Immunolocalization of Surface Antigens on Campylobacter jejuni using FESEM and a Back-Scatter Electron Detector
DELILAHF. WOOD,† ROBERTMANDRELL, ANNAH. BATES,
ANDPAULINEC. YU
U.S. Department of Agriculture, Agricultural Research Service, Western Regional Research Center. 800 Buchanan St., Albany, CA, USA
†Formerly published as Delilah W. Irving.
Campylobacter jejuni is one of the most common causes of food poisoning in humans in the US. Poultry are thought to be the main source of the bacteria, however; there is no definitive method for differentiating bacterial species on food surfaces. Most bacteria have surface molecules that are species specific and some of the molecules may be responsible for the adhesion of the bacteria to food surfaces.
Monoclonal antibodies (MAbs) can be produced that bind to antigens on these surface molecules, and they facilitate identification of bacterial species and strains in immunoas-says and complex environments. When coupled with an appropriate label, fluorescence and scanning electron microscopy (SEM) allow visualization of antibodies bound to antigens on bacteria in complex environments; e.g. C.
jejuni on food surfaces.
Specific surface antigens on C. jejuni have been identi-fied and MAbs that bind to them have been produced. The bacterial surface antigens have been labeled with MAbs and fluorescent anti-mouse secondary antibodies and the bac-teria observed by confocal microscopy. Confocal micro-scopy is an excellent tool; however, the resolution is insuf-ficient to recognize structural details. High-resolution images can be obtained of bacteria using field emission SEM, and, when combined with an appropriate electron-dense label (gold), the binding site of the antibody (there-fore, the specific antigen) is located.
Gold particles can be made in various sizes and conju-gated to an antibody. Smaller particles allow more precision in the localization of surface molecules. Gold particles were viewed using electron back-scattering techniques. Gold particles are visible using secondary electron detection.
However, secondary electrons only give superficial infor-mation and bacterial surfaces contain other particles that may be of similar dimensions as the gold particles, there-fore the gold particles are not distinct. Back-scattered elec-trons produce images of increasing brightness as a function of increasing molecular weight. Therefore, the heavier gold particles appear bright against the darker bacterium.
Campylobacter jejuni cells from culture were incubated with anti-C. jejuni MAbs, the cells were rinsed, then incu-bated with 15nm and 30nm gold-conjugated anti-MAbs.
The labeled bacteria were fixed overnight at 23%C in 3%
glutaraldehyde in phosphate buffered saline (PBS), pH 7.4, and rinsed three times in PBS. The samples were pel-leted between each step. The pellets were resuspended and post-fixed in 1% aqueous osmium tetroxide overnight at 4%C and rinsed three times in water (pelleted between each step). The pellets were re-suspended in water and filtered through 0.2 mm Nuclepore filters that had been pretreated in 0.1% poly-L-lysine to increase adherence of the bacte-ria to the filter. The filters containing bactebacte-ria were placed between hard plastic mesh, fastened together to keep the filters flat and maintain orientation of the bacteria and treated as a unit throughout subsequent steps. The mesh unit was dehydrated in ethanol and critical point dried. The fil-ters were removed from the mesh unit and mounted on specimen stubs and sputter coated with gold-palladium.
Samples were observed and photographed using both a standard secondary electron detector to obtain high reso-lution secondary images (Fig 1A), and a YAG (ytrium-alu-minum-garnet) back-scatter electron detector to locate the gold particles on bacterial surfaces (Fig 1B).
Spiral and coccoid forms of C. jejuni were observed (Fig.
1A), however, the immunogold labeled antibodies were located predominately on the coccoid forms (Fig. 1B).
These results indicate either that the MAb-defined antigen was not expressed on the spiral cells, or that the antigen is better exposed on the coccoid form rather than the spiral form. It has not been determined whether the specific anti-gen identified is associated with adhesion of C. jejuni to food surfaces. However, the results indicate that the form of the bacteria affects the surface availability of antigens;
which may be important in identifying antigens responsi-ble for the adhesion of C. jejuni to food. Most signifi-cantly, a method has been identified that will readily char-acterize the surface-available antigens.
FIG. 1 Campylobacter jejuni. A. Secondary electron image show-ing both the spiral (left) and coccoid (right) forms. B. Back-scatter electron image showing gold label on the coccoid form of the bac-terium. Bar = 500nm.
Acknowledgments
The authors wish to thank William Roth (Nissei-Sanyo America) for taking the photographs and Donald Becker and Daniel Becker (Hitachi) for facilitating the use of the equipment.
Particle Distribution in Microscopy Images of Cryo-sectioned Sausages as affected by Fat Content RAGNIOFSTAD, VIBEKEHØST, ØYVINDLANGSRUD ANDACHIMKOHLER
MATFORSK-Norwegian Food Research Institute, Osloveien, Norway
Introduction
It is previously reported that the type of meat and fat, the levels of moisture and fat, and the comminution process influence the microstructure as well as the sensory quality of finely comminuted meat products1–3. The precise quan-tification of structures will be the first essential step for pre-dictive modelling of sensory properties such as texture and juiciness.
The aim of this work was:
a) To quantify the fat particle distribution pattern in light microscopy images of cryo-sectioned sausages b) To use multivariate analyses of variance to identify
significant differences in the fat distribution pattern among treatments.
Materials and Methods
Nine different pork/fat trim sausages were prepared with three levels of fat (15, 20, 25%) and three levels of chop-ping (short, medium and long time) in dual trials (day 1 and 2). One sausage from each of the 18 samples was prepared for microscopy, i.e. frozen in liquid nitrogen, cryo-sec-tioned (10 µm), stained by Nile Red and examined with a fluorescent microscope (129×). A CCD camera (Hitachi KP-D50 colour digital) was used to capture five images from randomly selected regions.
Image processing was performed using Image Pro-Plus 3.0 according to Kohler et al.4With the area distribution method one gains an array of variables from each image.
It is therefore natural to use multivariate methods in order to classify the microscopy images. A new method for mul-tivariate analyses of variance5was used to identify signif-icant differences among treatments. This method (Fifty-fifty MANOVA) handles highly correlated responses and it can be used to analyse continuous curves.
Results
Figure 1 shows examples of images taken from the sausages. The pictures give a qualitative impression of
dif-ferences in fat content and particle distribution. All bright objects were counted and their areas were calculated. Based on these data, a smoothed histogram was obtained for each sample. A logarithmic scale was used on the x-axis (log area). To focus on the information on the images related to the fat particle distribution, and not differences related to the fat content, all the histograms were standardised (area=1). These distributions (curves) were used as responses in Fifty-fifty MANOVA. The difference in area distribution of the fat particles among the three fat levels was clearly significant (p<0.01). The effect of chopping time and the day-to-day variation was however not signif-icant.
(a)
(b)
(c)
FIG. 1 Images of cryo-sectioned sausages with different fat levels, 15% (a), 20% (b) and 25% (c) fat.
FIG. 2 Average area fraction distribution of the fat filled pores on images of cry-sectioned sausages with three levels of fat.
To illustrate the fat level differences, an average area fraction distribution within each fat level was calculated.
The distributions for the different fat levels are shown in Figure 2. At 15 % fat there is one peak, at 20% there are two peaks and at 25% fat content there is one peak and one saddle point, which indicates that there are two peaks over-lapping. Thus, in the high-fat batter (20, 25%) the fat is either emulsified to small droplets or remains as bigger droplets and/or agglomerates of droplets restricted by intact cell membranes and/or the protein-matrix. The higher aver-age diameter of the fat droplets in the low fat samples may be due to less restriction of the higher moisture protein matrix embedding them.1
Conclusion
1. In the experimental design performed, the fat distribu-tion pattern was only significantly dependent on the fat content.
2. The combination of area distribution method and mul-tivariate statistics seemed possible to use to find differ-ences in structures relevant for quality parameters.
References
Lee CM: Microstructure of meat emulsions in relation to fat stabili-sation. Fd Microstructure 4 63–72, 1985.
Koolmees PA, Moerman PC, Zidjderveld MHG: Image analysis of the fat dispersion in a comminuted meat system. Fd Microstruc-ture 8 81–90, 1989.
Colomenero FJ, Barreto G., Mota N, Carballo J: Influence of protein and fat content and cooking temperature on texture and sensory evaluation of bologna sausage. Lebensm.-Wiss. U. Technol. 28 481–487, 1995.
Kohler A, Hoest V, Ofstad R: Image analysis of particle distribution in microscopy images of cryo-sectioned sausages. Submitted, 2000.
Langsrud Ø: Fifty-fifty MANOVA: Multivariate Analysis of Variance for Collinear Responses. In preparation, 2000.
0.2
0.1
0
0 2
25% fat 20% fat 15% fat
4 6
log (area)
8 10
Ultrastructure of Fouled Microfiltration and Ultrafiltration Membranes
YANJING, BRYONYJ. JAMES
Research Center for Surface and Materials Science, Department of Chemical and Materials Engineering, The University of Auckland, Auckland, New Zealand
Membrane separation is a technology experiencing rapid growth, particularly in the field of food processing1. One of the major factors limiting the economical feasibility of membrane separation technologies is the problem of “foul-ing.” An understanding of the fundamental mechanisms of membrane fouling is essential if the optimum performance of this technique is to be realised.
Surface morphology and internal microstructure of a membrane have a great effect on fouling and separation per-formance. Unfortunately, because of the nature of proteins and the polymeric membrane, studying the structure of the filtration materials and fouled layer is difficult. This is especially true in the study of the internal membrane struc-ture2. In this study the microstructure and ultrastructure of the fouled membrane were examined using a number of techniques.
Scanning Electron Microscopy (SEM), Atomic Force Microscopy (AFM) and X-ray Photoelectron Spectroscopy (XPS) were used to investigate the physical evidence of the formation of a gel layer on the surface of the membrane, the internal fouling within the membrane structure and the effect of pressure on permeate flux and fouling. A new method has been developed to study internal membrane structure using high-resolution FEG-SEM with a low accel-erating voltage.
Two Osmonics©membranes were used: a microfiltration membrane (pore size of 3 µm) and an ultrafiltration mem-brane (molecular weight cutoff 3,500). Filtration of 10%
(w/w) skim milk powder was performed on a high-pressure flat sheet with cross-flow configuration. Characteristics of a fouled membrane were studied in terms of transmem-brane pressure and filtration time, some results are sum-marised here.
Surface morphology and internal microstructure of the unused membrane.
SEM and AFM were used in conjunction. The skin layer of the microfiltration membrane showed a finger like sub-structure and sieve like top layer. The membrane surface had a varied pore size with a maximum size of approxi-mately 0.35 µm. This was larger than the product specifi-cation.
The ultrafiltration membrane showed nodules, in vari-ous sizes, which spread across the entire surface. A cross section of the membrane revealed a smooth fouling resis-tant layer at the surface, a nodular layer in the middle and a foam like layer with voids connected to the support layer.
Fouling on the surface of the membrane
A gel layer had formed on the surface of both membranes after filtration. The structure of the surface of the fouled membrane was seen more clearly under AFM than SEM.
Fouling within the membrane structure
XPS analysis showed internal fouling occurred after fil-tration. It is believed this could be attributed to the inter-action between milk proteins and the membrane polymer and protein-protein reaction3. SEM cross sections, using a novel sample preparation technique, of the fouled micro-filtration membrane presented strong physical evidence to prove this assumption. These clearly showed protein-pro-tein interaction and interaction between proprotein-pro-tein and the pore walls of the membrane, thus forming agglomerates.
This led to pore plugging and narrowing and finally total blockage of the pores.
From the results of this study it may be concluded that:
1. AFM is an appropriate and convenient technique for examining the upper surface of fouled membranes.
2. SEM, using appropriate sample preparation, is the best technique for examining the internal structure of a fouled membrane.
In terms of the filtration parameters affecting fouling it would appear that in ultrafiltration permeate flux increases with pressure, whereas in microfiltration there is a critical pressure above which permeate flux decreases. This is due to the effect of concentration polarization and consolida-tion of the foulant, which is observed under SEM.
References
1. B. Hallstrom, G. Tragardh and J. L. Nilsson, “Membrane Tech-nology in the Food Industry”, Engineering and Food, vol. 3 Advanced Processes, Elsevier Applied Science, London, pp.
194-208
2. K. J. Kim, V. Chen and A. G. Fane, “Characterization of Clean and Fouled Membrane Using Metal Colloids”, Journal of Mem-brane Science, vol. 88, 1994, pp. 93-101.
3. George Belfort, Robert H. Davis, Andrew L. Zydney, “The Behavior of Suspensions and Macromolecular Solutions in Crossflow Microfiltration”, Journal of Membrane Science, vol.
96, 1994, pp1-58.
The Microstructure of Corn Tortillas During Baking
CASSANDRAMCDONOUGH, ELLYSUHENDRO,
ANDLLOYDW. ROONEY
Cereal Quality Lab, Soil and Crop Sciences, Texas A&M University, College Station, TX, USA
The basic structural changes that corn tortillas undergo during baking were documented with an Electroscan envi-ronmental scanning electron microscope. Samples were collected after the tortillas passed across each of the three tiers in the oven, placed in air-tight plastic bags and stored at –4°C until viewing. Corn is traditionally cooked in alkali to soften the kernel, partially gelatinize the starch, denature the protein, and remove the pericarp; it is then ground into masa. Masa consists of small pieces of germ, pericarp, aleurone and endosperm, and free starch granules, cell frag-ments, and dissolved or dispersed solids and lipids in water.
The masa is held together by a ‘glue’ that is composed of leached amylose and amylopectin from the starch, denatured protein and lipids. The masa is die-cut into a disk shape and baked in a 3-tier oven, with temperatures of tiers 1, 2, and 3 ranging from 280–470°C. The texture of the unbaked tor-tilla disks was very dense, with no air cells. Intact starch granules and small endosperm pieces were embedded in the starchy matrix of the masa. When the unbaked tortilla disk entered the oven, there was no expansion until the disk was heated. The heat flashed the moisture off the surface of the tortilla, so the starch on the surface remained ungelatinized.
After passing through the 1st tier of the oven, small chan-nels formed where the steam began to escape the interior, but the microstructure of the tortilla was not dramatically different from the masa disk. On the 2nd tier, the starch in the interior was further gelatinized, but that on both surfaces retained birefringence. Some expansion continued and one or more central air voids formed as the steam continued try-ing to leave the disk. On the 3rd tier, the tortilla flipped over once again so the first side was again exposed to heat. By this time, the starchy matrix was dehydrated enough to start to set its crumb structure. The tortilla crumb was char-acterized by large central voids due to expansion of the gases during heating (visually observed as ‘puffing’ like a bal-loon), and small air holes/channels that formed just beneath either surface, a few of which broke through to the surface.
Even though the center of the tortilla puffed in the oven, the outer edges of the disk remain sealed. In the crumb, cell walls were further degraded until they were no longer vis-ible, there was further loss of starch crystallinity to the point where roughly 60% of the starch was fully or partially gelatinized, and the protein was further degraded. As soon as the tortillas left the oven and fell to the cooling conveyer, they collapsed quickly as the heated gases cooled to near ambient temperature. The inner surfaces may fuse together if they are still moist enough at this point, though some will remain separated, which impacts the textural quality of the tortilla.
Use of Confocal Microscopy and the Green Fluo-rescent Protein in Ecological Studies of Salmonella on Plant Surfaces
M.T. BRANDL, R.E. MANDRELL
Food Safety and Health, USDA/ARS, WRRC, Albany, California, USA
In recent years , human infection with virulent strains of Salmonella enterica (S.e.) linked to contamination of food has been the source of great concern. Although generally associated with consumption of meat products, outbreaks of salmonellosis are increasingly linked to contaminated fruits and vegetables in the United States as well as in many foreign countries (3). In particular, large international out-breaks of Salmonella infections have originated from sprouts grown from contaminated seeds (4) and many out-breaks have been linked to cilantro, parsley and melon in the United States (2). The interaction of Salmonella with its vertebrate hosts has been the object of intense studies. In contrast, our understanding of how this enteric pathogen col-onizes plant surfaces is lacking. We have transformed a strain of S.e. serovar Thompson that was linked to an epi-demic from cilantro, with a gene encoding the green fluo-rescent protein (GFP) from Aequoria victoria (1). This GFP-labeled strain (S.e. Thompson pWM1007) was used for in situ investigation of colonization of cilantro plants in conjunction with confocal microscopy.
Young cilantro plants were inoculated with a suspension of S.e. Thompson pWM1007 at a final concentration of 1x104cells per g leaf tissue. The plants were incubated at room temperature in a dew chamber immediately follow-ing inoculation. Inoculated cilantro leaves were sampled at regular time intervals after inoculation, and bacterial pop-ulations on the leaves were observed directly with 40x and 63x objective lenses on a confocal laser scanning micro-scope (CLSM) (TCS4D, Leica Lasertechnik, Germany) equipped with argon, krypton, and He/Ne lasers. GFP/flu-orescein and LP590 filters were used to simultaneously monitor the fluorescent signals from the Salmonella cells and the leaf, respectively.
Cilantro leaves sampled immediately after inoculation harbored single cells of GFP-labeled S.e. Thompson dis-tributed at low density on the leaf surface. Major changes in colonization were observed as early as 2 days postinoc-ulation. The Salmonella population had reached high den-sities in the vein areas of the leaf, and distinct micro-colonies were present. Leaf surface areas between the veins also hosted Salmonella cells, but at much lower den-sity. Transversal (xz) scans revealed that the cells were located on the cuticle of healthy leaf tissue. However, high numbers of Salmonella cells were observed within the cuticle and epidermal cells of cilantro leaves where lesions were present. Examination of healthy senescent leaves, 9 days postinoculation, showed that Salmonella had reached very high populations on the veins and their vicinity.
We have investigated the spatial and temporal
We have investigated the spatial and temporal