Global spectral-kinetic analysis of room temperature ... Publications...chlorinaf2(clof2). These...

Global spectral-kinetic analysis of roomtemperature chlorophyll a uorescence from

light-harvesting antenna mutants of barley

Adam M Gilmore1 Shigeru Itoh2 and Govindjee3

1Photobioenergetics Group Research School of Biological Sciences Australian National University Canberra ACT 0200 Australia2Department of Cell Biology National Institute for Basic Biology Okazaki 444 Japan

3Center for Biophysics and Computational Biology University of Illinois at Urbana-Champaign Urbana IL 61801 USA

This study presents a novel measurement and simulation of the time-resolved room temperature chloro-phyll a pounduorescence emission spectra from leaves of the barley wild-type and chlorophyll-b-decentcientchlorina (clo) f2 and f104 mutants The primary data were collected with a streak-camera-basedpicosecond-pulsed pounduorometer that simultaneously records the spectral distribution and time dependenceof the pounduorescence decay A new global spectral-kinetic analysis programme method termed the doubleconvolution integral (DCI) method was developed to convolve the exciting laser pulse shape with amultimodal-distributed decay procentle function that is again convolved with the spectral emission bandamplitude functions We report several key results obtained by the simultaneous spectral-kineticacquisition and DCI methods First under conditions of dark-level pounduorescence when photosystem II(PS II) photochemistry is at a maximum at room temperature both the clo f 2 and clo f104 mutantsexhibit very similar PS II spectral-decay contours as the wild-type (wt) with the main band centredaround 685 nm Second dark-level pounduorescence is strongly inpounduenced beyond 700 nm by broad emissionbands from PS I and its associated antennae proteins which exhibit much more rapid decay kinetics andstrong integrated amplitudes In particular a 705^720 nm band is present in all three samples with a710 nm band predominating in the clo f2 leaves When the PS II photochemistry becomes inhibitedmaximizing the pounduorescence yield both the clo f104 mutant and the wt exhibit lifetime increases for theirmajor distribution modes from the minimal 250^500ps range to the maximal 1500^2500 ps range forboth the 685 nm and 740 nm bands The clo f 2 mutant however exhibits several unique spectral-kineticproperties attributed to its unique PS I antennae and thylakoid structure indicating changes in both PSII pounduorescence reabsorption and PS II to PS I energy transfer pathways compared to the wt and clo f104Photoprotective energy dissipation mediated by the xanthophyll cycle pigments and the PsbS protein wasuninhibited in the clo f104 mutant but as commonly reported in the literature signicentcantly inhibited inthe clo f2 the inhibited energy dissipation is partly attributed to its thylakoid structure and PS II to PS Ienergy transfer properties It is concluded that it is imperative with steady-state pounduorometers especiallyfor in vivo studies of PS II ecurrenciency or photoprotective energy dissipation to quantify the inpounduence of thePS I spectral emission

Keywords deconvolution pounduorescence lifetime distributions L1 robust minimization methodphotosystem II streak camera

1 INTRODUCTION

Measurements of chlorophyll a pounduorescence especiallyfrom photosystem II (PS II) are routinely used to gaugethe status and ecurrenciency of the photosynthetic electrontransport system (Baker 1996) They provide a practicalnon-intrusive measure of the various competing de-excitation pathways for the absorbed light energy namelyphotochemistry non-radiative or thermal dissipation andpounduorescence (Singhal et al 1999) Dark-level pounduorescence

conditions usually called Fo conditions are decentned whenpounduorescence is elicited by light too weak to drive photo-chemistry PS II photochemistry is at a maximal rate andthe primary quinone electron acceptor in PS II isoxidized Fo conditions also assume there is no thylakoidmembrane energization (see for example Van Kooten ampSnel 1990 Govindjee 1995 Gilmore amp Govindjee 1999)Measuring the dark-level pounduorescence conditions isimportant to understand the photobiophysical light-harvesting process because Fo indicates maximal photo-chemical dissipation by the PS II reaction centre in thepresence of minimal competition from other de-excitationpathways When measuring Fo it is important for tworeasons especially in vivo to be able to quantify any over-lapping spectral-kinetic signals from photosystem I

Phil Trans R Soc Lond B (2000) 355 1371^1384 1371 copy 2000 The Royal Society

doi 101098rstb20000699

Author for correspondence (gilmorersbsanueduau)Present address Department of Physics Laboratory of Photobioener-getics Graduate School of Science Nagoya University Chikusa-kuNagoya 4648601 Japan

(PS I) because (i) PS I exhibits much more rapid decaykinetics than PS II (Roelofs et al 1992 Schmuck amp Moya1994) and (ii) reabsorption of the shorter wavelengthpounduorescence from PS II decreases the amplitudes of thePS II bands relative to the longer wavelength PS I bandsin leaves (Govindjee amp Yang 1966 Weis 1985 PfIumlndel1998) This study reports both a novel in vivo measurementand a global time-resolved emission spectral analysismethod known as the `double convolution integralrsquo orDCI method as these are applied primarily to investigatethe so-called dark-level and maximal chlorophyll a pounduor-escence yields (Fm conditions) from both PS II and PS I atroom temperature

We analysed the in vivo lifetimes of chlorophyll apounduorescence from two well-characterized light-harvestingbarley mutants known as chlorina f104 (clo f104) andchlorina f2 (clo f 2) These mutants are distinguished byunique alterations in the content and composition of thelight-harvesting pigment^protein complexes associatedwith both PS II and PS I The chlorophyll-b-less clo f2mutant primarily lacks the major complement of Lhcb1and Lhcb6 proteins associated with PS II in addition tolacking the Lhca4 protein associated with PS I (Bossmanet al 1997 Knoetzel et al 1998) The light- andtemperature-sensitive clo f104 mutant lacks a largecomplement of Lhcb1 in addition to lacking the 23 kDLhca2 protein of PS I (Knoetzel amp Simpson 1991Bossman et al 1997) These mutants were chosen for thisstudy because they both exhibit strong variations in theirPS I chlorophyll emission spectral amplitudes at 77 K andsignicentcant diiexclerences in their room temperature steady-state pounduorescence behaviour (Simpson et al 1985 Bassiet al 1985) It was thus of interest to determine andcharacterize the probable relationships between the 77 Kspectral changes due to PS I and the room temperaturepounduorescence parameters More importantly it has alsobeen reported recently that both these mutants exhibitseveral PS-II-related activities that do not change asmight be predicted in conjunction with the loss of Lhcb1namely the light-limited quantum ecurrenciency andpounduorescence lifetimes of PS II and the well-knowntransthylakoid pH- and xanthophyll-cycle-dependentnon-photochemical quenching (Briantais et al 1996Gilmore et al 1996)

Here we report three main observations based on thesimultaneous spectral-kinetic acquisition and DCIanalysis of pounduorescence data collected using a streak-camera spectrograph (Schiller amp Alfano 1980 Tsuchiya1984 Davis amp Parigger 1992 BIumlhler et al 1998) Firstthat under conditions of dark-level pounduorescence both theclo f 2 and clo f104 mutants exhibit very similar PS IIspectral-decay contours to the wild-type (wt) with themain band centred around 685 nm Second that dark-level pounduorescence in the clo f 2 clo f104 and wt is stronglyinpounduenced beyond 700 nm by broad emission bandsfrom PS I that exhibit much more rapid decay kineticsand stronger integrated amplitudes than the main PS IIband at 685 nm Third that clo f 2 compared with clo f104and wt exhibits diiexclerent room temperature PS I emissionspectral bands unique Fo and Fm lifetime distributionsand an inhibited capacity for photoprotective thermalenergy dissipation We conclude that it is imperative forin vivo studies of PS II photochemistry and photo-

protective thermal energy dissipation using chlorophyll apounduorescence to quantify the inpounduence of the prominentPS-I-related spectral emission components

2 MATERIAL AND METHODS

(a) Plant material and experimental detailsSeeds of wt barley (Hordeum vulgare L) and the nuclear gene

mutants clo f 2 and clo f104 were obtained from Professor DSimpson of the Carlsberg Research Laboratories (CopenhagenValby Denmark) Plants were grown in a growth chamber at75 relative humidity under a day^night regime of 16 L 8 Dand 25 8C^20 8C The high (HL) medium (ML) and low (LL)light growth conditions were decentned as 250^270 130^160 and50^60 m molphotons m72 s71 respectively

The chlorophyll antenna sizes of PS II were determinedaccording to the oxygen poundash yield method of Chow et al (1991)The PS I content was obtained from P700 measurementsaccording to Schreiber et al (1988) Chlorophyll determinationsrelevant to the antenna size estimates were performed accordingto the method of Porra et al (1989) Quantitative analysis of thechloroplast pigments was according to the HPLC method ofGilmore amp Yamamoto (1991)

(b) Steady-state chlorophyll a pounduorescencemeasurements

Room temperature measurements of Fm and Fo were madewith a PAM 101-102-103 Model chlorophyll pounduorometer (Heinz-Walz Eiexcleltrich Germany) equipped with a standard 650 nmexcitation diode and RG-9 emission centlter that cuts oiexcl all lightbelow 690 nm and transmits ca 50 at 710 nm andca 90 4720 nm The PAM measuring beam settings andsaturating light pulse intensities for the Fm and Fo determina-tions were as used earlier (Gilmore et al 1996)

The 77 K spectra of intact leaf pieces were performed with anSLM 8100 spectropounduorimeter (Spectronic Institute RochesterNY USA) in photon counting mode Excitation was at 435 nm(peak chlorophyll a absorbance) through an 8 nm slit widthand emission was monitored with a slit width of 2 nm using amonochromator slew rate of 1nm s71 and integration time of 1sThe displayed spectra were centrst smoothed using the exponentialalgorithm by a factor of 04 then normalized to the peak emis-sion channel and centt by minimizing the sum of the squared resi-dual error (SSE) parameter

SSE ˆN

iˆ1

(Di iexcl Mi)2 (1)

where Di is the pounduorescence intensity at wavelength coordinatel i and Mi is the predicted intensity from a model comprisingthe sum of centve free-poundoating Gaussian bands Each Gaussianspectral band was decentned by the following amplitude function

I(l)i ˆ AMP pound exp iexcl12

l iexcl CTRWID

2

(2)

where AMP is the normalized maximum amplitude parameterCTR is the mean or central mode l parameter and WID thewidth is the standard deviation parameter in l units

(c) Time- and wavelength-resolved chlorophyll apounduorescence measurements

Time-resolved emission spectra were collected in photoncounting mode with a model C4334 Streakscope (Hamamatsu

1372 A M Gilmore and others Global spectral- kinetic analysis

PhilTrans R Soc Lond B (2000)

Photonics Hamamatsu City Japan) that was synchronized at afrequency of 1MHz with a pulsed (650 nm) laser diode excita-tion source (Model PLP-20 Hamamatsu Photonics) The sampleemission was dispersed with a Hamamatsu spectrograph using agrating of 50 groovesmm71 and a slit width of 150 m m Datawere collected using a 2 ns time window for the dark level and a10 ns window for the maximal level through a KC-17 centlter(4 630 nm) until 150 000 shots accumulated Trigger jitter wasestimatedby manufacturerrsquosspecicentcationsat 5 20 psThe charge-coupled device (CCD) camera image (640 pixelspound 480 pixels)was translated to an ASCII centle The integrated pulse procentle wasselected and sectioned separately from the sample image whichwas binned and integrated at a rate of centve pixel columns per(65) wavelength channel(s) and three pixel rows per (159) timechannel(s) prior to centtting the total of 10 335 data points Allimages were normalized to unity using the photoelectron countat the peak time wavelength (tl) channel coordinate Thetheory function and statistical analyses are decentned in Appen-dices A and B respectively

For the open PS II trap conditions primary or secondaryleaves from 7^10-day-old plants were dark adapted for 30 minthen vacuum incentltrated with distilled H2O and submersed in afront surface cuvette assembly The temperature-controlledsample compartment (20 8C) was kept dark and the illumina-tion by the pulsed laser diode was roughly one-half to one-centfthtoo low in intensity to cause spontaneous PS II trap closureStill data collection was interrupted every 25 000 shots tooxidize the PS II traps with 30 s of far-red illumination (WalzPAM 102 700 nm diode 5 10 m molphotons m72 s71) TheFmFo ratio was checked intermittently with the PAM inconjunction with the far-red illumination to ensure maximalsignal stability Measurement conditions for the closed PS IItrap conditions were the same except that a blue light (440 nmcentre 10 nm band-pass 100 m E m72 s71) was used to maintaintrap closure after vacuum incentltration of the leaves for 30 minwith 10 m M DCMU (3-(34-dichlorophenyl)-11-dimethylurea) Abackground steady-state pounduorescence image collected in theabsence of the pulsed laser excitation was collected and simplysubtracted from the image generated by the pulse in the presenceof the background illumination

3 RESULTS

(a) Photosystem II and I antenna sizes steady-statechlorophyll a pounduorescence xanthophylls

and PS II energy dissipationTable 1 shows that the total levels of chlorophyll (a + b)

per leaf area were reduced strongly in both the clo f 2(56 of wt) and the medium-light-grown clo f104 (44of wt) Furthermore the antenna size PS IIchl deter-mined by the oxygen poundash yield method (Chow et al

1991) decreased in the mutants to a similar degree as thechlorophyll per area (table 1) in clo f2 both were 58 ofthe wt whereas in clo f104 chlorophyll per area was 58of wt but PS IIchlorophyll was 45 no particularsignicentcance is assigned to this small diiexclerence Theamount of PS I estimated by the P700 chlorophyll absor-bance measurements in thylakoids (Schreiber et al 1988)appeared to increase slightly (136 of wt) in the clo f 2compared to the wt whereas it was slightly decreased toabout 93 of wt in the clo f104 Likewise the pattern ofthe PS IIPS I ratio (on a chlorophyll basis) was the sameas the P700 content

Figure 1 compares the changes in the photosystem chlantenna sizes (chlPS II) to the changes in the chl abratios (panel a) the levels of xanthophyll cycle de-epoxidation and energy dissipation (panel b) the roomtemperature chl a pounduorescence ratios of the maximal(Fm) and dark-level Fo pounduorescence (panel c) and thetwo main pools of xanthophylls namely lutein andviolaxanthin+ antheraxanthin+ zeaxanthin (panel d)Figure 1a compares the chlorophyll ab ratios in wt andclo f104 plants grown under three photon pounduxdensities namely LL (ca 50 m mol photons m72 s71)ML (ca 150 m mol photons m72 s71) and HL (ca250 m mol photons m72 s71) as well as the clo f 2 mutantplants grown under ML The clo f104 mutant chlorophyllab ratios responded sharply to the increasing photon pounduxdensity (PFD) reaching chlorophyll ab ratios 4 8 Thewt exhibited measurable but still much smaller changesthan the clo f104 in response to the PFD levels On theother hand the clo f2 mutant had virtually no chlorophyllb (at all the intensities data not shown) Figure 1c showsthe ratios of the pounduorescence FmFo yields under the samePFD conditions as in centgure 1a The clo f104 mutant hadthe largest (ca 8) FmFo at the medium and high PFDsbut this ratio was still high (ca 6) at the lowest PFD usedThe clo f 2 mutant exhibited a clearly diminished FmForatio (ca 4) compared with all the other wt andclo f104 sample conditions Figure 1b compares themaximal levels of xanthophyll cycle de-epoxidation asmeasured by DPS ˆ [zeaxanthin+ antheraxanthin][violaxanthin+ zeaxanthin+ antheraxanthin] and energydissipation as measured by the f2 fraction calculatedaccording to Gilmore et al (1998) in thylakoids extractedfrom wt plants grown under LL (solid upside-down trian-gles) to clo f104 thylakoids from LL (open upside-downtriangles) ML (open triangles) and HL (open squares)grown plants As the antenna size decreases the DPSclearly increases as does the level of energy dissipationcalculated as the f2 fraction However like the FmFo ratiosunder the high PFD there is a slight but signicentcant

Global spectral- kinetic analysis A M Gilmore and others 1373

Phil Trans R Soc Lond B (2000)

Table 1 Comparison of the antenna sizes of PS II and PS I in the leaves of wild-type (wt) clo f2 and clo f104 mutants of barleywith supporting analyses from isolated thylakoids a

typechlarea

( m molm7 2)PS IIchl

mmol mol chl71PS IIarea

( m mol m7 2) PS IIPS IabP700chl

(mmol mol chl71)

clo f 2 140 sect 3 (n ˆ 3) 552 sect 020 (n ˆ 4) 078 sect 004 (n ˆ 4) 236 234wt 248 sect 16 (n ˆ 3) 321 sect 002 (n ˆ 3) 080 sect 005 (n ˆ 3) 187 172clo f104 110 sect 4 (n ˆ 3) 548 sect 006 (n ˆ 5) 061 sect 002 (n ˆ 3) 342 160

b In the column of PSIIPSI data are expressedon a chl basis

downturn in the level of energy dissipation whencomparing the high to the medium PFD conditionsSince suitably active thylakoids were not recoveredfrom the clo f2 in this work we simply cite andconcentrm other literature that documents a clear (20^40) decrease in the maximum levels of energydissipation in the clo f2 mutant and other chlorophyll-b-less mutant leaves grown under constant butmoderate PFDs (Briantais 1994 Falk et al 1994 HIgravertelamp Lokstein 1995) Figure 1d documents the pool sizes(mol (PS II)iexcl1) of the xanthophyll cycle pigments(VAZ ˆ violaxanthin + antheraxanthin+ zeaxanthin) andlutein (L) in wt and clo f104 plants grown under a range ofPFDs generally inclusive of the high and low PFDs decentnedin the other panels of centgure 1 It is clear the loss of chloro-phyll b and decreasing PS II antenna size correlates with aloss of two-thirds of the maximum pool of L Howeverconsistent with the signicentcant levels of xanthophyll-cycle-dependent energy dissipation in centgure 1b the levels ofVAZ remained almost invariant over the range of mea-sured growth PFDs Elimination of the major pool of Lhcb1in the clo f104 reduces both L and chlorophyll b does notmuch reduce the VAZ but does increase the relativeDPS in conjunction with the energy dissipating f2 fraction

Figure 2 compares the 77 K steady-state chlorophyll apounduorescence spectra and component bands after Gaussiandeconvolution from intact leaves of the clo f2 (centgure 2a)wt (centgure 2b) and ML-grown clo f104 (centgure 2c) Alsoshown next to the spectral panels is a tabulation of the PSII and PS I protein components in the mutants and the77 K spectral bands that have been attributed to themThe spectra were normalized to unity at their respectiveemission peaks The clo f2 spectrum is characterized bylow amplitudes of the F685 and F695 bands from CP43and CP47 respectively relative to the peak emission fromthe F730 and F740 bands The clo f2 mutant is character-ized by a lack of Lhcb1 and Lhcb6 and reduced Lhcb2and Lhcb3 in PS II The clo f2 lacks Lhca4 from PS I andthe 730 nm emission is thus attributed to the remainingLhca1 2 and 3 complexes The wt spectrum is also char-acterized by low F685 and F695 amplitudes but the clo f2spectrum can easily be distinguished from the wt since itexhibits a maximum emission peak that is broadened inwidth and blue-shifted by about 10 nm compared to thewt The slight band resolved with a peak around 715 nmin the wt is tentatively attributed to PS I core complex ICC1 The clo f104 spectrum (centgure 2c) is distinguished byan increased ratio of the amplitudes of F685 and F695



8

10

wt HL

wt ML

wt LL

clo f104 HLclo f104 ML

clo f104 LLclo f2

(a)

(c)

(b)

(d )

yen

6

4

10

08

06

04

02

00

chl a

b

4

6

8

10

200 300 400 500

Ntot = chl (PS II) - 1

200 300 400 500

2

0

F

Fo

mm

aximum

50

40

30

20

10

0

mol (P

SII) -1

DPS

L = - 1833 + 014(Ntot)

VAZ = 557 + 002(Ntot)

f2

Figure 1 (a) Relationships between the PS II chlorophyll antenna size and the chl ab ratio (b) The xanthophyll cyclede-epoxidation state (DPS) and the level of xanthophyll-cycle-dependent energy dissipation ( f2) calculated according to Gilmoreet al (1998) (c) The ratio of the maximum (Fm) and dark (Fo) levels of chlorophyll a pounduorescence (d) Changes in the pools oflutein and the xanthophyll-cycle pigments DPS ˆ [zeaxanthin+ antheraxanthin][violaxanthin+ zeaxanthin+ antheraxanthin]f2 ˆ 7 125(F rsquomFm) + 1452 VAZ ˆ [violaxanthin+ antheraxanthin+ zeaxanthin] The data in (a) and (c) are from the same leafsamples the drop lines in (a) and (c) respectively show the coordination between the chlorophyll ab and FmFo and Ntot for thediiexclerent leaf samples The data in panels (b) and (d) represent composites from separate thylakoid preparations The data inpanel (d) are described with thick linear regression lines (and thin 95 concentdence intervals) and equations shown (r2 ˆ 0238 forVAZ and r2 ˆ 0932 for L) circles represent L for the clo f104 (open) and wt (closed) and squares represent VAZ for the clo f104(open) and wt (closed)

relative to F740 compared to the wt and clo f 2 Furtherthe clo f104 spectrum displays a prominent band centredaround 720 nm certainly attributed to CC1 The clo f104PS II protein composition is distinguished by a strong

reduction in Lhcb1 while the PS I protein compositiondenotes reduced levels of Lhca2 and 3 (also known asLHC I-680) Although some 77 K studies report a small680 nm band arising from Lhcb1 2 and 3 (Simpson et al



100

075

050

025

000

010

000

- 010

010

005

000680 700 720

(a)protein

PS II

PS I

Lhcb1

Lhcb2

Lhcb3

Lhcb6

PsbS

CP43

CP47

level

0

reduced

reduced

0wt

wt

wt

77K band

685 nm

695 nm

Lhca1ndash3

Lhca4

CC1

wt

0wt

730 nm

720 nm

PS I

Lhca1ndash4

CC1

wt

wt

740 nm

720 nm

fluo

resc

ence

inte

nsit

y

100

075

050

025

000

015

005

010

000680 700 720

(b)

fluo

resc

ence

inte

nsit

yR

i

010

000

- 010

Ri

100

075

050

025

000

(c)

fluo

resc

ence

inte

nsit

y

010

000

680 700 720

wavelength (nm)

740 760 780- 010

Ri

protein

PS II

Lhcb1ndash3

CP43

CP47

level

wt

wt

wt

77K band

weak 680 nm

685 nm

695 nm

PS I

Lhca1

Lhca2

wt

reduced740 nm

Lhca4

CC1

wt

wt 720 nm

protein

PS II

Lhcb1

PsbS

CP43

level

reducedwt

wt

77K band

685 nmCP47 wt 695 nm

Figure 2 Fluorescence spectral emission procentles and Gaussian deconvolution of the emission band components for barleychlorina f 2 (clo f 2) (a) wild-type (wt) (b) and chlorina f 104 (clo f 104) (c) mutant leaves at liquid nitrogen temperature (77 K) (leftpanels) Each spectrum was normalized to the peak emission and centtted with centve free-poundoating Gaussians The insets in (a) and (b)focus on the spectral region associated with short-wavelength antennae emissions mainly from the PS II core The subplotsbelow the main panels represent the residual errors (Ri) corresponding to the Gaussian model The data for the tables (rightpanels) were adapted from Bossman et al (1997) for (a) and (b) and Knoetzel amp Simpson (1991) for (c)

1985 Falbel et al 1994 Krugh amp Miles 1995) this bandwas not resolved in any samples in this study Perhaps itis hidden in the short-wave side of the F685 band in the wtassuming it may be narrower than this centt indicatesFurther work may answer this question

(b) Time-resolved chlorophyll a pounduorescenceemission spectra of the barley clo f2

wt and clo f104 leavesFigure 3 compares the contoured top-views and the

deconvolved spectral-kinetic components from the time-resolved dark-level chlorophyll a pounduorescence spectra andL1 model centts for the clo f 2 (centgure 3a^c) wt (centgure 3d^f )and clo f104 (centgure 3g^ i) barley leaves at roomtemperature The instrument pulse-response procentles ofthe streak-camera spectrograph are shown as subpanelsat the base of centgure 3adg In contrast to the large diiexcler-ences in the PS II antenna sizes described in frac12 3(a) (table

1 and centgure 1) each of pounduorescence spectra exhibitsimilar overall kinetics in the contoured region around685 nm this is one of the two major observable peaks inall leaves at room temperature The normalized pulse-induced pounduorescence rises and decays around 685 nm centtsimilarly within the same time-frame for each sample incentgure 3 Consistent with the 77 K spectra shown in centgure2 the clo f 2 spectral decay (centgure 3a) showed an overallblue-shift in its far-red region compared to the wt(centgure 3d) which showed its peak emission as a broadcontour around 740 nm Also consistent with the 77 Ksteady-state spectra in centgure 1 the clo f104 (centgure 3g)showed a strongly diminished intensity of its majorfar-red contour (of its 740^685 ratio) comparedwith the 685 nm contour as compared with the wt(centgure 3d )

The spectral-kinetic procentles are compared in sideviews in centgure 3beh and represent the kinetics of the



800w

avel

engt

h (n

m)

band

inte

nsity

integrated profile

band profilepu

lse

760

720

680

800

wav

elen

gth

(nm

)

760

720

680

10

05

00

10

08

06

04

02

00

80

60

40

20

0

6

5

4

3

2

1

0

6

5

4

3

2

1

0

6

5

4

3

2

1

0

80

60

40

20

0

80

60

40

20

0

10

08

06

04

02

00

10

08

06

04

02

00

puls

e

10

05

00

800

wav

elen

gth

(nm

)

760

720

680

puls

e 10

0500

500 1000time (ps)

1500 500 1000time (ps)

1500 680 720wavelength (nm)

760 800

(a)

(b)

(e)

(d)

f( )

(g)

(h)

(c)

(i)

Figure 3 Time and wavelength resolved pounduorescence spectra and their deconvolved components for barley chlorina f2 (clo f2)((a) (b) and (c)) wild-type (wt) ((d ) (e) and ( f )) and chlorina f104 (clo f104) ((g) (h) and (i)) mutant leaves at room temperaturemeasured with the streak-camera spectrograph The subplots under panels (a) (d) and (g) depict the instrument pulse^responseprocentles that were convolved with the respective model decay functions (see Appendix A) to yield the model curves (lines) The centtwas determined using the robust L1 method (see Appendix B) for minimizing the absolute deviations Panels (b) (e) and (h)depict the integral intensity procentles (right ordinate) for the complete spectral integral (680 to 800 nm) of the model (thickestline) and data points (open circles) the left ordinate refers to each of the four main component emission band amplitude procentles(red blue black and green) linked to the centre of gravity wavelength of the respective emission bands in panels (c) ( f ) and (i)Panels (c) ( f ) and (i) depict the Gaussian band procentles for the time integral (peak plus centve subsequent time channels) of the model(thickest line) actual data points (open circles) and each of the four component emission bands (red blue black and green)

intensity at the centre of gravity wavelength for thediiexclerent pounduorescence bands illustrated in centgure 3c fiwhich themselves represent the spectra in the front viewThe thick lines in the procentle plots represent the integratedmodel and the open circles represent the integral datapoints The front views are the integrated intensity of thepeak and centve subsequent time channels (integral ˆ 74 ps)The band intensity procentles for the clo f 2 (centgure 3bkinetics and 3c spectra) illustrate three major bandsnamely the F685 (red in the centgure) the F710 (blue) andthe F740 (black) The F685 band most of which is fromPS II decays with comparable kinetics to the F740 bandwhile the F710 band decays the fastest The large F710amplitude clearly dominates the decay spectrum incentgure 3a^c and determines the gradually sloping far-redspectral region that is clearly blue-shifted when comparedto the wt (centgure 3d^f ) A broad but minor far-red band(centred at 796 nm) was included to improve the centt in allthree images it exhibited similar decay kinetics as theF740 band and was interpreted as a satellite band It iswell known that chlorophyll a pounduorescence has a majorelectronic band followed by a satellite band It is thereforereasonable to assume that this band is a satellite band ofchlorophyll pounduorescing in the long-wave region (eg anF740 satellite band) just as a satellite band is expected inthe 740 nm region for F685

The wt band intensity procentle (centgure 3e f ) is initiallydominated by the F740 band (black in the centgure) thatdecays more rapidly than the F685 band (red) and crossesbelow it around the 1200 ps mark the F705 band (blue)has a low amplitude and has the most rapid overall decayIn contrast to the wt the clo f104 spectrum is dominatedat all times by the F685 band (red) the band intensityprocentle (centgure 3hi) shows the F740 band (black) does notexceed 75 of F685 at the peak of the pulse and decaysmore rapidly than F685 following the pulse peak Theminor F710 band showed almost identical decay kinetics asF705 in the wt

The clo f104 band spectral procentle (centgure 3g^ i) showsthat the integrated intensity of F740 is clearly depressedcompared with the wt while the kinetics and spectrum ofthe F705 band (blue in the centgure) is quite similar Theminor far-red band showed almost identical decay kineticsas F740 in the wt above A scatter band (tCTR ordm 000 ps)centred at 650 nm was also included in the model centts forall samples to account for a slight overlap of the diode lasertail emission with the sampleoumlbut because the integratedintensity of this band was very low it had little statisticaland no visual impact on the data or model presentationand hence is not illustrated in centgure 3

(c) Model decay distribution procentles and centttingparameters

Figure 4 compares the multimodal distribution procentlesthat represent the model decay functions for the dark-level (Fo) and maximal (Fm) pounduorescence conditions for clof2 (centgure 4ab) wt (centgure 4cd ) and clo f104 (centgure 4e f )The multimodal-distributed decay functions shown in themajor panels for the dark-level images are expressed asthe lifetime-weighted pre-exponential amplitude factornott(t) this is to emphasize those components mostheavily weighted in intensity The inset centgures show thedistributed pre-exponential amplitude factor not(t) they

illustrate signicentcant contributions of component ampli-tudes with faster pounduorescence lifetimes (5 1000 ps)Figure 4 shows that the major and most heavilyweighted mode of pounduorescence lifetime distribution forF685 is ca 400^500ps in all three samples with thetraps open The insets in centgure 4ace also show that forthe open trap conditions the lifetime distributions wereclearly multimodal for the F685 bands in all threesamples each exhibited one tailed modes centred below100 ps in addition to the more heavily weighted ca400^500ps modes it is possible that the faster compo-nent is the PS I F685 and the latter the PS II F685 TheF705 in the wt and clo f104 was clearly lower in ampli-tude compared with F710 in the clo f 2 The clo f 2 opentrap conditions were characterized by the very promi-nent single and short mode for F710 5 100 ps whereasF740 was bimodal and very similar to F685 in bothmodes again suggesting its origin in both PS I and PSII Compared to the clo f2 (centgure 4a) the wt (centgure 4b)and clo f104 (centgure 4c) exhibited similar major modesaround 250^350ps for F740

The clo f 2 closed trap conditions were characterized byextremely broad major modes for F685 and F740 withamplitudes peaking in the 1200^1500 ps range and longtails extending to 4 5 ns both the major F685 and F740modes integrated to lifetimes 4 2200 ps The anomalousF685 and F740 bandwidths and attenuated centre modeswere consistent with decay kinetics resolved by phase-modulation pounduorometry for the clo f2 under both minimaland maximal PS II pounduorescence conditions (data notshown) Consistent with the anomalous F685 and F740distributions a unique and relatively minor narrow modewas resolved in clo f2 for the F710 band around 500 psunder the maximal pounduorescence conditions

Compared with the anomalous clo f2 (centgure 4b) the wt(centgure 4d ) and clo f104 (centgure 4f ) exhibited remarkablysimilar major modes around 1900^2000 ps for both F685and F740 The major F740 modes were generally a fewhundred picoseconds faster than F685 in both the wt andclo f104 The major F740 mode in the clo f104 exhibited areduced integral amplitude compared to the wt The F740bands of the wt and clo f104 also exhibited broad minormodes in the sub 1000 ps rangeoumllikewise a similarminor mode was observed for F685 in the clo f104

(d) Residual error analyses for the w2vw2vw2 and

L1-robust minimization methodsTable 2 shows the results of the Runs tests the Durbin^

Watson d-statistic tests and the second w2-tests used tocomparatively evaluate the dark-level image centts obtainedfrom the w2

v and L1 methods The Runs tests compare theexpected and observed number of runs (consecutiveresidual signs between sign changes) as a global averagefor all the time-series (each l channel) by computing thestandard normal variates Ztf indicating too few and Ztm indicating too many runs per time-series For each imageand material it was clear the Runs tests indicated the L1method yielded more uniformly randomized residual signswhen compared to the w2

v method This was evident fromthe reduced mean and standard deviation values for boththe standard normal variates The sect frac14-values clearlyoverlapped the natural mean ( ˆ 2) of the d-statisticdistribution in each sample to indicate that the data



exhibit no extreme systematic serial trends The meanvalue was closer to 2 in the clo f 2 and wt L1 method centtsthan in the corresponding w2

v method centts but thed-statistic mean did not indicate an improved L1 method centtin the clo f104 The binned histogram analyses of theresidual errors from both the w2

v and L1 methods yieldedthe second w2-statistics The analyses were designed to testthe hypothesis that the residual error distribution wasnormal as assumed for the w2

v method as opposed to aLaplace distribution with a more peaked shape and moreheavily populated symmetrical tails as assumed for the L1method It was clear that for all three samples the centt ofthe residual error histogram to the error function

converged on the Laplace shape ˆ 2 (see footnote totable 2) for both the w2

v and L1 methods It was also clearthat in each image the L1 method led to the lowest andmost signicentcant second w2-statistic

4 DISCUSSION

(a) Time-resolved emission spectra of barleychlorina mutants

The most striking features of the images in this studywere the large variations in the PS-I-associated dark-levelpounduorescence spectral decay components in the clo f2 andclo f104 mutants compared to the wt It is clear that the



10

006

004

002

000

Fo

0 500(ps)t

( )t

a

1000

8

6

4

2

0

12

08

10

06

04

02

00

12

8

10

6

4

2

0

12

8

10

6

4

2

05000

(a)

10

006

004

002

0000 500

(ps)t

( )t

a

( )t

ta

1000

8

6

4

2

0

(c)

10

006

004

002

0000 500

(ps)t

(ps)t

( )t

a

1000

0 500 1000 1500 2000 0 1000 2000 3000 40002500

8

6

4

2

0

(e)

Fm

(b)

(d )

( f )

Figure 4 Chlorophyll a pounduorescence lifetime distribution procentles corresponding to the spectral components for the barleychlorina f 2 (clo f 2) (ab) wild-type (wt) (cd) and chlorina f104 (clo f104) (e f ) leaves Panels (a) (c) and (e) represent the dark-levelpounduorescence Fo and depict the lifetime-weighted distribution of the pre-exponential amplitude factor nott(t) for the respectivespectral bands whereas the insets in each panel depict the distribution of the pre-exponential amplitude factor not(t) Panels (b)(d ) and ( f ) represent the maximal pounduorescence Fm models and depict the lifetime-weighted distribution of the pre-exponentialamplitude factor (nott(t)) for the respective spectral bands The respective bands coloured in centgure 3 as red blue black and greenare denoted as black dashed grey and dark grey in centgure 4 The L1 statistics (see Appendix B) for the dark-level pounduorescenceimages (ace) are listed in table 2 while the L1 statistics for the maximal-level pounduorescence images for (b) (d) and ( f ) are 146165 and 122 respectively

steady-state spectral variations observed at 77 K for themutant leaves (Simpson et al 1985 Knoetzel et al 1997this study) are strongly echoed in the room temperaturespectral analysis It is also evident that self-absorption ofthe PS II pounduorescence in the leaves at room temperaturesincreased the amplitudes of the far-red bands which aremostly from PS I relative to the F685 band We concludethat reabsorption thus plays a major role in determiningthe amplitudes of the room temperature pounduorescencecomponents as it has been established at 77 K (Govindjeeamp Yang 1966Weis 1985)

The clo f 2 is of particular interest because of its anoma-lous spectral-kinetic properties compared to the clo f104and wt Lack of Lhca4 in clo f 2 as described by Knoetzelet al (1998) accompanies the far-red spectral changesMelkozernov et al (1998) have quanticented the pounduorescencelifetime behaviour of the Lhca4 and Lhca1 subunits invitro and also documented changes in energy transfer anddecay components that are attributable to changedpigment^protein interactions between the monomerscompared with the Lhca4^Lhca1 heterodimer thatcomprises the LHC I-730 complex

Both the dark and maximal pounduorescence levels of theclo f 2 are considerably anomalous compared with wt andclo f104 which are actually quite similar aside from therelative amplitudes of the F740 bands The clo f 2 anoma-lies compared to the wt and clo f104 primarily include(i) the blue-shifted broad emission spectrum (ii) theunique decay distribution at Fm for the F710 and (iii) thebroadened attenuated F685 and F740 distribution modes atFm It is clear that the F710 band originates mainly fromPS I and is likely to be of a similar nature to the PS Iband centrst observed by Lavorel (1963) at 712^718 nm Weconclude the spectral-kinetic behaviour of clo f2 iscomplex and strongly inpounduenced by altered energytransfer pathways and reabsorption patterns between PSII and PS I The observed changes must be partly attri-buted to changed absorption^emission overlap integralsand reabsorption procentles for the F685 and F710 bands andthe F685 (andor F710) and F740 bands respectively

The anomalous in vivo PS II pounduorescence lifetimes in clof2 clearly are physically related to the well-known Mg2+

eiexclects on isolated clo f2 thylakoids with respect tostacking and PS II versus PS I emission (Bassi et al 1985)Isolated clo f2 thylakoids compared to the wt and clo f104are quite labile with respect to stacking and require highMg2+ and chelating agents to induce and retain stackingin vitro Without the Mg2+ and chelation treatment there is

a strong attenuation in the lifetime centre of the 20 ns(F685) PS II distribution mode that results in a stronglyinhibited FvFm and inhibited xanthophyll-cycle-dependent energy dissipation (Gilmore et al 1996) Thewell-documented Mg2+-sensitive PS II quenching by PS Iprimarily aiexclects Fm and not Fo (Krause et al 1983) Toexplain the in vivo results of this study in light of the clo f 2phenotype we hypothesize that PS II units that structu-rally interface with PS I units in the outer poorly formedgranal stacks in the leaves may exhibit strong quenchingby energy transfer to PS I that shortens the Fm pounduorescencelifetimes Thus the broadening and attenuation of the Fmlifetime distributions in the clo f2 (centgure 4b) may indicatethe mixed emission from PS II units in the outer granalstacks (in excitonic contact with PSI) and weaklyquenchedunits in the inner granal regions disconnected from PS Iandthose PSII units in intermediate locations

Another interesting and important result of this studyis the similarity of the PS-II-associated spectral-kineticcontours of the 685 nm band at Fo for the clo f2 clo f104and wt It is well established and documented in table 1that both the clo f2 and the clo f104 grown under theprescribed conditions exhibit major decreases in theirperipheral PS II antennae components primarily due toreduced Lhcb1 levels However as shown previously thesechanges do not directly correlate with the changes in thepounduorescence lifetimes of the PS II 685 nm band under theFo conditions (Briantais et al 1996 Gilmore et al 1996)Much earlier Haehnel et al (1982) also reported similarPS II decay kinetics at Fo for intermittent-light-grown peachloroplasts depleted of Lhcb1 as for normal pea chloro-plasts The preponderance of data indicate that theperipheral antennae chlorophylls of the Lhcbs 1^3 are notabsolutely equilibrated kinetically with inner-coreantennae chlorophyll a molecules It is clear that thepounduorescence lifetimes are not linearly determined by thenumber of total chlorophyll as in the PS II unit at eitherthe Fo or Fm levels Similarly it appears that the PS Iantenna mutations in the chlorina mutants have muchmore signicentcant inpounduence on the amplitudes and emissionspectral energy levels than on the absolute decay kineticsThis can be explained by recognizing that loss ofchlorophyll b destabilizes certain Lhcs of PS I such thatthey decompose forcing the remaining Lhcs to reorga-nize and yield new heterocomplexes with alteredgeometric arrangements and energy transfer connectionsand hence changed emission amplitudes and energylevels



Table 2 Comparative statistical characterization of the dark-level pounduorescence images from the clo f2 wt and clo f104 leaves

(All statistical parameters are decentned in Appendix B including the standard normal variates for the Runs test the Durbin^Watson d-statistic the second chi-squared statistic (w2) the reduced chi-squared (w2

v ) and the robust L1 methods)

jZtfj jZtmj d second w2 a

leaf w2v L1 w2

v L1 w2v L1 w2

v L1 w2v L1

clo f 2 00238 173 0963 sect 0740 0897 sect 0662 1088 sect 0773 0983 sect 0701 1873 sect 0260 1910 sect 0261 1040 0778wt 00099 136 1147 sect 0823 0889 sect 0685 1292 sect 0846 0989 sect 0709 1872 sect 0215 1896 sect 0208 0885 0646clo f104 00074 137 1541 sect 1013 1323 sect 0817 1652 sect 1061 1418 sect 0872 1856 sect 0233 1837 sect 0229 0834 0707

aIn each case the best centt for the second w2-statistic converged on a shape parameter ˆ 2 indicating Laplace as opposed to normalresidual error distributions

With respect to PS II an absolute linear relationshipbetween t and N ˆ chlorophyll a (PS II)71 is in factpredicted only if one assumes that each chlorophyll of PSII has an equal probability for exciton residence during arandom walk of the Frenkel exciton through the entirechlorophyll matrix of the photosystem (Pearlstein 1982)This overly simplicented assumption would require thateach chlorophyll site in the lattice is uniformly spacedfrom and isoenergetic with its neighbour sites Althoughit still remains to be determined the lack of excitonequilibration between the Lhcb pool in the peripheralantennae and the PS II core-inner antennae is probablyprimarily due to entropic (structural) features as opposedto spectral (energetic) features (Holcomb amp Knox 1996Gilmore amp Govindjee 1999) This is proposed since thechlorophyll a energy levels of the Lhcbs are believed to benearly isoenergetic with the main chlorophyll a pool inthe PS II core antenna (Roelofs et al 1992)

Furthermore with respect to structural features ofthe antennae the xanthophyll-cycle-dependent non-photochemical quenching activity associated with PS IIis probably conserved in these mutants especially clof104 (Gilmore et al 1996 1998) because they eachcontain wild-type levels of the PsbS protein in the PS IIinner-core antennae (Bossman et al 1997) The PsbSprotein has recently been established as a requisite forthe conformational changes involved in xanthophyll-cycle-dependent non-photochemical quenching (Li et al2000) Inhibition of xanthophyll-cycle-dependent energydissipation in the clo f 2 or other completely chlorophyll-b-less mutants or etiolated systems might be expected tobe strongly inpounduenced by the PS-I-associated anomaliespointed out earlier In fact these changes could alter theabsorption^emission overlap integrals between PS II andPS I and or the pounduorescence reabsorption patterns toinpounduence directly and quench the emission from PS IIXanthophyll-cycle-dependent attenuation of the Fmpounduorescence yield would thus be inhibited by thecompeting de-excitation rate process of PS II energytransfer to PS I

It is self-evident that room temperature pounduorescencemeasurements of the clo f104 and clo f2 mutants usingemission centlters such as the Schott RG-9 centlter which cutsoiexcl 100 below 690 nm and 50 at 710 nm entirelyeliminates the majority of the longest lifetime componentfrom PS II namely the 685 nm band These cut-oiexclcentlters would thus be expected to have a signicentcant inpoundu-ence on the pounduorescence intensity readings since themain and slowest decaying PS II band (F685) is directlyand strongly obscured Therefore it is concluded that thediiexclerences observed for the PS II ecurrenciency and roomtemperature chlorophyll a pounduorescence intensity ratioswith the PAM between the clo f2 clo f104 and wt (seefor example Simpson et al 1985 Gilmore et al 1996) arestrongly inpounduenced by the PS I contributions to the Fmand Fo pounduorescence intensity readings This conclusion isconsistent with several other reports by PfIumlndel (1998)Genty et al (1990) Adams et al (1990) and Schmuck ampMoya (1994) A recommended way to measure andaccount for PS-I-related emission with a commercial PAMchlorophyll pounduorometer is to use a blue or green light-emit-ting diode for excitation in combination with either aband-pass centlter (F ˆ 685 nm) or a long-pass centlter

(F 4 660 nm) to capture the respective PS II or total leafchlorophyll emission

(b) Streak-camera spectrograph and doubleconvolution integral method

This study reports the centrst use of a new global dataanalysis method designed to facilitate a complete three-dimensional distribution-based simulation of the time-resolved emission spectrum from a time-correlatedsingle photon counting instrument The doubleconvolution integral (DCI) method developed hereincan be compared to both the conventional decay-associated spectra (DAS) and related time-resolvedemission spectra (TRES) methods (Hodges amp Moya1986 Holzwarth 1988 Knutson 1992 Roelofs et al1992 Hungerford amp Birch 1996) The DAS centttingmethod minimizes the residual error sum by searchingfor amplitude and decay parameters to simulate eachwavelength channel individually based on the assump-tion there is a physical relationship between adjacentwavelength channels in the form of variable amplitudefractions components with the same pounduorescence life-times The linked lifetime amplitudes decentne thespectral shape of a component and the lifetimecomponents are usually assumed to be discrete asopposed to distributed parameters The DCI methodin contrast to the DAS method works under theassumption that Gaussian spectral bands determine therelative distributed amplitudes of the lifetime compo-nents that may also exhibit distributed and possiblymultimodal kinetic features The DCI method isparticularly suited for simultaneously acquired spectral-kinetic data as obtained with the streak-cameraspectrograph and does not require independent deter-mination of the steady-state emission spectrum TheDAS method usually involves calculating so-called`localrsquo w2

v -statistics for each wavelength channel that arethen used to generate a composite global w2

v -statisticThe DCI method generates a single global statistic thattakes into account each and every (tl) channel coordi-nate simultaneously The DCI method signicentcantlyreduces the number of free-centtting parameters comparedto the DAS and TRES methods hence facilitating amore statistically signicentcant solution

With respect to the data in this paper it was clear thatthe w2

v method did not yield the most reproducible orstatistically signicentcant results In fact the resultantresidual error distributions from the w2

v method weredecidedly Laplace-like in shape The L1 method led toconsistently more signicentcant randomization of theresidual error signs in each time-series and in most casessimilar or better autocorrelation scores based on theDurbin^Watson d-statistic The use of the L1 method isstatistically justicented for the DCI method because the L1method is designed to resolve smaller amplitude signalswith low error amplitudes that may be convolved withlarger amplitude signals with respectively larger erroramplitudes because the former are not weighted asheavily In fact robust centtting is specially suited for datasets where the y-axis (amplitude) variable spans more thanfour log divisions as is common with time-correlatedsingle photon counting (Rundel 1991) We interpret theimproved signicentcance of the L1 method to imply that



intrinsic systematic errors possibly associated with thevarious instrument response convolutions and particu-larly the CCD camera lead to a Laplace error distribu-tion for the photoelectron counts recorded by theinstrument (Schiller amp Alfano 1980 Campillo amp Shapiro1983 Tsuchiya 1984 Davis amp Parigger 1992)

We thank the National Institute for Basic Biology (NIBB) fora Visiting Scientist Award (AMG) Professor N Murata andhis laboratory staiexcl for assistance in growing plants and pro-viding experimental and centnancial support and The AustralianDepartment of Industry Science and Tourism grant 970638and 970709 for travel funds and centnancial support of theinitial phase of the project Special thanks are due to Dr W SChow of the Australian National University (ANU) ResearchSchool of Biological Sciences (RSBS) Photobioenergetics(PBE) group for performing the photosynthetic unit sizemeasurements The present work was started at the Universityof Illinois at Urbana where AMG worked on the IntegrativePhotosynthesis Grant Govindjee was supported by an ANURSBS Visiting Fellowship SI was supported by Grants-in-Aid on Priority-Area-Research (molecular biometallics andsingle electron transfer devices) from the Ministry of Educa-tion Science Sports and Culture Japan

APPENDIX A THEORY FUNCTION FOR THE TIME-

AND WAVELENGTH-RESOLVED FLUORESCENCE

ANALYSIS

The time- and spectral-resolved pounduorescence modelassumes that the room-temperature chlorophyll a pounduores-cence decay from the higher plant photosyntheticapparatus comprises continuous multimodal distributionsof decay components for the following three reasons

(i) proteins exhibit conformational dynamics that mayinpounduence protein-bound pounduorophores (Alcala et al1987 Frauenfelder et al 1988)

(ii) the photosynthetic apparatus is heterogenous inpigment^protein composition (antennae and re-action centres)

(iii) the physiological state of the sample may pounductuateslightly during the acquisition with respect to re-action centre trap closure among other biochemicalchanges

The model derivation also assumes that the total chloro-phyll a pounduorescence emission comprises a number ofdistinct emission bands (Govindjee 1995) that arise fromparticular pigment^protein complexes (antennae andreaction centres of PS I and PS II) and are almost purelyGaussian in nature with respect to their spectral wave-length energy distribution (Lin amp Knox 1988 Mimuro etal 1989 Yamazaki et al 1984)

Individual model pounduorescence lifetime decay functionswere assigned to each spectral wavelength bandi n according to the following generalized function

F(tiexcl t 0)i ˆ

1

0

not(t)iexppermiliexcl(t iexcl t 0)tŠdt t4 t 0

0 t 4 t 0

(A1)

where the pounduorescence intensity at any time t7 trsquo is deter-mined by evaluating the integral when t 4 t rsquo but is 0when t4 t rsquo In equation (A1) not(t)i represents the contin-uous multimodal distribution of the pre-exponential

amplitude factor as a function of pounduorescence lifetime tnot(t)i is determined as the sum of varying contributions ofthe Gaussian^Lorentzian sum function decentned later Inthe following analyses the continuous not(t)i integralshown above was approximated as a discrete distributioncomprising 500 channels (15 ps channel71) over the decent-nite integral (0^15 ns) it was found that for these samplesthe centts obtained to not(t)i were uninpounduenced by thespecicented bounds The resulting integral decay functionF(t7 t rsquo)i was then numerically convolved with the instru-ment response procentle plus a time-axis shift L(trsquo7 tshift) toyield the following convolution integral that predicts thenormalized intensity at any time t for each bandi n ie

I(t)i ˆ

t

iexcl1

L(t 0 iexcl tshift) pound F(t iexcl t 0)i dt (A2)

decentned more compactly as

I(t)i ˆ L(t iexcl tshift) laquo F(t)i (A3)

The normalized continuous distribution of the intensityas a function of wavelength l for each bandi n wasdecentned by a Gaussian spectral line shape with the ampli-tude function (see equation (2)) the CTR and WID para-meters were globally linked through all time channels foreach bandi n It follows that in order to evaluate theintensity at any point in time t and at any wavelengthl I(t)i was convolved with I(l)i by multiplying thenormalized Gaussian AMP factor by I(t)i to yield thedouble convolution integral

I(tl)i ˆ I(t)t laquo I(l)i (A4)

for each bandi n The centnal simulated image was decentnedsimply as

T(tl) ˆ OFFSET(tl) DaggerN

iˆ1

I(tl)i (A5)

which represents the summed contributions of I(tl)i n for all bands in the t versus l grid plus a verticalOFFSET(tl) that was a global parameter for every (tl)coordinate to simulate and account for uniform back-ground noise contributions

The integrated distribution of the pre-exponentialamplitude factor not of the pounduorescence lifetime t isdecentned as the not(t)i component of F(t7 t rsquo)i in equation(A1) and was determined in our model centts by usingvarying summed combinations the Gaussian^Lorentziansum distribution function (Rundel 1991) This functionsums equal full-width at half-maximum Gaussian andLorentzian distributions and comprises four changingparameters namely the WID (full-width at half-maximum amplitude) the peak (modal) amplitude(AMP) the centre or mode (CTR) and the shape(SHAPE) The WID must be 4 0 and the SHAPE canonly vary between 0 and 1 where SHAPE ˆ 0 is a pureLorentzian and SHAPE ˆ 1 is a pure Gaussian



APPENDIX B STATISTICAL ANALYSIS OF THE TIME-

RESOLVED FLUORESCENCE SPECTRAL IMAGES

(a) w2w2w2 methodThe images were simulated and analysed using Micro-

soft1 Excel 97 and Visual Basic Macro programmingThe residual errors Ri ˆ Di7Mi were weighted assuminga Poisson distribution of errors at each (tl) coordinatethat can be approximated given a sucurrencient number ofcounts by a Gaussian distribution We recall that Di is thepounduorescence intensity at time-wavelength coordinate (tl)iand Mi is the predicted intensity for the same (tl)i co-ordinate The residual weighting is approximated as thestandard deviation or the square root of the intensityvalue at each (tl)i coordinate as frac14 Mi

p The least-

squares model centts were evaluated based on the random-ness of the weighted residuals and by comparing thereduced w2-statistic calculated according to Bevington(1969) as

w2v ˆ

1N iexcl m iexcl 1

N

iˆ1

(Di iexcl Mi)2

Mi (B1)

where N ˆ 10 335 is the number of (tl) coordinate chan-nels m is the number of dynamic model-centtting para-meters (Di7 Mi)2 represents the squared residual error ateach (tl)i

(b) L1 method for robust minimizationof the absolute deviations

The arbitrary nature of the natural spectral-kineticcontours among the varying samples dictates that theresidual weighting of the photoelectron count cannot bejusticentably normalized or referenced to any one given peak(tl) channel coordinate or intensity integral to aiexclord anabsolutely equitable statistical weighting either within oramong the images This is obvious because if one arbitra-rily chooses a peak channel or coordinate to normalize forone image one would necessarily use diiexclerent residualweighting (based on the assumption that frac142 Mi) whencomparing any other (t) or (l) channel or (tl) coordinatechannels or total integrated intensity Tint(tl) of that orany other image The robust L1 method (Draper amp Smith1998) minimizes the sum of the absolute deviations and isdecentned as

L1 ˆN

iˆ1

jRi j (B2)

The L1 method assumes the residual error distribution isbest decentned by a Laplace distribution function (Evans etal 1993) which has a sharper peak and more heavilypopulated tails than a normal distribution

(c) Analyses of the randomness of the residual errorsThe residual error distributions resulting from the w2

vand L1 centts were comparatively analysed by constructinghistograms comprising N ˆ 500 bins each with a value of(Ri ˆ j00026j) The distributions were compared to thebest centtting error distribution function (Evans et al 1993)using a second w2-statistic

second w2 ˆ1

N iexcl m iexcl 1

N

iˆ1

(Di iexcl Mi)2

Mi (B3)

where m represents the centtting parameters for the symme-trical error distribution function decentned as

y ˆ AMP pound exp iexcl 12

jx iexcl CTRj(2=SHAPE)

WID (B4)

with the SHAPE parameter that varies between 1 and 2where 1 determines a Gaussian distribution and 2 deter-mines a Laplace or double-exponential distribution TheSHAPE parameter was critical for evaluating thesecomparative analyses because the w2

v method assumes aGaussian residual distribution whereas the robust L1method assumes a Laplace error distribution The mode(CTR) of the error was centxed at 0 to test for symmetryabout the ideal mean while the WID AMP and SHAPEparameters were free poundoating

The randomness of the residual distribution about theassumed mean middotx ˆ 0 was also determined using the Runstest as decentned by Draper amp Smith (1998) The Runs testanalyses only the sign and not the amplitude of the resi-dual value but was still a useful indicator of successfulmodel minimization in these analyses The number ofconsecutive signs or runs in the residuals time-series foreach l channel was compared with the expected numberof runs (and a standard deviation) that was determinedby counting the number of positive and negative signs ineach time-series and assuming a completely randomsequence Standard normal variates were then calculatedfor each time channel to test the assumptions that therewere either too few (jZtfj ) or too many (jZtmj) runs in thetime-series compared with a completely random residualpattern The global middotx sect frac14 was then calculated for allwavelength channels and mean values of jZtfj (or jZtmj)that were less than or equal to three standard deviationswere accepted as a general `rule of thumbrsquo (Straume ampJohnson 1992)

Additionally each time-series for each wavelengthchannel was analysed for autocorrelation trends by use ofthe Durbin^Watson d-statistic (Draper amp Smith 1998)and a global d ˆ middotx sect frac14 for all wavelength channels wascalculated (Straume amp Johnson 1992) It is generallyaccepted that values approaching the lowest theoreticalvalue of d ˆ 0 indicate positive serial correlations whereasvalues approaching the maximum theoretical d ˆ 4indicate negative serial correlations Thus the test of thenull hypothesis for no autocorrelation was viewed`qualitativelyrsquo as being meaningful if the d ˆ middotx sect frac14 rangeencompassed the natural mean of the d-statistic distri-bution that equals two

REFERENCES

Adams III W W Demmig-Adams B Winter K amp SchreiberU 1990 The ratio of variable to maximum chlorophyll pounduor-escence from photosystem II measured in leaves at ambienttemperatures and at 77 K as an indicator of the photon yieldof photosynthesis Planta 180 166^174

Agati G Cerovic Z G amp Moya I 2000 The eiexclect ofdecreasing temperature up to chilling values on the in vivoF685F735 chlorophyll pounduorescence ratio in Phaseolus vulgarisand Pisum sativum The role of photosystem I contribution tothe 735 nm pounduorescence band Photochem Photobiol 72 75^84



Alcala J R Gratton E amp Prendergrast F G 1987Resolvability of pounduorescence lifetime distributions using phasepounduorometry Biophys J 51 587^596

Baker N (ed) 1996 Photosynthesis and the environment Advances inphotosynthesis research vol 5 Dordrecht The NetherlandsKluwer

Bassi R Hinz U amp Barbato R 1985 The role of the lightharvesting complex and photosystem II in thylakoid stackingin the chlorina-f2 barley mutant Carlsberg Res Commun 50347^367

Bevington P R 1969 Data reduction and error analysis for thephysical sciences NewYork McGraw-Hill

Bossmann B Knoetzel J amp Jansson S 1997 Screening ofchlorina mutants of barley (Hordeum vulgare L) with antibodiesagainst light-harvesting proteins of PS I and PS II absence ofspecicentc antenna proteins Photosynth Res 52 127^136

Briantais J-M 1994 Light-harvesting chlorophyll a^b complexrequirement for regulation of photosystem II photo-chemistry by non-photochemicalquenching Photosynth Res 40287^294

Briantais J-M Dacosta J Goulas Y Ducruet J-M ampMoya I 1996 Heat stress induces in leaves an increase of theminimum level of chlorophyll pounduorescence Fo a time-resolved analysis Photosynth Res 48 189^196

BIumlhler C A Graf U Hochstrasser R A amp Anliker M 1998Multidimensional pounduorescence spectroscopy using a streakcamera based pulse pounduorometer Rev Sci Instr 69 1512^1518

Campillo A J amp Shapiro S L 1983 Picosecond streak pounduoro-metryoumla review IEEE J Quantum Elec 19 585^603

Chow W S Hope A B amp Anderson J M 1991 Furtherstudies on quantifying photosystem II in vivo by poundash-induced oxygen yield from leaf discs Aust J Plant Physiol 18397^410

Davis L M amp Parigger C 1992 Use of a streak camera fortime-resolved photon counting pounduorimetry Meas Sci Technol3 85^90

Draper N R amp Smith H 1998 Applied regression analysis 3rdedn NewYork Wiley

Evans M Hastings T amp Peacock B 1993 Statistical distribu-tions 2nd edn NewYork Wiley

Falbel T G Staehelin A amp Adams III W W 1994 Analysis ofxanthophyll carotenoids and chlorophyll pounduorescence in lightintensity-dependent chlorophyll-decentcient mutants of wheatand barley Photosynth Res 42 191^202

Falk S Krol M Maxwell D P Rezansoiexcl D A Gray G Ramp Huner N P A 1994 Changes in in vivo pounduorescencequenching in rye and barley as a function of reduced PS IIlight harvesting antenna size Physiol Plants 91 551^558

Frauenfelder H A Parak F amp Young R D 1988 Conform-ational substates in proteins A Rev Biophys Biophys Chem 17451^479

Genty B Wonders J amp Baker N R 1990 Non-photochemicalquenching of Fo in leaves is emission wavelength dependentconsequences for quenching and its interpretation PhotosynthRes 26 133^139

Gilmore A M amp Govindjee 1999 How higher plants respondto excess light understanding energy dissipation strategies forphotosystem II In Concepts in photobiology photosynthesis andphotomorphogenesis (ed G S Singhal G Renger S K SoporyK-D Irrgang amp Govindjee) pp 513^548 New Delhi IndiaNarosa Publishing House

Gilmore A M amp Yamamoto H Y 1991 Resolution of luteinand zeaxanthin using a lightly carbon-loaded C18 high-performance liquid chromatographic column J Chromatogr543 137^145

Gilmore A M Hazlett T L Debrunner P G amp Govindjee1996 Photosystem II chlorophyll a pounduorescence lifetimes areindependent of the antenna size diiexclerences between barleywild-type and chlorina mutants photochemical quenching and

xanthophyll-cycle dependent nonphotochemical pounduorescencequenching Photosynth Res 48 171^187

Gilmore A M Shinkarev V P Hazlett T L amp Govindjee1998 Quantitative analysis of the eiexclects of intrathylakoid pHand xanthophyll cycle pigments on the chlorophyll a pounduores-cence lifetime distributions and intensity in thylakoidsBiochemistry 37 13 582^13 593

Govindjee 1995 Sixty-three years since Kautsky chlorophyll apounduorescence Aust J Plant Physiol 22 131^160

Govindjee amp Yang L 1966 Structure of the red pounduorescenceband in chloroplasts J Gen Physiol 49 763^780

HIgravertel H amp Lokstein L 1995 Relationship between quenchingof maximum and dark-level chlorophyll pounduorescence in vivodependence on photosystem II antenna size Biochim BiophysActa 1228 91^94

Haehnel W Nairn J A Reisberg P amp Sauer K 1982Picosecond pounduorescence kinetics and energy transfer inchloroplasts and algae Biochim Biophys Acta 680 161^173

Hodges M amp Moya I 1986 Time-resolved chlorophyll pounduores-cence studies of photosynthetic membranes resolution andcharacterization of four kinetic components Biochim BiophysActa 849 193^202

Holcomb C T amp Knox R S 1996 The relationship ofintercompartmental excitation transfer rate constants tothose of an underlying physical model Photosynth Res 50117^131

Holzwarth A R 1988 Time resolved chlorophyll pounduorescenceIn Applications of chlorophyll pounduorescence (ed H K Lichtenthaler)pp 21^31 DordrechtThe Netherlands Kluwer

Hungerford G amp Birch D J S 1996 Single-photon timingdetectors for pounduorescence lifetime spectroscopy Meas SciTechnol 7 121^135

Knoetzel J amp Simpson D 1991 Expression and organisation ofantenna proteins in the light- and temperature-sensitivebarley mutant chlorina f104 Planta 185 111^123

Knoetzel J Bossmann B amp Grimme H 1998 Chlorina andviridis mutants of barley (Hordeum vulgare L) allow assignmentof long-wavelength chlorophyll forms to individual Lhcaproteins of photosystem I in vivo FEBS Lett 436 339^342

Knutson J R 1992 Alternatives to consider in pounduorescencedecay analysis Meth Enzymol 210 357^374

Krause G H Briantais J-M amp Vernotte C 1983Characterization of chlorophyll pounduorescence quenching inchloroplasts by pounduorescence spectroscopy at 77 K I centpH-dependent quenching Biochim Biophys Acta 723 169^175

Krugh B W amp Miles D 1995 Energy transfer for low tempera-ture pounduorescence in PS II mutant thylakoids Photosynth Res44 117^125

Lavorel J 1963 Indications drsquoordre spectroscopique sur lrsquohetero-geneite de la chlorophylle in vivo Colloques Internationaux deCentre National de la Recherche Scienticentque 119 161^176

Li X P BjIcircrkman O Shih C Grossman A Rosenquist MJansson S amp Niyogi K K 2000 A pigment-binding proteinessential for regulation of photosynthetic light harvestingNature 403 391^395

Lin S amp Knox R S 1988 Time resolutionof a short-wavelengthchloroplast pounduorescence component at low temperature JLuminescence40^41 209^210

Melkozernov A N Schmid V H R Schmidt G W ampBlankenship R E 1998 Energy redistribution in hetero-dimeric light-harvesting complex LHC I-730 of photosystemI J Phys Chem 102 8183^8189

Mimuro M Yamazaki I Tamai N amp Katoh T 1989Excitation energy transfer in phycobilisomes at 7196 8Cisolated from the cyanobacterium Anabaena variabilis (M-3)evidence for the plural transfer pathways to the terminal emit-ters Biochim Biophys Acta 973 153^162

Pearlstein R M 1982 Exciton migration and trapping in photo-synthesis Photochem Photobiol 35 835^844



PfIumlndel E 1998 Estimating the contribution of photosystem Ito total leaf chlorophyll pounduorescence Photosynth Res 56185^195

Porra R J Thompson W A amp Kriedemann P E 1989Determination of accurate extinction coecurrencients and simulta-neous equations for assaying chlorophyll a and b with fourdiiexclerent solvents vericentcation of the concentration of chloro-phyll by atomic absorption spectroscopy Biochim Biophys Acta975 384^394

Roelofs T A Lee C-H amp Holzwarth A R 1992 Globaltarget analysis of picosecond chlorophyll pounduorescence kineticsfrom peas chloroplasts Biophys J 61 1147^1163

Rundel R 1991 Technical guide to peak centt nonlinear curve-centtting soft-ware Corte Madera CA Jandel Scienticentc

Schiller N H amp Alfano R R 1980 Picosecond characteristicsof a spectrograph measured by a streak cameravideo readoutsystem Optics Comm 35 451^454

Schmuck G amp Moya I 1994 Time-resolved chlorophyll pounduor-escence spectra of intact leaves Remote Sens Environ 47 72^76

Schreiber U Klughammer C amp Neubauer C 1988 MeasuringP700 absorbance changes around 830 nm with a new type ofpulse modulation system Z Naturforsch43c 686^698

Simpson D J Machold O HIcircyer-Hansen G amp VonWettstein D 1985 Chlorina mutants of barley (Hordeum vulgareL) Carlsberg Res Commun 50 223^238

Singhal G S Renger G Sopory S K Irrgang K-D ampGovindjee (eds) 1999 Concepts in photobiology Photosynthesis andphotomorphogenesis New Delhi India Narosa PublishingHouse

Straume M amp Johnson M L 1992 Analysis of residualscriteria for determining goodness-of-centt Meth Enzymol 21087^105

TsuchiyaY 1984 Advances in streak camera instrumentation forthe study of biological and physical processes IEEE JQuantum Elec 20 1516^1528

Van Kooten O amp Snel J F H 1990 The use of pounduorescencenomenclature in plant stress physiology Photosynth Res 25147^150

Weis E 1985 Chlorophyll pounduorescence at 77 K in intact leavescharacterization of a technique to eliminate artifacts relatedto self-absorption Photosynth Res 6 73^86

Yamazaki I Mimuro M Murao T Yamazaki TYoshihara K amp Fujita Y 1984 Excitation energy transferin the light harvesting antenna system of the red algaPorphyridium cruentum and the blue-green alga Anacystisnidulans analysis of time-resolved pounduorescence spectraPhotochem Photobiol 39 233^240

DiscussionH Y Yamamoto (Department of Plant Molecular PhysiologyUniversity of Hawaii USA) What constitutes a saturatingconcentration of zeaxanthin or anthereaxanthin (orlutein) with respect to the non-photochemical quenchingactivity of PS II

A M Gilmore Aside from minor changes associated withpleiotropic eiexclects as may be observed in some of thexanthophyll mutants of Arabidopsis one can generallypredict given a saturating level of thylakoid lumen acidi-centcation that as little as 2 or 3 mols of the xanthophyllszeaxanthin or antheraxanthin per PS II will be saturating

H Y Yamamoto If 2 or 3 mols of xanthophylls are sucurren-cient to saturate the PS II non-photochemical quenchingactivity then why do plants tend to accumulate morexanthophylls and why do they then need to undergo de-epoxidation at all

A M Gilmore It is my point of view that much likepollen generation plants naturally tend to overcompen-sate for this particular biosynthetic pathway The analogyis very clear because it only takes one xanthophyll toquench one PS II unit and it only takes one pollen grain tofertilize and ovum to make a seed yet plants tend to makean enormous excess of pollen that is usually simply lost inthe wind It is quite curious that no-one ever ascribes aprofound signicentcance to the natural tendency to generateexcess pollen whereas I have been asked this questionregarding the xanthophyll cycle many times

E Hideg (Institute of Plant Biology Biological Research CenterSzeged Hungary) Please comment on the physiology ofMantoniella squamata is it grown under special lightconditions

A M Gilmore We grow M squamata under about60 m molphotons m72 s71 However our actinic light treat-ments were usually up to 1000 m mol photons m7 2 s71 foras long as 15 min After the end of the light treatment weobserved little if any irreversible quenching of PS II pounduor-escence that we could use or interpret as an indicator ofphotoinhibition

E Hideg Is it light sensitive as compared to other algae

A M Gilmore I would conclude that in the short termas pertains to the experiments presented in the DiscussionMeeting but not reported here M squamata is not extre-mely or obviously more light sensitive than other algae Ihave worked with including diatoms and Chlamydomonasgrown under similar light conditions ie 60 to100 m molphotons m72 s71

E Garcia-Mendoza (Depatment of Microbiology University ofAmsterdam The Netherlands) What controls the poundip poundop ofantheraxanthin in Mantoniella squamata

A M Gilmore My hypothetical interpretation is that inthe presence of lumen acidicentcation with the PS II innerantennae proteins protonated the antheraxanthin mol-ecule binds to the transmembrane proteins more tightly(in the position where its de-epoxidized end-group facesthe lumen) than does the violaxanthin molecule Thusantheraxanthin rarely and almost never poundips to formzeaxanthin with a pH gradient

E Garcia-Mendoza What about epoxidation characteris-tics

A M Gilmore My hypothesis is that when the lumenacidity is neutralized in the dark antheraxanthin bindingis less ecurrencient and the molecule is free to poundip in themembrane to interact with the epoxidase enzyme on thestroma side Hence it is the protonation of the proteins bythe pH gradient that causes antheraxanthin to stop poundip-ping and this inhibits its conversion to zeaxanthinLikewise without the pH gradient antheraxanthinconverts back to violaxanthin with similar kinetics as seenin higher plant leaves or chloroplasts

Note added in proof Agati et al (2000) independentlyshowed that the F735 band emitted from the PS I antennacompartment exhibits a much shorter pounduorescence life-time than F685 from PS II in leaves The inpounduence of F735

on the average pounduorescence lifetime is most stronglyexhibited under the Fo conditions when PS II photo-chemistry is at a maximum



(PS I) because (i) PS I exhibits much more rapid decaykinetics than PS II (Roelofs et al 1992 Schmuck amp Moya1994) and (ii) reabsorption of the shorter wavelengthpounduorescence from PS II decreases the amplitudes of thePS II bands relative to the longer wavelength PS I bandsin leaves (Govindjee amp Yang 1966 Weis 1985 PfIumlndel1998) This study reports both a novel in vivo measurementand a global time-resolved emission spectral analysismethod known as the `double convolution integralrsquo orDCI method as these are applied primarily to investigatethe so-called dark-level and maximal chlorophyll a pounduor-escence yields (Fm conditions) from both PS II and PS I atroom temperature

We analysed the in vivo lifetimes of chlorophyll apounduorescence from two well-characterized light-harvestingbarley mutants known as chlorina f104 (clo f104) andchlorina f2 (clo f 2) These mutants are distinguished byunique alterations in the content and composition of thelight-harvesting pigment^protein complexes associatedwith both PS II and PS I The chlorophyll-b-less clo f2mutant primarily lacks the major complement of Lhcb1and Lhcb6 proteins associated with PS II in addition tolacking the Lhca4 protein associated with PS I (Bossmanet al 1997 Knoetzel et al 1998) The light- andtemperature-sensitive clo f104 mutant lacks a largecomplement of Lhcb1 in addition to lacking the 23 kDLhca2 protein of PS I (Knoetzel amp Simpson 1991Bossman et al 1997) These mutants were chosen for thisstudy because they both exhibit strong variations in theirPS I chlorophyll emission spectral amplitudes at 77 K andsignicentcant diiexclerences in their room temperature steady-state pounduorescence behaviour (Simpson et al 1985 Bassiet al 1985) It was thus of interest to determine andcharacterize the probable relationships between the 77 Kspectral changes due to PS I and the room temperaturepounduorescence parameters More importantly it has alsobeen reported recently that both these mutants exhibitseveral PS-II-related activities that do not change asmight be predicted in conjunction with the loss of Lhcb1namely the light-limited quantum ecurrenciency andpounduorescence lifetimes of PS II and the well-knowntransthylakoid pH- and xanthophyll-cycle-dependentnon-photochemical quenching (Briantais et al 1996Gilmore et al 1996)

Here we report three main observations based on thesimultaneous spectral-kinetic acquisition and DCIanalysis of pounduorescence data collected using a streak-camera spectrograph (Schiller amp Alfano 1980 Tsuchiya1984 Davis amp Parigger 1992 BIumlhler et al 1998) Firstthat under conditions of dark-level pounduorescence both theclo f 2 and clo f104 mutants exhibit very similar PS IIspectral-decay contours to the wild-type (wt) with themain band centred around 685 nm Second that dark-level pounduorescence in the clo f 2 clo f104 and wt is stronglyinpounduenced beyond 700 nm by broad emission bandsfrom PS I that exhibit much more rapid decay kineticsand stronger integrated amplitudes than the main PS IIband at 685 nm Third that clo f 2 compared with clo f104and wt exhibits diiexclerent room temperature PS I emissionspectral bands unique Fo and Fm lifetime distributionsand an inhibited capacity for photoprotective thermalenergy dissipation We conclude that it is imperative forin vivo studies of PS II photochemistry and photo-

protective thermal energy dissipation using chlorophyll apounduorescence to quantify the inpounduence of the prominentPS-I-related spectral emission components

2 MATERIAL AND METHODS

(a) Plant material and experimental detailsSeeds of wt barley (Hordeum vulgare L) and the nuclear gene

mutants clo f 2 and clo f104 were obtained from Professor DSimpson of the Carlsberg Research Laboratories (CopenhagenValby Denmark) Plants were grown in a growth chamber at75 relative humidity under a day^night regime of 16 L 8 Dand 25 8C^20 8C The high (HL) medium (ML) and low (LL)light growth conditions were decentned as 250^270 130^160 and50^60 m molphotons m72 s71 respectively

The chlorophyll antenna sizes of PS II were determinedaccording to the oxygen poundash yield method of Chow et al (1991)The PS I content was obtained from P700 measurementsaccording to Schreiber et al (1988) Chlorophyll determinationsrelevant to the antenna size estimates were performed accordingto the method of Porra et al (1989) Quantitative analysis of thechloroplast pigments was according to the HPLC method ofGilmore amp Yamamoto (1991)

(b) Steady-state chlorophyll a pounduorescencemeasurements

Room temperature measurements of Fm and Fo were madewith a PAM 101-102-103 Model chlorophyll pounduorometer (Heinz-Walz Eiexcleltrich Germany) equipped with a standard 650 nmexcitation diode and RG-9 emission centlter that cuts oiexcl all lightbelow 690 nm and transmits ca 50 at 710 nm andca 90 4720 nm The PAM measuring beam settings andsaturating light pulse intensities for the Fm and Fo determina-tions were as used earlier (Gilmore et al 1996)

The 77 K spectra of intact leaf pieces were performed with anSLM 8100 spectropounduorimeter (Spectronic Institute RochesterNY USA) in photon counting mode Excitation was at 435 nm(peak chlorophyll a absorbance) through an 8 nm slit widthand emission was monitored with a slit width of 2 nm using amonochromator slew rate of 1nm s71 and integration time of 1sThe displayed spectra were centrst smoothed using the exponentialalgorithm by a factor of 04 then normalized to the peak emis-sion channel and centt by minimizing the sum of the squared resi-dual error (SSE) parameter

SSE ˆN

iˆ1

(Di iexcl Mi)2 (1)

where Di is the pounduorescence intensity at wavelength coordinatel i and Mi is the predicted intensity from a model comprisingthe sum of centve free-poundoating Gaussian bands Each Gaussianspectral band was decentned by the following amplitude function

I(l)i ˆ AMP pound exp iexcl12

l iexcl CTRWID

2

(2)

where AMP is the normalized maximum amplitude parameterCTR is the mean or central mode l parameter and WID thewidth is the standard deviation parameter in l units

(c) Time- and wavelength-resolved chlorophyll apounduorescence measurements

Time-resolved emission spectra were collected in photoncounting mode with a model C4334 Streakscope (Hamamatsu





3 RESULTS









typechlarea




(mmol mol chl71)







8

10

wt HL

wt ML

wt LL


clo f104 LLclo f2

(a)

(c)

(b)

(d )

yen

6

4

10

08

06

04

02

00

chl a

b

4

6

8

10

200 300 400 500


200 300 400 500

2

0

F

Fo

mm

aximum

50

40

30

20

10

0

mol (P

SII) -1

DPS

L = - 1833 + 014(Ntot)

VAZ = 557 + 002(Ntot)

f2






100

075

050

025

000

010

000

- 010

010

005

000680 700 720

(a)protein

PS II

PS I

Lhcb1

Lhcb2

Lhcb3

Lhcb6

PsbS

CP43

CP47

level

0

reduced

reduced

0wt

wt

wt

77K band

685 nm

695 nm

Lhca1ndash3

Lhca4

CC1

wt

0wt

730 nm

720 nm

PS I

Lhca1ndash4

CC1

wt

wt

740 nm

720 nm

fluo

resc

ence

inte

nsit

y

100

075

050

025

000

015

005

010

000680 700 720

(b)

fluo

resc

ence

inte

nsit

yR

i

010

000

- 010

Ri

100

075

050

025

000

(c)

fluo

resc

ence

inte

nsit

y

010

000

680 700 720

wavelength (nm)

740 760 780- 010

Ri

protein

PS II

Lhcb1ndash3

CP43

CP47

level

wt

wt

wt

77K band

weak 680 nm

685 nm

695 nm

PS I

Lhca1

Lhca2

wt

reduced740 nm

Lhca4

CC1

wt

wt 720 nm

protein

PS II

Lhcb1

PsbS

CP43

level

reducedwt

wt

77K band











800w

avel

engt

h (n

m)

band

inte

nsity

integrated profile

band profilepu

lse

760

720

680

800

wav

elen

gth

(nm

)

760

720

680

10

05

00

10

08

06

04

02

00

80

60

40

20

0

6

5

4

3

2

1

0

6

5

4

3

2

1

0

6

5

4

3

2

1

0

80

60

40

20

0

80

60

40

20

0

10

08

06

04

02

00

10

08

06

04

02

00

puls

e

10

05

00

800

wav

elen

gth

(nm

)

760

720

680

puls

e 10

0500

500 1000time (ps)

1500 500 1000time (ps)


760 800

(a)

(b)

(e)

(d)

f( )

(g)

(h)

(c)

(i)























4 DISCUSSION





10

006

004

002

000

Fo

0 500(ps)t

( )t

a

1000

8

6

4

2

0

12

08

10

06

04

02

00

12

8

10

6

4

2

0

12

8

10

6

4

2

05000

(a)

10

006

004

002

0000 500

(ps)t

( )t

a

( )t

ta

1000

8

6

4

2

0

(c)

10

006

004

002

0000 500

(ps)t

(ps)t

( )t

a

1000

0 500 1000 1500 2000 0 1000 2000 3000 40002500

8

6

4

2

0

(e)

Fm

(b)

(d )

( f )















leaf w2v L1 w2

v L1 w2v L1 w2

v L1 w2v L1




















ANALYSIS







F(tiexcl t 0)i ˆ

1

0


0 t 4 t 0

(A1)



I(t)i ˆ

t

iexcl1








iˆ1

I(tl)i (A5)









p The least-


w2v ˆ

1N iexcl m iexcl 1

N

iˆ1

(Di iexcl Mi)2

Mi (B1)




L1 ˆN

iˆ1

jRi j (B2)




second w2 ˆ1

N iexcl m iexcl 1

N

iˆ1

(Di iexcl Mi)2

Mi (B3)




WID (B4)





REFERENCES
















































































3 RESULTS









typechlarea




(mmol mol chl71)







8

10

wt HL

wt ML

wt LL


clo f104 LLclo f2

(a)

(c)

(b)

(d )

yen

6

4

10

08

06

04

02

00

chl a

b

4

6

8

10

200 300 400 500


200 300 400 500

2

0

F

Fo

mm

aximum

50

40

30

20

10

0

mol (P

SII) -1

DPS

L = - 1833 + 014(Ntot)

VAZ = 557 + 002(Ntot)

f2






100

075

050

025

000

010

000

- 010

010

005

000680 700 720

(a)protein

PS II

PS I

Lhcb1

Lhcb2

Lhcb3

Lhcb6

PsbS

CP43

CP47

level

0

reduced

reduced

0wt

wt

wt

77K band

685 nm

695 nm

Lhca1ndash3

Lhca4

CC1

wt

0wt

730 nm

720 nm

PS I

Lhca1ndash4

CC1

wt

wt

740 nm

720 nm

fluo

resc

ence

inte

nsit

y

100

075

050

025

000

015

005

010

000680 700 720

(b)

fluo

resc

ence

inte

nsit

yR

i

010

000

- 010

Ri

100

075

050

025

000

(c)

fluo

resc

ence

inte

nsit

y

010

000

680 700 720

wavelength (nm)

740 760 780- 010

Ri

protein

PS II

Lhcb1ndash3

CP43

CP47

level

wt

wt

wt

77K band

weak 680 nm

685 nm

695 nm

PS I

Lhca1

Lhca2

wt

reduced740 nm

Lhca4

CC1

wt

wt 720 nm

protein

PS II

Lhcb1

PsbS

CP43

level

reducedwt

wt

77K band











800w

avel

engt

h (n

m)

band

inte

nsity

integrated profile

band profilepu

lse

760

720

680

800

wav

elen

gth

(nm

)

760

720

680

10

05

00

10

08

06

04

02

00

80

60

40

20

0

6

5

4

3

2

1

0

6

5

4

3

2

1

0

6

5

4

3

2

1

0

80

60

40

20

0

80

60

40

20

0

10

08

06

04

02

00

10

08

06

04

02

00

puls

e

10

05

00

800

wav

elen

gth

(nm

)

760

720

680

puls

e 10

0500

500 1000time (ps)

1500 500 1000time (ps)


760 800

(a)

(b)

(e)

(d)

f( )

(g)

(h)

(c)

(i)























4 DISCUSSION





10

006

004

002

000

Fo

0 500(ps)t

( )t

a

1000

8

6

4

2

0

12

08

10

06

04

02

00

12

8

10

6

4

2

0

12

8

10

6

4

2

05000

(a)

10

006

004

002

0000 500

(ps)t

( )t

a

( )t

ta

1000

8

6

4

2

0

(c)

10

006

004

002

0000 500

(ps)t

(ps)t

( )t

a

1000

0 500 1000 1500 2000 0 1000 2000 3000 40002500

8

6

4

2

0

(e)

Fm

(b)

(d )

( f )















leaf w2v L1 w2

v L1 w2v L1 w2

v L1 w2v L1




















ANALYSIS







F(tiexcl t 0)i ˆ

1

0


0 t 4 t 0

(A1)



I(t)i ˆ

t

iexcl1








iˆ1

I(tl)i (A5)









p The least-


w2v ˆ

1N iexcl m iexcl 1

N

iˆ1

(Di iexcl Mi)2

Mi (B1)




L1 ˆN

iˆ1

jRi j (B2)




second w2 ˆ1

N iexcl m iexcl 1

N

iˆ1

(Di iexcl Mi)2

Mi (B3)




WID (B4)





REFERENCES


















































































8

10

wt HL

wt ML

wt LL


clo f104 LLclo f2

(a)

(c)

(b)

(d )

yen

6

4

10

08

06

04

02

00

chl a

b

4

6

8

10

200 300 400 500


200 300 400 500

2

0

F

Fo

mm

aximum

50

40

30

20

10

0

mol (P

SII) -1

DPS

L = - 1833 + 014(Ntot)

VAZ = 557 + 002(Ntot)

f2






100

075

050

025

000

010

000

- 010

010

005

000680 700 720

(a)protein

PS II

PS I

Lhcb1

Lhcb2

Lhcb3

Lhcb6

PsbS

CP43

CP47

level

0

reduced

reduced

0wt

wt

wt

77K band

685 nm

695 nm

Lhca1ndash3

Lhca4

CC1

wt

0wt

730 nm

720 nm

PS I

Lhca1ndash4

CC1

wt

wt

740 nm

720 nm

fluo

resc

ence

inte

nsit

y

100

075

050

025

000

015

005

010

000680 700 720

(b)

fluo

resc

ence

inte

nsit

yR

i

010

000

- 010

Ri

100

075

050

025

000

(c)

fluo

resc

ence

inte

nsit

y

010

000

680 700 720

wavelength (nm)

740 760 780- 010

Ri

protein

PS II

Lhcb1ndash3

CP43

CP47

level

wt

wt

wt

77K band

weak 680 nm

685 nm

695 nm

PS I

Lhca1

Lhca2

wt

reduced740 nm

Lhca4

CC1

wt

wt 720 nm

protein

PS II

Lhcb1

PsbS

CP43

level

reducedwt

wt

77K band











800w

avel

engt

h (n

m)

band

inte

nsity

integrated profile

band profilepu

lse

760

720

680

800

wav

elen

gth

(nm

)

760

720

680

10

05

00

10

08

06

04

02

00

80

60

40

20

0

6

5

4

3

2

1

0

6

5

4

3

2

1

0

6

5

4

3

2

1

0

80

60

40

20

0

80

60

40

20

0

10

08

06

04

02

00

10

08

06

04

02

00

puls

e

10

05

00

800

wav

elen

gth

(nm

)

760

720

680

puls

e 10

0500

500 1000time (ps)

1500 500 1000time (ps)


760 800

(a)

(b)

(e)

(d)

f( )

(g)

(h)

(c)

(i)























4 DISCUSSION





10

006

004

002

000

Fo

0 500(ps)t

( )t

a

1000

8

6

4

2

0

12

08

10

06

04

02

00

12

8

10

6

4

2

0

12

8

10

6

4

2

05000

(a)

10

006

004

002

0000 500

(ps)t

( )t

a

( )t

ta

1000

8

6

4

2

0

(c)

10

006

004

002

0000 500

(ps)t

(ps)t

( )t

a

1000

0 500 1000 1500 2000 0 1000 2000 3000 40002500

8

6

4

2

0

(e)

Fm

(b)

(d )

( f )















leaf w2v L1 w2

v L1 w2v L1 w2

v L1 w2v L1




















ANALYSIS







F(tiexcl t 0)i ˆ

1

0


0 t 4 t 0

(A1)



I(t)i ˆ

t

iexcl1








iˆ1

I(tl)i (A5)









p The least-


w2v ˆ

1N iexcl m iexcl 1

N

iˆ1

(Di iexcl Mi)2

Mi (B1)




L1 ˆN

iˆ1

jRi j (B2)




second w2 ˆ1

N iexcl m iexcl 1

N

iˆ1

(Di iexcl Mi)2

Mi (B3)




WID (B4)





REFERENCES


















































































100

075

050

025

000

010

000

- 010

010

005

000680 700 720

(a)protein

PS II

PS I

Lhcb1

Lhcb2

Lhcb3

Lhcb6

PsbS

CP43

CP47

level

0

reduced

reduced

0wt

wt

wt

77K band

685 nm

695 nm

Lhca1ndash3

Lhca4

CC1

wt

0wt

730 nm

720 nm

PS I

Lhca1ndash4

CC1

wt

wt

740 nm

720 nm

fluo

resc

ence

inte

nsit

y

100

075

050

025

000

015

005

010

000680 700 720

(b)

fluo

resc

ence

inte

nsit

yR

i

010

000

- 010

Ri

100

075

050

025

000

(c)

fluo

resc

ence

inte

nsit

y

010

000

680 700 720

wavelength (nm)

740 760 780- 010

Ri

protein

PS II

Lhcb1ndash3

CP43

CP47

level

wt

wt

wt

77K band

weak 680 nm

685 nm

695 nm

PS I

Lhca1

Lhca2

wt

reduced740 nm

Lhca4

CC1

wt

wt 720 nm

protein

PS II

Lhcb1

PsbS

CP43

level

reducedwt

wt

77K band











800w

avel

engt

h (n

m)

band

inte

nsity

integrated profile

band profilepu

lse

760

720

680

800

wav

elen

gth

(nm

)

760

720

680

10

05

00

10

08

06

04

02

00

80

60

40

20

0

6

5

4

3

2

1

0

6

5

4

3

2

1

0

6

5

4

3

2

1

0

80

60

40

20

0

80

60

40

20

0

10

08

06

04

02

00

10

08

06

04

02

00

puls

e

10

05

00

800

wav

elen

gth

(nm

)

760

720

680

puls

e 10

0500

500 1000time (ps)

1500 500 1000time (ps)


760 800

(a)

(b)

(e)

(d)

f( )

(g)

(h)

(c)

(i)























4 DISCUSSION





10

006

004

002

000

Fo

0 500(ps)t

( )t

a

1000

8

6

4

2

0

12

08

10

06

04

02

00

12

8

10

6

4

2

0

12

8

10

6

4

2

05000

(a)

10

006

004

002

0000 500

(ps)t

( )t

a

( )t

ta

1000

8

6

4

2

0

(c)

10

006

004

002

0000 500

(ps)t

(ps)t

( )t

a

1000

0 500 1000 1500 2000 0 1000 2000 3000 40002500

8

6

4

2

0

(e)

Fm

(b)

(d )

( f )















leaf w2v L1 w2

v L1 w2v L1 w2

v L1 w2v L1




















ANALYSIS







F(tiexcl t 0)i ˆ

1

0


0 t 4 t 0

(A1)



I(t)i ˆ

t

iexcl1








iˆ1

I(tl)i (A5)









p The least-


w2v ˆ

1N iexcl m iexcl 1

N

iˆ1

(Di iexcl Mi)2

Mi (B1)




L1 ˆN

iˆ1

jRi j (B2)




second w2 ˆ1

N iexcl m iexcl 1

N

iˆ1

(Di iexcl Mi)2

Mi (B3)




WID (B4)





REFERENCES






















































































800w

avel

engt

h (n

m)

band

inte

nsity

integrated profile

band profilepu

lse

760

720

680

800

wav

elen

gth

(nm

)

760

720

680

10

05

00

10

08

06

04

02

00

80

60

40

20

0

6

5

4

3

2

1

0

6

5

4

3

2

1

0

6

5

4

3

2

1

0

80

60

40

20

0

80

60

40

20

0

10

08

06

04

02

00

10

08

06

04

02

00

puls

e

10

05

00

800

wav

elen

gth

(nm

)

760

720

680

puls

e 10

0500

500 1000time (ps)

1500 500 1000time (ps)


760 800

(a)

(b)

(e)

(d)

f( )

(g)

(h)

(c)

(i)























4 DISCUSSION





10

006

004

002

000

Fo

0 500(ps)t

( )t

a

1000

8

6

4

2

0

12

08

10

06

04

02

00

12

8

10

6

4

2

0

12

8

10

6

4

2

05000

(a)

10

006

004

002

0000 500

(ps)t

( )t

a

( )t

ta

1000

8

6

4

2

0

(c)

10

006

004

002

0000 500

(ps)t

(ps)t

( )t

a

1000

0 500 1000 1500 2000 0 1000 2000 3000 40002500

8

6

4

2

0

(e)

Fm

(b)

(d )

( f )















leaf w2v L1 w2

v L1 w2v L1 w2

v L1 w2v L1




















ANALYSIS







F(tiexcl t 0)i ˆ

1

0


0 t 4 t 0

(A1)



I(t)i ˆ

t

iexcl1








iˆ1

I(tl)i (A5)









p The least-


w2v ˆ

1N iexcl m iexcl 1

N

iˆ1

(Di iexcl Mi)2

Mi (B1)




L1 ˆN

iˆ1

jRi j (B2)




second w2 ˆ1

N iexcl m iexcl 1

N

iˆ1

(Di iexcl Mi)2

Mi (B3)




WID (B4)





REFERENCES



































































































4 DISCUSSION





10

006

004

002

000

Fo

0 500(ps)t

( )t

a

1000

8

6

4

2

0

12

08

10

06

04

02

00

12

8

10

6

4

2

0

12

8

10

6

4

2

05000

(a)

10

006

004

002

0000 500

(ps)t

( )t

a

( )t

ta

1000

8

6

4

2

0

(c)

10

006

004

002

0000 500

(ps)t

(ps)t

( )t

a

1000

0 500 1000 1500 2000 0 1000 2000 3000 40002500

8

6

4

2

0

(e)

Fm

(b)

(d )

( f )















leaf w2v L1 w2

v L1 w2v L1 w2

v L1 w2v L1




















ANALYSIS







F(tiexcl t 0)i ˆ

1

0


0 t 4 t 0

(A1)



I(t)i ˆ

t

iexcl1








iˆ1

I(tl)i (A5)









p The least-


w2v ˆ

1N iexcl m iexcl 1

N

iˆ1

(Di iexcl Mi)2

Mi (B1)




L1 ˆN

iˆ1

jRi j (B2)




second w2 ˆ1

N iexcl m iexcl 1

N

iˆ1

(Di iexcl Mi)2

Mi (B3)




WID (B4)





REFERENCES



























































































leaf w2v L1 w2

v L1 w2v L1 w2

v L1 w2v L1




















ANALYSIS







F(tiexcl t 0)i ˆ

1

0


0 t 4 t 0

(A1)



I(t)i ˆ

t

iexcl1








iˆ1

I(tl)i (A5)









p The least-


w2v ˆ

1N iexcl m iexcl 1

N

iˆ1

(Di iexcl Mi)2

Mi (B1)




L1 ˆN

iˆ1

jRi j (B2)




second w2 ˆ1

N iexcl m iexcl 1

N

iˆ1

(Di iexcl Mi)2

Mi (B3)




WID (B4)





REFERENCES































































































ANALYSIS







F(tiexcl t 0)i ˆ

1

0


0 t 4 t 0

(A1)



I(t)i ˆ

t

iexcl1








iˆ1

I(tl)i (A5)









p The least-


w2v ˆ

1N iexcl m iexcl 1

N

iˆ1

(Di iexcl Mi)2

Mi (B1)




L1 ˆN

iˆ1

jRi j (B2)




second w2 ˆ1

N iexcl m iexcl 1

N

iˆ1

(Di iexcl Mi)2

Mi (B3)




WID (B4)





REFERENCES


















































































p The least-


w2v ˆ

1N iexcl m iexcl 1

N

iˆ1

(Di iexcl Mi)2

Mi (B1)




L1 ˆN

iˆ1

jRi j (B2)




second w2 ˆ1

N iexcl m iexcl 1

N

iˆ1

(Di iexcl Mi)2

Mi (B3)




WID (B4)





REFERENCES














































































Global spectral-kinetic analysis of room temperature ... Publications...chlorinaf2(clof2). These...

Documents

Transcript of Global spectral-kinetic analysis of room temperature ... Publications...chlorinaf2(clof2). These...