Development of Immunosensor for Early Detection of ...

Journal of Materials Science and Engineering B 7 (5-6) (2017) 127-134 doi: 10.17265/2161-6221/2017.5-6.006

D

DAVID

PUBLISHING

Development of Immunosensor for Early Detection of Aquatic Harmful Algal Blooms (HABs) Faridah Salam1, Hazana Razali1, Gayah Abdul Rahman1, Roziawati Mohd Razali2, Mohd Nor Azman Ayub2, Syah Noor Muhamad Ramli1 and Norhafniza Awaludin1 1. Biotechnology and Nanotechnology Research Centre, Malaysian Agricultural Research and Development Institute (MARDI), MARDI Headquarters, Persiaran MARDI-UPM, Serdang 43400, Selangor, Malaysia 2. Fisheries Research Institute(FRI), Batu Maung 11960, Penang, Malaysia Abstract: Blooms of toxic and harmful microalgae, known as “red tides”, represent a significant and tremendous threat to human health and fisheries resources throughout the Southeast Asia region and the world. A harmful microalga which produces secondary metabolite known as saxitoxin (STX) is a neurotoxin produced by a variety of genera of dinoflagellates. The ingestion of this toxin through contaminated shellfish will inhibit the neuron depolarizations and action potentials which result in respiratory failure in human and known as PSP (paralytic shellfish poisoning). PSP takes effect as soon as five minutes depending on the species and the concentration consumed. Currently, the identification of dinoflagellates cells is done by taxonomy, which is based on a broad expertise of specially trained staff, expensive equipment like electron microscopes and is very time consuming. In many countries, for the toxin monitoring, they use mouse bio-assay which is bulky, expensive, loss of animal life and high variability and higher limit of detection (40 µg per 100 g meat) and HPLC methods required many pre-treatment steps for the analysis. Therefore, in order to overcome this problem, development of early detection of harmful algae bloom and bio-toxin in aquatic ecosystem is needed. The detection is based on screen-printed carbon working electrode with onboard carbon counter and silver-silver chloride pseudo-reference electrode. A direct ELISA (enzyme-linked immunosorbent assay) format was developed and optimized on the surface of a carbon screen-printed working electrode by immobilizing the capture antibody using electro-deposition of gold nanoparticles conjugated with polyclonal anti-toxin antibodies. The detection reagent, toxin-HRP (horseradish peroxidase) conjugate was used as an enzyme label. Electrochemical detection was then carried out using 3,3’,5,5’-tetramethylbenzidine dihydrochloride (TMB)/H2O2 as the enzyme mediator/substrate system and conducted using chronoamperometry at 100 mV vs. onboard screen-printed Ag-AgCl pseudo- reference electrode. The minimum detection limits for harmful micro algae about 1 cell per mL. The immunosensor is very selective which is able to detect Alexandrium minitum with 100% selectivity and minimal cross reaction with others nontoxic algae (< 2%). Key words: Gold nanoparticles, immunosensor, screen-printed carbon electrode, Alexandrium minutum, saxitoxin, rapid detection, harmful algal blooms.

1. Introduction HABs (harmful algal blooms) have become a significant threat to fisheries, aquaculture industries, tourism, public health and worldwide economies. In the case of long run effect on the massive blooms can cause human food poisoning by accumulating of algal-origin toxins in filter-feeding shellfish. The economic impact of harmful algae is increasing sharply than in the past due to the drastic increase of Corresponding author: Faridah Salam, Ph.D., research field: biosensor technology.

consumption of seafood, growth of coastal inhabitants and tourism industries. A report published from USA indicated the massive impact of HABs on huge climate change and ocean acidification. The recurrence phenomenon of HABs is one of the big problems in Malaysia since it is now affecting Peninsular Malaysia. The demand for fish is expected to increase from 1.3 million tons in 2010 to 1.9 million metric tons in the year 2020 with a growth of 3.8% annually, therefore a great loss results in depletion of aquaculture and fishery productions in

128

Development of o Immunose ensor for Earlly Detection of o Aquatic Ha armful Algal B Blooms (HAB Bs)

long term peeriod. This iss a crucial issue for Malaaysia as the aquacculture industtry contributeed RM 7.3 billlion or 17.9% too the agro-foood industry. Scientists have h characterized 40 speciees of harmfuul microalgaee in Malaysia annd majority of these alggae will threeaten human and livestock. Cuurrently, the identification of dinoflagellattes cells is done d by taxoonomy, whicch is based on a broad experttise of speciaally trained staff, s expensive eqquipment likee electron miicroscopes annd is very time--consuming. Due to this problem, aquaculture industries neeed a simple and real timee for in-situ analyysis of dinofllagellates prooducing saxitooxin detection. Currently, biosensor devices cann offer a very v attractive alternative a technology for f contaminant detection sinnce they can be b rapid, sennsitive and sim mple to perform [1] as well as a can provide real-time and on-site analyysis [2]. Reaal-time detection of chem mical contaminantts is importannt since it proovides immeddiate interactive information i regarding thhe sample being tested and enables foood manufacturers to take corrective measures m befoore the produuct is releasedd for consumptionn. This methhod has beenn widely used in many appliccations such as in polluution control and monitoring of mining, industrial and a toxic gaases, environmenttal monitoriing, qualityy control, drug d developmennt, agriculturaal and veterinnary analysis and crime detecttion [2].

hyd drogen peroxiide concentraations and yiields a linearr resp ponse with thhe concentrattions of HRP (horseradishh pero oxidase) usuually emplooyed in im mmunologicall assaays. It also coontains stabiliizers. 2.2 Isolation and a Propagaation of Diinoflagellatess Alexxandrium sppp. The T pure cultture of the Alexandrium minutum m wass isollated from cuulture collectiion in Fisherries Researchh Insttitute. The biio-toxin prodduced algae, Alexandrium m min nutum was cuultivated in 22-10 L flask (Fig. 1) withh norm mal aerationn for 2-6 w weeks in the laboratoryy con nducted by Fisheries R Research Insstitute, Batuu Mau ung, Pulau Pinang. P Afterr 6 weeks off cultivation,, the Alexandriium cells were harrvested byy cen ntrifugation (55,000 rpm, 100 min) and th he pellet cellss wass collected and a freeze-dried before th he extractionn process of bio-tooxin and proteein marker was w proceed. n 2.3 Dinoflagellaates Bio-Toxinn Purification Bio-toxin B puurification w was conducteed followingg Ced dric Robillott et al. [3] m methods. Th he 5 g driedd Aleexandrium ceells was grindded with 50 mL of 80% % ethaanol (pH 2 inn HCL) ratio 1:1 (w/v) and was heatedd at 100 1 oC for 5 h under refluux to convert the majorityy of C-toxins to gonyautoxinn. The extraact was thenn app plied to the activated a chaarcoal and th he toxin wass elutted with 20% ethanol contaaining 6% aceetic acid.

2. Materiaal and Meth hods 2.1 Source of Chemicals PBS (phoosphate bufffered saline), comprisingg of 0.13 mM NaH2PO4, 0.5 mM Na2HPO O4 and 0.51 mM NaCl, pH 7.4 was prepaared by dissoolving five buuffer tablets in 1 L distilleed-deionised water. For the analysis, reeady-made T TMB (3,3’,5,5’-tettramethylbennzidine) soluution was used. u According to t the manuffacturer (Eurropa Bioprodducts Ltd., Ely, UK), U the TM MB solution is stable at rooom temperature and is not sensitive s to normal n laboraatory light. It is optimized with respecct to TMB and

Fig.. 1 flask k.

Alexandriium minutum aat 4-6 weeks culture in 10 L

Elu uted gonyau--toxins weree further pu urified on a


Bio-Gel P2 column and then were eluuted isocraticcally with 0.05% % acetic accid. Toxic fractions were w combined and a concentrrated by rottary evaporaation under vacuuum and lyophhilized for connversion to STX. S The toxin sample was re-dissolved r i 0.05 M accetic in acid and theen was injectted on a Bioo-Rex 70 coluumn and the toxiin was elutedd with a graddient of 0.05-3 M acetic acid with 10 mL L fractions collected. c Figg. 2 showed the chromatograam of the bio-toxin b fracction after purificcation using the t three collumns (Charccoal, Bio-Gel P2 and Bio-Rexx 70). The chromatogram c m of HPLC analyysis of the peeak that wass eluted from m the three colum mns indicated the type of bio-toxin (S STX) obtained from the Alexanndrium minutum. 2.4 Dinoflaggellates Proteein Biomarkerr Purificationn The 5 g drried Alexandrrium cells waas grinded witth 50 mL phosphaate buffer salinne (PBS, 0.01 M, pH 7.4)) and centrifuged for 15 min at a 6,000 rpm. The supernaatant was collecteed and preccipitated withh 80% saturrated ammonium sulphate forr 30 min. This pellet was re-dissolved with PBS annd was dialyzzed three timees in 5 L PBS (00.01 M, pH 7.4). Alexanndrium cell wall protein from m pellet (from the PBS extrraction stage) was extracted agaain with 50 mL m of 50 mM phosphate buuffer containing 1 mM ED DTA and 1 mM DTT by re-grinding, precipitated and centrifuugation follow wing the similar PBS extracction step. Extracted E prootein marker wass further purrified using Sephadex75 gel chromatograaphy with 0.2 MPa and 0.22 mL·min-1 off the flow rate using ACTA prrotein purifieer system. Prootein sample was then eluted by PBS andd its content was

1299

CA proceduree for each off the fractionn tested using BC usin ng BSA (boviine serum albuumin) as stan ndard protein. 2.5 Antibody Prroduction agaainst Protein n Bio Markerr and d Bio-Toxin The T polyclonnal antibodiies were produced p byy imm munizing New N Zealannd white rabbits viaa subcutaneous injjection with aan emulsion consisting off 0.5 mL of bio-tooxin (5 µg·mL L-1) in 0.5 mL L of PBS andd an equal voluume of CF FA (complete Freund’ss adju uvant). The injections w were repeated d three timess weeekly after thhe initial injjection, subsstituting IFA A (inccomplete Freuund’s adjuvannt) for compllete adjuvant.. A booster b injecction was givven one mon nth after thee initial injection and was repeeated at montthly intervalss therreafter. The rabbits weree bled for antibody a titerr deteerminations two t weeks aafter each boost. Antiseraa agaainst bio-toxiin was dilutted with disstilled waterr (1:1 10) and theen precipitated with 80 0% saturatedd amm monium sullphate with continuous stirring too preccipitate serum m proteins. The serum mixture wass cen ntrifuged (Avvanti J-26 XP P, Beckman Coulter, Inc.. USA A) at 20,000 rpm for 15 m minutes at 4 oC. The pellett wass resuspendedd in PBS andd dialyzed 3X in PBS too rem move ammonnium salt andd then eluteed through a protein A coluumn (manuaally pack) using u AKTA A purifier instrum ment (Pharrmacia Ltd.., Sweden).. Fractions givingg the highesst absorbance reading att OD D 280 nm were w collected and freezee-dried (IgG G stocck). As the freeze-dryinng did not affected thee antiibody activitiies and also will prolong g the storagee stab bility [4]. IgG G titers for annti-bio-toxin antibody a wass

(a) (b) Fig. 2 Schem matic diagram m of the antibod dy-antigen reaaction on the carbon c workingg electrode usiing immunoassay format (a)) Indirect comp petitive formatt (b) Sandwich h format.

130

Development of Immunosensor for Early Detection of Aquatic Harmful Algal Blooms (HABs)

determined by indirect ELISA method. 2.6 Preparation of Conjugated Gold Nano-Particles with Antibody and Horseradish Peroxidase Enzyme Gold nano-particles with average diameter of approximately 20 nm were purchased from BBInternational, Cardiff, UK. The colloidal gold solution was stored in a dark bottle at 4 °C and used directly without any pre-treatment. The particle size of colloidal gold was ~20 ± 3.2 nm and the number of particles is ~7  1011 particles per mL. The UV-vis spectrometer showed an absorption peak at 525 nm. The antibody-colloid gold conjugate was prepared according to the procedure described by Chen et al. [5]. The antibody-colloidal gold conjugate was prepared by adding 100 µL of anti-tetracycline antibody (1.0 mg·mL-1) to 1.0 mL of pH-adjusted colloidal gold solution (pH 9.0), followed by slow shaking for one hour at room temperature. This allowed the antibody to adsorb onto the gold nanoparticles through a combination of ionic and hydrophobic interactions. Then, 100 µL of 10% BSA solution was added to the gold-antibody mixture and left for further 30 minutes at room temperature. The mixture was then centrifuged at 9,000 rpm for 30 minutes. The supernatant solution was then discarded and the soft sediment of immuno-gold was dissolved in 100 µL of 1% BSA solution and stored at 4 °C until used. The BSA was used to stabilize the immuno-gold colloid and minimize the non-specific adsorption during the assays and also to block the unoccupied site of the gold surface. The amount of antibody bound per nano-particle was then calculated based on protein assay results achieved from analysing the supernatant for unbound proteins using Bradford protein assay of Sigma, UK. 2.7 Screen-Printed Carbon Electrode Fabrication and Electrochemical Assays SPCEs (screen-printed carbon electrodes) consisting of carbon working electrode, carbon

counter electrode and silver-silver chloride pseudo-reference electrode were fabricated by Screen Print Technology Sdn Bhd. Sungai Petani, Malaysia. The SPCE used in this work consisted of a carbon working electrode with a 5 mm diameter giving a 19.6 mm2 planar area. All electrodes were then tested using a multimeter before use. The sensor edge connector was purchased from DropSense Ltd. UK through MetrohmSdnBhd, its product distributor is in Malaysia. Electrochemical measurements were carried out by placing a 50 μL solution onto the electrode, covering the 3-electrode area. Each measurement was carried out in triplicates using a new strip in a non-deaerated and unstirred solution. Measurements were performed using the Autolab Type II (Eco Chemie, The Netherlands) with NOVA 1.6 software. Cyclic voltammetric measurements were carried out by scanning at 50 mV s-1 between -1.0 and 1.0 V relative to Ag/AgCl reference electrode. Stock solutions of 50 mM potassium ferrocyanide were prepared in 0.1 M KCl. For the selection of optimal potential for TMB-H2O2-HRP system, chrono-amperometry was conducted similar to that reported previously [6]. 2.8 Immunosensor Development The carbon surface of the screen-printed electrode was directly used for the electro-deposition of polypyrrole-antibody-nano gold treatment. The freshly cleaned carbon was then covered with a 75 mM solution of polypyrrole plus antibody-nano gold and kept unstirred at room temperature. Chronoamperometry measurements were carried out by setting 1.0 V fix potential for 20 minutes onboard screen-printed Ag-AgCl pseudo-reference electrode. The antibody-nano gold coated electrode was then washed with 0.1 M phosphate buffer to remove the excess of unbound antibody-nano gold. The layer is very stable if kept dry (under silica gel) for several weeks at 4 oC [7]. For saxitoxin or Alexandrium assays, various dilutions of saxitoxinstandard (0-250


ppb) or Alexandrium A cells and saxitoxin-H HRP conjugate or anti-A Alexandrium antibody-H HRP conjugate (220 uL) were added to thhe electrodes and incubated foor 20-30 minnutes at 37 oC. C The electrrode was washed 3 times withh PBS-T and dried d gently. The assay was thhen performeed by addingg 50 uL of TMB T solution usiing chronoam mperometry at 100 mV for 100-200 s. Calibrationn curve wass fitted witth a non-linear regression using 4-parrameter loggistic equation annd the detecction limit (L LOD) was then calculated based on the t followinng equationn as described byy Tijssen [8]:

where, s is for f standard deviation d of the zero valuue, a and d are the maximuum and miniimum valuess of calibration curve, c x is thhe concentrattion at the EC50 E value and k is the hill sloppe. 2.9 Cross Reeactivity Studdies The sensiitivity and specificity s off the sensor was investigatedd in relation to t other algaae such Pavllova, Teraselmis, VaunoBatan, Chaelo, Scipsiella, SK, Alexandrium m affine annd Alexandrrium spp. The (biotoxin) sensor specificity of o this toxic Alexandrium A with otherss nontoxic algae was conducted by replacing Alexandriumm A minitum witth above liisted algae.

3. Result and a Discusssion The eleectrochemicaal immunoosensor sysstem developed inn this work for f saxitoxin and a Alexandrrium minutum deetection was based on diirect competiitive and sandwicch immunoasssay format with w HRP useed as the enzym me label and TMB/H2O2 as the substrate/meediator system m as illustrateed in Fig. 2. The higher saxittoxin in the sample, the lower the siignal achieved frrom the electrochemicall immunosennsor. While for thhe Alexandriium minutum m, the higher cell numbers arre in the saample, the higher h signaal is achieved in the electrochhemical measuurement.

131

3.1 Anti-Alexanddrium minutuum and Bioto oxin Antibodyy Cha aracterizationnand Standarrd Curve Fab brication The T total prottein from exttracted Alexa andrium from m drieed cells usingg ammonium m sulphate preecipitation iss abo out 29.88% and a after beiing further purified p withh Sep phadex75 gel chromatoggraphy, the pure p proteinn con ntain is about 8.45%. Proteein profile from extractedd sam mple was characterized c d and estim mated aboutt 115 5,000 KD. For the bio-toxiin purification n, HPLC andd ELIISA analysis is determineed about 10.2 2% with 0.900 g STX S from 5 g dried alggae sample. Protein andd bio--toxin compllex were thenn used as an n immunogenn in immunization proceduure for th he antibodyy production. The primary andd booster imm munization inn rabb bit with bio--toxin compllex with thee addition off Freu und’s adjuvant is an iinexpensive strategy forr poly yclonal antibbody producction. Freund d’s adjuvantt (parraffin oil bassed) has beenn used for sttimulation off the immune sysstem by Myccobacterium in Completee Freu und’s Adjuvaant to generatte high antibo ody titers [9].. Thee initial expperiment wass to screen the purifiedd poly yclonal anttibodies by titration with algaee (Aleexandrium minutum) aas the targ get harmfull bio--toxin. Thiss assay invvolved the titration off mon nthly-purifiedd antiserum ssamples usin ng algae cellss adsorbed to miicrotitre plattes. Control experimentss usin ng purified prre immune seerum showed insignificantt bind ding levels. The T whole ppattern of anttibodies levell in rabbit r withinn six monthss immunizatiion schedulee wass obtained. There T is a ssignificant biinding and a grad dual increasee in the apparrent binding activity withh boo osting injectiions over thhe period teested. Thesee initial tests provide informatiion on the exttent of serum m matturation folloowing the tyypical pattern n on primaryy and d secondary response, w where specifiic antibodiess app pear after repeating r im mmunizations. From thee imm munization schedule, bleed 3 and 4 showed thee high hest activity where the anntibody titer of 1:0.00001 mg was obtaineed. The antibbody concenttration at 0.1 mg··mL-1 was ussed for the deevelopment of o algae-toxinn stan ndard curve using u direct ssandwich EL LISA method..

132


The ELISA A result for standard currve developm ment -1 shows a lim mit of detectioon about 5 cells mL that was obtained (Fiig. 3a). Immuunosensor tecchnique was then developed using u sandwicch ELISA forrmat by replaacing anti-algea toxin t compleex antibody with nano-gold coated antibbody on the seensor platform m. Fig. 3b shhows immunosenssor standard curve that was develooped using puree A. minuttum cells. The LOD for Alexandrium m detection iss 1 cells mL-1. 3.2 Specificiity of the HAB Bs Immunoseensor The sensiitivity and specificity s off the sensor was investigatedd in relation to t other algaae such Pavllova, Teraselmis, VaunoBatan, Chaelo, Scipsiella, SK, Alexandrium m affine annd Alexandrrium spp. The (biotoxin) sensor specificity of o this toxic Alexandrium A

with h others nontoxic n alggae was co onducted byy repllacing Alexaandrium minnitum with above listedd algaae. The reesults show wed that th here is noo crosss-reactivity with non-toxxic tested allgae and thee imm munosensor is i able to deetect all the Alexandrium m spp. tested with more than 500%. 3.3 Immunosenssor Method V Verification Using U Actuall Sam mple Algea A samplee was collected from CherohPaloh, C , Pek kan, Pahang. Nine samplees out of 19 samples aree positive Alexanndrium meassured by im mmunosensorr strip p and was correlated c w with microsco opic analysiss (Taable 1). Thee data are the initial step of thee app plication of HABs im mmunosensorr for earlyy diag gnosis of HA AB in aquatic ecosystem.

Fig. 3

(a) Stand dard curve for Alexandrium using ELISA (a) ( and immun nosensor (b).

Fig. 4

Crosss-reactivity of immunosensor i r with others non-toxic n algaee.

(b)

133

Development of Immunosensor for Early Detection of Aquatic Harmful Algal Blooms (HABs)

Table 1

Comparative data of water samples analysis using immunosensor and microscopic analysis.

Sampling point

No. of Alexandrium minutum cells (cells/ml) Rep 2 Mean Mean * made to integer

Name of sample Rep 1

1

2

3

4

5

1.0 Plankton at 1m deep

11.58

‐9

5.79000

6

No

‐15

0.61750 0.05900

0 0

No No

‐8

48.68675

48

Yes

‐8

1.78350

2

No

‐12

0.05437

0

No

‐5

0.03063

0

No

‐6

0.11575 0.00750 4.63500

0 0 5

No No No

‐9

0.00107 1.23500 12.21600 33.43000 0.02200 0.77100 4.51400 10.17900 0.21100

0 1 12 33 0 0 5 10 0

No No Yes Yes No No No Yes No

8.07 x 10

1.1 Water sample at 1m 1.2 Water sample at 2m

1.235 0.104

3.25 x 10 0.014

1.3 Water sample at 3m

96.52

1.02 x 10

2.0 Plankton at 1m deep

3.567

6.73 x 10


0.073

1.32 x 10


0.04

2.16 x 10

2.3 Water sample at 3m 3.0 Plankton at 1m deep 3.1 Water sample at 1m

1.34 x 10 0.015 9.153

3.2 Water sample at 2m 3.3 Water sample at 3m 4.0 Plankton at 1m deep 4.1 Water sample at 1m 4.2 Water sample at 2m 5.0 Plankton at 1m deep 5.1 Water sample at 1m 5.2 Water sample at 2m 5.3 Water sample at 3m

9.99 x 10 1.235 12.216 33.43 0.022 0.771 4.514 10.179 0.211

‐5

5.25 x 10 0 0.117

‐5

1.98 x 10 0.541 16.49 96.52 0.706 3.17 1.374 6.428 4.906

4. Conclusion One of the advantages of this immunosensor can be

funding this research (02-03-08-SF0267/RB5018SF10).

packed with customized portable digital reader and bio-reagent for easy use and handling. Another advantage of the present invention provides detection for the presence of toxic microalgae before HAB occurred and its toxin molecules in various fluids or solid food. A great advantage of direct detection of

References

bio-toxin and toxic microalgae using antibody-nanogold-polymer electrode is direct sample application with minimal sample preparation/extraction. This is opposed to other bio-toxin and toxic microalgae determinations using established instrumentation

[3]

(HPLC/LC-MS/Microscope) and mouse bioassay. In addition, the device components in the apparatus of the present invention are more robust than those used for example in optical measurements, facilitating the manufacture of the low cost reusable meter for use with disposable test devices. Further, the present invention can provide for very rapid test times.

[1]

[2]

[4]

[5]

[6]

[7]

Acknowledgements The authors would like to thank the MOSTI for

Microscope analysis

[8]

project

Alocilja, E., and Radke, S. M. 2003. “Market Analysis of Biosensors for Food Safety.” Biosensors & Bioelectronics 18: 841-6. Tothill, I. E. 2011. “Biosensors and Nanomaterials and Their Application for Mycotoxin Determination.” World Mycotoxins Journal 4: 361-74. Cedric, R., Daniel, K., James, B., and Lyndon, E. L. 2009. “Synthesis of bifunctional saxitoxin analogues by biotinylation.” Toxicon 53: 460-5. Dráber, P., Dráberová, E., and Nováková, M. 1995. “Stability of Monoclonal Igm Antibodies Freeze-Dried in the Presence of Trehalose.” Journal of Immunological Methods 181 (1): 37-43. Chen, Z-P., Peng, Z-P., Zhang, P., Jin, X-F., Jiang, J-H., Zhang, X-B., Shen, G-L., and Yu, R-Q. 2007. A Sensitive Immunosensor Using Colloidal Gold as Electrochemical Label Talanta 72: 1800-4. Salam, F., and Tothill, I. E. 2009. “Detection of Salmonella Typhimurium Using an Electrochemical Immunosensor.” Biosensors and Bioelectronics 24: 2630-6. Chaki, N., and Vijayamohanan, K. 2002. “Self-assembled Monolayers as a Tunable Platform for Biosensor Applications.” Biosens. Bioelectron. 17: 1-12. Tijssen, P. 1985. Practice and Theory of Immunoassay.

134

[9]

Development of Immunosensor for Early Detection of Aquatic Harmful Algal Blooms (HABs) Amsterdam: Elsevier. Trott, D. L., Hellestad, E. M., Yang, M., and Cook, M. E. 2008. “Additions of Killed Whole Cell Bacteria

Preparations to Freund Complete Adjuvant Alter Laying Hen Antibody Response to Soluble Protein Antigen.” Poultry Science 87: 912-7.