eggs; range 2–12, mean 9.72 ± 0.26). We measured concentrations of ovotransferrin, lysozyme, and avidin because they are the most abundant and well-characterized antimicrobials in egg albumen. We used the total iron-binding capacity assay of Yamanishi et al. (2002) to measure concentration of ovotransferrin in each egg. This assay measures the amount of iron necessary to saturate ovotransferrin in the sample and thus correlates well with ovotransferrin concentration (Yamanishi et al. 2002). We added 125 μl of a 1:500 dilution of an iron-standard solution (1000 mg/ml; Sigma, St Louis, MO) in buffer (pH 8.4) containing 300 mmol/l Tris (Fisher, Pittsburgh, PA), 150 mmol/l sodium hydrogen carbonate (EMD, Darmstadt, Germany), and 4.2 g/l Triton X-100 (Sigma) to 24 μl of each albumen sample in wells of a 96-well plate (Nunc MaxiSorp, Rochester, NY). After 5 min of incubation at 37 °C, we added 25 μl of a second reagent (pH 4.0) containing 10 mmol/l ferrozine (Baker, Phillipsburg, NJ) and 32.6 mmol/l L-ascorbic acid (Acros, Geel, Belgium) in 50 mmol/l Tris buffer to each well and (Yamanishi et al. 2002) incubated plates at 37 °C for 5 min before adding 100 μl of a third reagent containing 600 mmol/l citric acid (Baker) and 25.6 mmol/l thiourea (Baker). We then measured absorbance (color change) every 20 s at 570 and 660 nm for 6.2 min using a Versamax microplate reader (Molecular Devices, Sunnyvale CA). To calculate ovotransferrin concentration, we determined the difference in absorbance at 570/660 nm at the beginning and end of the 6.2-min period. We then determined absolute ovotransferrin concentration by comparing these values with those in a standard curve. To determine avidin concentration, we used a modified version of the colorimetric method of Gan and Marquardt (1999). We diluted each albumen sample 4-fold in carbonate–bicarbonate buffer (Sigma) and added serial dilutions (100 μl) to columns 1 through 11 of each row. The bottom row contained a serial dilution of avidin (Sigma, 5–0.002 μg/ml) in carbonate–bicarbonate buffer. We incubated these plates at 4 °C overnight and then rinsed them 3 times with phosphate-buffered saline (PBS)/0.05% Tween-20 (Sigma). Nonspecific binding was prevented by adding Superblock buffer (Pierce, Rockford, IL) and incubating for 30 s at room temperature 3 times. We then added 100 μl of a 1:4000 dilution of biotin/horseradish peroxidase (Sigma) in Superblock/0.05% Tween-20 to each well and incubated plates for 25 min at room temperature. Finally, we washed the plates 5 times with PBS/0.05% Tween-20 and added 100 μl of blue peroxidase (POD) substrate (Roche, Basel, Switzerland) to each well before incubating the plates at room temperature for 30 min. We measured absorbance at 450 nm with a Versamax microplate reader. Concentration of avidin in each sample was calculated by comparison of absorbance values to those in the standard curve. To measure lysozyme concentration, we used a version of the lysoplate method of Osserman and Lawlor (1966) modified for 96-well plates (Millet et al. 2007). We added 25 mg dried Micrococcus lysodeikticus (Sigma) to 50 ml 1% agar (Difco, Detroit, MI) and kept the suspension at a temperature of 50–60 °C. Then 150 μl of this suspension was added to each 10-μl albumen sample in a well of a 96-well plate. We obtained a standard curve by adding the bacterial suspension to serial dilutions of a standard lysozyme solution. Plates were incubated overnight at room temperature and absorbance was measured with a Versamax microplate reader at 850 nm. Concentration of lysozyme in each sample was calculated by comparison of absorbance values to those in the standard curve. All assays were run in duplicate by an observer blind to the identity of the sample being tested and were highly repeatable (sensu Lessels and Boag [1987] using values taken from the duplicate measurements; repeatability = 0.86 for ovotransferrin, 0.89 for avidin, and 0.79 for lysozyme). All statistical tests were performed in SPSS 13. Only avidin data required log transformation to obtain normality. Concentrations of the 3 antimicrobials were not strongly correlated with one another in any of the species (Pearson correlation coefficient; all P > 0.09), so we ran separate analyses for each antimicrobial. To test for covariation between laying order and antimicrobial concentration, we used a separate analysis of covariance (ANCOVA) (general linear model univariate function in SPSS) for each species with ovotransferrin (milligrams per milliliters), lysozyme (micrograms per milliliters), or avidin (log scale from data measured in micrograms per milliliters) concentration as the dependent variable, nest of origin as a random factor, and clutch size and egg-laying order as We clutch size in because we predicted that differential would be in as first-laid eggs in clutches are for periods of time than in clutches. The were run with the clutch size laying order In this was not = the were then run this Because of the number of a be to P which would the for significance to = However, of in such a manner may be in with sample such as our study. We with and to to their the significance of the of our were not with or All was carried with of the of the of California at under from the and California and North Dakota and Alabama of and Kansas of and and in with the we collected complete clutches of eggs, we on the by all dummy eggs after clutch In all the birds nested we not that the of these eggs have a on the of these species. on from an increase in of eggs to clutch is more than by the number of offspring to the number of nest and in the by laying order 2–4, 24 only 5 of differential antimicrobial eggs were before and only after with laying order in pearly-eyed thrashers = Table and the between laying order and clutch size for ovotransferrin in pearly-eyed thrashers was = that this was in clutches. However, were data from each clutch size were analyzed all P with the ovotransferrin with laying order in violet-green swallows at the = = Table but this not remain after to avidin levels with laying order in bluebirds, violet-green swallows, and tree swallows = and Table and the after All other for clutch size and laying and their were 2–4, and of and of concentration of the antimicrobial ovotransferrin in the egg albumen of 8 bird species by clutch size laying order and nest of origin The were run with the clutch size In this was not the were then run this and P values for and are from these the was = and P values for nest are from the = P values are listed in a was the of significance for all analyses would be = of and of and of concentration of the antimicrobial ovotransferrin in the egg albumen of 8 bird species by clutch size laying order and nest of origin The were run with the clutch size In this was not the were then run this and P values for and are from these the was = and P values for nest are from the = P values are listed in a was the of significance for all analyses would be = of and of and of concentration of the antimicrobial ovotransferrin in the egg albumen of 8 bird species by clutch size laying order and nest of origin The were run with the clutch size In this was not the were then run this and P values for and are from these the was = and P values for nest are from the P values are listed in a was the of significance for all analyses would be = P values at this are with an of and of and of concentration of the antimicrobial ovotransferrin in the egg albumen of 8 bird species by clutch size laying order and nest of origin The were run with the clutch size In this was not the were then run this and P values for and are from these the was = and P values for nest are from the P values are listed in a was the of significance for all analyses would be = P values at this are with an of of concentration of the antimicrobial ovotransferrin by laying order in eggs of 8 species of birds The within each the concentration, the and of each are the and and the and are the and are at the bottom of each are with in antimicrobial protein concentration was the most before of 24 nests in parrotlets, eastern bluebirds, and tree swallows (Table nests in tree swallows, violet-green swallows, and blue-winged teal (Table nests in tree swallows and violet-green swallows (Table However, only variation in avidin teal after (Table of not any of these and of and of concentration of the antimicrobial ovotransferrin in the egg albumen of 8 bird species by clutch size laying order and nest of origin The were run with the clutch size In this was not the were then run this and P values for and are from these the was = and P values for nest are from the P values are listed in a was the of significance for all analyses would be = of and of and of concentration of the antimicrobial ovotransferrin in the egg albumen of 8 bird species by clutch size laying order and nest of origin The were run with the clutch size In this was not the were then run this and P values for and are from these the was = and P values for nest are from the P values are listed in a was the of significance for all analyses would be = of an test for effects of laying we eggs within a clutch as and laid and ran an analysis of with nest as a random and laying as a effects This analysis the of our only avidin deposition in tree eggs and ovotransferrin concentration in violet-green eggs were = and but of these after between clutches was only for avidin in blue-winged teal eggs after as in our We the of antimicrobial in the eggs of a of wild bird species, but our for the hypothesis that birds higher concentrations of antimicrobials to than later-laid eggs. after the only were in the the concentration of avidin with laying order in eggs of species. These 1) by them with more antimicrobial or by higher the vitamin by avidin, is also in growth and is by the nestling swallows the albumen the day before hatching et al. levels of avidin to levels of to In any this not to be but than these concentration of antimicrobials to be not our our data that some bird species may in antimicrobial at the Only variation teal after but the of a of variation in antimicrobial these species suggests some in antimicrobial deposition clutches. in the concentrations of egg albumen be by in environmental or factors (Saino et al. some evidence suggests a between temperature and lysozyme deposition (Saino et al. et al. suggested that in lysozyme deposition can occur because lysozyme in eggs by barn swallows with the evidence is Our with the only study of of lysozyme in the albumen of wild birds (Saino et al. 2002) and with some of other factors in egg such as and et al. 2006, but as a et al. These and the risk of infection for unincubated eggs (Cook et al. 2003, to our prediction. However, the production cost of these antimicrobials in albumen to be and of their function may their differential of concentration of the antimicrobial avidin by laying order in eggs of 8 species of birds The within each the concentration, the and of each are the and and the and are the and are at the bottom of each are with of concentration of the antimicrobial lysozyme by laying order in eggs of 8 species of birds The within each the concentration, the and of each are the and and the and are the and are at the bottom of each are with and deposition of antimicrobials may be costly than deposition of yolk in the yolk can in et al. and et al. et al. and such as vitamin may be costly to to other yolk et al. 2006). However, the amount of needed to and antimicrobial proteins is greater than for other albumen from a between and egg lysozyme between embryo and as by et al. because egg and lysozyme are by than from a lysozyme is by egg lysozyme is by in the and et al. and Thus, is not to the of our evidence for a between and egg lysozyme et al. and et al. (2002) between and offspring lysozyme levels but not egg and lysozyme only 1 suggests a between lysozyme and egg (Saino et al. 2002). the other lysozyme and egg 1 a and 1 et al. and not them Thus, production of egg lysozyme not to and the cost of lysozyme deposition in albumen may be in the and the is also by et al. et al. and et al. 2002). production is by in the et al. by iron in the et al. and by infection in the et al. 2002). production is by in the of the et al. but is known of in the and therefore, not to be the of antimicrobial deposition in eggs. Thus, birds may have to antimicrobial concentrations of eggs within a adding more antimicrobials the amount of and protein in the albumen to the and to the embryo and and acid is for growth 1998). adding more antimicrobials a in other of albumen and of the and of albumen are required to determine the effects of for proteins on embryo and amount of in the egg may also because of variation in albumen egg size et al. et al. 2007). deposition of antimicrobials within clutches may be because birds can increase the of antimicrobials through incubation (Board and Fuller birds before the onset of full incubation and 1997; and Beissinger 1999). incubation may the temperature of eggs to levels at which antimicrobial are more incubation to be in and as well as other functions, may birds to increase the of antimicrobials in their eggs their should the effects of incubation on microbial infection and egg Finally, variation in antimicrobial concentration species was ranging of for all 3 This for lysozyme by et al. should sampling and to the of this nest and and microbial and may all the of concentrations for species. concentrations were between eastern and western eggs, despite in subtropical and temperate respectively, between violet-green and tree swallows in the not However, the concentrations of some antimicrobials for teal and parrotlet eggs, be a for of eggs is a function in other animal those that their eggs in such as and bacterial is many times greater than in the et al. 2007). the of antimicrobial at and in into the of parental a study that antimicrobials to their eggs, a (Giacomello et al. 2006). Microbes may in the of parental and their effects greater National We for for and of for to and of the for for to for in the and the Beissinger and the bird for on the