Proteomic study of a model causative agent of harmful red tide, Prorocentrum triestinum I:...

Proteomic study of a model causative agent ofharmful algal blooms, Prorocentrum triestinum II:The use of differentially expressed protein profilesunder different growth phases and growth conditionsfor bloom prediction

Leo Lai Chan1, 3, Ivor John Hodgkiss1, 3*, Jennifer Man-Fan Wan2, John Hon-Kei Lum3,Abby Sin-Chi Mak2, Wai-Hung Sit2 and Samuel Chun-Lap Lo3*

1Department of Ecology and Biodiversity2Department of Zoology, The University of Hong Kong, HKSAR, Hong Kong, P. R. China3Proteomic Task Force, Department of Applied Biology and Chemical Technology,The Hong Kong Polytechnic University, HKSAR, Hong Kong, P. R. China

Simultaneous comparison of differentially expressed protein profiles of Prorocentrum tri-estinum grown under different growth phases and growth conditions indicated the pres-ence of phase-specific and stress-responsive proteins, respectively. Correlation studieson these proteins in relation to cell division phasing patterns and to models of phytoplank-ton growth inferred the possible functions. Most notable among these proteins weregroups of proteins thought to trigger or mediate cells through specific phases of divisionof this alga, e.g., BP1, BP2, PB1, PB2, and PB3. Other proteins (e.g., group 1 proteins)thought to be responsible for maintaining and supporting cell concentration under adverseconditions were found. Furthermore, another group of proteins (group 2 proteins) thoughtto be stress-responsive were also detected. Taken overall, these differentially expressedproteins provided important information for uncovering various protective and adaptivemechanisms in the dinoflagellate’s life cycle. These proteins have the potential to serveas “indicator proteins” for rapid assessment of the nutritional or metabolic status ofthese phytoplankton cells,and monitoring the differential expression of these phase-spe-cific proteins and stress-specific proteins could be an important biomarker for bloom pre-diction.

Keywords: Cell cycle / Environmental stress / Growth phases / Harmful algal bloom (HAB) species / Two-dimensional gel electrophoresis

Received 12/8/03Revised 3/2/04Accepted 3/3/04

3214 Proteomics 2004, 4, 3214–3226

1 Introduction

Dinoflagellates are unicellular organisms that cause harm-ful algal blooms (HABs) and are one of the best studiedmicroalgal groups at the cellular level of organization.The harmful effects and the toxicity of many dinoflagellate

species have attracted the attention of many researchers.However, the mechanism of HAB generation is poorlyunderstood and HABs are considered to be unexpectedevents. Although many studies have found that HABoccurrences were correlated with an increase in nutrientload, relatively little is known about the complex bio-chemical and molecular interactions that brought aboutblooming. Proteins are the actual machinery that bringsabout cell growth, proliferation, and homeostasis. There-fore, it is simply logical that one should study proteins inorder to help uncover the mechanism underlying bloom-

Correspondence: Dr. Samuel C.-L. Lo, The Proteomic TaskForce, Department of Applied Biology and Chemical Technology,The Hong Kong Polytechnic University, HKSAR, People’s Repub-lic of ChinaE-mail: bcsamlo@inet.polyu.edu.hkFax: 1852-2364-9932

Abbreviations: HAB, harmful algal bloom; PB, preblooming;PMF, peptide mass fingerprint * These authors contributed equally.

2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.proteomics-journal.de

DOI 10.1002/pmic.200300838

Proteomics 2004, 4, 3214–3226 2-DE proteome reference maps for HAB bloom prediction 3215

ing. In the field of marine biology, 2-DE analysis has beenused in the characterization of the water-soluble proteinfraction from a commercial flat fish species [1] and thestudy of mutualistic associations of a sea anemone [2].We have described previously the optimized technicalprocedures for proteomic study of harmful algal blooms[3] and have reported that signature proteins could befound in 2-DE proteome reference maps of known algalspecies grown at optimal conditions in the laboratorywhich can serve as reference standards for species re-cognition [4].

Growth of dinoflagellate cells is limited primarily bymacronutrients [5] and light [6]. Only with the provision ofsurplus nutrients and adequate irradiance, the causativeorganisms can grow to the concentration necessary toproduce discoloration of the seawater (red tides). For abloom to begin, the “seed population” must be presentto act as an inoculum. Only a small number of dinoflagel-lates found in subtropical Hong Kong produce benthiccysts. Most of them usually develop directly from vegeta-tive cells under suboptimal conditions to form the “seedpopulation” and replicate rapidly under favorable condi-tions and are terminated by the depletion of nutrients.Therefore, we are mainly focusing on subtropical dinofla-gellates in the present study. The outbreaks of subtropicalHABs include three distinct phases: (i) initiation; (ii) devel-opment, and (iii) dissipation [7]. Nutrient limitation altersgene expression in many algae and bacteria [8] andchanges in irradiance and wavelength can induce a qual-itative change in components of the alga’s light-harvest-ing complexes as well as modulating the quantity of theconstituent polypeptides [9]. Based on the postulationthat blooming is manifested by changes in protein ex-pression, 2-DE patterns of a model dinoflagellate, Proro-centrum triestinum under different growth phases (i.e.,from initiation to dissipation) and under various cultureconditions (i.e., under light-, nitrogen-, phosphate-limitedbalanced growth) as well as under continuous illuminationwere examined. Results showed that P. triestinum dis-played extreme tolerance to environmental stresses andis able to survive in nutrient depleted conditions, undercontinuous illumination and continuous darkness. Thus,it provides a suitable model for uncovering various pro-tective and adaptive mechanisms important in the dino-flagellate’s life cycle. Through identification of those pro-teins of relevance and their involvement in various meta-bolic pathways, the molecular mechanisms involved inthe initiation and decline of algal blooms can be uncov-ered by correlating these changes with physiological andtoxicological studies. The ultimate goal of the presentstudy is to establish the biological background for the elu-cidation of the mechanism of P. triestinum blooms andtheir prediction.

2 Materials and methods

2.1 Materials

Unless stated otherwise, all chemicals were purchasedfrom Sigma (St. Louis, MO, USA). Standard 2-DE markerswere purchased from Bio-Rad (Hercules, CA, USA). Allsolvents were at least of AR-grade while most were ofHPLC-grade.

2.2 Cultivation of a seed population

Unialgal culture of the vegetative cells of P. triestinum,which was isolated from Hong Kong waters, was ob-tained from The University of Hong Kong Algal CultureCollection. Stock cultures of these cells were kept inexponential growth phase by transferring to a new medi-um every week. The “seed population” for laboratory-induced blooming was prepared by inoculating 200 mLof an exponentially growing culture into 10 L of f/2 culturemedium [10]. The cultures were grown at 207C under a16:8 h light: dark cycle at a light intensity of approximately100–150 mmol photons m22s21 provided by fluorescentlamps in a Conviron growth chamber (Model S10H; Con-viron Controlled Environments, Winnipeg, Canada) for14 days until the exponential growth phase was reached.These exponentially growing cells were then collected bycentrifugation at 10006g for 15 min at 227C (himac CR22f; Hitachi High-Speed Refrigerated Centrifuges, Japan)and the pellets were rinsed twice with sterilized seawaterto avoid any carry-over of nitrogen, phosphorus or inhibi-tors in the inoculum. The collected cells were then inocu-lated into either the normal f/2 medium or nitrogen-limitedf/2 medium or phosphorus-limited f/2 medium to achievean initial cell density of 46105 cells mL21 (seed popula-tion) for subsequent analysis.

2.3 Experimental conditions

2.3.1 Proteomic analysis of algae in differentgrowth phases

While cultured in normal f/2 medium, the growth of P. tri-estinum was analyzed during three distinct phases (initia-tion, development, and dissipation, respectively). Cul-tures representing the three phases were prepared asfollows: (i) The initiation culture was prepared by inoculat-ing the “seed population” into normal f/2 medium and itwas incubated under a normal dark/light cycle. The celldensities were constantly monitored until the equilibriumgrowth phase was reached. (ii) The development culture

3216 L. L. Chan et al. Proteomics 2004, 4, 3214–3226

was prepared by inoculating the initiation cultures intofresh normal f/2 medium and incubated at a normal dark/light cycle. The cell densities were constantly monitoreduntil exponential phase was reached with a growth rate(m) of 0.7 (see Section 3.1.1 for details of calculation).(iii) The dissipation culture was prepared by growing theinitiation cultures at a normal dark/light cycle until thedissipation phase was reached.

2.3.2 Proteomic analysis of samples takenduring a 24 h cell cycle

Synchronization of the experimental cultures was achiev-ed by maintaining the cells in continuous darkness for48 h. The cells were then entrained to the same photope-riod regime of 8D/16L. Light was turned on eventuallyat the 56th hour and a sample was taken immediately.This sample was designated as T = 8 h sample since thetime of collection was at 8 a.m. P. triestinum grown underoptimal condition in the exponential growth phase wasanalyzed at six scheduled points (0 h, 4 h, 8 h, 14 h, 18 h,and 22 h, respectively) during a 24 h cell cycle. The 0 hand 4 h samples were collected in the dark while the 8 hsample was collected immediately when the light wasturned on. The other three samples (14 h, 18 h, and 22 h)were collected under light.

2.3.3 Proteomic analysis of protein expressionchanges to environmental stresses

P. triestinum cultures in light-excessive and light-, nitro-gen-, or phosphate-limited balanced growth were ana-lyzed by 2-DE. The cultures at different conditions wereprepared as follows: (i) The starting cultures were pre-pared by inoculating the “seed population” into normalf/2 medium and incubating them at a normal dark/lightcycle for at least two days until a steady state wasreached. The light-limited and light-excessive cultureswere prepared by incubating the starting cultures incontinuous darkness and continuous illumination re-spectively for 72 h. (ii) The nitrogen-limited and phos-phorus-limited cultures were prepared by inoculatingthe “seed population” into nitrogen-limited and phos-phate-limited f/2 medium, respectively, and the cultureswere incubated at a normal dark/light cycle. The celldensities were constantly monitored until the stationaryphase was reached. The purpose of using stationaryphase cultures was to ensure that all carry-over of nitro-gen and phosphorus in the inoculum have been usedup.

2.4 Preparation of extracts and proteindetermination

2.4.1 Cell disruption

Approximately 26106 P. trestinum cells were collectedby centrifugation at 10 0006g for 20 min at 227C (himacCR 22f; Hitachi High-Speed Refrigerated Centrifuges,Japan). Water-soluble proteins were isolated as pre-viously described [3]. Briefly, with a Microtip-probe soni-fier (Model 250; Branson Ultrasonics, Danbury, CT, USA),cells were lysed in 0.5 mL 40 mM prechilled (47C) Tris buf-fer at pH 8.7 containing 30 units of endonuclease (benzo-nase isolated from Serratia marcescens; Sigma E8263).Cell debris and unbroken cells were removed by centrifu-gation at 22 2206g for 15 min at 47C (Mikro 22R; Hettich,Germany). The supernatants were concentrated by ultra-filtration through an Amicon YM-3 membrane (Amicon,Bedford, MA, USA) following the manufacturer’s instruc-tions. The extracts were then applied to a Micro BioSpin 6Column (Bio-Rad) previously equilibrated with a Tris buf-fer (10 mM Tris-HCl, pH 7.4) containing 0.02% sodiumazide following the manufacturer’s instructions. Flowthrough from the column was collected.

2.4.2 Protein determination

As the salts and other interfering substances present inthe extracts interfered with the protein assay, proteinextracts must be desalted prior to the assay by the proce-dure described above. Protein quantification of non-ureacontaining samples was performed by the Bradford assay(Bio-Rad) [11].

2.5 2-DE

Exactly 20 mg of each sample was mixed with a rehydra-tion buffer before being loaded onto IPG strips of linearpH gradient of either 3–6 (Bio-Rad) or 3–10 (AmershamBiosciences, Hong Kong, China). Rehydration, isoelectricfocusing, and equilibration were performed as previouslydescribed [3]. Subsequently, SDS-PAGE was performedand proteins on the 2-DE gels were visualized by silverstaining. Treatment time for each step of silver stainingwas standardized and the intensities of staining of proteinspots in each gel were normalized with 2-DE standardmarkers before image analysis. Three 2-DE gels were per-formed for each condition. Unless stated otherwise, the2-DE gels shown were representative of the three gelsperformed.

2.6 MALDI-TOF-MS and N-terminal amino acidsequencing by Edman degradation

Protein spots were selected to determine the peptidemass fingerprinting by a MALDI-TOF-MS (Autoflex; Bru-ker Daltonics, Bremen, Germany) if they were consistentlyvisible in all samples from preblooming and blooming cul-tures. Protein spots of interest separated by 2-D PAGEwere picked before digested in gels according to themethod described by Shevchenko and coworkers [12].The digests were cleaned up with ZipTip (Millipore, Bos-ton, MA, USA) and subjected to analysis by MALDI-TOF-MS. Calibration of the instrument was performed withinternal standards, namely angiotensin, substance P,bombesin, trypsin autolysis fragment, and adrenocortico-tropic hormone with the respective monoisotopic massesof 1046.542 m/z, 1347.736 m/z, 1620.807 m/z, 2211.105m/z, and 2465.105 m/z. For each sample, spectra from150 shots at several different positions were combined togenerate a peptide mass fingerprint (PMF). PMF obtainedfor each protein of interest were searched against theNCBI nonredundant database using the search engineMASCOT available at http://www/matrixscience.com. Thesearch was limited with a mass tolerance of 6 0.2 Da.One missed cleavage per peptide was allowed andcysteines were assumed to be carbamidomethylatedwith acrylamide adducts and methionine in oxidizedform. The minimum number of peptides required for eachmatch was 4. Proteins unidentified after MALDI-TOF-MSwere further characterized by N-terminal sequencing. Pro-teins separated by 2-D PAGE were electrotransferred ontoPVDF membranes as previously described [13]. The PVDFmembrane-bound proteins were visualized by stainingwith 0.1% Coomassie Brilliant Blue R-250 in 50% aqueousmethanol for 2 min, and destained in 40% methanol and10% acetic acid. Selected protein spots were excised andsubjected to N-terminal amino acid sequencing using aProcise 492 cLC Model 610A Protein sequencer (AppliedBiosystems, Hong Kong, China). Amino acid sequencesobtained were searched against either the Protein DataBank (PDB) or Swiss-Prot by BLAST. Settings for query-ing short sequences for nearly exact matches of peptidewere used.

3 Results

3.1 Proteomic analysis of samples taken duringdifferent growth phases

3.1.1 Population dynamics

Starting from a cell density of 46104 cells mL21, the initia-tion culture grew rapidly after the first day and develop-ed into a bloom in the new medium after the second

Figure 1. Population growth of the “initiation culture” ofP. triestinum over a 7 day experimental period after inocu-lating into fresh normal f/2 medium.

day. Cell number reached its maximum at the 4th dayand then dropped sharply after depletion of nutrients(Fig. 1). The lag phase lasted only for one day (the 1st

day), and an exponential growth phase started at thesecond day with the fastest growth rate (m), i.e., 0.7,and shortest doubling time, i.e., 0.99 day21 (Table 1).Population growth then slowed down and shifted mark-edly to the onset of the dissipation phase at the 4th dayof initiation phase and lasted for 14 days. The maximumcarrying capacity was approximately 1.76105 cellsmL21.

Table 1. Growth rate (m) (in doublings day21) of the “initia-tion culture” of P. triestinum over a 7 day experi-mental period after inoculating into fresh normalf/2 medium

Day C1 C0 m D.T. (day)

1 44 125 34 100 0.26 2.682 88 600 44 125 0.70 0.993 122 200 88 600 0.32 2.164 169 200 122 200 0.33 2.135 161 400 169 200 20.0476 136 800 161 400 20.1657 100 000 136 800 20.313

C0 and C1 indicate the cell numbers at time T0 and T1,respectively; D.T., doubling time in days.

The growth rate (m) was estimated from cell counts in theexponential phase using the formulation of Guillard [14]:

m ¼ (lnC1 � lnC0)/T1 � T0) (1)

where C0 and C1 indicate the cell numbers at time T0 andT1, respectively.

The doubling time (D.T.), in days, was calculated using thefollowing equation:

D.T. ¼ ln2/m

3.1.2 Differential protein expression patternsof P. triestinum during different growthphases

Vegetative cells of P. triestinum during different growthphases (initiation, development, and dissipation) wereharvested and analyzed by 2-D PAGE. The 2-DE proteinprofiles shared a majority of common protein spots

Figure 2. 2-DE protein profiles of 20 mg soluble proteins of P. triestinum extracted with 40 mM Trisbase from: (A) initiation culture at day 0; (B) blooming culture at day 2; (C); stationary culture at day 4;(D) late stationary culture at day 7. The IEF of the first dimension was over a pH range of 3.0 to 10.0.Boxes in (B) show the protein clusters which were clearly resolved when the IEFs were run acrossnarrow pH IPG strips (i.e., pH 4.5–5.5 for BP1 and BP2, pH 5.0–6.0 for PB3 and pH 5.5–6.7 for PB1and PB2) as shown in (E). The second dimension was a SDS-PAGE in a 15.0% polyacrylamide gel.

(Fig. 2). However, given that equal amounts of proteinswere loaded onto each gel and the intensities of silverstaining were normalized with standard 2-DE markers,several groups of differentially expressed preblooming(PB) proteins were found. PB1, PB2, and PB3, with anapparent molecular mass of 21 kDa and pI ranging from5.0 to 6.0, were barely visible in the initiation and station-ary stages (Figs. 2A, D) but they were greatly enhanced inthe blooming stage (Fig. 2B). Another two series ofblooming proteins (BP), BP1 (with an apparent molecularmass (MW) of 55 kDa and pI about 5.3) and BP2 (with anapparent MW of 21 kDa and pI 4.8) were found in theblooming stage only (Fig. 2B). These groups of proteinscould only be clearly resolved when the IEF were runusing IPG strips of very narrow pH range (Fig. 2E).Furthermore, in the late stationary stage, P. triestinumexhibited a markedly different protein expression pattern(with fewer protein spots mainly accumulated in the acidicregion with pI ranging from 3.0 to 5.0 and apparent molec-ular mass ranging from 31 to 76 kDa) when comparedwith that of the other stages.

3.2 Effect of environmental stresses on theprotein expression of P. triestinum

With the hypothesis that global protein expression ofP. triestinum will change with environmental stress, acomparison was made between the 2-DE protein profilesof P. triestinum balanced growth cultures either grownunder nitrogen-, phosphate-, and light-limited conditionsor under continuous illumination (Fig. 3C–F), with theinitiation (Fig. 3A) and blooming culture (Fig. 3B). Despitethe fact that the 2-DE protein profiles of cultures grownunder different stress conditions revealed minor differ-ences amongst each other, significant changes in theirprotein profiles were observed when compared with theblooming culture. Two groups of proteins exhibiting differ-ent response patterns were observed. With the differenttypes of stress exerted, group 1 proteins (with pI rangingfrom 4.0 to 5.0 and apparent molecular mass between31 and 66.2 kDa) showed a dramatic increase in relativeabundance under stress (Figs. 3C–F). Using IPG stripsof a more narrow range, pH 3–6, these proteins could bebetter resolved (Figs. 5E, F). The expression of anothergroup of proteins, group 2 proteins (with pI ranging from5.0 to 6.0 and apparent molecular mass ranging between31 and 43 kDa) clearly increased from the initiation stage(Fig. 3A) to the blooming stage (Fig. 3B) together with adecline in the expression of group 1 proteins. However,this group of proteins declined in abundance and wasbarely visible under these given environmental stresses.It should be emphasized that, among those abundantproteins unique in the vegetative periods, PB1, PB2, and

PB3 were still expressed at high abundance and re-mained constant under these suboptimal conditions.On the other hand, expression level of BP1 proteins wasvery low. The other protein, BP2, was only found in the2-DE gels of the nitrogen-limited balanced growth culture(Fig. 3C).

3.3 Proteomic analysis of samples taken duringa 24 h cell cycle

Proteomic analysis of cultures grown under optimal con-dition in the exponential growth phase was conducted atsix scheduled points (0 h, 4 h, 8 h, 14 h, 18 h, and 22 h,respectively) during the 24 h cell cycle with a 8D/16Lphotoperiod cycle (Fig. 4). Our results showed that thecell density increased slowly in the 8 h of darkness. Therewas a slight increase of cell numbers in the first 6 h afterthe onset of the light phase (i.e., up to the 14th hour).Thereafter, the cell density increased rapidly within thenext 2 h (i.e., within the 14th and 16th hour) and thenremained constant for the rest of the light phase. Further,it can bee seen from the growth curve (which is in lnscale), that cell counts of P. triestinum increased from9.26 to about 10 during the 24 h cell cycle and that corre-sponds to a twofold increase in cell number. It can also beseen that cell division was completed in the light phaseand the growth rate of P. triestinum is one doubling/day.

3.3.2 Differential protein expression patterns ofP. triestinum during a 24 h cell cycle

Differential changes in the proteome during a 24 h cellcycle of the exponential growth culture of P. triestinumwere investigated (Fig. 5). Although the general proteinpatterns are quite consistent, there were significant fluc-tuations in the relative abundance of the BP1 group ofproteins. When we tried to correlate the expression ofBP1 proteins with the different growth phases, it wasfound that BP1 were differentially expressed in three dis-tinct periods. The expression level of BP1 was found toincrease significantly in the first 6 h of the light period(from T = 8th h to T = 14th h). Maximal levels of BP1 expres-sion were seen in the 18th hour samples. Thereafter, itsexpression declined rapidly towards the end of the lightphase (from T = 22nd h to T = 0 h). From the results inSection 3.3.1 above, it is evident that cytokinesis ofP. triestinum took place in the middle of the light phase(from T = 14th h to T = 16th h, Fig. 4). Hence, it is temptingto hypothesize that BP1 expression is related to cyto-kinesis. Further studies are required to investigate theroles of the BP1 group of proteins.

Figure 3. 2-DE protein profiles of 20 mg soluble proteins of P. triestinum extracted with 40 mM Trisbase from: (A) initiation culture; (B) blooming culture; (C) nitrogen-limited balanced growth culture;(D) phosphate-limited balanced growth culture; (E) 72nd h of darkness for the light-starved culture;(F) 72nd h of continuous illumination culture. The two groups of proteins, group 1 and group 2, whichexhibit reverse patterns under stresses are circled. The IEF of the first dimension was over a pHrange of 3.0 to 10.0. Second dimension, SDS-PAGE in a 15.0% polyacrylamide gel.

Figure 4. Diurnal changes in cell density of P. triestinumin the exponential growth phase during a 24 h cell cycle,given a photoperiod of 8D/16L. The line over the hoursindicates the dark period.

3.4 Protein identification by MALDI-TOF-MSand N-terminal amino acid sequencing byEdman degradation

Attempts to identify groups of protein spots of interest(four for PB1, six for PB2, four for PB3, three for PB 2,and four for PB1; Fig. 2E) were carried out usingMALDI-TOF-MS. After tryptic digestion, PMFs of eachprotein of interest were obtained. Results revealed thatmultiple protein spots of the same group are isoforms of

the same proteins as they have nearly identical PMFs(data not shown). Further, the PMFs of PB1, PB2, andPB3 were highly similar, showing almost identical pep-tide masses and relative peak intensities. The massspectrum of the peptide tryptic digest of PB1 was shownas a representative spectrum for these three proteins(Fig. 6A). The mass spectra of peptide tryptic digest ofBP1 and BP2 are shown in Figs. 6B and C, respectively.Bioinformatic searches aiming to identify these proteinsusing the PMFs obtained were not successful. Hence,further attempts to identify these protein spots were per-formed using N-terminal amino acid sequencing. After2-DE of samples containing these proteins, they wereelectroblotted onto PVDF membranes before stainingwith Coomassie Brilliant Blue R-250. The protein spotswere carefully excised and loaded into the amino acidsequencer. The NH2-terminal amino acid sequences ofPB1, PB2, PB3, and BP2 are shown in Table 2 whileBP1 is N-terminal-blocked. However, bioinformaticsearch on these partial amino acid sequences returnedno similarity to any protein known in the existing proteindatabases. Accession numbers were assigned to thesenovel proteins on 16th February 2004 by the Swiss-ProtProtein Knowledgebase (Table 2). These data will bemade available under the indicated accession numberin Swiss-Prot/TrEMBL. It should be noted that attemptsto identify BP1 using post-source decay in the MALDI-TOF setup were not successful. Attempts are currentlyin progress to obtain a sufficient enough amount of BP1for de novo sequencing.

Table 2. N-Terminal amino acid sequences of preblooming proteins PB1, PB2, PB3 and bloomingproteins, BP1 and BP2

Accession No. inSwiss-Prot/TrEMBL

Proteinfull name

Proteinabbreviation

N-Terminal sequence

P83764 Prebloomingprotein 1

PB1 AEYDVSDADIEAFYQ

PB2 AEYDVSDADIEAFYQXXTMTWa)

PB3 AEYDSDADIEAFYQ

Nil Blooming-relatedprotein 1

BP1 N-Terminal block

P83767 Blooming-relatedprotein 2

BP2 VSAEYLERQGPKDDXDCFDDa)

a) X, ambiguities in the sequence

Figure 5. Diurnal changes in protein expression of P. triestinum in the exponential growth phaseduring a 24 h cell cycle. (A) T = 8 h, (B) T = 14 h, (C) T = 18 h, (D) T = 22 h (dark period = 0:00–8:00;light period = 8:00–24:00). The IEF of the first dimension was run over a pH range of 3.0 to 10.0.Circles in (A) and (B) showed the regions that were clearly resolved when the IEF was run acrossthe narrow pH range of 3.0 to 6.0, as shown in (E) and (F), respectively. Second dimension, SDS-PAGE in a 15.0% polyacrylamide gel.

Figure 6. MALDI-TOF PMFs of (A) protein cluster PB1;(B) protein cluster BP1, and (C) protein cluster BP2.

4 Discussion

4.1 Proteomic analysis during different growthphases

In natural waters, P. triestinum can survive under sub-optimal conditions to form a “seed population”, whichcan replicate quickly when a suitable temperature occursand when the seawater is enriched with suitable nutrients.Such blooms normally last for one week and are dissi-pated by the 14th day. In our experiment, the initiationculture was prepared by harvesting the “seed population”after it reached the stationary phase. These cultures wereused to represent the bloom initiation stages. With thereplenishment of nutrients by a new supply, the initiationculture grew rapidly into a bloom after two days butdropped significantly after depletion of the nutrients. Pop-ulation dynamics found in this present study using labora-tory cultures concurred with that observed in dinoflagel-late blooms in natural waters. Furthermore, P. triestinumexhibits a specific regularity in cell replication, which hasbeen termed phased division [15]. According to Costas etal. [16], cytokinesis of this species took place at the

beginning of the light phase. However, our results inFig. 4 showed that cell division occurred predominantlyin the middle of the light period and was completed onthe 18th hour. These discrepancies in division could beexplained if the circadian chronotype varied betweendifferent strains of the same species [17].

4.1.2 Differential protein expression patternsof P. triestinum during different growthphases and cell cycles

The commonly accepted model to describe changes inpopulation density in dinoflagellate cells assumes thatcell division is rapid during exponential growth phases,but slows down dramatically when the population ages.According to Mitchison [18], the cell cycle is composedof two parallel cycles: (i) the DNA cycle which encom-passes genome replication and nuclear division, and(ii) the growth cycle which includes the macromolecularsynthesis necessary to ensure the doubling of all cellcomponents between two divisions. The initiation cul-tures in our experiment were prepared by collecting the“seed population” after it reached the stationary phase.It replicated rapidly when transferred into fresh f/2 medi-um and a very short lag period was observed from initia-tion phase to development phase (Fig. 1). This showedthat no adaptive phase was required to “wake up” thesecells to divide again. From analyses of the relationshipsbetween differential protein expressions during differentgrowth phases, some phase-specific proteins were de-tected. PB1, PB2, and PB3 were prominently found inthe vegetative stages (Figs. 2A–C) but BP1 and BP2were expressed at high abundance and greatly enhancedonly in the blooming stages (Fig. 2B). This agreed withthe hypothesis of Alberghina et al. [19] that the growthcycle controls the DNA cycle and this control is mediatedthrough the attainment of a level of protein necessary forDNA replication to begin. A drastic reduction in abun-dance of these phase-specific proteins (i.e., PB1, PB2,PB3, BP1, and BP2) and some other cellular proteins(i.e., group 2 proteins), in pI ranging from 5.0 to 6.0 andapparent molecular masses ranging between 31 and43 kDa, were detected in the dissipation culture (Fig. 2D).It appeared that this phenomena of “cellular aging” indinoflagellates (as they are in the dissipation phase) re-sulted in a global decrease in protein content. This sug-gested that these phase-specific proteins are probablytransient proteins expressed at high abundance duringthe course of progression across the cell cycle in nutri-ent-replete, steady state growth conditions. It is well-known that by replenishment of nutrients with a new sup-ply, aged cells require more time for metabolic adapta-tion, and that the resumption of cell division and popula-

tion growth are normally delayed in comparison to thosein nutrient replete, steady-state growth conditions (i.e.,the initiation culture). It appears that time is required formacromolecular synthesis in these aged cells to regener-ate all the proteins necessary for cell division, supportingand maintaining cell concentration. Our results suggestthat these proteins (PB1, PB2, PB3, BP1, and BP2) mighthave functional roles in modulating cell growth and cellproliferation rather than a response to changing environ-mental factors.

In many cell types, a restriction point (or transition point,or “start” point) for cell division has been identified.Beyond this restriction point, the cell is committed to gothrough the DNA cycle to mitosis and cytokinesis [20].The determining factor for attaining this restriction pointwas suggested to be size-related (critical cell mass orprotein content) [21] or to be a function of a specific groupof proteins [22]. This group of proteins is structurally simi-lar and is called “cyclin”. They are the ultimate and func-tional operators of the cell cycle checkpoints. Cyclinlevels oscillate with specific phases of the cell cycle[23]. A detailed analysis of the proteomic changes ofP. triestinum in the exponential growth phase during a24 h cell cycle revealed that most of the proteinsexpressed remained constant throughout the cell cycle.However, levels of BP1 varied in accordance with thecell cycle (Figs. 4, 5). BP1 was upregulated during theonset of mitosis and downregulated after mitotic activi-ties were completed. As its periodic up- and downregu-lations paralleled the diurnal phased division of thetested species, they are speculated to act as “specifictrigger proteins” that regulate cell cycle progression inthis tested alga.

4.2 Effect of environmental stresses on thedifferential protein expression profiles ofP. triestinum

The effects of nutrient and light deprivation as well asprolonged illumination on the expression of phase-specif-ic proteins were examined (Figs. 3C–F). Differential pro-tein expression patterns in samples cultured under thesesuboptimal conditions were found to be very similar tothose in the initiation culture (Fig. 3A) and this showedthat growth of dinoflagellate cells was halted either bymacronutrients depletion or light deprivation. But drasticchanges in the relative abundance of several groups ofabundant proteins were observed when compared withthe blooming stage (Fig. 3B). The constancy of proteinprofiles of P. triestinum under these different suboptimalconditions could possibly be used as a potential“biomarker” to indicate these preblooming or suboptimalstages. The protein profile of the nitrogen-limited culture

was quite similar to that of the blooming stage and BP2 ishighly expressed and unique in the nitrogen-limited con-dition. This suggested that the nitrogen-limited culturepossesses a higher potential for growth and P. triestinumcontains enough N from intracellular pools, i.e., remobili-zation of nitrogen from degraded organelles or macromo-lecules, that can be utilized for cell maintenance and sup-port. Therefore, phosphorus and light become the primegrowth-limiting factors. It is this result which might wellexplain why dinoflagellates grow best at lower N:P ratios,where the amounts of phosphorus available are relativelyhigh. This finding is consistent with bottle bioassay, whichshowed that P. triestinum is primarily limited by phos-phorus and light [24]. The differential accumulation of pro-teins under different conditions could be attributed to dif-ferences in the availability of nutrients, irradiance, andlight quality and/or reflects the production of specificsignals that control the cell division state or adaptationsto nutritional and environmental conditions. Olson andChisholm [25] postulated that the cell cycle might be con-trolled directly by the photocycle at specific “transitionpoints” that were crossed when a threshold value of cellmass, protein or RNA content, or specific “target protein”were achieved. It appeared that these differentially ex-pressed proteins have their specific functions during dis-tinct periods of the cell cycle. PB1, PB2, and PB3 are un-likely to be the “trigger proteins” but rather act as the“mediators” of cell division since they were constantlyexpressed in high abundance under these suboptimalconditions. BP2, on the other hand, was neither foundin phosphorus-limited nor light-stressed cells. This sug-gested that the synthesis of BP2 might require the pres-ence of PO4

32 under optimal illumination (mediatedthrough photosynthetic pathways). Therefore, both pro-longed sunshine and darkness could inhibit this speciesfrom initiating a bloom. Based on our observations, it isspeculated that BP2 is one of the “checkpoint” proteinsthat ensures the cells, prior to cell division, hold enoughinternal P stores to successfully complete DNA replicationor for synthesis of the major cellular components requiredfor division.

Another two groups of proteins with different responsepatterns were also noticed in response to these limita-tions. Group 1 increased in expression as the alga accli-mated in the environment while group 2 exhibited thereverse pattern. The synthesis of the group 1 proteinsseems not to be dependent on light or an immediatesource of common macronutrients, i.e., nitrate and phos-phorus (mediated through a nonphotosynthetic pathway).We hypothesize that group 1 proteins might be involvedin scavenging limiting resources, elevating the efficiencywith which those resources are used, or are involved inaltering key metabolic pathways of the cell (e.g., produc-

tion of glycerol from fixed carbon). Therefore, retentionand enhanced expression of certain proteins can beexpected as the stresses progress. For example, pro-teases and enzymes involved in degradation of macro-molecules and remobilization of their components will beenhanced. Mitochondrial proteins will be conserved sinceretention of mitochondrial function is expected to providerespiratory activity in the stressed alga. Most of these pro-teins were probably mainly used for supporting and main-taining the cell concentration and integrity under adverseconditions. Their expression might show apparent in-creases in relative abundance when the amounts of in-dividual proteins are expressed as a fraction of the quan-tified total. On the contrary, the synthesis of group 2proteins was dependent on both light and commonmacronutrients (mediated through photosynthetic path-ways) and they were speculated to be cytosolic proteins,i.e., regulatory proteins including transcription factorsand protein kinases or phosphatases involved in signal-ing pathways. This is because these proteins wereexpressed at high abundance during the stage in whichcell division is rapid and under conditions with a suffi-cient supply of nutrients and light irradiance. Therefore,they were probably responsible for cell proliferation.Degradation of these proteins during stress may providea source of carbon, nitrogen and amino acids for reusesince organelles and macromolecules constitute a largeproportion of cellular proteins. Downregulation of theseproteins during stress might help to protect this organ-ism by switching from a state devoted to cell divisionand growth, to a state of rapid starch synthesis andaccumulation.

The relative abundance of these phase-specific and/orstress-responsive proteins assumed great interest in ourstudy because of their potential role in either directly orindirectly controlling bloom decline or initiation. Monitor-ing the differential expression of these phase-specificproteins could be an important biomarker for bloom pre-diction.

4.3 Protein identification by MALDI-TOF-MSand N-terminal amino acid sequencing byEdman degradation

Identification was limited by the paucity of sequence dataavailable for dinoflagellates. Searching using PMFs ob-tained against NCBI non-reductant database, Swiss-Protand TrEMBL databases, and partial amino sequencesagainst the Protein DataBank (PDB) and Swiss-Prot byBLAST revealed no similar protein in the database. How-ever, close matching of the PMFs of PB1, PB2, and PB3indicated that they are isoforms of the same protein (data

not shown). Their differences in pI could be accounted foreither because of artifacts generated or because of post-translational modification of the proteins, which couldresult in a variation in the relative amounts of acidic tobasic amino acids. In the present study, all five groups ofproteins were protein clusters containing multiple spotsand one protein seems to have three isoforms. All of theseproteins identified are novel ones. Since there is littlegenomic sequence data currently available for dinoflagel-lates for sequence comparison and alignment purposes,identification of these proteins has not progressed veryfar. Nevertheless, using these N-terminal sequence data,degenerate PCR primers should be designed for amplifi-cation of the genes. This would allow confirmatory experi-ments to be performed to validate the functional roles ofthese novel proteins. However, these experiments areoutside the scope of this paper.

In conclusion our proteomic study of the differential pro-tein profiles under different environmental conditionsfound that P. triestinum showed both constancy of pro-tein profiles and variability of the relative abundance ofsome abundant proteins under different environmentalconditions and culture conditions. The differentially ex-pressed protein patterns of P. triestinum under differentgrowth conditions showed conclusively that P. triestinumgrown under nitrogen-limited conditions possessed highergrowth potential since its protein expression profiles aremore similar to that of the blooming stage. Our resultsare consistent with the observations that it is possible torapidly trigger a bloom by the addition of a supply ofnitrates. A more detailed analysis of the differential syn-thesis of proteins in these various stages and studies toelucidate the functions of these proteins will be criticalfor understanding the molecular mechanisms involved inblooming. Monitoring the differential expression of thesephase-specific and stress-specific proteins could be animportant biomarker for bloom prediction and can beused to rapidly assess the nutritional or metabolic statusof these phytoplankton cells.

This work is partially supported by a research grant fromthe Dean’s reserve (FAST) of the Hong Kong PolytechnicUniversity (Account No. 875D) awarded to S.C.L. Theauthors would also like to express their gratitude to KouHing Hong Scientific Supplies Ltd., Hong Kong, for mak-ing a private donation to the department in supporting thiswork. Chan L.L. is a PhD candidate of the Dept. of AppliedBiology and Chemical Technology, The Hong Kong Poly-technic University (A/C: G-V986). This paper is dedicatedto John Hodgkiss, who spent much of the past 25 yearsworking on the interactive dynamics between eutrophica-tion (particularly N:P ratios) and the formation mechan-isms of HABs.

5 References

[1] Pineiro, C., Barros-Velazquez, J., Perez-Martin, R. I. et al.,Electrophoresis 1999, 1425–1432.

[2] Weis, V. M., Levine, R. P., 1996, J. Exp. Biol. 1996, 199, 883–892.

[3] Chan, L. L., Lo, S. C. L., Hodgkiss, I. J., Proteomics 2002, 2,1169–1186.

[4] Chan, L. L., Hodgkiss, I. J., Lo, S. C. L., Proteomics 2004, 4,180–192.

[5] Ho, K. C., Hodgkiss, I. J., Asian Marine Biol. 1993, 10, 77–94.[6] Spudich, J. L., Sager, R., J. Cell Biol. 1980, 85, 136–145.[7] Ho, K. C., Hodgkiss, I. J., Asian Marine Biol. 1991, 8, 5–23.[8] Dumont, F., Loppes, R., Kremers, P., Planta 1990, 182, 610–

616.[9] Jovine, R. V. M., Triplett, E. L., Nelson, N. B., Prézelin, B. B.,

Plant Cell Physiol. 1992, 33, 733–741.[10] Guillard, R. R. L., in: Stein, J. R. (Ed.), Handbook of

Phycological Methods, Cambridge University Press, Cam-bridge 1975, pp. 289–312.

[11] Bradford, M. M., Anal. Biochem. 1976, 72, 248–254.[12] Shevchenko, A. et al., Anal. Chem. 1996, 68, 850–858.[13] Butt, Y. K. C., Lum, J. H. K., Lo, S. C. L., Planta 2003, 216,

762–771.

[14] Guillard, R. R. L., in: Smith, W. L., Chanley, M. H. (Eds.), Cul-ture of Marine Invertebrate Animals, Plenum Press, NewYork 1995, pp. 181–194.

[15] Hastings, J. W., Sweeney, B. M., in: Zeuthen, E., (Ed.), Syn-chrony in Cell Division and Growth, Interscience Publication,New York 1964, pp. 307–321.

[16] Costas, E., González, S., Aguilera, A., López-Rodas, V.,Botanica Marina 1992, 35, 429–436.

[17] Partensky, F., Vaulot, D., Videau, C., J. Phycol. 1991, 27,733–741.

[18] Mitchison, J. M., The Biology of the Cell Cycle, CambridgeUniversity Press, Cambridge 1971.

[19] Alberghina, L., Sturani, E., Microbiol. Rev. 1981, 45, 99–122.

[20] Yanishevshy, R. M., Stein, G. H., Int. Rev. Cytol. 1981, 69,223–259.

[21] Fantes, P. A., J. Cell Sci. 1977, 24, 51–67.[22] Rossow, P. W., Riddle, V. G. H., Pardee, A. B., Proc. Natl.

Acad. Sci. USA 1979, 76, 4446–4450.

[23] Darzynkiewicz, Z., Gong. J., Juan, G., Ardelt, B., Traganos, F.,Cytometry 1996, 25, 1–13.

[24] Ho, K. C., Hodgkiss, I. J., Proc. 2nd Int. Conf. on the MarineBiology of the South China Sea, China 1993, pp. 111–118.

[25] Olson, R. J., Chisholm, S. W., J. Plankton Res. 1986, 8, 785–793.

Proteomic study of a model causative agent of harmful red tide, Prorocentrum triestinum I:...

Documents

Somali piracy problems and the causative factors

Interested consumers' awareness of harmful chemicals in

Perkinsosis in the Manila clam Ruditapes philippinarum affects responses to the harmful-alga, Prorocentrum minimum

Collective Detection of Potentially Harmful Requests Directed at Web Sites

workplace health promotion at rhodes university: harmful use of

HARMFUL CULTURAL AND TRADITIONAL PRACTICES

Are Prorocentrum hoffmannianum and Prorocentrum belizeanum (DINOPHYCEAE, PROROCENTRALES), the same species? An integration of morphological and molecular data

Children & Young People Harmful Sexual Behaviour (HSB

Harmful sexual behaviour in schools: a briefing on the findings

Harmful Algae

The ecology of harmful diatoms

Pathology and immune response of the blue mussel (Mytilus edulis L.) after an exposure to the harmful dinoflagellate Prorocentrum minimum

Addressing Harmful Bias and Eliminating Discrimination in

Alcohol consumption and harmful drinking: Trends and social

Anthropogenic nutrients and harmful algae in coastal waters

Blue cone monochromacy: Causative mutations and associated phenotypes

Spontaneous hip fractures in fluoride-treated patients: Potential causative factors

On the Causative/Anti-Causative Alternation as Principle of Affix Ordering in the Light of the Mirror Principle, the Lexical Integrity Principle and the Distributed Morphology

The Role of Allelopathy for Harmful Algae Bloom Formation

A preliminary analysis of causative verbs in English* - UFMG