Proc. Natl. Acad. Sci. USAVol. 92, pp. 636-640, January 1995Plant Biology

Cyanobacterial protein with similarity to the chlorophyll a/bbinding proteins of higher plants: Evolution and regulationNADIA A. MIROSHNICHENKO DOLGANOV*, DEVAKI BHAYAt, AND ARTHUR R. GROSSMAN*t*Department of Plant Biology, The Carnegie Institution of Washington, 290 Panama Street, Stanford, CA 94305; and tCentre for Biotechnology, JawaharlalNehru University, New Delhi 110067, India

Communicated by Olle Bjorkman, Carnegie Institution of Washington, Stanford, CA, September 8, 1994

ABSTRACT We have isolated, from the prokaryotic cya-nobacterium Synechococcus sp. strain PCC 7942, a gene en-coding a protein of 72 amino acids [designated high lightinducible protein (HLIP)J with similarity to the extendedfamily ofeukaryotic chlorophyll a/b binding proteins (CABs).HLIP has a single membrane-spanning av-helix, whereas boththe CABs and the related early light inducible proteins havethree membrane-spanning helices. Hence, HLIP may repre-sent an evolutionary progenitor of the eukaryotic members ofthe CAB extended family. We also show that the gene encodingHLIP is induced by high light and blue/UV-A radiation. Theevolution, regulation, and potential function of HLIP arediscussed.

The light-harvesting antenna of photosynthetic organismshave evolved to bind an array of different pigment moleculeswith various absorption maxima and form a multiproteinordered structure in the membranes. These complexes effi-ciently harvest available light energy and, subsequently, trans-fer that energy to the photosynthetic reaction centers withminimal energy loss (1).The most abundant membrane proteins in plastids are the

light-harvesting chlorophyll (Chl) a/b binding proteins(CABs), associated with photosystem (PS) I and II; thoseassociated with PSII are designated light-harvesting CABs ofPSII (LHCII). The CABs are a highly conserved family ofnuclear-encoded proteins that contain three membrane-spanning helices (MSHs) and bind a minimum of 12 Chls andtwo xanthophylls (2). The recent crystal structure determina-tion of LHCII shows that the first and third MSHs are heldtogether by ion pairs formed by charged residues that also bindChla (3). Residues involved in Chla binding are strictly con-served in all members of the CAB family and in the moredistantly related fucoxanthin Chla/c binding proteins (FCPs)of the chromophytic algae (4). The two xanthophylls arethought to be vital for the prevention ofphotodamage resultingfrom the generation of singlet oxygen species in the PSIIreaction center.The early light induced proteins (ELIPs) are another group

of proteins that are homologous to CABs (5-8). ELIP mRNAincreases transiently during the early stages of the maturationof an etioplast to a chloroplast. ELIPs can also accumulate inmature plastids during exposure of the plants to high light andappear to be associated with PSII reaction centers (9, 10),suggesting a role for ELIPs in the acclimation of plants tohigh-intensity light (high light) (10).We have recently identified in Synechococcus sp. strain PCC

7942 (hereafter, Synechococcus sp.), a small gene (hliA, forhigh light inducible)§ that encodes a polypeptide [designatedthe high light inducible protein (HLIP)] with similarity toMSHs of the CABs. The HLIP has a hydrophobic domain thatcould constitute a membrane-spanning a-helix and has several

residues that are conserved among all of the members of theextended CAB family. Furthermore, the expression of the hliAgene is controlled by light.

MATERIALS AND METHODSCulture Conditions. Cells of Synechococcus sp. were cul-

tured as described (11) at a light intensity of 50 ,tmol m-2.s'1(standard conditions). For high light treatments, culturesgrown to 2-5 x 108 cells per ml were diluted to 5 x 107 cellsper ml and the light intensity was increased to 500 or 1000,tmol.m-2_s-'. For experiments in which j3-glucuronidase(GUS) activity was quantitated, cultures were preincubated for18 h at a fluence of 10 ,umol m2 s' (to reduce backgroundGUS activity), before transfer to high light.DNA Analysis and Sequence Determination. Genomic DNA

was prepared from Synechococcus sp. as described (12). South-ern blot hybridizations were as described by Conley et al. (13).A restriction map of the region containing the hliA gene wasgenerated (see Fig. 1), and the 1.2-kb Pst I fragment wassubcloned into the Bluescript vector (Stratagene). DNA se-quencing was by the dideoxynucleotide chain-terminationmethod (14). Analyses of the nucleotide and the deducedamino acid sequences were performed using the GCG (Mad-ison, WI) software package.RNA Isolation and Primer-Extension Analysis. RNA iso-

lated from Synechococcus sp. (15) was used for Northern blothybridizations (16). A DNA fragment extending from 119 bpupstream of the AUG start codon of the hliA gene to 97 bpdownstream of the translation termination codon was used asa gene-specific probe. The psbAIII gene probe was the anti-sense transcript derived from plasmid pAM090 (17) andsynthesized using an in vitro transcription kit with [a-32P]UTP(New England Nuclear). Primer-extension analysis was asdescribed by Sambrook et al. (16). The oligonucleotide 5'-TTGAGGTTCGATGGCGAAATTATTCTGACGACC-G-3', complementary to nt 57-90 (see Fig. 2A), was used as aprimer for hliA mRNA. The oligonucleotide 5'-CACGGG-TTGGGGTTTCTACAGGACGTAACATAAGGG-3', com-plementary to nt -5 to +31 in the uidA gene (encoding GUS),was used as a primer for the hliA-GUS chimeric mRNA.

Construction of Plasmids Containing hiiA Promoter-GUSFusions. A 2.4-kb EcoRI-HindIII fragment containing theuidA gene from pBIN101.3 (Clontech) was blunt-end-ligatedinto the Sma I-Stu I sites within the polycloning site of theshuttle vector pCB4 (18) to create pGUS267-1. A PCR wascarried out to amplify a 508-bp region of Synechococcus sp.DNA containing the promoter of hliA gene (positions -467 to

Abbreviations: Chl, chlorophyll; CAB, Chla/b binding protein; ELIP,early light inducible protein; FCP, fucoxanthin Chla/c binding protein;GUS, ,3-glucuronidase; HLIP, high light inducible protein; PS, pho-tosystem; LHCII, light-harvesting CABs of PSII; MSH, membrane-spanning helix; ORF, open reading frame.tTo whom reprint requests should be addressed.§The sequence reported in this paper has been deposited in theGenBank data base (accession no. U12333).

636

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement" inaccordance with 18 U.S.C. §1734 solely to indicate this fact.

Dow

nloa

ded

by g

uest

on

Mar

ch 1

4, 2

020

Proc. Natl. Acad. Sci. USA 92 (1995) 637

+41 in Fig. 2A) and was ligated into the Sal I-BamHI sites ofpGUS267-1 at a position 24 bp upstream of the AUG initiatorcodon of uidA to create pHLIP-GUS. The translational fusioncontained 467 bp upstream of the transcription start site, 27 bpof the transcribed untranslated 5' leader, and 15 bp encodingthe first 5 aa of HLIP fused in-frame with the AUG start codonof uidA.

Transformations and GUS Assays. Synechococcus sp. cells,grown to midlogarithmic phase, were transformed with pH-LIP-GUS and pGUS267-1 as described (11). Individual trans-formants were selected on solid BG-11 medium supplementedwith ampicillin at 2 ,ug/ml. GUS activity was determined by theprotocol of Wilson et al. (19) and protein concentrations of celllysates were measured using the Micro BCA protein assay kit(Pierce).

Radiation Sources. Photon fluences for visible and UV-radiation were measured with a quantum photometer (modelLi-185B; Li-Cor, Lincoln, NE) and a portable Spectroradiom-eter (model Li-1800; Li-Cor), respectively. Cells were placed ina thermostated glass chamber (maintained at 32°C) and ex-posed to different intensities of white light (650-W tungstenlamp, Type DWY; General Electric). Blue and red light wereobtained by passing light from a 300-W ELH bulb (Kodak)through blue (no.5113 with maximal transmission at 436 nm;Corning) and red (type B-40 with maximal transmission at 660nm; Balzers) interference filters, respectively. UV-A radia-tion, from 320 to 410 nm with a peak at 366 nm, was obtainedusing self-filtering black-light blue bulbs (no. F1ST8/BLB;General Electric). UV-B radiation, from 290 to 350 nm witha peak at 310 nm, was obtained using a UV-B-specific lamp(no. 3-4404; Fotodyne, New Berlin, WI).

RESULTSSequencing of an 8-kb Sal I genomic fragment isolated duringthe characterization of a filamentous Synechococcus sp. mu-tant (11) revealed the presence of an open reading frame(ORF), designated hMiA, encoding a small polypeptide of 72 aa(HLIP) with significant similarity to the first and third MSHsof the CABs and ELIPs of higher plants. A map of the regioncontaining the hliA gene is shown in Fig. 1. A Southern blotof genomic DNA digested with five endonucleases and hy-bridized at moderate stringency (30% mismatch) with the0.7-kbp EcoRI-Pst I fragment, containing hMiA, is also shown.

H

x

PP B P P EWE P..I. I I .. I

X -

H B E

I I I

kb

10

-4.6

-2.7

-2.3

-1.2

FIG. 1. (Upper) Restriction endonuclease map of the Synechococ-cus sp. genomic region containing the hliA gene. Arrow indicates theposition and transcriptional orientation of hliA. B, BamHI; E, EcoRI;H, HindlIl; P, Pst I; X, Xho I. (Lower) Southern blot hybridization ofSynechococcus sp. genomic DNA. Genomic DNA (3 j.Lg) was digestedwith EcoRI (E), BamHI (B), HindIII (H), Xho I (X), or Pst I (P);separated on a 1% agarose gel; and hybridized with the 0.7-kbpEcoRI-Pst I fragment (see map).

In all five lanes, there is only one hybridizing fragment,suggesting that the hliA gene exists as a single copy on thecyanobacterial genome.The nucleotide sequence of hliA and the deduced amino

acid sequence of HLIP is shown in Fig. 24. A potentialribosome-binding site precedes the translation start site and aninverted repeat is located at the 3' end of the ORF. AKyte-Doolittle hydropathy plot and Chou-Fasman analysisfor a-helices indicate that there is a potential MSH within theregion extending from aa 24 to 55.The hliA gene has similarity to genes encoding characterized

members of the CAB extended family of proteins and to smallORFs present in the GenBank data base. The ORF to whichthe hliA gene bears the strongest similarity (49% identity, 61%similarity) is on the chloroplast genome of the unicellular redalga Cyanidium caldarium (20); it encodes a putative polypep-tide of 50 aa (Fig. 2B). This ORF is 52 bp downstream of thepsbC termination codon and is transcribed in the oppositedirection of the psbD-psbC operon. In the filamentous cya-nobacterium Anabaena sp. strain PCC 7120, an ORF (59 aa)similar to that of hliA (42% identical; Fig. 2B) is 121 bpdownstream of the sigA gene (sigA encodes RNA polymeraseo- factor) (21). An alignment of the Synechococcus sp. HLIPand theAnabaena and Cyanidium ORFs shows that the regionof strongest similarity among the sequences lies between aa 37and 52 (numbering based on Synechococcus sp. HLIP), whichis within the MSH of HLIP. Therefore, the cyanobacterial andCyanidium genes encoding the small CAB-like polypeptidesmay form an HLIP subfamily of the CAB extended family.Only the Synechococcus sp. HLIP has a somewhat hydrophobicdomain of 20 aa at the N terminus.The three sequences discussed above are similar to both

MSH1 and MSH3 of the CABs, ELIPs, FCPs, and PsbS. PsbSis a 22-kDa member of the CAB extended family that isassociated with PSII and has four MSHs (23, 24). MSH1 andMSH3 of the CABs are also similar to each other. Analignment of HLIP with MSH1 and MSH3 of representativemembers of the CAB extended family is shown in Fig. 2C. Thegreatest similarity occurs within a stretch of 16 aa in the middleof MSH1 (aa 63-78 of LHCII) and MSH3 (aa 178-193 ofLHCII). The residues in MSH1 and MSH3 known to bind Chla(3) are Glu-65, His-68, and Gly-78 (of MSH1) and Glu-180,Asn-183, and Gln-197 (of MSH3). The residues Glu-65-Arg-185 and Glu-180-Arg-70 are involved in forming ion pairs thatlink MSH1 and MSH3.

Since hliA appeared to be related to a family of genes ofwhich many are regulated by both light quality and intensity,we examined its expression with respect to light. RNA wasisolated at various times during the exposure of cells to 4 h ofhigh light (500 ,umol m2 s') followed by 2 h of low light (50pmolm-2s-1). Northern blot hybridizations were performedwith ahliA orpsbAIIIgene-specific probe (Fig. 3A). As noted (25,26), the psbAIII gene is strongly induced under high light. ThehliA transcript, -250 bases long, also accumulates after exposureto high light. However, the kinetics of accumulation of the twotranscripts are different. hliA mRNA levels increase within 15min ofthe transfer of cells to high light, accumulate to a maximumwithin 1 h, and decline by 2 h to 25% of the maximum value. Incontrast, thepsbAIII transcript accumulates gradually over a 2-hperiod and remnains at a high level during the entire high lighttreatment. After shifting the cells to low light, both the psbAIIIand hliA transcripts decay rapidly.To investigate wavelengths and intensity-dependent regula-

tion of hMlA, a shuttle vector was constructed in which the uidAreporter gene was fused to the hliA promoter (pHLIP-GUS).Synechococcus sp. cells transformed with pHLIP-GUS or withthe promoterless pGUS267-1 were grown at a light intensity of50 1molm2s1 and then transferred to 500 nolm2s'.Primer-extension analysis was carried out on RNA isolated fromtransformants after 0, 1, 2, 4, and 7 h in high light (Fig. 3B). The

Plant Biology: Dolganov et al.

.........

Dow

nloa

ded

by g

uest

on

Mar

ch 1

4, 2

020

638 Plant Biology: Dolganov et al.

A -467 gggaacattcagcacttcgccagtagtcgccacgtagcccgccaaactat -418

-417

-3 67

-3 68

-3 18

A 1 2 3 4 5 6 7 8 910B W

hijA

-317 cagctggaagatactgcctgcatccgcttgggtcaacgcgatcgccgtat -268

-267 -218 psbAIII

-217 tcccgaaactgagtcagtgggttttcggcgggattgagagggagcgtcat -168

-167 cgcaggcatgaggaaggcgatcgctgcgctcagtcgaggaggcgcaagca -118

-117 gtcgcagcttaccgccagcggcgagaaatcgcagtagggatcaggcttag -68

-67

-17 tgtgttacacttcaaacagaacaaccaaccaaaaggaattcattatgcgt+1 M R

-18

33

34 agcggacggacagttgacgaattcggtcgtcagaataatttcgccatcga 83S G R T V D E F G R Q N N F A I E

84 acctcaagtggttgttgctgaagcagacgcatcctggggcttccatgacc 133P Q V V V A R A D A S W G F H D R

134 gcgccgagaaactgaatggtcgcctggcgatgatcggctttgtggctttgA E K L N G R L A M I G F V A L

183

184 atcctgactgaagttgctttgggtcaaggcctgctgcctttcctcgcgag 233I L T E V A L G Q G L L P F L A S

234 cggcttgagctaattcagtagcagtcaacaacttaagatgatgagtaggg 283G L S *

284 acttactoaagtatcac scagcctctatgatttcggctcagtaccctttt 333

1 50mrsgrtvdefgrqnnfaIepqVVvaeadaSWGFhdrAEklNGRLAMIGFV................... .mqnkDRNtWSWGFTtgAEnWNGRLAMIGFV........... mtdttkIsasVVeDRNsWrWGFTpqAEiWNGRLAMIGF1----------------- I---VV-DRN-WSWGFT--AE-WNGRLAMIGFV

80

4051 . . 72Syn HLIP AliLTEvalGQGILpFLasg1s*Cya ORF sAlvTELitGkGvLHFLG1v*Ana ORF AAtLiELfsGQGfLHFwGi1*

Consensus AA-LTEL--GQG-LHFLG--*

PeaELIP

DunCBRP

SpiPSBS

PhaFCPC

PeaLHCII

(45)

(31)(21)(57)(55)

dlmasgp r a 79)evrsg. it s (64)gginken . 1fv:FS itk| ( 54 )qekfdrlry k gvrlp (91)pet fsknrel ihs clgcvf lsrnj (89)

SynHL IP (29) aswndr5a.UcE3Wiii1tI7alMi (62)

PeaELIPDunCBRPSpiPSBSPhaFCPCPeaLHCII

(117)

( 99)

(125)(160)

(170)

s ksimssdc..sfw aftfv: :Sam 1pj. i LIap1 fEksn. f fs it <1eetic kr i nc fEmv lgvslpeafaelkv smfgffvqaivt&k

** * ~~*A

(150)

(132)

(158)

(194)(204)

FIG. 2. (A) Nucleotide and deduced amino acid sequence ofSynechococcus sp. DNA containing the hliA gene. The arrow indicatesthe transcription start site (position + 1). The potential ribosomebinding site is in boldface type and a stem-loop structure that couldserve in transcription termination is underlined. (B) Comparison ofHLIP from Synechococcus sp. (Syn) to ORFs from Cyanidium cal-darium (Cya) (20) and Anabaena sp. strain PCC 7120 (Ana) (21). Theconsensus sequence is at the bottom. Residues conserved in at leasttwo of the ORFs are capitalized and those conserved in all three are

in boldface type. An asterisk indicates a stop codon. (C) Comparisonof the amino acid sequence of Synechococcus sp. HLIP with MSH1 andMSH3 of representative members of the CAB extended family. Aminoacid sequences of the MSH1 (35 aa) and MSH3 (30 aa plus 5 additionalresidues at the C terminus) of pea ELIP, Dunaliella carotenoid biosyn-thesis related protein (Dun CRBP), spinach (Spi) PsbS protein, Phae-odactylum FCP (Pha FCPC), and pea LHCII are the five lines above andbelow, respectively, the Synechococcus HLIP (Syn HLIP) sequence.Residues conserved in at least five of the sequences are boxed. Thenumbers in parentheses indicate the position of the amino acid residue inthe mature protein. Gaps, represented by dots, have been inserted tooptimize alignments. Stars mark conserved residues involved in Chlbinding and the caret marks the Glu residue conserved in most of theMSHs (see text). GenBank accession nos. for the sequences are as follows:pea ELIP, X05979 (6); Dun CRBP, L23871 (7); Pha FCPC, Z23153 (4);Pea LHCII, K02067 (22); Spi PSBS, X68552 (23).

hliA-GUS transcript was initiated at the same start site as theendogenous transcript (data not shown) and the kinetics of

Time, min

B 5

41

0.0 2.0 4.0Time, h

6.0

FIG. 3. (A) Changes in psbAIII and hliA transcript levels inresponse to light intensity. Cells grown in standard light conditionswere transferred to 500 ,umol m-2-s-1 for 4 h and then placed back inlow light. RNA was extracted from cells harvested at various timesafter the light shifts. Duplicate RNA samples (10 ,ug per lane) were

hybridiied with radiolabeled probes specific for the hliA or psbAIIItranscript. Lanes 1-6 represent RNA from cells after a shift from lowlight to high light for the following times. Lanes: 1, 0 min; 2, 15 min;3, 30 min; 4, 60 min; 5, 120 min; 6, 240 min. After 240 min in high lightthe cells were transferred to low light for the following times. Lanes:7, 15 min; 8, 30 min; 9, 60 min; 10, 120 min. The signal from psbAIII(-) and hliA (0) hybridizations were quantitated using a Phosphor-Imager and are presented as a graph. (B) Primer-extension analysis ofhliA-GUS mRNA. Synechococcus sp. transformied with pHLIP-GUSwas transferred from 50 ,molFm-2 s-1 to 500 ,umol m-2-s-1 and RNAwas isolated 0, 1, 2, 4, and 7 h after the transfer. Primer extension wasperformed with a GUS-specific primer and RNA was isolated at eachtime point. Signals were quantitated using a Phosphorlmager and are

presented as a graph.

accumulation of the hhiA-GUS mRNA were similar to those ofthe endogenous hliA transcript (compare Fig. 3 A and B). ThehliA-GUS transcript does, however, appear to be more stablethan the hliA transcript; the latter is barely detectable after 4in the light while the former is still present at 30% of the maximallevel. Based on this data, it was feasible to use the levels of GUSactivity to measure hliA gene expression.We investigated the effect of various light intensities on hliA

gene expression. Cultures maintained at 10 ,umol m-2_S-1 for18 h were transferred to white light of 100, 500, or 1000,umol m-2.s-1 and the GUS activity was measured 0, 2, 4, and6 h after the various light treatments (Fig. 44). There is adiscernible induction of activity after 2 h at 100 ,umol m-2.s-1with no further increase between 2 and 6 h after the transfer.At 500 ,umol m-2_s-1, there is a 4-fold increase after 2 h andan additional 2-fold increase after 4 h but practically no

increase between 4 and 6 h. Exposure of the cells to 1000gmol-m2-S1 caused an 8-fold induction in GUS activitywithin 2 h and a 27-fold increase by 6 h.

BSyn HLIPCya ORFAna ORFConsensus

C

Proc. NatL Acad ScL USA 92 (1995)

I

r-

a)

1 2 3 4

Dow

nloa

ded

by g

uest

on

Mar

ch 1

4, 2

020

Proc. Natl. Acad Sci. USA 92 (1995) 639

500

OhO2h

* 4h300 - 6h

100P

~.0

WL-100 WL-500 WL-1000

B300 Oh

X2h*4h

200 *6h

100

0R-20 R-100 B-20 B-100 UVA-11 UVA-22

FIG. 4. (A) GUS activity in Synechococcus sp. transformed withpHLIP-GUS or pGUS3267-1 and exposed to various intensities ofwhite light. Cells grown under standard conditions were transferred towhite light at 10 ,Lmol m-2 s-1 for 18 h. Cultures were diluted to 2-5x 107 cells per ml and illuminated with white light at 100 ,mol.m-2 s-I(WL-100), 500 ,molm-2-s-1 (WL-500), or 1000 ,umol-m-2 s-1 (WL-1000). GUS activity was measured 0, 2, 4, and 6 h (as indicated) afterexposure to the different light levels. Data are the average of threeexperiments in which the backgrQund (25-30 nmol per min per mg ofprotein), measured in pGUS267-1 transformants, was subtracted. (B)The effects of light quality and fluence on GUS activity in Synecho-coccus sp. transformed with pHLIP-GUS or pGUS267-1. Cells weregrown as described in Fig. 3, diluted to 2-5 x 107 cells per ml, andilluminated with red, blue, or UV-A light. R-20 and R-100 refer to redlight at 20 and 100 ,umol-m-2s-1, B-20 and B-100 refer to blue lightat 20 and 100 ,umolFm-2.s-', and UVA-11 and UVA-22 refer to UV-Alight at 11 and 22 tLmolm-2s-1, respectively. GUS activities weremeasured 0, 2, 4, and 6 h after the various light treatments andrepresent the average of three experiments in which the background(25-30 nmol per min per mg of protein), measured in pGUS267-1transformants, was subtracted.

We also investigated the effects of different wavelengths onexpression of the hliA-GUS reporter gene. Red light at 20 and100 ,umol.m-2 s-1 caused little induction, whereas blue light at20 and 100 ,mol m2 s1 caused a 3- to 4-fold increase in GUSactivity. The most dramatic induction (10-fold) was caused by6 h of UV-A radiation at 22 gmolm-2s-1. UV-A at 11ILmolm-2s-I also caused induction, but to a lesser extent(6-fold). UV-B light caused extensive bleaching within 2 h andwas not used in further studies.

DISCUSSIONThere has been considerable interest in understanding theevolution of and relationship among the various members ofthe CAB extended protein family (27). This family includesELIPs of higher plants and green algae, FCPs of the chromo-phytic algae, and the PsbS polypeptide of PSII (27). All of theseproteins have sequence similarities and contain three a-helicalMSHs, except for PsbS, which has four. MSH1 and MSH3 aremost conserved among the different family members and arealso similar to each other, suggesting that they arose via a geneduplication (27).The similarity between HLIP and the CABs suggests that

the former is a single helix progenitor of the latter. CABs mayhave evolved from single helix polypeptides in at least twoways. Duplication of an hliA-like gene could have generated a

gene encoding a protein with two similar MSHs. The relativeorientation of two-helix MSHs in the membranes would havebeen opposite to what is observed in CABs. To attain theconformation of the three-helix CABs, an additional MSH(equivalent to MSH2 in CABs) must have arisen later inevolution (3). Alternatively, two genes encoding single-helixproteins (one with similarity to hliA) may have fused to yielda gene encoding a two-helix protein. The gene for the two-helixprotein may have then undergone a tandem duplication yield-ing a sequence encoding a four-helix protein (similar to PsbS).The similarity between CAB and the first three helices of PsbSsuggests that they had a common four-helix ancestor (27). Lossof the fourth MSH could then have yielded the three-MSHstructure of CABs and ELIPs. The evolution of light-harvesting proteins with multiple MSHs that bind pigmentsmay have allowed for more efficient dissipation of excess lightenergy, which would be critical to the development of oxygenicphotosynthesis (3).The amino acids that bind Chla and that form ion pairs

between MSH1 and MSH3 have been determined from thethree-dimensional structure of LHCII (3). It is striking thatfive of the eight residues involved in Chla binding, Glu-65,Arg-70, Glu-180, Arg-185, and Asn-183 (numbering based onLHCII), are conserved in the ELIPs, FCPs, and CABs. TheHLIPs also have the conserved Glu and Arg residues. Of theother Chla binding residues, His-68 is present in the FCPs butis replaced by an Asn in the HLIPs and ELIPs; Gly-78 andGln-197 are not conserved among the family members. How-ever, a conserved Glu residue just 1 aa away from Gln-197 ispresent in MSH3 of ELIPs and PsbS and is at an analogousposition in MSH1 (position 83) and the MSH of the HLIP; thisGln -> Glu substitution may represent a functionally conser-vative change. Hence, there is strong conservation, among theCAB family members, of specific residues crucial for both Chlabinding and the ion pairing that stabilizes interactions betweenMSH1 and MSH3. Therefore, it seems likely that ELIPs andHLIP bind Chla. It has been established that PsbS can bind Chl(28). It is also conceivable that the HLIP forms a homodimerin the thylakoid membranes via the Glu and Arg residuesanalogous to those that facilitate the pairing of MSH1 andMSH3 of LHCII polypeptides. Since the ion pairing would beintermolecular for HLIPs, rather than intramolecular, theassociation might be more labile (Chla binding might also bemore labile).A number of genes encoding FCPs and CABs are regulated

by light (29, 30). ELIPs, which appear to be closely associatedwith PSII (9, 31), accumulate in mature pea leaves subjectedto high light; the most pronounced induction is in high-fluenceblue/UV-A radiation. An ELIP from Dunaliella increases inabundance when the alga is exposed to high light (7) and maybe a zeaxanthin-binding protein (32). The desiccation-tolerantplant Craterostigma exhibits light-dependent synthesis of anELIP during dessication; it may help stabilize the thylakoidmembranes during dessication. ELIPs also accumulate whenplants are treated with inhibitors of carotenoid biosynthesis,which might make the plants more susceptible to high lightdamage (33). These results suggest that ELIPs are involved inthe acclimation of plants to high light (10, 33, 34).

Like ELIPs, the level of HLIP may also increase in high lightsince high light triggers the elevated accumulation of tran-scripts from hMiA. Since the hliA transcript accumulates tran-siently after exposure to high light, ULIP is probably onlysynthesized at high levels during the initial exposure to highlight. Once the HLIP has accumulated in the membranes,perhaps saturating its site of assembly and function, there maybe diminished transcription from hliA.

Increased expression from hliA in high light suggests a rolefor HLIP in the acclimation of cyanobacteria to phototoxicconditions. However, the role of HLIP in photoprotection isunclear since a strain in which the hliA gene was disrupted

Plant Biology: Dolganov et aL

Dow

nloa

ded

by g

uest

on

Mar

ch 1

4, 2

020

640 Plant Biology: Dolganov et al.

exhibited no unusual phenotype when grown in high or lowlight (unpublished data).Genes that are regulated similarly by light may have com-

mon regulatory elements. In cyanobacteria the genes encodingDl (psbA) and D2 (psbD) of the PSII reaction centers arepresent in multiple copies (26, 35); some copies of these genesare specifically activated by a short exposure to high light (35).The single copy psbD-psbC operon on the plastid genome ofbarley was shown to be induced by blue/UV-radiation (36-38).Transcription from the psbD-psbC operon is controlled by alight-responsive promoter (LRP) that is conserved betweendicots and monocots (39). The region upstream of the hliAtranscription start site has no similarity to the promoterregions of the cyanobacterial light-regulated psbA and psbDgenes or to the conserved LRP elements.

We thank Olle Bjorkman and David Fork for helpful discussion,Emmauel Liscum and Julie des Rosiers for technical advice, andKrishna Niyogi for help with preparing one of the figures. SusanGolden provided pAM090 and David Laudenbach providedpGUS267-1. The work was partly supported by National ScienceFoundation Grant DCB 8916301 to A.R.G. This is Carnegie Instituteof Washington Publication 1226.

1. Zuber, H. & Brunisholz, R. A. (1991) in Chlorophylls, ed. Scheer,H. (CRC, Boca Raton, FL), pp. 627-703.

2. Thornber, J. P., Cogdell, R. J., Chitnis, P., Morishige, D. T.,Peter, G. F., Gomez, S., Anandan, S., Preiss, S., Welty, B., Lee,A., Takeuchi, T. & Kerfeld, C. (1994) in Advances in Molecularand Cell Biology, ed. Barber, J. (JAI, Greenwich, CT), Vol 10, pp.55-118.

3. Kuhlbrandt, W., Wang, D. N. & Fujiyoshi, Y. (1994) Nature(London) 367, 614-621.

4. Bhaya, D. & Grossman, A. R. (1993) Nucleic Acids Res. 21,4458-4468.

5. Grimm, B. & Kloppstech, K. K. (1987) Eur. J. Biochem. 167,493-499.

6. Kolanus, W., Scharnhorst, C., Kuehne, U. & Herzfeld, F. (1987)Mol. Gen. Genet. 209, 234-239.

7. Lers, A., Levy, H. & Zamir, A. (1991) J. Biol. Chem. 266,13698-13705.

8. Bartels, D., Hanke, C., Schneider, K., Michel, D. & Salamini, F.(1992) EMBO J. 11, 2771-2778.

9. Adamska, I. & Kloppstech, K. (1991) Plant Mol. Biol. 16,209-223.

10. Adamska, I., Ohad, I. & Kloppstech, K. (1992) Proc. Natl. Acad.Sci. USA 89, 2610-2613.

11. Dolganov, N. & Grossman, A. R. (1993) J. Bacteriol. 175, 7644-7651.

12. Tandeau de Marsac, N., Borrias, W. E., Kuhlemeier, C. J.,Castets, A. M., van Arkel, G. A. & van den Hondel, C. A. M. J. J.(1982) Gene 20, 111-119.

13. Conley, P. B., Lemaux, P. G. & Grossman, A. R. (1985) Science230, 550-553.

14. Sanger, F., Nicklen, S. & Coulson, A. R. (1977) Proc. Natl. Acad.Sci. USA 74, 5463-5467.

15. Golden, S. S., Brusslan, J. & Haselkorn, R. (1987) MethodsEnzymol. 153, 215-231.

16. Sambrook, J., Fritsch, E. F. & Maniatis, T. (1989) MolecularCloning: A Laboratory Manual (Cold Spring Harbor Lab. Press,Plainview, NY), pp. 5.78-5.79.

17. Bustos, S. A., Schaefer, M. R. & Golden, S. S. (1990) J. Bacteriol.172, 1998-2004.

18. Gendel, S., Straus, N., Pulleybank, D. & Williams, J. (1983) J.Bacteriol. 156, 148-154.

19. Wilson, K. J., Hughes, S. G. & Jefferson, R. A. (1992) in GUSProtocols: Using the GUS Gene as a Reporter of Gene Expression,ed. Gallagher, S. R. (Academic, San Diego), pp. 1-20.

20. Maid, U. & Zetsche, K. (1992) Plant Mol. Biol. 19, 1001-1010.21. Brahamsha, B. & Haselkorn, R. (1991) J. Bacteriol. 173, 2442-

2450.22. Cashmore, A. R. (1984) Proc. Natl. Acad. Sci. USA 81, 2960-

2964.23. Wedel, N., Klein, R., Ljungberg, U., Andersson, B. & Herrmann,

R. G. (1992) FEBS Lett. 314, 61-66.24. Kim, S., Sandusky, P., Bowlby, N. R., Aebersold, R., Green,

B. R., Vlahakis, S., Yocum, C. F. & Pichersky, E. (1992) FEBSLett. 314, 67-71.

25. Kulkarni, R. D., Schaefer, M. R. & Golden, S. S. (1992) J.Bacteriol. 174, 3775-3781.

26. Kulkarni, R. D. & Golden, S. S. (1994)J. Bacteriol. 176, 959-965.27. Green, B. R. & Pichersky, E. (1994) Photosynth. Res. 39,149-162.28. Funk, C., Schroeder, W. P., Green, B. R., Renger, G. & Ander-

son, B. (1994) FEBS Lett. 342, 261-266.29. Tobin, E. M. & Kehoe, D. M. (1994) Semin. Cell Biol. 5,335-346.30. Friedman, A. L. & Alberte, R. S. (1986) Plant Physiol. 80, 43-51.31. Kruse, E. & Kloppstech, K. (1992) Eur. J. Biochem. 208, 195-202.32. Levy, H., Tal, T., Shaish, A. & Zamir, A. (1993) J. Biol. Chem.

268, 20892-20896.33. Adamska, I., Kloppstech, K. & Ohad, I. (1992)1. Biol. Chem. 267,

24732-24737.34. Potter, E. & Kloppstech, K. (1993) Eur. J. Biochem. 214,779-786.35. Bustos, S. A. & Golden, S. S. (1992) Mol. Gen. Genet. 232,

221-230.36. Gamble, P. E. & Mullet, J. E. (1989) EMBO J. 8, 2785-2794.37. Sexton, T. B., Christopher, D. A. & Mullet, J. E. (1990)EMBO J.

9, 4485-4494.38. Christopher, D. A. & Mullet, J. E. (1994) Plant Physiol. 104,

1119-1129.39. Christopher, D. A., Kim, M. & Mullet, J. E. (1992) Plant Cell 4,

785-798.

Proc. NatL Acad ScL USA 92 (1995)

Dow

nloa

ded

by g

uest

on

Mar

ch 1

4, 2

020