EVOLUTION AND POPULATION GENETICS OF PROCHLOROCOCCUSMARINUS

by Ena Urbach B.S. Molecular Biophysics and Biochemistry Yale University, 1980

Submitted in partial fulfillment of the requirements of the

DEGREE OF DOCTOR OF PHILOSOPHY in Biology and Civil and Environmental Engineering at the Massachusetts Institute of Technology February 1995 @ Massachusetts Institute of Technology 1995

Signature of Author ,_ _ _ Departments of Biology and Civil and Environmental Engineering

Certified by Sallie W. Chisholm, Professor BiQlogv and Civil and Environmental Engineering

Accepted b) Departmental Committee on Graduate Studies Civil and Environmental Engineering Joseph M.Sussman

Accepted b3 I A- gl 'I. )"

tal Committee on Graduate Stuaies Departm Biology Jo-Ann Murray

r..'I : i i EVOLUTION AND POPULATION GENETICS OF PROCHLOROCOCCUSMARINUS by Ena Urbach Submitted to the Departments of Biology and Civil and Environmental Engineering February, 1995, in partial fulfillment of the requirements for the Degree of Doctor of Philosophy in Biology and Civil and Environmental Engineering

ABSTRACT Prochlorococcusmarinus is a tiny, photosynthetic marine prokaryote containing the unusual photosynthetic pigments divinyl chlorophylls a and b, and lacking phycobilisomes, the light harvesting apparatus found in most cyanobacteria. The unique pigments of this organism suggest a phylogenetic affinity with the Prochlorophyta, a recently described group of (monovinyl) chlorophyll a and b-containing prokaryotes which also lack phycobilisomes. The Prochlorophyta have been proposed as a monophyletic group sharing a common ancestry with "green" chloroplasts in algae and higher plants. Phylogenetic analysis of relationships among P. marinus, the prochlorophytes Prochloronsp. and Prochlorothrixhollandica, and other members of the cyanobacterial radiation indicates that the distribution of prochlorophytes within the cyanobacterial radiation is polyphyletic, and that none of the known chlorophyll b-containing prokaryotes is specifically related to chloroplasts. This result, inferred from 16S ribosomal RNA (rRNA) sequences, has been confirmed by a parallel, independent study using rpoC gene sequences. The polyphyletic distribution of prochlorophytes and green chloroplasts among the cyanobacteria implies that photosynthetic systems employing (monovinyl and divinyl) chlorophyll b arose several times during the evolution of photosynthetic organisms. However, since the publication of these results it has been pointed out that they are also consistent with the existence of an ancestral cyanobacterium containing both chlorophyll b and phycobilisomes, with subsequent loss of one or the other light harvesting system in all known surviving lineages. P. marinus is a major constituent of the photosynthetic picoplankton across wide regions of the tropical and subtropical open oceans, with cell densities of 105 cells/ml frequently encountered in both the Sargasso Sea and the Pacific. Cultured strains of this ecologically important organism isolated from Sargasso Sea, north Atlantic, Mediterranean and Pacific waters were found by phylogenetic analysis of 16S rRNA, psbB and petB and D sequences to belong to a single lineage within the cyanobacteria. P. marinus shares this lineage with strains of marine A Synechococcus, a planktonic cyanobacterium which also shares its open ocean habitat. An investigation into the genetic structure of P. marinus field populations in depth profiles from the Sargasso Sea and the Gulf Stream revealed a high degree of genetic heterogeneity within water samples, detected by partial sequencing of cloned PCR products containing portions of the single-copy genes petB and D and their intergenic region, amplified from flow cytometrically sorted cells. Populations within water samples contained a minumum of six alleles recovered from a sample of 23 clones. The presence of numerous additional alleles were implied by rarefaction analyses, which indicated that diversity was incompletely sampled in datasets containing 19 to 28 clones per water sample. Overlapping sets of alleles were recovered from the two water columns, from different depths within each water column and from flow cytometrically distinguishable subpopulations within water samples, suggesting that each of these populations drew its membership from a single gene pool. Thesis Subervisor: Prof. Sallie W. Chisholm Title: Professor of Biology and Civil and Environmental Engineering ACKNOWLEDGEMENTS

I would like to thank my thesis advisor, Penny Chisholm, for allowing me an unusual degree of intellectual freedom in pursuing what I wanted to study. Penny's intellectual courage has allowed her to develop a laboratory with an extraordinary range of skills and interests. I would also like to thank the members of my thesis committee, past and present, who have helped me with sharp critical eyes and large amount of patience: Alan Grossman and Ethan Signer of the Biology Department, Bill Thilly of the Department of Civil and Environmental Engineering, Bruce Levin of the Zoology Departement at University of Massachusetts at Amherst (and currently at Emory University), John Waterbury of the Woods Hole Oceanographic Institution and Mitchel Sogin of the Center for Molecular Evolution at the Marine Biological Laboratories. Special thanks to Bill Thilly for his generous contributions of laboratory space, equipment and supplies and for his intellectual input and to Mitchel Sogin for generous contributions of his phylogenetic experitise and computer time. Over the years at MIT I have been fortunate to learn from a number of talented individuals who have been generous in sharing their technical expertise. Brian Binder, Phouthone Keohavong, Jasper Reese, Jason Seaman, Michael Way and Sheila Frankel have all served as both teacher and friend, and without them I would never have been able to complete this thesis. Samantha Roberts has been a great companion and source of information as the other "molecular person" in the Parsons lab. I would like to thank Cremilda Dias for her excellent technical help in obtaining four of the 16S ribosomal gene seuqneces used in Chapter Two. I have also had the good fortune to interact with the unflaggingly friendly and stimulating members of the Chisholm laboratory: Ginger Armbrust, Lana Arras, Kent Barres, Brian Binder, Jim Bowen, Jeff Dusenberry, Sheila Frankel, Mark Gerath, Karina Gin, Liz Mann and Lisa Moore. Finally, I would like to acknowledge the financial support provided by NSF (BSR-9020254) and by NIH through a training grant to the Department of Biology. TABLE OF CONTENTS

Page A BSTRA CT ...... 2 ACKNOWLEDGEMENTS ...... 4 INTRODUCTION ...... 11 R eferences ...... 16 CHAPTER ONE: MULTIPLE EVOLUTIONARY ORIGINS OF ...... 20 PROCHLOROPHYTES WITHIN THE CYANOBACTERIAL RADIATION

CHAPTER TWO: PHYLOGENETIC RELATIONSHIPS AMONG ...... 25 CULTURED STRAINS OF PROCHLOROCOCCUSMARINUS ISOLATED FROM DIVERSE OCEANIC PROVINCES

Abstract ...... 26 Introduction ...... 26 Background ...... 28 M ethods ...... 35 Results ...... 37 Discussion ...... 58 References ...... 61 CHAPTER THREE: GENETIC DIVERSITY IN NATURAL ...... 67 POPULATIONS OF PROCHLOROCOCCUS MARINUS IN THE SARGASSO SEA AND THE GULF STREAM

Abstract ...... 68 Introduction ...... 68 M ethods ...... 73 Results ...... 82 Discussion...... 108 References ...... 113

CHAPTER FOUR: EPILOGUE: RESPONSES IN THE LITERATURE ... 116 TO THE PUBLICATION OF CHAPTER ONE: MULTIPLE ORIGINS OF PROCHLOROPHYTES WITHIN THE CYANOBACTERIAL RADIATION

References ...... 120 Page APPENDIX I: PRELIMINARY ANALYSIS OF GENETIC ...... 122 DIVERSITY IN GULF STREAM POPULATIONS OF PROCHLOROCOCCUSMARINUS BY DIRECT SEQUENCING OF PCR PRODUCTS AMPLIFIED FROM FLOW CYTOMETRICALLY SORTED CELLS

Reference ...... 129

APPENDIX II: JUSTIFICATION FOR THE • 3.5% SEQUENCE ...... 130 DIFFERENCE CRITERION FOR SEQUENCES ASSIGNED TO A SINGLE ALLELE References ...... 147 APPENDIX III: COMPARISON OF NUCLEOTIDE ...... 148 DIFFERENCES AMONG CLONES ASSIGNED TO ALLELE 1 TO KNOWN PATTERNS OF NUCLEOTIDE SUBSTITUTION ERROR BY TAQ POLYMERASE: FOR DATA OF CHAPTER THREE

References ...... 152

APPENDIX IV: ALIGNED SEQUENCE DATA...... 153 LIST OF FIGURES Page CHAPTER ONE

Figure 1. Nucleic-acid sequence phylogenies, inferred from 16S ...... 20 rRNA data, illustrating the apparent independent appearance of chlorophyll b-containing organisms in the cyanobacterial lineage.

Figure 2. Phylogenetic tree illustrating relationships among ...... 21 prochlorophytes, members of the Synechococcus group and shotgun-cloned sequences from Sargasso Sea plankton(SAR) and Pacific Ocean picoplankton (ALO). CHAPTER TWO

Figure 1. Phylogenetic relationships among cultured strains of ...... 38 P. marinus, cloned sequences from Sargasso Sea picoplankton and other members of the cyanobacterial lineage, inferred from 16S rRNA gene sequences using Agrobacteriumtumefaciens as an outgroup.

Figure 2. Phylogenetic relationships among cultured strains of ...... 40 P. marinus and other members of the cyanobacterial lineage inferred using psbB sequences. Figure 3. Phylogenetic relationships among cultured strains of ...... 42 P. marinus and other members of the cyanobacterial lineage inferred using petB/D sequences. Figure 4. Phylogenetic relationships among P. marinus cultured ...... 50 strains and diverse cyanobacterial taxa, inferred by neighbor joining using 16S rRNA gene sequences.

Figure 5. Comparison of petB/D intergenic region sequences for ...... 56 cultured strains of P. marinus and other members of the oxygenic photosynthetic radiation. CHAPTER THREE

Figure 1. Depth profiles of temperature and in vivo chlorophyll ...... 74 fluorescence at the Sargasso Sea and Gulf Stream sampling sites.

Figure 2. Contour plots for flow cytometric distributions of...... 77 P. marinus sorted from Sargasso Sea and Gulf Stream field populations. Page Figure 3. Intergenic region sequences for 71 petBID alleles recovered ...... 86 from the Sargasso Sea and Gulf Stream field samples, cultured P. marinus strains Med4, SS120, FP5, MIT9107 and MIT9313, Synechococcus strains WH8103 and PCC7002, cyanobacteria Prochlorothrixhollandica and Nostoc PCC7906 and chloroplasts from Chlorella protothecoides,Marchantia polymorpha and Zea maize. Figure 4. Frequency distributions for nucleotide mismatch in all ...... 91 pairwise comparisons of allele prototype sequences. Figure 5. Phylogenetic relationships among 20 petBID alleles cloned ...... 96 from flow cytometrically sorted populations in the Sargasso Sea and the Gulf Stream, cultured strains of P. marinus and other members of the oxygenic phototroph radiation, inferred by neighbor joining using 163 nucleotides at the first two codon positions of petB and petD. Figure 6. Rarefaction curves for clones recovered from a) individual ...... 102 sorted samples, b) lumped samples containing both members of flow cytometric double populations or constituents of entire water columns and c) a lumped sample containing the entire dataset. Figure 7. Frequency of Allele 1-like alleles in sorted samples...... 111

APPENDIX I Figure 1. Autoradiograms from direct sequencing of petB/D PCR ...... 127 products amplified from flow cytometrically sorted P. marinus from natural populations in the Gulf Stream and from unsorted, cultured P. marinus MIT9313, isolated from the 135 m Gulf Stream water sample. APPENDIX II Figure 1. Ratio of third codon position mismatch to all codon ...... 139 position mismatch, plotted against total mismatch for comparisons of clones with alignable intergenic regions in the dataset of Chapter 3. Figure 2. Optimization of the criterion identifying the upper bound ...... 142 for total sequence mismatch considered indistinguishable from PCR+SD error.

Figure 3. Data from Figure 1 divided into two groups by the criterion .... 144 drawn at 3.5% total mismatch. Page APPENDIX IV Figure 1. 16S rRNA sequence data used to infer trees in Chapter ...... 154 Two, Figure 1. Figure 2. psbB sequence data used to infer trees in Chapter Two...... 162 Figure 3. petB/D sequence data used to infer trees in Chapter Two...... 170 Figure 4. petB/D sequence data for alleles cloned from flow ...... 178 cytometrically sorted cells from the Gulf Stream and the Sargasso Sea and from cultured cells, used for analyses in Chapter Three. LIST OF TABLES Page CHAPTER ONE Table 1. Evolutionary distance and fractional similarity matrix for ...... 20 16S rRNA sequences from A. tumefaciens and members of the prokaryotic oxygenic phototroph radiation Table 2. Evolutionary distance and fractional similarity matrix for...... 21 16S rRNA sequences from A. tumefaciens, members of the prokaryotic oxygenic phototroph radiation and shogun clones from marine plankton communities. CHAPTER TWO Table 1. Prochlorococcusmarinus strains used in this study...... 30 Table 2a. Genetic distance and fractional similarity for pairwise ...... 44 comparisons of 16S rRNA sequences. Table 2b. Evolutionary distance and fractional similarity at all, first ...... 45 two and third codon positions for pairwise comparisons of psbB sequences. Table 2c. Evolutionary distance and fractional similarity at all, first ...... 47 two and third codon positions for pairwise comparisons of petBID sequences. Table 3. G+C base compositions of sequences used in these ...... 55 analyses. CHAPTER THREE Table 1. Frequency of alleles among cloned amplification ...... 83 products from flow cytometrically sorted samples. Table 2. Pairwise comparisons among alleles identified by ...... 84 subdivision of sets of clones with similar intergenic regions. Table 3. Fractional mismatch at first two codon positions for all ...... 100 alleles not included in the phylogenetic tree with sequences from cultured organisms, chloroplasts and selected alleles included in the tree. Table 4. Presence of alleles found in multiple samples...... 107 APPENDIX II Table 1. Theoretical statistics for error in PCR product sequences...... 135 for the range of reported Taq polymerase error rates and an estimated range of sequence determination errors. Page APPENDIX III

Table 1. Nucleotide mismatches between clones assigned to Allelel ..... 151 and the Allele 1 prototype sequence, excluding G->A at position 1243. INTRODUCTION G. Ledyard Stebbins has pointed out that "the ultimate aim of scientific endeavor in any discipline is to obtain facts by reductionist methods and use them to synthesize broadly integrated theories that provide new insights into the world of nature" (Stebbins 1982). The evolutionary and population genetic studies that comprise this thesis have been designed to obtain, by the reductionist methods of DNA sequencing and phylogenetic analysis, information leading to new theories about the evolution of light harvesting pigments in oxygenic photosynthetic organisms, and about the population structure of a prokaryotic phytoplankter which inhabits the largest and possibly best- mixed environment on earth, the tropical and subtropical open oceans.

Procholorococcusmarinus is a eubacterial photosynthetic plankter less than 0.8 pm in diameter which contains the unusual photosynthetic pigments divinyl chlorophylls a and b (chl a2 and b2), and lacks phycobilisomes, multicomponent protein and pigment complexes used for light harvesting by most cyanobacteria (Chisholm et al 1988, 1992, Goericke and Repeta 1992). Upon its discovery in 1988, the unusual pigments found in P. marinus suggested a taxonomic affinity with the newly recognized Prochlorophyta (Lewin 1977), photosynthetic prokaryotes containing "normal" chlorophylls a and b (chl al and bl) and lacking phycobilisomes (Chisholm et al. 1988).

According to the endosymbiotic hypothesis proposed by Mereschkowsky (1905, 1910), plastids of higher plants and green algae originated in a free living prokaryote containing green pigments, and which came to live endosymbiotically within the cytoplasm of a eukaryote. Over the course of the symbiosis, the chl a and b-containing prokaryote degenerated into an organelle incapable of independent survival (Margulis 1970, 1981, Raven 1970). At the time research for this thesis began, the evolutionary origin of photosynthetic organelles within the cyanobacteria had become well established by molecular phylogenetic analyses (Fox et al. 1980, Woese 1987, Giovannoni et al. 13 1988, Van den Eynde et al. 1988, Witt and Stackebrandt 1988, Ludwig et al. 1990, Valentin and Zeitsche 1990a, b), but whether the chlorophyll b photosynthetic antenna had been "invented" by an intracellular prokaryote containing phycobilisomes or arose by endosymbiosis of a free-living, chl b-containing prokaryote remained to be established

(Cavalier-Smith 1982). The recently discovered chlorophyll b-containing prokaryotes Prochlorondidemni (Lewin 1977, Lewin and Cheng 1989), and Prochlorothrix hollandica (Burger-Wiersma et al. 1986) had been proposed as descendants of the putative free-living, chlorophyll b-containing ancestor to green chloroplasts, but P. hollandica had subsequently been shown to branch separately from cholorplasts in 16S rRNA and psbA phylogenetic trees (Turner et al. 1989, Morden and Golden 1989a, b).

Chapter One of this thesis contributed to the formulation of new hypotheses about the origin of chlorophyll b-containing photosynthetic antennae by providing a 16S rRNA phylogenetic analysis of relationships among the prochlorophytes P. marinus,

Prochloronsp., P. hollandica,chloroplasts and other cyanobacteria (Urbach et al. 1992). The results were congruent with those of an independent study using rpoC sequences

(Palenik and Haselkorn 1992). Chapter Four is an Epilogue discussing responses in the literature to this publication.

Chapter Two of this thesis is a phylogenetic study which infers evolutionary relationships among P. marinus cultures isolated from diverse ocean provinces, expanding on the previous work. Phylogenetic patterns inferred from three sequences, 16S rRNA, psbB and petBID, are used to address questions of whether oceanic prochlorophytes derive from multiple lineages dispersed among the cyanobacteria, whether phylogenetic patterns for the three genes are consistent, and whether they reflect geographic relationships among sites of P. marinus culture isolation. In addition, this study provides sequence data which can be used to link the properties of P. marinus in 14 culture to those of populations in the sea and a phylogenetic framework for interpretation of the evolution of characteristics among P. marinus cultures.

P. marinus is a major constituent of the picophytoplankton in tropical and subtropical regions of the open ocean, with depth integrated cell abundances in the Sargasso Sea and north Pacific of 5 x 108 and 2 x 109 cells cm -2, respectively (Olson et al. 1990, Cambell and Vaulot 1993) and a global population of approximately 1026 cells (according to a back of the envelope calculation which assumes a tropical and subtropical open ocean area of 2 x 1014 m2, an average P. marinus cell density of 104 cells/ml and a euphotic zone depth of 150 m). Experiments with cloned, cultured isolates have established that individual P. marinus genotypes can grow over wide ranges of light intensity and temperature, but differ in their growth potential at environmentally relevant extremes of high and low light (Partensky et al. 1993, Moore et al. 1994). In order to enable future workers to interpret the patterns of distribution and abundance in P. marinus field populations and to use the properties of P. marinus in culture to predict the growth and succession of populations in the wild, it is necessary to assess the patterns of distribution of P. marinus genotypes in natural populations (Wood 1988, Wood and Leatham 1992, Weisse 1993). Specifically, it is necessary to answer the questions, are P. marinus natural populations genetically heterogeneous, and what is the temporal and spatial scale of genetic variability among populations?

Chapter Three is a direct analysis of P. marinus populations in which the questions of genetic heterogeneity and differences among populations at different depths and in different water masses are addressed by sequencing cloned petBID PCR products amplified from natural populations sorted by flow cytometry. With the basic outline of the population structure of P. marinus known, it becomes possible to formulate improved hypotheses about properties and processes involving this important component of the marine food web. REFERENCES Bryant, D.A. (1992). Puzzles of chloroplast ancestry. Curr. Biol. 2:240-242. Burger-Wiersma, T., Veenhuis, M., Korthals, J., Van der Weil, C.C.M. and Mur, L (1986). A new prokaryote containing chlorophylls a and b. Nature 320:262-294. Campbell, L., and Vaulot, D. (1993). Photosynthetic picoplankton community structure in the subtropical North Pacific Ocean near Hawii (station ALOHA). Deep-Sea Res. 40:2043-2060. Chisholm, S.W., R.J. Olson, E.R. Zettler, R. Goericke, J. Waterbury, and N. Welschmeyer. (1988). A novel free-living prochlorophyte abundant in the oceanic euphotic zone. Nature, 334(6180):340-343. Chisholm, S.W., Frankel, S.L., Goericke, R., Olson, R.J., Palenik, B., Waterbury, J.B., West-Johnsrud, L. and Zettler, E.R. (1992). Prochlorococcus marinus nov. gen nov. sp.: a marine prokaryote containing divinyl chlorophyll a and b. Arch. Microbiol. 157:297-300. Cavalier-Smith, T. (1982). The origins of plastids. Biol. J. Linn. Soc. 17:289-306> Fox, G., Stackebrandt, E., Hespell, R., Gibson, J., Maniloff, J., Dyer, T., Wolf, R., Balch, W., Tanner, R., Magrum, L., Zablen, L., Blekemore, R., Gupta, R., Bonen, L., Lewis, B., Stahl, D., Luehrsen, K., Chen, K. and Woese, C. (1980). The phylogeny of prokaryotes. Science 209:457-463. Giovannoni, S.J., Turner, S., Olsen, G.J., Barnes, S., Lane, D.J. and Pace, N.R. (1988). Evolutionary relationships among cyanobacteria and green chloroplasts. J. Bacteriol. 170:3584-3592. Goericke, R. and Repeta, D.J. (1992). The pigments of Procholorcoccus marinus: the presence of divinyl chlorophyll a and b in a marine prokaryote. Limnol. Oceanogr. 37:425-433. Lewin, R.A. (1977). Prochloron, type genus of the Prochlorophyta. Phycologia 16:217. Lewin, R.A. (1981). Prochloron and the theory of symbiogenesis. Ann. N.Y. Acad. Sci. 361:325-329 Lewin, R.A. and Cheng, L. (1989). Introduction. In Prochloron: a microbial enigma. Lewin, R.A., and L. Cheng, eds. Chapman and Hall. Ludwig, W., Weizenegger, D., Betzel, D., Leidel, E., Lenz, T., Ludvigsen, D., Mollenhoff, D., Wenzig, P. and Schleifer, K.H. (1990). Complete nucleotide sequences of seven eubacterial genes coding for the elongation factor Tu: functional, structural and phylogenetic evaluations. Arch. Microbiol. 153:241- 247. Margulis, L. (1970). Origin of eukaryotic cells. New Haven, Yale University Press. Margulis, L. (1981). Symbiosis in Cell Evolution. Freeman. Martin, W., Somerville, C.C. and Loiseaux-de Goer, S. (1992). Molecular phylogenies of plastid origins and algal evolution. J. Mol. Evol. 35:385-404. Mereschkowsky, C. (1905). Uber Natur und Ursprung der Chromatophoren im Pflanzenreiche. Biol. Centralbl 25:593-604. Mereschkowsky, C. (1910). Theorie der zwei Plasmaarten als Grundlage der Symbiogenesis, einer neuen Lehre von der Entstehung der Organismen. Biologische Centralblatt 30:278-303; 321-347; 353-367. Moore, L.R., Goericke, R., and Chisholm, S.W. (1994). The comparative physiology of Synechococcus and Prochlorococcus: influence of light and temperature on growth, pigments, flourescence and absorptive properties. Mar. Ecol. Prog. Ser., in press. Morden, C.W. and Golden, S.S. (1989a). psbA genes indicate common ancestry of prochlorophytes and chloroplasts. Nature 337:382-385. Morden, C.W. and Golden, S.S. (1989b). Corrigendum: psbA genes indicate common ancestry of prochlorophytes and chloroplasts. Nature 339:400. Olson, R.J., Chisholm, S.W., Zettler, E.R., Altabet, M.A., and Dusenberry, J.A. (1990) Spatial and temporal distributions of prochlorophyte picoplankton in the North Atlantic Ocean. Deep Sea Res. 37, 1033-1051. Palenik, B. and Haselkorn, R. (1992). Multiple evolutionary origins of prochlorophytes, the chlorophyll b-containing prokaryotes. Nature 355:265-267. Partensky, R., Hoepffner, N., Li, W.K.W., Ulloa, 0. and Vaulot, D. (1993). Photoacclimation of Prochlorococcus sp. (Prochlorophyta) strains isolated from the north Atlantic and the Mediterranean Sea. Plant Physiol. 101:285-296. Raven, P. (1970). A multiple origin for plastids and mitochondria. Science 169:641- 646. Stebbins, G.L. (1982). Modal themes: a new framework for evolutionary syntheses. In Perspectives on Evolution (R. Milkman, ed.). Sinaur pp. 1-14. Turner, S., Burger.Wiersma, T., Giovannoni, S.J., Mur, L.R., and N.R. Pace. (1989). The relationship of a prochlorophyte Prochlorothrixhollandica to green chloroplasts. Nature 337:380-382. Urbach, E., Robertson, D. and Chisholm, S.W. (1992). Multiple evolutionary origins of prochlorophytes within the cyanobacterial radiation. Nature 355:267-269. Valentin, K. and Zetsche, K. (1990a). Rubisco genes indicate a close phylogenetic relationship between the plastids of Chromophyta and Rhodophyta. Plant Mol. Biol. 15:575-584. Valentin, K. and Zetsche, K. (1990b). Structure of the Rubisco operon from the unicellular red alga Cyanidium caldarium: evidence for a polyphyletic origin of plastids. Mol. Gen. Genet. 222:425-430. Van den Eynde, H., De Baere, R., De Roeck, E., Van de Peer, Y., Vandenberghe, A., Willekens, P. and De Wachter, R. (1988). The 5S ribosomal RNA sequences of a red algal chloroplast and a gymnosperm chloroplast. Implications for the evolution of plastids and cyanobacteria. J. Mol. Evol. 27:126-132. Witt, D. and Stackebrandt, E. (1988). Disproving the hypothesis of a common ancestry for the Ochromonas danicachrysoplast and Heliobacteriumchlorum. Arch. Microbiol. 150:244-248. Woese, C.R. (1987). Bacterial Evolution. Microbiol. Rev. 51:221-271. Weisse, T. (1993). Dynamics of autotrophic picoplankton in marine and freshwater ecosystems. Adv. Microbial. Ecol. 13:327-370. Wood, A.M. (1988). Molecular biology, single cell analysis and quantitative genetics: new evolutionary genetic approaches in phytoplankton ecology. In Immunochemical Approaches to Coastal, Estuarineand Oceanographic Questions (C.M. Yentsch, F.C. Mague and P.K. Horan, eds.) Springer-Verlag pp. 41-73. Wood, A.M., and Leatham, T. (1992). The species concept in phytoplankton ecology. J. Phycol. 28:723-729. Chapter One

MULTIPLE EVOLUTIONARY ORIGINS OF PROCHLOROPHYTES WITHIN THE CYANOBACTERIAL RADIATION © Nature 1992. Reprinted by permission.

Multiple evolutionary origins of prochlorophytes within the cyanobacterial radiation Ena Urbach, Deborah L. Robertson* & Sale W. Chisholm

Ralph M.Parsons Laboratory, 48-425 Massachusetts Institute of Technology, Cambridge. Massachusetts 02139. USA * Department of Molecular Genetics and Cellular Biology, University of Chicago. Chicago, Illinois 60637, USA

THE taxonomic group Prochlorales (Lewin 1977) Burger- Wiersma, Stal and Mur 1989 was established to accommodate a set of prokaryotic oxygenic phototrophs which, like plant, green algal and euglenoid chloroplasts, contain chlorophyll b instead of phycobiliproteins. Prochlorophytes were originally proposed (with concomitant scepticism' - ) to be a monophyletic group sharing a common ancestry with these 'green' chloroplasts". Results from molecular sequence phylogenies, however, have suggested that Prochlorothrixr oliandica7 is not on a lineage that leads to plas- tidss- 2. Our results from 16S ribosomal RNA sequence com- parisons, which Include new sequences from the marine picoplank- ter Prockhorococmn marinus13" and the Lissocdiam patella symbiont Prochloron sp."s, indicate that prochiorophytes are polyphyletic within the cyanobacterial radiation, and suggest that none of the known species is specifically related to chloroplasts. This implies that the three prochlorophytes and the green chloro- plast ancestor acquired chlorophyll b and its associated structural proteins in convergent evolutionary events. We report further that the 16S rRNA gene sequence from Procklorococcus is very similar to those of open ocean Synechococcus strains (marine cluster A (ref. 16)), and to a family of 16S rRNA genes shotgun-cloned from plankton in the north Atlantic and Pacific Oceanso'". 7 The order Prochlorales subsumes two formally described gen- era: Prochloron, a symbiont of marine ascidians 20 , and Pro- chlorothrix, a free-living, filamentous, fresh-water plankter7 . Like green chloroplasts, these prokaryotes have appressed thylakoid membranes containing chlorophylls a and b and lacking phy- cobilisomes. The recently discovered Prochlorococcus, a free- living, tiny unicell that is widespread and abundant in the oceans", shares these traits with its 'relatives', but differs in that it contains divinyl chlorophylls a and b (exclusively), and a- rather than 9-carotene'4"2 . NATURE VOL 355 1~N.'UARY 1992 LETTERS TO NATURE

TABLE 1 Evolutionary distance and fractional similarity matrix for 16S rRNA sequences from A. tumefaciens and members of the prokaryotic oxygenic Dhototroph radiation

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 1. Agrobacternum tumefaciens 797 2. SARIO0 885 3. Synechooccus PCC6301 886 4. Prochlorococcus mannus 886 5. Prochloron sp. 910 6. Synechocystis PCC6308 888 7. Synechocystis PCC6906 890 8. Gloeobacter PCC7421 884 9. Gloeothece PCC6501 896 10. Myxosarcina PCC7312 893 11. Lyngbya PCC7419 910 12. OscillatoriaPCC6304 897 13. Anabaena PCC7122 905 14. Prochlorothrix hollanoica 883 15. Cyanophora cyanelle 16. Marchantia chloroplast 127 17. Synechococcus WH7805 133 18. Synechococcus WH8103 130

29 Estimated evolutionary distances (xl.000. below the diagona, and fractional similarity values (xl.000. above) for pairs of 16S riRNAsequences for 22 Prochlorothrix hollandica.e other prochlorophytes (this work). cyanooacteria 30 (D.L. Distel and J. B. Waterbury, manuscript in preparation). photosynthetic 22". l 3 organelles , a full-length, shotgun-cloned sequence from the Sargasso Sea' . and A. tumefaciens . used in the construction of the phylogenetic tree 23, in Fig. la. Sequences were aligned according to their conserved primary and secondary structures ano compared by the computer program of Olsen using 788 nucleotides of sequence unambiguously aligned for all organisms considered. Gap values were set at 0.5 nucleotide substitution. Marchantia Marchantia polymorpha: Cyanophora, Cyanophora paradoxa. Genomic DNA from Prochlorococcus(isolated from the Sargasso Sea, 30 May 1988, 28* 58.9' N.64" 21.5' W. 33 at a depth of 120 m: ref. 14) was extracted using standard tecnniques from a dilution culture inoculated with an average cell density of one cell per culture (isolate SSW5. 58% probability of clonality). Prochloron cetis were collected in Fpbruary 1990 from reef flats in the Kamori Channel. Koror. Palau. West Caroline Islands (7" 25'N. 134* 30'E). Collection and method of isolating the symbiont from the host, Lissoclinum patella are descnribed in ref. 34. 5 DNA was extracted from frozen cells using standard protocols' . except that the extraction buffer was adjusted to 0.25 M EDTA. 16S rRNJAgenes were amplified from genomic DNA of Prochlorococcususing the polymerase chain reaction (PCR) with the primers PLG1.1 (ACGGGTGAGTAACGCGTRA) and PLG2.1 (CTTATGCAGGCGAGTTGCAGC). designed to be specific for members of the oxygenic phototroph lineage and thus select against sequences from heterotrophic contaminants in the culture. Prochloron sequences were amplifea using the primers P1 (AGAGTTTGATCCTGGCTCAG) and P2 (CTTGTTACGACTTCACCCC). which amplify 16S rRNA sequences from all eubacteria. as there was no significant contamination from heterotrophs in this preparation PCRreactions contained 10 ng genomic DNA. 100 nM each primer, 200 pM eacn dNTP. 2.25 mM MgCI2 . 50 mM KCI.10 mM Tris (pH 8.4), and 5 units Tao polymerase (Perkin Elmer Cetus). Reaction volumes were 200 L1.covered with 100 pl sterile mineral oil. Cycle parameters were: 94 TCfor 1 min. Tafor i min. and 72 "C for 2 min. Samples were processed for 40 to 50 cycles. T. was 66 'C for the PLG1.1/PLG2.1 primer pair. and 58 *Cfor the P1/P2 pair. PCR products were purified from preparative polyacrylamide gels by electro-elution (D-gel. EpiGene), and reamplified asymmetrically using 2 nM of one limiting primer. Asymmetric PCR products from both coding and complementary strands were ethanol-precipitated twice with ammonium acetate and used for dideoxy-nucleotide sequencing with Sequenase version 2 (US Biochemical Corporatioa, in parallel reactions containing OGTP and dlTP. plus and minus single-strand binding protein. 1.147 bases of continuous sequence (Escherichia coli 16S rRNA nucleotide positions 128 to 1.312) were determined for Prochlorococcus and 1.377 bases (E coli positions 28 to 1.452) for Prochloron. Sequences are available through GenBank. accession numbers X63140 and X63141.

a Agrobacter/um Agrobacterlum Glosobacter PCC7421 Gloeobacter PCC7421 F -Synechococcus WH7805 Cyanophora cyanelle lat_ SAR100 Marchantia chloroplast Synechoccus WH85103 Syneechococcus WH7805 ___ PROCHLOROCOCCUS SARIOG Synechooccus PCC6301 Synechococcus WH8103 PROCHLOROTHRIX SAnabaena PCC7122 PROCHLOROCOCCUS Synechococcus PCC6301 Cyanophora cyanelle 1- PROCHLOROTHRIX Marchantia chloroplast - Lyngbya PCC7419 -- Lyngbya PCC7419 I-- Oscillatoria PCC6304 Oscillatoria PCC6304 , - Synechocyotis PCC6906 Gloeothece PCC6501 - 4 . Gloeothece PCC6501 Synechocystis PCC6906 L_ Anabaena PCC7122 PROCHLORON PROCHLORON Synechocystis PCC6308 LSyechocystas PCCC6308 hMyxosarcina PCC7312 - Myxzosarclna PCC7312

0 a nlo a0s15 branches immediately to the left of Prochlorococcus with 100% bootstrap Fixed point mutations per sequence position confirmation in both distance matrix and parsimony trees (not shown). Numbers inoicate per cent confirmation by bootstrap analysis, summarizing FIG.1 Nucleic-acid sequence phylogenies, inferred from 16S rRA data, the tree output from 100 resampled datasets, of the phylogenetic group to illustrating the apparent independent appearance of chlorophyll b-cortaining the right. Only values greater than 50% are labelled. a Distance matrix organisms in the cyanobacterial lineage. The phylogenetic position of a analysis. Horizontal distances are proportional to evolutionary distances 1 9 full-length, shotgun-cloned 16S rRNA gene from Sargasso Sea plarKton (Table 1).b. Parsimony analysis of the same data using the heuristic tree I (SAR100) is also shown. A sequence for the freshwater Synecnaococus search option of PAUP". An equally parsimonious tree reverses the relative strain PCC6307 (D L.Distel and J.B. Waterbury. manuscript in precarationl positions of Prochlorococcus and Synechococcus WH7805. NATURE - VOL 355 • 16 JANUARY 1992 LETTERS TO NATURE

TABLE 2 Evolutionary distance and fractional similarity matrix for 16S rRNA sequences from A. tumefaciens. memoers of the prokaryotic oxygenic phototroph radiation and shotgun clones from marine plankton communities 1 2 3 4 5 6 7 8 9 10 . Agrobactenum tumefacens 782 808 821 797 812 805 799 812 805 2. SAR6 258 967 954 918 964 912 951 951 971 3 SAR7 222 33 987 925 977 919 964 984 971 4.SAR100 205 47 13 931 977 932 964 984 971 5. Synec~ococcus PCC6301 236 87 79 72 935 974 922 915 935 6. Prochlorococcusmarinus 217 37 23 23 68 942 987 981 994 ,. Prochlorothrix hollandica 226 94 86 72 27 61 929 922 942 8. Synechococcus WH7805 234 51 37 37 83 13 75 981 981 9. Synechococcus WH8103 217 51 16 16 90 20 82 20 974 3. AL037 226 30 30 30 68 7 61 20 26 Estimated evolutionary distances (x1.000, below the diagonal) and fractional similarity values (x1,000. above) for pairs of 16S rRNlAsequences for ProchNorothrixg. other prochloroohytes (this workl. cyanobacteriaX 1D.L. Distel and 8. Waterbury, manuscript in preparation), shotgun-cloned 7- J. 16S rRNA geres" 19.and Agrooacterium 2, used in the construction of the phylogenetic tree in Fig. 2. The 186 bases of aligned sequence corresponding to position 306 to 514 in the E coli 16S r•NA sequence were analysed as aescribed in the legend to Table 1.

We compared 16S rRNA gene sequences for both Pro. with the Cyanophora paradoxa cyanelle (95% bootstrap chlorococcus and Prochloron with sequences from the oxygenic confirmation , in accord with evidence that green chloroplasts 8s 22 phototroph database iD. L. Distel and J. B. Waterbury, and the cvanobacterium-like cyanelle descend from a common manuscript in preparationI using both distance matrix-' (Fig. ancestor ': . Branching patterns for Prochioron and Prochloroth- Ia: Table 11 and parsimony2 • (Fig. I b) algorithms. Bootstrap rix are less than 95% confirmed, thus their descent from separate, resampling:5 was used to estimate reliability of the inferred trees phycobilisome-containing ancestors. though likely, is less cer- (Fig. la, b). For simplicity we refer only to the results of the tain than for Prochlorococcus and the chloroplasts. distance matrix analysis in our discussion, although results from Prochlorococcus is closely related not only to the open-ocean parsimony analysis are in essential agreement. strains, Synechococcus WH7805 and WH8103, but also to a The 16S rRNA tree strongly supports the derivation of both diversity of shotgun-cloned 16S rRNA gene sequences from the Prochlorococcus and green chloroplasts (represented in our Sargasso Sea"' ' 9 and the north Pacific Ocean' g (Figs 1 and 2; analysis by the Marchantia polymorpha chloroplasti from Tables I and 2). Individual samplings from these sites, where different, phycobilisome-containing ancestors among the both Prochlorococcus and Svnechococcus are common, have cyanobacteria (Fig. la). Prochlorococcus forms a shallowly routinely netted a puzzling multiplicity of sequences. exhibiting branching cluster with marine A Synechococcus strains WH7805 greater than 95% similarity, interpreted as belonging to open- and WH8103 (ref. 16) (confirmed in 100% of bootstrap resamp- ocean Synechococcus "".' This diversity can now be partially lings), before which branch the cyanobacteria Svnechococcus attributed to sympatric populations of Synechococcus and Pro- PCC6307 (D. L. Distel and J. B. Waterbury, manuscript in chlorococcus. which, despite their sequence similarity, are separ- preparationi (not shown: confirmed in 100% of bootstrap ate species by ecological as well as phenotypic criteria2'. 730 resamplings) and Synechococcus PCC6301 (confirmed in /9 We conclude that the phycobilisome-/chlorophyll b÷ (green of bootstrap resamplings .The Marchantia chloroplast branches chloroplast Iphenotype seems to have evolved de now at least three, and possibly four times during the evolution of cyanobac- teria: once in the ancestry of the green chloroplast lineage, again Agrobactertum in that of Prochlorococcus,and once or twice in the ancestries Synechococcus PCC6301 of Prochloron and Prochlorothrix. the mutual independence of which has not positively been demonstrated. The shallowness .----. cPROCHLROTHRI of the Prochlorococcus cluster suggests a particularly recent SARS origin for this organism's pigment phenotype. In the light of these results, and s -PROCHLOROCOCCUS those of Palenik and Haselkonm, the order Prochlorales can no longer be justified as a natural grouping. -ALo37 The taxon should therefore be abandoned, and prochlorophytes Syneehococcus WH7Is05 reclassified in the cyanobacteria. O - SARiMOO - SAR7 Recewed27 Augustacceted 18 October1991. --_68 1. Cavaier-Smiti 7 "o i Linn.Soc. LOn 17. 289-306 (1962). - Synechococcus WH8103 2. Whitton.B. A & Carr.N. G. in TheBiol~y of we Cyanonacene• edsCar. N. G &Whitton. a. A.) 1-8 (Unnvers,.of California.Benrely, 1982). 3 LewmnR A &C-eng. L in Proch•oron a Atcromw Engnma(eds Lewin. R.A. & Cr',w L 1-7 (Chapan an Ill.New York1989) S0.f 004 006 O.8 010 0.12 4. Va Valen. L M. & Maorna V C. Nature 2. 248-250 (191O). 5. Lewm.Ra Arnn1Y Y.. Aaa Sc,31 325-329 ;1981) F xed point mutations per sequence position 6. Margull.L Snr•osis in CellEvolution (Freeman.Sa Francisco.191). 7. Burger.Wersma T et a lnt J Syst Bacterol 3. 250-257 (1989). F:G 2 Phylogenetic tree illustrating relationships among prochloroDnytes. 8 Turner,S e: a %aure337. 380-382 11989). rre-ioers of tne Synecnococcas group ano shotg.n-clonea sequences from 9 Morden.C n ,. Golden. S S 'ature 337. 382-3e5 1989) 71 : 10. MorOen.C W Goldoen.S 5 Natune 33. 400 1989. Sargasso Sea planktone' (SAR) and Pacific Ocean picoplankton a iALO). 11. Morden.C W & Golden. S S. 1 mosec Evoal 3. 379-395 (1991). The tree was constructed using distance matrix analysis and 16S rCRNA 12i Kishlno.H.. Myata. T.& Hasegawa.M. J molec Evc 31a151-160 (1990). seoJence data (Table 2). Differences in the branching order between this 13. Ctisholm. S W. e at Nature 334 340-343 (1988 tree and the one shown in 14. Chisholm.S WN t at. Arolv ecrobiolt(in the ress). Fig. la are presumaoay due to the short length 15. Lewin.R. A & Withers.N. W Nature 26 735-737 (1975). of :he shotgun-cloneo sequences. which limits the resolution of this analysis. 16. Wateury. . B. a.& oa. R.ain Beraey' s Manual of SystematicBacteriloy v 3 (3ds Starley. Bootstrap analysis indicates that Procnlorococcus. Synechococcus strains J. T. el al. 1728-1746 (Wilolimsand WIens. Balm•re. 1989). WH8101. WH7805. and the cloned sequences form a coherent phylogenetic 17 GiovaninonS et af Nature 341. 60-62 (1990-. c!.s:er distinct from Prochloromrix ano Synechococcus 18 Schmidtt.T M DeLon E. F & Pace. N R. 1 Bacr I73. 4371-4378 1991) PCC6301. out the 19 BrItschg7T B . Govannon, S J AO. enrw M~oao. 57. 1707-1713 U199 tr'•icning oroer in tris cluster Is Indeterminate. 20 Lewin R A 'rcoroga1. 217 !1977) 1;'1URE - VOL 355 - 16 JANUARY 1992 LETTERS TO NATURE

21 Goecke. R & Repeta, D Limnol. Oceanogr. lnthe press) 22. Giovannon. S Jert ai. J Bact 170. 3584-3592 (1988,. 23. Olsen.G. J. Meth Enzym 164, 793-838 (1988). 24 Swofford.D L. Paup Version 3.0 (Illinois Natural tstory Survey Champaign.Illinois. 1989). 25 Felsenstein.J Evolution 39. 783-791 (19851. 26 Douglas. S. E. & urner. S . moec. Evol 33, 267-273 (1991). 27 Olson R. J..Chisholm. S W..Zettler. E. R.. Altabet. M A & Dusernerry. J A. Deep Sea Res. 37, 1033-1051 (1990) 28. Palenk B & Haselkorn.R Nature 358. 265-267 (1992) 29 Jukes T. H & Cantor. C. R in Mammalian Protein Metabolism led. Munro. H N! 21 tAcademic. New York.1969) 30. Tomioka.N. & Sugiura.M. Molec gen Genet 191. 45-50 (1983) 31. Ohyama K at . Pt molec S•/ Reporter 4. 148-175 (1986) 32. Yang.D. et a Proc nam. Acao Sc. U.SA 58I 4443-4447 (1985) 33. Samorook.J. Frntsch. E F & Manlatis.T Molecular Conrrnga Laboratory Manus IColdSpring Harbor Laboratory Press.Cold Spring Harbor.New York 1989) 34. Swift. H & Robertson. 0 L. Symbrosas 10. 95-113 (1991) 35. DzelkalmS, V. A. Szekeres M & Mulligan B. J. in Plant MalecularBology le. Shaw C H.) 277-299 (IRL. WashingtonDC 1988).

ACKNOWLEDGEMENTSWe thank M L. SogBnfor assistance with phylogenetic analyses 0 L. Distel andJ. B. Waterburyfor unpulishedsequences. and W.Thdly and P. Keohavong fortecrncal support. Thiswork was suDporteo by grantsfrom the NSF.EPA ano Offce of NavalResearch. D.L.R was supported by a training grant from the NIH

NATURE - VOL 355 - 16 JANUARY 1992 Chapter Two

PHYLOGENETIC RELATIONSHIPS AMONG CULTURED STRAINS OF PROCHLOROCOCCUSMARINUS ISOLATED FROM DIVERSE OCEANIC PROVINCES ABSTRACT

Based on relationships inferred from 16S rRNA, psbB and pet B and D sequences, cultured strains of Prochlorococcusmarinus were found to belong to a single lineage within the cyanobacteria, which they share with strains of marine A Synechococcus. Branching patterns inferred for P. marinus strains using 16S rRNA, psbB and pet B and D sequences were consistent, and revealed no correlation between genetic distance and the geographic distance between sites of culture isolation, with the most closely related strains originating in surface waters of the Mediterranean and the Pacific Oceans. The branching order of deeply diverging lineages within the P. marinuslmarineA Synechococcus cluster, including one strain of P. marinus and two strains of marine A Synechococcus, appear inconsistent in 16S rRNA gene trees inferred by neighbor-joining, parsimony and maximum likelihood methods of analysis. This result suggests a near- simultaneous radiation of P. marinus and marine A Synechococcus lineages.

INTRODUCTION

Prochlorococcusmarinus is a tiny, photosynthetic marine prokaryote containing divinyl chlorophylls a and b (chl a2 and chl b2) and which recently has been recognized as a major constitutent of the photosynthetic picoplankton in wide regions of the Atlantic, Pacific, Mediterranean and Red Seas and in the Banda Sea of Indonesia (Chisholm et al. 1988, Gieskes et al. 1988, Li and Wood 1988, Olson et al. 1990, Vaulot et al. 1990, Li et al. 1992, Vaulot and Partensky 1992, Cambell and Vaulot 1993, Goericke and Welschmeyer 1993, Goericke and Repeta 1993, Lantoine and Neveux 1993, Veldhuis and Kraay 1993, Vaulot et al. 1994). P. marinus comprises 35% of the seasonally averaged chl a at north Pacific station ALOHA (Letelier et al. 1993) and 30% at Sargasso Sea station OFP (Goericke and Welschmeyer 1993). The ecological importance of this ubiquitous and unusual organism has sparked efforts to characterize its physiological capabilities in pure culture (Partensky et al. 1993, Morel et al. 1993, Moore et al. 1994).

Phylogenetic analysis using 16S ribosomal RNA gene sequences has established that a strain of P. marinus isolated from the Sargasso Sea (strain SSW5, descended from SARG, the original P. marinus isolate) is closely related to the marine A Synechococcus (Waterbury and Rippka 1989), a group of cyanobacterial strains including isolates from open ocean habitats in which P. marinus also is found (Urbach et al. 1992). Cyanobacterial 16S rRNA gene sequences cloned from natural populations in the Sargasso Sea and north Pacific (Giovannoni et al. 1990, Britschgi and Giovannoni 1991, Schmidt et al. 1991) also fall into this cluster (Urbach et al. 1992). An independent study using rpoC gene sequences found that the genetic distance between SARG and a second P. marinus isolate originating in the Mediterranean Sea (MED), was greater than the distance between two heterocyst-forming cyanobacteria assigned to different genera (Palenik and Haselkorn 1992), suggesting that differences between P. marinus isolates may be significant. rpoC analyses also cluster P. marinus with Synechococcus WH8103 (Swift and Palenik 1992). The 16S rRNA and rpoC studies each reported, in addition, that P. marinus is unrelated to the "normal" chlorophyll a and b (chl al and bl) - containing prokaryotes Prochloronsp. and Prochlorothrixhollandica, and to green plant and algal chloroplasts. These conclusions have been cast into some doubt, however, by the discovery that base substitution biases at rapidly evolving nucleotide sites are correlated with branching patterns in these studies' phylogenetic trees (Lockhart and Penny 1992).

To help resolve these controversial relationships and examine isolates of P. marinus from additional oceanographic provinces, we present here the results of an investigation into the phylogenetic relationships among cultured isolates of P. marinus, addressed by comparisons of three gene sequences and including marine A Synechococcus and cloned 16S rRNA gene sequences from the Sargasso Sea. Questions to be addressed are (1) whether oceanic prochlorophytes represent more than one independent lineage of chlorophyll b2-containing prokaryotes dispersed among the cyanobacteria, (2) whether the inferred phylogenetic patterns for each of the three genes are consistent, (3) whether a different phylogenetic pattern is inferred by methods suggested as insensitive to nucleotide substitution bias, and (4) whether evolutionary relationships reflect geographic relationships among sites of P. marinus culture isolation. Detailed knowledge of the evolutionary relationships among P. marinus cultured strains is, in addition, useful for organizing information on the physiological differences among these strains. Also, if gene transfer does not play a major role in the evolution of P. marinus, then knowledge of the phylogenetic relationships will provide a link between phenotypes of cultured cells and sequences recovered by molecular cloning experiments exploring genetic diversity in field populations (Giovannoni et al. 1990, Schmidt et al. 1991, Britschgi and Giovannoni 1991, DeLong et al 1993, Fuhrman et al. 1993, this thesis Chapter Three).

BACKGROUND

Characteristics of P. marinus strains. Cultures have been established for P. marinus isolated from a variety of geographic and hydrographic regimes (Chisholm et al. 1992, Partensky et al. 1993, Moore et al. 1994, L. Moore, pers. comm.). For this study we selected cultures drawn from widely separated locales within P. marinus' geographic range: the Sargasso Sea, Mediterranean Ocean, north Atlantic and south Pacific Oceans. The P. marinus type strain, Sargasso Sea clone SS120 (designated CCMP-1375 at the Center for the Culture of Marine Phytoplankton, Bigelow Laboratory for Ocean Sciences, West Boothbay Harbort, ME) originated in the deep euphotic zone during spring stratification and is descended from SARG, the original P. marinus isolate (Chisholm et al. 1992). The Sargasso Sea isolate used for 16S rRNA analysis, SSW5, was isolated from SARG by dilution to a mean cell density of one cell per culture tube, and is assumed to be identical to SS120 1. Mediterranean clone Med4 (CCMP1378) originated in mixed surface waters of the Mediterranean in winter and is descended from strain MED (Chisholm et al 1992, Partensky et al. 1993). North Atlantic isolate FP5 is an uncloned culture collected from 30 m in the North Atlantic by F. Partensky, and Pacific strain MIT9107, also not cloned, came from 25m in the mixed surface layer of the oligotrophic south Pacific, and was isolated by J. Dusenberry and M. DuRand. In addition to these strains, Sargasso Sea culture MIT9303, isolated from 100 m in the Sargasso Sea during summer stratification by L. Moore, was included in some of the analyses. Synechococcus clone WH8103, a high phycourobilin (PUB) strain isolated from the surface of the Sargasso Sea, and the reference culture for marine A Synechococcus (Waterbury and Rippka 1989), was selected to represent marine A Synechococcus (Table 1).

Physiological studies have characterized P. marinus clones SS120 and Med4 (or their parent cultures) according to pigment content and respose to variation in temperature and illumination (Partensky et al. 1993, Morel et al. 1993, Moore et al. 1994). These studies confirm the presence in both strains of divinyl chlorophylls a and b

(chl a2 and b2) as well as accessory pigments zeaxanthin, a-carotene and an unknown pigment which elutes with chlorophyll cl by HPLC (Chisholm et al. 1988, Chisholm et al. 1992, Goericke and Repeta 1992). In addition, it is now known that SS120 (and

1In light of the results of Chapter Three, which indicate that P. marinus field populations are genetically heterogeneous, the possibility exists that the primary isolate SARG could have contained more than one genetic variant. Since SSW5 and SS 120 were independently isolated from SARG, it therefore is possible that they are not genetically identical. Omitting the SSW5 sequence from the dataset does not change the conclusions of this study. ?ýNNilýN OO0 0 00 0 OV) • mt \

4 m

W)\?0 0•

e1

F- 1-4 O3 0 C) C .N tf 2 SARG) grown at high light (>20 giE m- s-1) contains chl b1, which may comprise up to

55% of total chlorophyll b (chl b, equal to chl b1 + chl b2) (Partensky et al. 1993, Morel et al. 1993, Moore et al. 1994). In contrast, clone Med4 (and MED) contains no detectable chl bl under any growth conditions tested. Both strains are capable of photoadaptation by changing the ratio of their chl b to chl a2 , but chl b/a 2 ratios for Med 4 are one tenth those of SS120 at all growth irradiances (Partensky et al. 1993, Morel et al. 1993, Moore et al. 1994).

It has been hypothesized that the different chl b/a2 ratios for SS 120 and Med4 are genetic adaptations to the low and high light environments from which the two clones were collected, respectively. The low light SS120 clone is capable of acclimating to conditions at the bottom of the euphotic zone in oligotrophic waters by elaborating extensive, chl b-containing photosynthetic antennae capable of absorbing blue light, while the high light-adapted Med4 fails to elaborate an extensive antenna (Partensky et al. 1993, Moore at al. 1994). This hypothesis is consistent with the results of growth experiments which have shown that SS120 is capable of growth at low light intensities (2 to 6 tE m- 2 s-1) at which no growth was observed for Med4, and that Med4 is capable of growth at high light intensities (>100 igE m-2 s- 1) at which SS120 fails to grow. The compensation light intensity (Icomp), which predicts the minimum illumination for cell survival, was significantly lower for SS120 than for Med4 (Moore et al. 1994), again consistent with the hypothesis. Other cultures included in the phylogenetic study are much less well characterized than SS120 and Med4.

Genetic loci. For this study, evolutionary relationships among the four P. marinus cultures SS120, Med4, FP5 and MIT9107 and Synechococcus WH8103 were examined using three sequences exhibiting varying degrees of evolutionary conservation: the 16S ribosomal RNA genes, psbB, and petB and D. Strain MIT9303 was evaluated using only 16S rRNA, as psbB and petB/D sequences were unavailable for this culture due to time limitations. Future work after the completion of this thesis may include psbB and petB/D sequences for this strain.

16S ribosomal RNA (rRNA) genes are the standard tool for phylogenetic reconstructions (Hillis and Dixon 1991, Olsen and Woese 1993) and have provided phylogenetic information for many diverse taxa (e.g. Woese 1987, Sogin et al. 1989, Medlin et al. 1993). 16S rRNA analyses have the advantage of compatibility with a large and well-characterized database, containing sequences from hundreds of organisms and organelles, and a dedicated computerized clearinghouse which provides automated sequence alignment and similarity searching (Larsen et al. 1993). Several cloning experiments exploring the diversity of microbial communities in the sea have provided 16S rRNA gene sequences, some of which are likely to derive from wild P. marinus (Giovannoni et al. 1990, Schmidt et al. 1991, Britschgi and Giovannoni 1991, DeLong et al. 1993, Fuhrman et al., 1993).

psbB and petB and D encode proteins of the photosynthetic apparatus and are much less constrained evolutionarily than 16S rRNA. These loci were selected to infer phylogenies sensitive to evolutionary differences on a finer scale. Each has the additional advantage of being a single-copy sequence unique to photosynthetic organisms (Vermaas and Ikeuchi 1991), and so is easily amplified from P. marinus cultures, which contain heterotrophic bacteria. psbB encodes the chlorophyll a-binding antenna protein CP47 and petB and D encode subunits of the photosynthetic b6/fcomplex responsible for transferring electrons between Photosystem II and Photosystem I (Widger and Cramer 1991). petB and D are neighboring cistrons with a consistent tandem orientation in oxygenic photosynthetic prokaryotes and organelles (Vermaas and Ikeuchi 1991, Greer and Golden 1992) and provide PCR priming sites which permit amplification of a fragment containing the 3' end of petB and the 5'end of petD, plus the highly variable intergenic region between them (the amplified locus will be referred to as "petB/D"). The psbB and petB/D loci are not close to each other on cyanobacterial chromosomes (Vermaas and Ikeuchi 1991).

Problems posed for molecular phylogenetic analysis by nucleotide substitution biases. An unusal feature of cyanobacterial and chloroplast sequences is the wide variation in their DNA base compositions (Herdman et al. 1979, Lockhart et al. 1992a), which is especially evident at rapidly evolving sites such as noncoding regions and silent third codon positions (Lockhart and Penny 1992, Lockhart et al. 1992a, b). An additional property of the cyanobacterial phylogeny is the presence of many deeply branching lineages which apparently originated within a short period of evolutionary history (Giovannoni et al. 1988). The juxtaposition of these two unusual features creates a situation in which the inferred branching pattern is susceptible to the artefactual clustering of sequences converging towards similar G+C content (Lockhart et al 1992a, b, c, Lockhart and Penny 1992, Lockhart et al. 1993), and there is as yet no accepted method for assessing the relative contributions of nucleotide substitution bias "noise" and true phylogenetic "signal" to trees inferred from such data (Lockhart et al. 1993). Bootstrap resampling analyses, which are frequently used to assess the sensitivity of phylogenetic inferences to sampling error in the nucleotide positions chosen for the analysis (Felsenstein 1988), cannot detect systematic distortions in branching patterns due to nucleotide substitution biases (Lockhart et al. 1992a).

Lockhart et al. (1992a, b) used G+C content at third codon positions to characterize nucleotide substitution biases in the evolution of different taxa, noting that branching patterns inferred from protein-encoding sequences grouped organisms having similar nucleotide substitution biases. They suggest that in instances when nucleotide substitution biases differ among lineages, standard phylogenetic methods will infer incorrect phylogenetic relationships. Lockhart et al. (1992a) further assert that the influence of nucleotide substitution bias detected at third codon positions in protein encoding genes may determine branching patterns inferred for the same organisms using 16S rRNA genes, even though the 16S rRNA sequences may exhibit little variation in G+C content (c.f. Table 3).

Several molecular phylogenetic methods have been suggested as being relatively immune to nucleotide substitution biases, although none are yet generally accepted as such (Sogin et al. 1989, 1993). Woese et al. (1991) suggested that distance and parsimony methods applied to transversion mismatches would have generally reduced sensitivity, and so would be less susceptible to these artefacts. Hasegawa and Hashimoto (1993) suggested that phylgenetic analysis of amino acid sequences would have this property as well (see also Hashimoto et al. 1994), and amino acid parsimony has been applied to plastid and cyanobacterial data (Morden et al. 1992). Most recently, two similar, new algorithms have been devised which are reported to be insensitive to variation in G+C bias among lineages, LogDet (Lockhart et al. 1994) and paralinear distances (Lake 1994). However, computer programs which impliment these algorithms are not yet publicly available.

For our analysis of P. marinus phylogenetic relationships we have employed standard methods of phylogenetic inference: neighbor-joining, parsimony and maximum likelihood. In addition, we explored branching patterns inferred by transversion distance analysis and amino acid parsimony to assess whether these methods, suggested as relatively insensitive to nucleotide substitution biases, infer a phylogeneic pattern consistently different from those inferred by standard analyses. METHODS

Cell culture and DNA isolation. P. marinus clones and isolates were grown in modified K/10 - Cu medium as previously reported (Chisholm et al. 1992). DNA was prepared from dense cultures as described (Urbach et al. 1992), or by a modification of the alkaline lysis protocol of Li et al. (1991). For this protocol, 1 ml of culture containing 106 to 108 P. marinus was concentrated by spin filtration at 2,000 x g using 0.2 Apm pore Ultrafree-MC filter units (Millipore) and washed twice with cell suspension buffer (0.5 M NaCl, 10 mM Tris [pH 8.0] 10 mM EDTA). Cells were resuspended in a final volume of 100 pl cell suspension buffer and lysed by addition of 12 .l 0.5 M DTIT and 6 .l 10 M NaOH followed by a 10 minute incubation at 650C. After addition of 120 ptl neutralization buffer (90 mM Tris [pH 8.0] plus 0.5 mol/1 HC1), DNA was precipitated by addition of 1 pl 20 mg/ml glycogen (Boehringer) and 0.6 ml cold 100% ethanol and storage overnight at -20 0C. Precipitates were collected by centrifugation at 16,000 x g for 30 min at 40C, washed with 70% ethanol and resuspended in TE (10 mM Tris [pH 8.0] 1 mM EDTA). DNA was stored at -200C.

Synechococcus WH8103 was grown in S/N Medium according to Waterbury et al. (1986). DNA was prepared by CsCl density gradient centrifugation (Sambrook et al. 1989).

PCR and DNA sequencing. 16S rRNA genes from P. marinus Med4, FP5, MIT9107 and MIT9303 were amplified using oxygenic phototroph-specific primers and protocols as described (Urbach et al. 1992), except that one primer in each PCR reaction was labelled with biotin (Hultman et al. 1989). psbB and petBID were amplified using biotinlyated/unbiotinylated primer pairs containing inosine (Knoth et al. 1988) under the following conditions: psbB primers PPSBB 1353 (5'GTIGCIGGIACIATGTGGTA) and 0 PPSBB 1928R (5'GCRTGICCRAAIGTRAACCA), 2.25 mM MgC12, 2 min. 94 C, followed by 1 min 940 C, 1 min 510 C, 1 min 720C. (5 cycles), 1 min 940 C, 1 min 560 C, 1 min 72 0C (10 cycles), 1 min 940C, 1 min 620 C, 1 min 720C (25 cycles), followed by 10 min 720C; petB/D primers PPETBD314 (5'PPETBD314 (5'ATGATGGTIYTIATGATGAT) and PPPETBD 1160R 0 (5'CCRTARTARTTRTGICCCAT), 1.5 mM MgCl2 , 2 min 94 C followed by 1 min 940 C, 1 min 450 C, 1 min 72 0C (5 cycles), 1 min 940C, 1 min 500C, 1 min 720 C (10 cycles), 1 min 940C, 1 min 570C, 1 min 720 C (25 cycles), followed by 10 min 720 C.

100 Wl PCR reactions contained MgCI2 concentrations as indicated, plus lx Mg-free Taq polymerase buffer (Promega), 200 gM each dNTP, 100 nM each primer and 5 units Taq polymerase (Promega). PCR products from three or more replicate reactions were pooled and separated by electrophoresis in 1% agarose gels. Bands were purified using GeneClean (Bio 101) and single stranded sequencing templates prepared using Dynabeads (Dynal) according to the manufacturers instructions.

DNA sequences were determined using Sequenase (United States Biochemical), according to the manufacturer's protocol. 16S rRNA gene sequences were determined from both strands using amplification primers and internal primers as described (Urbach et al. 1992). 90.5% of 16S rRNA gene sequences were determined on both strands. psbB and petB/D sequences were determined using amplification primers and internal primers PPSBB1527 (5'ITTYTAYGAYTAYGTIGG), PPSBB 1508R (5'CCIAGRTARTGRTARAAIGC), PPSBB 1898R (5'CGAAAIACICCRTC), and PPETBD532R (5'CCIACRCTYTCICCICC). psbB sequences were redundantly determined over 92% of their length, with 60% of nucleotides determined on both strands. petB/D sequences were determined from one strand only, as the biotinylated version of PPETBD1160R did not function well as a PCR primer. 94% of petB/D sequences were redundantly determined.

Sequence alignment and phylogenetic analysis. Sequences were aligned according to conserved regions of 16S rRNA primary and secondary structure or according to conserved regions in amino acid translations (Lang and Haselkorn 1989, Widger and Cramer 1991) using the Olsen sequence alignment editor (Olsen 1990). Mismatch frequency and base compositions were calculated using the program of Olsen (1988, 1990). Neighbor-joining (Saitou and Nei 1987) and protein parsimony were performed using Kimura two-parameter genetic distance estimates (2:1 transition transversion ratio) and the DNADIST, NEIGHBOR and PROTPARS programs in PHYLIP verision 3.4 (Felsenstein 1991). DNA parsimony was performed with PAUP's branch and bound and MULPARS options (Swofford 1991). DNA maximum likelihood calculations were done using fastDNAml (Olsen et al. 1992), and likelihood comparisons between different phylogenetic trees were performed using PHYLIP's DNAML program. Log likelihood values and standard deviations are from DNAML comparisons.

RESULTS

Relationships inferred by standard methods: molecular phylogenetic branching patterns and bootstrap calculations. Relationships among cultured strains of P. marinus, marine A Synechococcus and other members of the oxygenic photosynthetic radiation were investigated using three standard methods for phylogenetic analysis: neighbor-joining (Saitou and Nei 1987), parsimony, and maximum likelihood (Felsenstein 1988). All strains of P. marinus and marine A Synechococcus fell into a single lineage in trees inferred by all analyses in this study (Figures 1, 2, 3). This Figure 1. Phylogenetic relationships among cultured strains of P. marinus, shotgun cloned sequences from Sargasso Sea picoplankton ("SAR" sequences) and other members of the cyanobacterial lineage, inferred from 16S rRNA gene sequences using Agrobacterium tumefaciens as an outgroup. (a), Distance matrix tree inferred by neighbor-joining with Kimura two parameter genetic distance estimates (Kimura 1980) (Table 2a). Numbers indicate the percent of trees inferred from 100 bootstrap datasets which contained the phylogenetic group to the right, with numbers above the lines from neighbor-joining analysis and below from parsimony. Bootstrap values below 50% are omitted. (b), The most parsimonious tree (668 steps) inferred by parsimony. (c), Maximum likelihood tree (In L = -4767.84277), not significantly better than the tree inferred by neighbor-joining (In L = -4800.28223, S.D. 19.0879, Aln L = -32.43945). Transversion analysis inferred a branching pattern identical to that in the maximum likelihood tree. Agrobacterium tumefaciens P.marinusMed4 SAR6 P.marinusMIT9107 P.marinusFP5 narinus SSW5 nus MIT9303 echococcus WH8 103 R139 ;AR7 :hococcus WH7805 PCC6301 rena PCC7120 rochloron sp. fnrrhnntin nnlvmnrnhn chln 1.~,.,.,.,I,, yv Y/··~V·y·rrr rr-rry·

Genetic distance

' •amejaclenz Agrobacterium twnefaciens utmefaciens P.marinusMed4 SAR6 P.marinus MIT9107 ,marinus FP5 P.marinusSSW5 Synechococcus WH8103 SAR139 SAR7 ynechococcu W17805 P.marinus MIT9303 Synechcoccus PCC6301 Anabaena PCC7120 ymorpha chlp. Marchantiapolymorphachip.

. ,.• , a r,. Prochloron sp. Figure 2. Phylogenetic relationships among cultured strains of P. marinus and other members of the cyanobacterial lineage inferred using psbB sequences. Trees are arbitrarily rooted to Anabaena PCC7120 to facilitate comparison to those inferred from 16S rRNA gene sequences (Figure 1). (a), Distance matrix tree inferred by neighbor- joining with Kimura two parameter genetic distance estimates (Table 2b). Numbers in parentheses indicate G+C base composition at third codon positions; other numbers are as in Figure la. Identical branching patterns were inferred by parsimony, maximum likelihood (In L = -1925.90128) and transversion neighbor-joining analyses, with some deep branches rearranged in the transversion tree (not shown). (b), Consensus for the four most parsimonious trees (267 steps) inferred by protein parsimony. r, ,,,,~,,,~,·An,~A /~ ~r\ _-. rtrUFLFLUs IVILCU'L+ •L.ZL)

P. marinus MIT9107 (0.27)

'. marinus FP5 (0.26)

. marinus SS120 (0.34)

us WH8103 (0.66)

tis PCC6803 (0.52)

Vrix hollandica (.58)

PCC7942 (0.61)

>last (0.26)

AIAf "'Uc',54 " AX I XJ.Y _-WJ I I I 0.2 0.4

Genetic distance a

P. marinus Med4 P. marinus MIT9107

- P. marinus FP5

Synechococcus WH8 103

Zea maize chloroplast

Synechococcus PCC7942

"'----'----' --- Figure 3. Phylogenetic relationships among cultured strains of P. marinus and other members of the cyanobacterial lineage inferred using petB/D sequences. Trees are arbitrarily rooted to Nostoc PCC7906 to facilitate comparison to those inferred from 16S rRNA gene sequences (Figure 1). (a), Distance matrix tree inferred by neighbor-joining with Kimura two parameter genetic distance estimates (Table 2c). Numbers in parentheses indicate G+C base composition at third codon positions; other numbers are as in Figure la. An identical branching pattern was inferred by protein parsimony, and for P. marinus and Synechococcus WH8103 by neighbor-joining analysis of transversion distances, which inferred a different pattern for more deeply branching lineages (not shown). (b), Consensus of the two best trees inferred by DNA parsimony (129 steps). (c), Tree inferred by maximum likelihood (In L = -1031.58031), which was not significantly better than the tree inferred by neighbor-joining (In L = -1038.59363, S.D. = 6.7154, Aln L = -7.01331 ). P. marinus Med4 (0.18)

P. marinus MIT9107 (0.21)

P. marinus FP5 (0.29)

SL- P.marinus S 20 (0.32)

Synechococcus WH8103 (0.74

Synechococcus PCC7002 (0.55)

Prochlorothrixhollandica (0.67)

rarchantiapolymorpha chlp. (0.11)

!amaize chlp. (0.29)

-·------·--- \-· · //

I I I m I Irm 0.05 0.10 0.15

Genetic distance

rSS1•D

awdupdymopAacdp.

amarschlp

I I I " 0 0.05 0.10 0.15 ck..I dkm

43 0 '

O U) -- 0 0 m CDI m m 1DC n C> %u m r- "IT I 0- o Cu 4 U 0

S m-mIm.- -4 -4 -4- -4 m -4 4 00 0 Cu4 0 0)- rZQ0 a) ) 4 0H 'r- 0 v -O u0 w 0 O : 0.0 0 0 0 0 CDOD0-4 00000 mODco00 - ON m H H HH H HA H H-H .-IW WM Z 4 C : 0 0 -i 00 E U) 0 I U SammmmmCmem Na 9: - 0 0 -1 C 0 4 0 -H %oCu 0 Ow P m " m - C'-H I -40CC) 0 CO0 0 0 0 U) 4 - *H(5 0C-4H *H0-H 0a

00 0 U)1 Ho a' 00C 04r0 ·r M Os a OZ( U)0 IN 000DI•0 ' Y 3b a aOO rn 0 0 U) *4- C 4 - -H4) CuH (1 0 *4 H 0-Hw \ awomoooooooo Coooo oo i m mqr0 H 0r- 00 r mCC) 0C m m U)C0 4) MC 0' H mmm o4mcNmcN i a'qaq 04-4 0 ) 00Ic) -1ý4 00 -4 000000000Cq C 0 -- 1 E CCJ 0 0

a, 0 m>C>C>>mI c lmm 4- I 54CU M -0 U .w O 0 O0U 0 (-n * 0 I: .04 0 0Cu C 00O .11S E n -0-0raU4>4 E 0 0(1 .0 -1-40Cu **0O -O 0 0-H a 04M 4 000

aWLO map a, moooooomC C) C N C HCoC N ' a 0 U CD-H 00 O m-tzINONC O 4mmN4-4-4D -4l O mmv C) IOD0DLfl-VMraam H 0 H-H U1)0 0 0 04-.010 -i a0 C">000000 -: QHC 00 : $ a o a'')"zIm U0m m ·:3: 5.L 0 m 0 am 0 -Hu -H-H U) H -H U) X S 00 40) ) ( a'-m m mMma ON m c E40U0)4 U)\ C • AA Immamm M~~0 Cmmm. 4 O-4a'a'm C 00000 04) 'M 0 rUE -H( ý a a- 0 9Q a 0 =%-0 0: a 0 f a 0 0-E4O 1 4a -1 Cu Cu 0 C O Cu 40000 E ->00U) U)-H 4aU 0* .R 0 4-i.0 0 0 0U)0 000 0to -4 CD ,14H a m' Oa' O a ý4 (Z -j c -H0a 0 0 Cu 0 ~ ' C -H Cu 00 Cu Cu 0- 004 CH:-H U) Cu 04 0 41Cu .E.0ýc -4; OH0Cu-1 04040-Q -Ha0) 'CoC)u0I 4)U4-4 0 -H00WNO - - 04 0 -i00 0, m m0 m O )I-V Dn " O o CH0X (Vý4wuu 0 0 0)Cu. -- 0 U) 0 0o 0 Cu -ý4- H r r 5 0)>4 a, 4-4 U Cu -H 4-) ON O ?L 1 3 m e ·.) H e -4 Cu-H -H -0 > 0 0 0 ' @HOC

4- -4r A-) 004-i U) a)C 44 ' , HJ U) 4-1-4 E0- 00OrO OaOH 0

Cu.0- 0-1'0 00.ý CuH 0 000)rn 4 4-)-H 4-4 Cu 04-0 0 -H -04.H U)· U) x u ' 0>0E 0 000 inCuHQO roC0 EOuOHC0 U 14.. 0a)U) 0 H0 ) Cu ) 0 0 to -4 a'0O U)0 0) Oa) 0r 0 -<1-H a)a tn 0 - 4 r E u ::I mm ra 8010 0 W 00 E4 oaHaO -UCC H OC0CiL 0 X Cu 00 0 O'0 4- 0 m -n' He a' (1 U -) r0 r uo0E a k Co- -H -atH0

*-4 U)H'OCu U CH 4 a' 0 a-C CU'004 0-4 ' E-4 L 1E- U, E; )HU)U4 0 0 N' a' 0 *C O0 -H Ea u:3 u 4-4 4--H 4-i -H Cu H 0IOemH 'n 0O 0000 0 O Cu C4--H:)

0d E , Q co 9 0, 0 > 000 )- 0) o mmo m U 0 ( 00 0O p4 00-H

* * -Ha'>4 OH0 -4 4 -1 -A ( C -( H -04 Cu - 0

-4 m4 to 0

m oO C O -a44 -H04 0 Oa 'cr LfDlOD'q, ('4r'4

O0 a, 10U 0 NN I Ln ch L(' rl - l c)I.'Qeq 4 U

m- qoo ( .0~ Lf)-4WC') I.) nto mGo mo0 U) L - m r- D n r- OeODN '.P- en - Dn cT %mOD( r P rl)m '-I

uL r-I 4-4o CDl LIOD 00ho 'oCe ~~~9 aCCre~ ~ ~~ r

J..-0 I m cyMtn 7 D qo) q, kDNI r- q D OD ' I4 N 04 c- ()i ,t r -I Nc o c q0 -4 -4 ',0 ODrn kot- r) -,o t, rn ý,otý rn ý.otý r-) I'Dý (1)

( 0 ILfCCIa e nr- en N C4OO0 T"t N 'U 0r- - 'e (1 (7 C r- aDU) ',)( DC'

OD o I ' # ~ e - aO 4-4 m al4o 0 m co 1 L OQ c0c I Ci n1 m L QO _4(C

4-1-440 Lo0 -4 -4

Nto

.0 -H

.00) ('4 o -4S .0-. m SN m N a o 0 -A U]r. 0 0 tU -4 a)

4J*H0 -4 v-e ý4- 4 O -4

0-4I 0 O4 -L) Ow k toC 4-4 0 04- 00 0COMu O) OOc' 10 " Uoc) 00 -4 Z(Uk 0 -4

O cU O 0 4 QCL Z0r~ 04 O Sa) 4-J (L) Q Q) -4 a 0 -H E(0 4.J U)~0 O m -c0

0O 4-4 to cci 4 cO 0 4O 4 O O E uO 0 Qua,

4-j' -(0 :3 X C 00, a4 -AO4 -a 0 U :)j 0)

E, 0 -4 -

3,CC(0 - 4 .OccciH-4 4) 00OO cm a, 4- -40 -4 U)0 a .0

0)0 0 co U cciO-U dO H 0 O0 (dC 0 U 0 0:*0 to O-4o-F0 -4(0(U]C '1l ci (cc>kD a 4-4cci cci C14

0oa, U,0 >l 1a rl > ,Cz OOW OO -4an -4u 4W z U4-1 En C0 -44- ro & '0j -0 -AQI-H -40 u)ci 0 doo:30C jo 04.) (0c 0 QC Z 0 '0c 4O) 0 U]0 0 -44- .0 *) -4 0 COa1 OU4 - 4 U 0 *En

ci)0 - 'H 40-44 -4 - u4- 4 rci >1 EOw Eca$04 Ul OD 0 w004-404-4~ to 4 004 '0 U>1Lo 0j L) 04-0 0 41004-40 4 4- 04004-) 0 0 x0 ý .CO 0 .C(U)O ci 0rci0 -4a)U)ý (0440 ci 44 .00 ci 0 0 (N 1-4 OD Nr N oL) -4 -4 r- 0

(N ro o inC Lx -1

r- -)C I -4OO%a " 1%q r-OD %o D'-i

r0

.-4 rl)NQ CNk 1D.0-0%D IV m0DQC) .0 O (N - r-- O cM 1 Oo en0Oc -4 -4 \ome men( 4-1 ~ODr .i)OU

I'MD r CCNO.C -14' Ir4k -1 (, r- -4m'T0%U r ODmrn r-OD .0 r- 1 N04-"4) NO4'z % ) KNQ lp NC> Oat " 0OD 0 me-4r, I eIý UE r- co LI) % D 00< N O~ b~ VM~ om 0 aG em~ weN men~~

O

0 )0 CN cD CoQ"n lo-0 (n o %nMr a •4 In 0 ODNo4 l -, ::NQr OD q, N0e '.0 '.0 me( 0-•0 " s0 c%L t zrn 0 t- 0 00 .q ill

a40 IiOo kDycoe mobo eq " rc I) rI q 4 mTOo N cq C4 " 0 4" U-4 r- at 'IT Chmnme k OD Dz

r0 CDIn -1 L,- c)omf i , ) l O0 o 0em maI om N M. CIN r4 ca 0a0 Q) 4 0N %D ODC m 0am an r- OD r- ODV on q

%D N I m am~~ mam m mh ob4 (N4N ~ ~ mIOoN%0 . ~ nODco OD OD ~ ~ON'oTN 0> .Ol oatc as n OD()I 0%%-0 r- atI k-0a.14 ' GoN r--4wf OD ' L)0 C)w4 0NOI- LJNr'- 4 0 o-N qLi'q

U) 1-4 mn1 1NOWNr-tot o4a% 1,0Ot ac..eN o O) ý4O : 'A m- 1: r. C N v r, (1y4C' rN40 G Ln DDLI)1 OI% c Lfl-0 Wn Co 4 co atLt 0c)at Lfl -a *: - %r r-(NO 0%' (N r- D ') N OD .D OD () r- DU) r- co'I CO

-0

4-H(

CI Z

S4

' U) '0 Ný (X In 0 -4 - 4-4 0 0z

ci) ý4) -4(

to 0, -H Qi) O0 a 0 0LO El ý4) 04 ý44 a) 0 c)44 ro a 0 0 0 4-1c:

14- 0o 0L9 r 0 0 to4O a 0-1 0-4O U:JF 'H0 L) 0 o to E; rý >0 ci) O4 ci) u u 0 u to, 41-'ci 0 4. - 4-) 'ci4- k ci) C- 0,H 0 4. 4-)0 0 0 U) vi) z>o -4 440

O(15Lo -4

U cU::s mtoto ~cci4-.

0 (-4 '- OD0, U 0 tp -14ti1 0C: O a to

0)a -4(c '0 ci 0 Ua) to 4-3

to a, OOu)-4.ci *cc 1

'ci S0 ci) -H EOr4 z 4 0o'0 - 044 c) '0to o to 'H -4 04 -Hto -ro0 OU) WU) aU 4-) H C00 c 0O H

00)

ci) 4-) ) .0 ci) 4-1 q kci COto O4 0 4.)" ý4 ) w-0 O 4-104 -H 0 Oto- I SC 00 L) O 004Ql --1c 0 d -H -H O clustering was consistent in neighbor-joining and parsimony analyses, whether or not chloroplast sequences were included in the datasets, and persisted in a 16S rRNA gene tree which included diverse cyanobacterial taxa (Figure 4).

P. marinus strains Med4, MIT9107, FP5 and SS120 (the "P.marinus four culture clade") formed a phylogenetic group distinct from Synecyococcus WH8103 in all analyses. However, the phylogenetic position of the deeply branching P. marinus MIT9303 was different in 16S rRNA trees inferred by different methods.

Identical branching patterns were inferred for relationships among members of the P. marinus four culture clade and Synechococcus WH8103 for all three genes using the neighbor-joining method, and for 16S rRNA and psbB genes using the parsimony and maximum likelihood methods as well (Tables 2a, b, c, Figures la, b, c, 2a, 3a). For petB/D sequences, however, relationships among the four P. marinus cultures could not be resolved using parsimony (Figure 3b), and the maximum likelihood tree, though not significantly more likely than the neighbor-joining tree, exhibited a slightly altered branching order (Figure 3c).

Bootstrap analysis (100 resampled datasets each for neighbor-joining and parsimony calculations) strongly supported the pattern inferred by all three methods for 16S and psbB sequences for branching order among members of the P. marinus four culture clade and Synechococcus WH8103, except that the relative positions of strains FP5 and SS120 received varying levels of support (Figures la, 2a). Consistent with the generally lower resolving power in the petBiD dataset, bootstrap values were lower in trees inferred from these sequences (Figure 3a). The lower resolution in petB/D trees may be due to the shorter length of petB/D sequences; after intergenic regions and third Figure 4. Phylogenetic relationships among P. marinus cultured strains and diverse cyanobacterial taxa, inferred by neighbor joining using 16S rRNA sequences. Differences in the branching pattern between this tree and the tree in Figure la are presumably due to the smaller number of nucleotides used in the analysis, due to short length and sequence ambiguities of some sequences. SAgrobacterium tumefaciens P. marinus Med4 P. marinus MIT9107 P. marinus FP5 P. marinus Sargasso Sea isolate P. marinus MIT9303 Synechococcus WH7805 Synechococcus WH8103 Synechococcus PCC6301 Synechococcus WH7335 Phormidium PCC7375 Prochlorothrixhollandica DL1,,l P rochoro rtNp. - Synechocystis PCC6308 Myxosarcina PCC7312 l1oeothece PCC6501 ,yngbyaPCC7419

U II )scillatoriaPCC6304 us 10mfx ora paradoxacyanelle archantiapolymorpha chip. 'hlamydomonasreinhardii chip. Euglena gracilis chip. Cyanidium caldarium 14-1-1 chlp. 4 - rhromrnonna Annira chln Plectonema PCC73110 LP PseudanabaenaCCC OL-75-PS Anabaena PCC7120 FischerellaPCC7414 -- GloeobacterPCC7421 codon positions were removed from the analysis, petBID phylogenetic inferences were drawn from 258 nucleotide positions, while psbB and 16S rRNA gene analyses considered 333 and 1085 nucleotide postitions, respectively.

The phylogenetic position of P. marinus strain MIT9303 was ambiguous. While 16S rRNA gene trees inferred by the three standard methods consistently recovered the P. marinus four culture clade as a coherent group, they were inconsistent in their depictions of the relative branching order of five more deeply branching lineages in the

P. marinus/marineA Synechococcus lineage, including P. marinus MIT9303, the four culture clade, the two marine A Synechococcus strains and SAR7, a cloned sequence from the Sargasso Sea (Giovannoni et al. 1990) (Figure la, b, c). In bootstrap calculations, the branching order of these deeply branching lineages received less than 50% support (Figure la). The discordance in the branching order of P. marinus MIT9303 and marine A Synechococcus lineages inferred by different phylogenetic methods and their low bootstrap values suggests a near-simultaneous radiation of P. marinus and marine A Synechococcus lineages during evolutionary history (Hoelzer and Melnick 1994).

Effects of nucleotide substitution biases. We explored the possible effects of nucleotide substitution bias on our phylogenetic trees by labelling the psbB and petB/D neighbor-joining trees according to their third codon position G+C content (Lockhart and Penny 1992, Lockhart et al. 1992b) (Figures 2a, 3a). It is immediately apparent from inspection of these labelled trees that P. marinus isolates fail to cluster with chloroplasts despite similarities in their third codon position G+C content (Figures 2, 3), and that nucleotide substitution bias is therefore not the sole organizing factor in these trees. 16S rRNA gene trees also fail to group P. marinus with chloroplasts (Figure 1). However, within the P. marinuslmarineA Synechococcus cluster the phylogenetic branching pattern inferred by the majority of analyses is correlated with G+C content at third codon positions, with the Med4/MIT9107 phylogenetic pair showing considerably lower third codon position G+C content (0.18 to 0.27) than SS120 (0.32 to 0.34) and Synechococcus WH8103 (0.66 to 0.72). Although this clustering may reflect actual phylogenetic relationships in which substitution biases simply differ among lineages, it is also consistent with a systematic error in phylogenetic inference (Lockhart et al. 1992a). Again, there is no accepted molecular criterion for determining which hypothesis is correct.

To further explore the possible effects of nucleotide substitution biases, phylogenetic branching patterns were assessed using two methods of analysis thought to be insensitive to them (Woese 1991, Hasegawa and Hashimoto 1993): transversion analysis with neighbor-joining and, for the two protein-encoding sequences, amino acid parsimony. Transversion analysis inferred an identical branching pattern for Synechococcus WH8103 and the members of the P. marinus four culture clade as was found by standard neighbor-joining for all three gene sequences, although the relative branching order for deep branches in the P. marinus/marineA Synechococcus clade agreed with the maximum likelihood calculation in the 16S rRNA transversion tree (Figures 1, 2, 3). Relationships inferred by protein parsimony were also largely congruent with trees inferred by standard methods, except that P. marinus strains SS120 and FP5 were joined to form a sister clade to P. marinus Med4 and MIT9107 in the psbB protein parsimony tree (Figure 2b).

In sum, transversion and protein parsimony analyses do not suggest a consistent, alternative phylogenetic branching pattern, but instead reinforce the conclusions of the standard neighbor-joining, parsimony and maximum likelihood analyses: P. marinus Med4 is most closely related to P. marinus MIT9107, below which diverge P. marinus strains FP5 and SS120 in somewhat indeterminate order, with the most deeply branching culture being P. marinus MIT9303. The branching order of five deeply branching lineages in the P. marinuslmarineA Synechococcus cluster is not resolved.

Intergenic region and third codon position mismatches. Intergenic "nonsense" regions and third codon (silent) positions in protein-encoding genes are rapidly evolving, and therefore sensitive indicators of small amounts of evolutionary change. However, when compared to the first two codon positions in protein-encoding genes or to 16S rRNA genes, intergenic regions and third codon positions become rapidly saturated with nucleotide substitutions and (in intergenic regions) insertions and deletions. These tend to obscure evolutionary relationships when homologous nucleotide positions cannot be aligned or when the proportion of superimposed substitutions becomes large. Comparison of aligned sequences of petB/D and psbB from cultured strains of P. marinus, Synechococcus WH8103, chloroplasts and other cyanobacteria illustrates this point. In comparisons of P. marinuspetB/D intergenic regions, which range from 34 to 90 basepairs in length, homologous nucleotide positions could be identified only for a portion of the intergenic regions of strains Med4 and MIT9107 (Figure 5). Third codon position mismatches ranged from 41.1% to 64.3% (similarity 0.589 to 0.357) for psbB and from 39.5 to 66.1 percent (similarity 0.65 to 0.339) for petB/D, and were considered too large to give reliable phylogenetic information (Tables 2b, c) (for a similar analysis, see Bhattacharya et al. 1991). In addition, third codon positions were highly heterogeneous in G+C content (Table 3), which may bias phylogenetic inferences. Intergenic regions and third codon positions were therefore omitted from phlyogenetic analyses. Table 3. G+C Base compositions of sequences used in these analyses.

DetB/D Codon Dosition(s) All 1st2 3rd P. marinus Med4 0.37 0.47 0.18 P. marinus MIT9107 0.38 0.46 0.21 P. marinus FP5 0.40 0.45 0.29 P. marinus SS120 0.41 0.46 0.32 Synechococcus WH8103 0.57 0.49 0.72 Prochlorothrixhollandica 0.54 0.47 0.67 Synechococcus PCC7002 0.50 0.48 0.55 Nostoc PCC7906 0.49 0.49 0.49 Marchantia chloroplast 0.32 0.43 0.11 Zea maize chloroplast 0.41 0.47 0.29

DSbB Codon position(s) All ist2 3rd P. marinus Med4 0.39 0.47 0.25 P. marinus MIT9107 0.40 0.47 0.27 P. marinus FP5 0.41 0.48 0.26 P. marinus SS120 0.45 0.51 0.34 Synechococcus WH8103 0.58 0.54 0.66 Synechocystis PCC6803 0.51 0.50 0.52 Prochlorothrixhollandica 0.56 0.55 0.58 Synechococcus PCC7942 0.55 0.52 0.61 Anabaena PCC7120 0.48 0.53 0.39 Zea maize chloroplast 0.42 0.49 0.26

16S rRNA G+C composition

Agrobacterium tumefaciens 0.54 P. marinus Med4 0.53 P. marinus Pac7 0.53 P. marinus FP5 0.54 P. marinus SSW5 0.54 P. marinus MIT9303 0.54 Synechococcus WH8103 0.54 Synechococcus WH7805 0.54 SAR6 0.53 SAR7 0.54 SAR139 0.54 Synechococcus PCC6301 0.54 Prochloron sp. 0.52 Anabaena PCC7120 0.53 Marchantia polymorpha chloroplast 0.54 Figure 5. Comparison of petB/D intergenic region sequences for cultured strains of P. marinus and other members of the oxygenic photosynthetic radiation. Sequences are aligned according to amino acid translations in coding regions, and, with the exception of the sequence for P. marinus MIT9107, intergenic regions are arbitrarily represented as continuous with petB. An alignment gap has been inserted into the P. marinus MIT9107 intergenic region sequence in order to accentuate the similarity between the 5' end of this sequence and that from P. marinus Med4. Intergenic region sequences from Prochlorothrixhollandica, Nostoc PCC7906 and Marchantia polymorpha and Zea maize chloroplasts have been truncated. Sequence data for P. marinus MIT9303 are from this work, other attributions are as in Table 2c. rI

I i (II II I g

I

;I IC U PI u

E- L E,) I iI I

I~c I3I U 5I t : E~5

I I

I u~~ e I I

I I i IU

I I I$

Eg-

ElE

4 I r

EU): (iI 14

IU: U C, U A E-.ou

C,

tnt z E % aa 0 0 0) ;I £I £;U '- 2 8 U .t U 0 IIIU U7 12 j 0 ) '- 1 0~ CII 0 ). V

3 to to 'Ell

. 0" 0 0N I ** ~EI -q1 U C DISCUSSION

Chlorophyll a2 and b2-containingmarine prokaryotes arise from a single lineage within the cyanobacteria. The cultured P. marinus strains included in this study are phylogenetically restricted to a single lineage within the cyanobacteria, which they share with marine A Synechococcus strains WH8103 and WH7805. Neighbor-joining, parsimony and maximum likelihood analyses of nucleotide sequences, parsimony analysis of amino acid sequences, and analysis of transversions by neighbor-joining, all replicate this result, which links strains of P. marinus and marine A Synechococcus having very different G+C substitution biases, despite the presence in the dataset of chloroplast sequences having third codon position G+C content similar to those of P. marinus. The consistency of these results argues that the P. marinuslmarineA Synechococcus phylogenetic grouping represents a true evolutionary lineage, and not an artefactual cluster.

Branching patterns among four cultured isolates of P. marinus and Syenchococcus WH8103 are highly consistent. The branching order inferred for P. marinus Med4, MIT9107, FP5 and SS120 and Synechococcus WH8103 in phylogenies inferred from three gene sequences was consistent. It is therefore likely that gene transfer does not play a major role in the evolution of these organisms and that P. marinus exhibits clonal inheritence similar to that in Escherichia coli (Selander et al. 1987). This finding buttresses the assumption that prokaryotic organisms exhibiting closely related sequences at a single locus are likely to be genetically similar at most loci, and hence phenotypically similar. This assumption is implicit in most analyses of sequences cloned from natural populations (Young 1989). Phylogenetic relationships among P. marinus isolates do not correlate with geography. P. marinus Mediterranean strain Med4 and Pacific strain MIT9107, both isolated from mixed surface waters, were found to be closely related, while north Atlantic strain FP5 and Sargasso Sea deep chlorophyll maximum cultures SS120 and and MIT9303 fell into progressively more deeply branching lineages. There was thus no correlation between genetic distance and the geographic distance between sites of culture isolation, with the two Sargasso Sea isolates exhibiting less sequence similarity than the Med4/MIT9107 pair.

The different possible branching positions for P. marinus MIT9303 suggest different evolutionary scenarios for P. marinus and marine A Synechococcus pigments. The ambiguity in the branching position of P. marinus MIT9303 leaves open the question of the origins of P. marinus and marine A Synechococcus pigments. If the correct placement of P. marinus MIT9303 were as the deepest branch of the P. marinuslmarineA Synechococcus cluster, as indicated by the 16S rRNA gene parsimony and maximum likelihood trees (Figures lb, c), this would suggest that the chl a2 and b2 P. marinus phenotype was ancestral to the whole P. marinus/marineA Synechococcus cluster, implying further that cells capable of this phenotype arose from an ancestor shared with Synechococcus PCC6301 and gave rise to Synechococcus WH8103 (both strains being chl a l and phycobilisome-containing cyanobacteria, neither of which contains chl a2 or chl b (Rudiger and Schoch 1988, Waterbury et al. 1979, L. Moore, pers. comm.). If, on the other hand, the correct placement of P. marinus MIT9303 were at the base of a unitary P. marinus lineage, as inferred by neighbor-joining using 16S rRNA gene sequences (Figure la), then the phylogeny would be consistent with a single origin of the chl a2 and b2 phenotype without "reversion" to the chl al and phycobilisome-containing phenotype. The characterization of the P. marinus photosynthetic apparatus and its genes, with comparison to those of marine A Synechococcus is likely to yield the most conclusive data pertaining to this question.

Ambiguity in the branching order of deeply diverging lineages in the P. marinuslmarineA Synechococcus cluster precludes identification of some cloned sequences from natural populations. Most studies investigating the phylogenetic diversity of uncultured microorganisms in the sea have employed 16S rRNA gene sequences, many of which fall into the P. marinuslmarineA Synechococcus cluster (Giovannoni et al. 1990, Schmidt et al. 1991, Britschgi and Giovannoni 1991, DeLong et al., 1993, Fuhrman et al, 1993). While some of these sequences form close phylogenetic associations with cultured P. marinus or marine A Synechococcus strains (e.g. SAR6 and SAR139), others (e.g. SAR7) form deep branches in the P. marinuslmarineA Synechococcus cluster and are not specifically related to sequences from characterized cells (Figure 1). Because the branching order of P. marinus and marine A Synechococcus lineages in the cluster cannot at present be resolved, these cloned sequences cannot confidently be assigned to either taxonomic group. Assignment of phenotypes to these sequences must await the discovery of closely related sequences in phenotypically characterized cells, results of in situ hybridization experiments, or future refinements in the methods of phylogenetic inference. REFERENCES Bhattacharya, D., Stickel, S.K. and Sogin, M.L. (1991). Molecular phylogenetic analysis of actin genic regions from Achlya bisexualis (Oomycota) and Costariacostata (Chromophyta). J. Mol. Evol. 33:525-536. Brand, S.N., Tan, X. and Widger, W.R. (1992). Cloning and sequencing of the petBD operon from the cyanobacterium Synechococcus sp. PCC7002. Plant Mol. Biol. 20:481-491. Britschgi, T.B. and Giovannoni, S.J. (1991) Phylogenetic analysis of a natural marine bacterioplankton population by rRNA gene cloning and sequencing. Appl. Environ. Microbiol. 57:1707-1713. Campbell, L., and Vaulot, D. (1993). Photosynthetic picoplankton community structure in the subtropical North Pacific Ocean near Hawii (station ALOHA). Deep-Sea Res. 40:2043-2060. Chisholm, S.W., R.J. Olson, E.R. Zettler, R. Goericke, J. Waterbury, and N. Welschmeyer. (1988). A novel free-living prochlorophyte abundant in the oceanic euphotic zone. Nature, 334(6180):340-343. Chisholm, S.W., Frankel, S.L., Goericke, R., Olson, R.J., Palenik, B., Waterbury, J.B., West-Johnsrud, L. and Zettler, E.R. (1992). Prochlorococcusmarinus nov. gen nov. sp.: a marine prokaryote containing divinyl chlorophyll a and b. Arch. Microbiol. 157:297-300. DeLong, E.F., Franks, D.G. and Alldredge, A.L. (1993) Phylogenetic diversity of aggregate-attatched vs. free-living marine bacterial assemblages. Limnol. Oceanogr. 38:924-934. Felsenstein, J. (1988). Phylogenies from molecular sequences: inference and reliability. Annu. Rev. Genet. 22:521-65. Felsenstein, J. (1991). Phylip version 3.4 (University of Washington, Seattle, WA). Fitch. W.M. and Margoliash, E. (1967). Construction of phylogenetic trees. Science 155:279-284. Fuhrman, J.A., McCallum, K. and Davis, A.A. (1993). Phylogenetic diversity of subsurface marine microbial communities from the Atlantic and Pacific Oceans. Appl. Environ. Microbiol. 59:1294-1302. Geiskes, W.W.C., G.W. Kraay, A. Nontji, D. Setiapermana, and Sutomo. 1988. Monsoonal alternation of a mixed and a layered structure in the phytoplankton of the euphotic zone of the Banda Sea (Indonesia): A mathematical analysis of algal pigment fingerprints. Neth. J. Sea Res. 22:(2):123-137. Giovannoni, S.J., Turner, S., Olsen, G.J., Barnes, S., Lane, D.J. and Pace, N.R. (1988). Evolutionary relationships among cyanobacteria and green chloroplasts. J. Bacteriol. 170:3584-3592. Giovannoni, S.J., Britschgi, T.B., Moyer, C.L., and K.G. Field. (1990). Genetic diversity in Sargasso Sea bacterioplankton. Nature 345:60-62. Goericke, R. and Repeta, D.J. (1992). The pigments of Procholorcoccusmarinus: the presence of divinyl chlorophyll a and b in a marine prokaryote. Limnol. Oceanogr. 37:425-433. Goericke, R. and Repeta, D.J. (1993). Chromatographic analysis of divinyl-chlorophylls a and b in samples from the subtropical north Atlantic Ocean. Mar. Ecol. Prog. Ser. Goericke, R. and Welschmeyer, N.A. (1993). The marine prochlorophyte Prochlorococcuscontributes significantly to phytoplankton biomass and primary production in the Sargasso Sea. Deep-Sea Res. 40:2283-2294. Greer, K.L. and Golden, S.S. (1992). Conserved relationship between psbH and petBD genes: presence of a shared upstream element in Prochlorothrixhollandica. Plant Mol. Biol. 19:355-365. Hasegawa, M. and Hashimoto, T. (1993). Ribosomal RNA trees misleading? Nature 361:23. Hashimoto, T., Nakamura, Y., Nakamura, F., Shirakura, T. Adachi, J., Goto, N., Okamoto, K-i. and Hasegawa, M. (1994) Mol. Biol. Evol. 11:65-71. Herdman, M., Janvier, M., Waterbury, J.B., Rippka, R. and Stanier, R.Y. (1979) Deoxyribonucleic acid base composition of cyanobacteria. J. Gen Microbiol. 111:63-71. Hillis, D.M., and Dixon, M.T. (1991). Ribosomal DNA: Molecular evolution and phylogenetic inference. Quarterly Rev. Biol. 66:411-453. Hoelzer, G.A and Melnick, D.J. (1994) Patterns of speciation and limits to phylogenetic resolution. Trends Ecol. Evol. 9:104-107. Hultman, T., Stahl, S., Homes, E. and Uhlen, M. (1989). Nucl. Acids Res. 17:4937- 4946. Kallas, T., Spiller, S. and Malkin, R.C. (1988). Characterization of two operons encoding the cytochrome b6-f complex of the cyanobacterium Nostoc PCC7906 and highly conserved sequences but different gene organization than in chloroplasts. J. Biol. Chem. 263:14334-14342. Kimura, M. (1980). A simple method for estimating evolutionary rate of base substitutions through comparative studies of nucleotide sequences. J. Mol. Evol. 16:111-120. Knoth, K., Roberds, S., Poteet, C. and Tamkun, M. (1988). Highly degenerate, inosine- containing primers specifically amplify rare cDNA using the polymerase chain reaction. Nuleic Acids Res. 16:10371. Lake, J.A. (1994). Reconstructing evolutionary trees from DNA and protein sequences: paralinear distances. Proc. Natl. Acad. Sci. USA 91:1455-1459. Lang, J.D., and Haselkom, R. (1989). Isolation, sequence and transcription of the gene encoding the phtotsystem II chlorophyll-binding protein, CP-47, in the cyanobacterium Anabaena 7120. Plant Mol. Biol. 13:441-456. Lantoine, F. and Neveux, J. (1993). Spectrofluorometric analysis of photosynthetic pigments in the Gulf of Lions (Mediterranean Sea) with an emphasis on the autotrophic picoplankton. Ann. Inst. Oceanogr., Paris 69:173-175. Larsen, N., Olsen, G.J., Maidak, B.L., McCaughey, M.J., Overbeek, R., Macke, T.J., Marsh, T.L., and Woese, C.R. (1993) The ribosomal database project. Nucleic Acids Res. 21 Supplement:3021-3023. Letelier, R.M., Bidigare, R.R., Hebel, D.V., Ondrusek, M., Winn, C.D. and Karl, D.M. (1993). Temporal variability of phytoplankton community structure based on pigment analysis. Limnol. Oceanogr. 38:1420-1437. Li, H., Cui, X. and Arnheim, A. (1991). Analysis of DNA sequence variation in single cells. Methods 2:49-59. Li, W.K.W., Dickie, P.M., Irwin, B.D., Wood, A.M. (1992). Biomass of bacteria, cyanobacteria, prochlorophytes and photosynthetic eukaryotes in the Sargasso Sea. Deep-Sea Res. 39:501-519. Li, W.K.W., and M. Wood. (1988). Vertical distribution of North Atlantic ultraphytoplankton: analysis by flow cytometry and epifluorescence microscopy. Deep Sea Res. 35:1615-1638. Lignon, P.J.B., Meyer, K.G., Martin, J.A. and Curtis, S.E. (1991). Nucl. Acids Res. 19:4553. Lockhart, P.J., Penny, D. (1992). The problem of GC content, evolutionary trees and the origins of chl-a/b photosynthetic organelles: are the prochlorophytes a eubacterial model for higher plant photosynthesis? in Research in Photosynthesis, VoL III (N. Murata, ed.). Kluwer Academic, pp. 499-505. Lockhart, P.J., Howe, C.J., Bryant, D.A., Beanland, T.J. and Larkum, A.W.D. (1992a). Substitutional bias confounds inference of cyanelle origins from sequence data. J. Mol. Evol 34:153-162. Lockhart, P.J., Penny, D., Hendy, M.D., Howe, C.J., Beanland, T.J. and Larkum, A.W.D. (1992b). Controversy on chloroplast origins. FEBS Lett. 301:127-131. Lockhart, P.J., Beanland, Howe, C.J. and Larkum, A.W.D. (1992c). Sequence of Prochloron didemni atpBE and the inference of chloroplast origins. Proc. Natl. Acad. Sci. USA 89:2742-2746. Lockhart, P.J., Penny, D., Hendy, M.D., and Larkum, A.D. (1993). Is Prochlorothrix hollandicathe best choice as a prokaryotic model for higher plant chl a/b photosynthesis? Photosynthesis Res. 37:61-68. Lockhart, P.J., Steel, M.A., Hendy, M.D., and Penny, D. (1994). Recovering evolutionary trees under a more realistic model of sequence evolution. Mol. Biol. Evol. 11:605-612. Medlin, L.K., Williams, D.M. and P.A. Sims (1993). The evolution of the diatoms (Bacillariophyta). I. Origin of the group and assessment of the monophyly of its major divisions. Eur. J. Phycol. 28:261-275. Moore, L.R., Goericke, R., and Chisholm, S.W. (1994). The comparative physiology of Synechococcus and Prochlorococcus: influence of light and temperature on growth, pigments, flourescence and absorptive properties. Mar. Ecol. Prog. Ser., in press. Morden, C.W., Delwiche, C.F., Kuhsel, M., and Palmer, J.D. (1992). Gene phylogenies and the endosymbiotic origin of plastids. Biosystems 28:75-90. Morel, A., Ahn, Y.-H., Partensky, F., Vaulot, D. and Claustre, H. (1993). J. Mar. Res. 51:617-649. Neveux, J., Vaulot, D., Courties, C., and E. Fukai. (1989). Green photosynthetic bacteria associated with the deep chlorophyll maximum of the Sargasso Sea. C.R. Acad. Sci. Paris 308:9-14. Ohyama, K., Fukuzawa, H., Kohchi, T., Shirai, H., Sano, T., Sano, S., Umesono, K., Shiki, Y., Takeuchi, M., Chang, Z., Aota, S., Inokuchi, H. and Ozeki, H. (1986). Chloroplast gene organization deduced from complete sequence of liverwort Marchantiapolymorpha chloroplast DNA. Nature 322:572-574. Olsen, G.J. (1988). Phylogenetic analysis using ribosomal RNA. Methods Enzymol. 164:793-838. Olsen, G.J. (1990). Sequence editor and analysis package (University of Illinois, Urbana, IL). Olsen, G.J., Matsuda, H., Hagstrom, R and Overbeek, R. (1992). FastDNAml (Argonne National Laboratory, Argonne, IL). Olsen, G.J. and Woese, C.R (1993). Ribosomal RNA: a key to phylogeny. FASEB J. 7:113-123. Olson, R.J., Chisholm, S.W., Zettler, E.R., Altabet, M.A., and Dusenberry, J.A. (1990) Spatial and temporal distributions of prochlorophyte picoplankton in the North Atlantic Ocean. Deep Sea Res. 37, 1033-1051. Palenik, B. and Haselkorn, R. (1992). Multiple evolutionary origins of prochlorophytes, the chlorophyll b-containing prokaryotes. Nature 355:265-267. Partensky, R., Hoepffner, N., Li, W.K.W., Ulloa, 0. and Vaulot, D. (1993). Photoacclimation of Prochlorococcussp. (Prochlorophyta) strains isolated from the north Atlantic and the Mediterranean Sea. Plant Physiol. 101:285-296. Rock, C.D., Barkan, A. and Taylor, W.C. (1987). The maize plastid psbB-psbF-petB- petD RNAs encode alternative products. Curr. Genet 12:69-77. Rudiger, W., and S. Schoch. (1988). Chlorophylls. In Plant Pigments. T.W. Goodwin, ed. Academic Press. 1-60. Saitou, N. and Nei, M. (1987). The neighbor-joining method: a new method for reconstructing phylogenetic trees. Mol. Biol. Evol. 4:406-425. Sambrook, J., Fritsch, E.F. and Maniatis, T. (1989). Molecular Cloning: a laboratory manual. Cold Spring Harbor Laboratory Press, New York. Schmidt, T.M., DeLong, E.F., and Pace, N.R. (1991). Analysis of a marine picoplankton community by 16SrRNA gene cloning and sequencing. J. Bacteriol. 173:4371- 4378. Selander, R.K., Caugant, D.A., and Whittam, T.S. (1987). Genetic structure and variation in natural populations of Escherichiacoli. In Escherichia coli and Salmonella typhimunium: cellular and molecular biology (F.C. Neidhardt, ed.). Am. Soc. Microbiol. pp. 1625-1648. Sogin, M.L., Gunderson, J.H., Elwood, H.L., Alonso, R.A. and Peattie, D.A. (1989). Phylogenetic meaning of the kingdom concept: an unusual ribosomal RNA from Giardialamblia. Science 243:75-77. Sogin, M.L., Hinkle, G., Leipe, D.D. (1993). Universal tree of life. Nature 362:795. Swift, H. and Palenik, B. (1992). Prochlorophyte evolution and the origin of chloroplasts: morphological and molecular evidence. In Origins of Plastids (R.A. Lewin, ed.) Chapman and Hall, pp. 123-139. Swofford, D.L. (1991). PAUP version 3.0 (Illinous Natural History Survey, Champaign, IL). Tomioka, N. and Sugiura, M. (1983). The complete nucleotide sequence of a 16S ribosomal RNa gene from a blue-green alga, Anacystis nidulans. Mol. Gen. Genet. 191:45-50. Turner, S., Burger.Wiersma, T., Giovannoni, S.J., Mur, L.R., and N.R. Pace. (1989). The relationship of a prochlorophyte Prochlorothrixhollandica to green chloroplasts. Nature 337:380-382. Urbach, E., Robertson, D. and Chisholm, S.W. (1992). Multiple evolutionary origins of prochlorophytes within the cyanobacterial radiation. Nature 355:267-269. Vaulot, D., Marie, D., Olson, R.J. and Chisholm, S.W. (1994). Prochlorococcusis highly synchronized to the diel cycle in the equatorial Pacific and divides up to once a day. submitted. Vaulot, D., and Partensky, F. (1992). Cell cycle distributions of prochlorophytes in the northwestern Mediterranean Sea. Deep-Sea Res. 39:727-742. Vaulot, D., Partensky, F., Neveux, J., Mantoura, R.F.C. and C. Llewellyn. (1990). Wintertime presence of prochlorophytes in surface waters of the North-Western Mediterranean Sea. Limnol. Oceanogr. 35:1156-1164. Veldhuis, M.J.W. and Kraay, G.W. (1993). Cell abundance and Fluorescence of picoplakton in relation to growth irradience and nitrogen availability in the Red Sea. Netherlands J. Sea Res. 31:135-145. Veldhuis, M.J.W. and Kraay, G.W. (1990). Vertical distribution and pigment composition of a picoplanktonic prochlorophyte in the subtropical N. Atlantic: a study of HPLC-analysis of pigments and flow cytometry. Mar. Ecol. Prog. Seq. 68:121-127. Vermaas, W.F., Williams, J.G. and Arntzen, C.J. (1978). Sequencing and modification of psbB, the gene encoding the CP-47 protein of photosystem II in the cyanobacterium Synechocustis 6803. Plant Mol. Biol. 8:317-329. Vermaas, W.F.J., and Ikeuchi, M. (1991). Photosystem II. in The Photosynthetic Apparatus: molecular biology and operation (L. Bogorad and I.K. Vasil, eds.). Academic Press pp. 26-111. Waterbury, J.B. and Rippka, R. (1991). Subsection 1. Order Croococcales Wettstein 1924, emend. Rippka et al., 1979. in Bergey's Manual of Systematic Bacteriology, Vol. 3. J.T. Stanley, et al., eds. Williams and Wilkins. Waterbury, J.B., Valois, F.S., Franks, D.G. (1986). Biological and ecological characterization of the marine unicellular cyanobacterium Synechococcus. in Photosynthetic Picoplankton (T. Platt and W.K.W. Li, eds.) Can. Bull. Fish Aquatic Sci. 214:71-120. Waterbury, J.B., Watson, S.W., Guillard, R.R.L. and Brand, L.E. (1979) Widespread occurrence of a unicellular, marine, planktonic, cyanobacterium. Nature 277:293- 294. Widger, W.R. and Cramer, W.A. (1991). The b6/f complex. in The Photosynthetic Apparatus: molecularbiology and operation (L. Bogorad and I.K. Vasil, eds.). Academic Press pp. 149-224. Woese, C.R. 1987. Bacterial Evolution. Microbiol. Rev. 51:221-271. Woese, C.R., Achenbach, L., Rouviere, P. and Mandelco, L. 1991. Archaeal phylogeny: reexamination of the phylogenetic position of Archaeolobusfulgidus in light of certain composition-induced artifacts. System. Appl. Microbiol. 14:364-371. Yang, D., Oyaizu, Y., Oyaizu, H., Olsen, G.J., Woese, C.R. (1985). Mitochondrial origins. Proc. Natl. Acad. Sci. USA 82:4443-4447. Young, J.P.W. (1989). The population genetics of bacteria. in Genetics of Bacterial Diversity (D.A. Hopwood and K.F. Chater, eds.). Academic Press pp. 415-438. Chapter Three

GENETIC DIVERSITY IN NATURAL POPULATIONS OF PROCHLOROCOCCUSMARINUS IN THE

SARGASSO SEA AND THE GULF STREAM ABSTRACT

An investigation into the genetic structure of Prochlorococcusmarinus field populations in depth profiles from the Sargasso Sea and the Gulf Stream revealed a high degree of genetic heterogeneity within water samples, detected by partial sequencing of cloned PCR products amplified from flow cytometrically sorted cells. Overlapping sets of alleles were recovered from the two water columns, from different depths within each water column and from flow cytometrically distinguishable subpopulations within water samples, suggesting that each of these populations drew its membership from a single gene pool.

INTRODUCTION

Prochlorococcusmarinus is a unicellular prokaryote which makes up a high proportion of the photosynthetic picoplankton throughout wide regions of the tropical and subtropical marine environment (Chisholm et al 1988, Giskes et al. 1988, Olson et al. 1990, Vaulot et al. 1990, Li et al. 1993, Campbell and Vaulot 1993, Veldhuis and Kraay 1993). Sometimes referred to as a prochlorophyte (Lewin 1981, Chisholm et al. 1988), P. marinus is recognized by its tiny size (0.6 to 0.8 ptm) and unique set of photosynthetic pigments, which include divinyl chlorophylls a and b (chl a2 and b2) and specifically do not include phycobilipigments, typical of most other cyanobacteria.

In efforts to understand the role of this organism in the marine microbial community, a number of recent studies have characterized the water column distribution of P. marinus in terms of cell number, biomass, productivity, chlorophyll fluorescence and DNA content at several locations (Giskes et al. 1988, Li and Wood 1988, Olson et al. 1990, Vaulot et al. 1990, Li et al. 1992, Vaulot and Partensky, 1992, Cambell and Vaulot 1993, Goericke and Repeta 1993, Goericke and Welschmeyer 1993, Vaulot et al. 1994) and over time (Olson et al. 1990, Cambell and Vaulot 1993). These studies have established that P. marinus is present throughout the euphotic zone at all times of the year and numerically dominates the picophytoplankton at oligotrophic, deep chlorophyll maxima during thermal stratification. Other, parallel studies have analyzed physiological and pigment differences among cultured P. marinus isolated from different depths and locations (Morel et al. 1993, Partensky et al. 1993, Moore et al. 1994), identifying differences in their growth rate response to varying levels of illumination which suggest that some genetic variants may have relatively greater growth rates in brightly lit surface waters, while others may have growth advantages at dimly lit depths (Moore et al. 1994).

Med4 and SS120, a pair of cloned isolates from 5 m and 120 m in the in the Mediterranean and the Sargasso Sea, respectively, are the most extensively characterized P. marinus strains (Partensky et al. 1993, Morel et al. 1993, Moore et al. 1994). Each can grow over a wide range of light intensities, but consistent with having been isolated at the surface, Med 4 is able to grow at high light intensities (>100 kE m-2 s- 1) at which SS120 is unable to grow, and SS120, isolated from a deep chlorophyll maximum (DCM) at 120 m, grows at low light intensities (2 - 6 gE m- 2 s-1) at which Med4 shows no growth (Moore et al. 1994). The two isolates differ in their compliment of photosynthetic pigments as well, with SS120 elaborating "normal" chlorophyll b (chl bl) at light -2 intensities above 20 pE m s- 1 and exhibiting a tenfold higher chlorophyl b (chl b1 + chl b2) to chl a2 ratio (chl b/a2 ratio) than Med4 over the entire range of growth light intensities (0.05 to 0.15 for Med4 and 0.4 - 2.4 for SS120) (Morel et al. 1993, Partensky et al. 1993, Moore et al. 1994). Med4 contains no detectable chl b1 under any conditions tested (Morel et al. 1993, Partensky et al. 1993, Moore et al. 1994). Phylogenetic analyses using four gene sequences, 16S ribosomal RNA (rRNA), psbB, petB/D and rpoC, have shown that cultured strains of P. marinus fall into a single phylogenetic cluster which they share with the marine A Synechococcus, phycobilisome- containing cyanobacteria with which they also share their marine habitat (Swift and Palenik 1992, this thesis Chapter Two). P. marinus Mediterranean strain (Med4), derived from strain MED (Chisholm et al. 1992, Partensky et al. 1993) and Pacific strain (MIT9107), both isolated from mixed surface waters (Chisholm et al. 1992, J. Dusenberry, pers. comm.), were found to be closely related, while DCM culture SS120 and others fell into more deeply branching lineages (this thesis Chapter Two). There was no correlation between genetic distance and the geographic distance between sites of culture isolation, with two Sargasso Sea isolates exhibiting less sequence similarity than the Mediterranean and Pacific pair. Fractional sequence similarity among four P. marinus isolates ranged from 0.994 to 0.989 for 16S rRNA and from 0.973 to 0.934 at the first two codon positions of petB and D (0.838 to 0.788 at all codon positions). Intergenic region sequences between petB and D were similar for Med4 and MIT9107, while sequences from other cultures were so divergent that homologous nucleotide positions could not be identified (this thesis Chapter Two).

Data from HPLC and flow cytometric studies suggest that genetic differences may distinguish P. marinus populations at different depths in stratified water columns, and that different genetic variants may coexist within water samples. At sites in the

Atlantic, Pacific and in the Red Sea, ratios of chl b/a2 were found to be lower near the surface than at the DCM (Veldhuis and Kraay 1990, 1993, Cambell and Vaulot 1993,

Goericke and Repeta 1993), with the range of chl b/a2 ratios within a water column (0.15 to 2.9 in the Sargasso Sea) being greater than those observed in individual cultures (0.05 to 0.15 for Med4 and 0.4 - 2.4 for SS120) (Goericke and Repeta 1993). HPLC analysis of samples from from DCM's in the subtropical Pacific has shown that cells passing through a 0.65 gtm filter exhibit 1.7-fold lower ratios of chl b2/a2 than cells retained by the filter, suggesting that subpopulations with different pigment ratios may coexist in the water sample (Letelier et al. 1993). While the flow cytometric study provided data on fluorescence ratios for individual P. marinus cells excited at 457 and 488 nm, exploiting the different fluorescence excitation properties of chl a2 and b at these wavelengths to estimate pigment ratios (Cambell and Vaulot 1993), the HPLC studies examined bulk chlorophyll, and so could have included chl b from cells other than P. marinus.

Differences in chl b/a2 ratios in these HPLC reports are therefore not direct evidence of genetic heterogeneity between P. marinus at different depths.

Flow cytometry of field samples from the Pacific, the Red Sea and the Sargasso Sea have occasionally identified "multiple populations," consisting of discrete distributions of chlorophyll autofluorescence and light scatter from individual cells (Cambell and Vaulot 1993, Veldhuis and Kraay 1993, B. Binder, R. Olson, J. Dusenberry, E. Zettler, unpublished observations). These flow cytometric subpopulations may derive from genetically distinct subpopulations which express different pigment phenotypes when exposed to uniform conditions, but they could also derive from samples containing recently mixed, genetically similar populations acclimated to different conditions of light or nutrient supply (Dusenberry and Chisholm in preparation).

Data consistent with the possibility of genetic heterogeneity within P. marinus populations at discrete depths have also come from molecular cloning experiments (Giovannoni et al. 1990, Britchgi and Giovannoni 1991, Schmidt et al. 1991, DeLong et al. 1993, Fuhrman et al. 1993). Investigations in which 16S ribosomal RNA gene sequences were cloned from uncharacterized marine picoplankton have recovered a number of sequences which fall into the P. marinuslmarineA Synechococcus cluster. While some of these sequences (e.g. SAR6 and SAR139) can be phylogenetically linked to cultured P. marinus or marine A Synechococcus, others (e.g. SAR7) form deep branches that are not robustly affiliated with either taxon, and so cannot be identified as P. marinus or Synechococcus with the data at hand (Urbach et al. 1992, this thesis Chapter Two). In order to assess the the genetic structure of P. marinus populations (as defined by their flow cytometric signature), therefore, it is necessary to physically separate P. marinus from other members of the community.

We report the results of a direct analysis of P. marinus populations using cells sorted by flow cytometry as starting material. P. marinus were functionally defined for this study as cells exhibiting the P. marinus flow cytometric "signature:" characteristic values for red fluorescence and forward light scatter (measured relative to standard 0.57 pm beads), in the absence of orange fluorescence, when excited by 488 nm laser light (Chisholm et al. 1988). Genotypes were sampled from sorted P. marinus by PCR amplification of the petBID locus 1 with cloning and single run, partial sequencing of the resulting clones (Adams et al. 1991, Adams et al. 1992, Kahn et al 1992). Preliminary analysis ofpetB/D amplification products by direct sequencing revealed the presence of superimposed sequences differing mostly in the intergenic region, suggesting the presence of multiple petB/D alleles (Appendix I). 21 to 28 clones were analyzed for each of eight sorted samples from different depths in the Sargasso Sea and the Gulf Stream, and including subpopulations of two flow cytometrically defined "double populations" in the Gulf Stream depth profile. Questions addressed are whether P. marinus field populations at discrete depths are genetically homo- or heterogeneous, and whether genetic differences can be detected between populations at different depths, between

1The petB/D locus includes the 3' end of petB, an intergenic region, and the 5' end of petD, (this thesis Chapter Two). These genes encode subunits of the photosynthetic bff complex and are single-copy, linked cistrons with a consistent orientation in all photosynthetic prokaryotes and organelles in the literature (Vermaas and Ikeuchi 1991, Greer and Golden 1992). subpopulations in multiple populations (as differentiated by their fluorescence and forward light scatter per cell), and between water columns in adjacent, but hydrographically distinct water masses. Gene sequences obtained from field populations are compared to sequences from cultured cells.

METHODS

Characteristics of sampling sites and flow cytometric signatures. P. marinus were collected from depth profiles in the Sargasso Sea (330 58.65'N, 660 6.63'W, 10:00 local time, 11 July 1993) and the Gulf Stream (370 30.68'N, 680 13.69'W, 08:00 local time, 17 July 1993) during cruise 9306 of RV Columbus Iselin. The Sargasso Sea sampling site exhibited a typical summer hydrographic profile, with a chlorophyll-poor, mixed upper layer overlying thermally stratified waters and a pronounced DCM (as indicated by in vivo chlorophyll fluorescence measurements) at 70 m. The Gulf Stream profile was more complex, with relatively low concentrations of chlorophyll at the surface and a pair of subsurface chlorophyll maxima at 85 m and 150 m (Figure 1). Samples were collected at depths predicted to contain the most different P. marinus populations, avoiding samples at the immediate surface which are difficult to visualize using the sorting flow cell. Sargasso Sea samples were collected at 40 m, 70 m (the DCM) and below the DCM at 120 m. Gulf Stream samples were collected at 50 m, 85 m (the shallower DCM) and at 135 m, between the two DCM's (Figure 1).

Flow cytometric characteristics of P. marinus populations varied with depth according to patterns familiar from previous investigations (Olson et al. 1990): at each depth in the Sargasso Sea P. marinus populations formed a single cluster on two dimensional scatter plots of chlorophyll fluorescence versus forward light scatter for Figure 1. Depth profiles of temperature (thin lines) and in vivo chlorophyll fluorescence (heavy lines) at the Sargasso Sea (330 58.65'N, 660 6.63'W, 10:00 local time, 11 July 1993) and the Gulf Stream (370 30.68'N, 680 13.69'W, 08:00 local time, 17 July 1993) sampling sites. Open squares on the in vivo cholrophyll fluorescence plot mark depths at which samples were collected. 0oo -

e,4 " 0I 0a -. ) ;- o *-~ c I-0 0 Q

0 0 o

c I- bQ 0 U ct 00 r-. 0

oo

(u) qldaG individual cells, with the distribution of values for each parameter approximating a lognormal distribution. Modes for chlorophyll fluorescence and forward light scatter showed an increase with depth, reflecting photoacclimative increase in chlorophyll per cell, in addition to possible genetic differences between populations at different depths. In the Gulf Stream, the P. marinus surface population again showed a unimodal distribution, but populations at 85m and 135m were double ("multiple") populations (Figure 2).

Sample concentration and flow cytometric sorting. Samples were collected using Niskin (30 1)or Go-Flo (10 or 30 1)bottles. Subsamples (300 to 750 ml) were concentrated to approximately 7 ml over 0.45 gpm Durapore filters (Millipore) under light vacuum (5" Hg) and cells adhering to the filter resuspended by vigorous pipetting. P. marinus cell recoveries as measured by flow cytometry in similar concentrates were approximately 50%. Concentrates were immediately sorted by flow cytometry aboard ship using an Epics V flow cytometer (Coulter Electronics) or frozen in liquid nitrogen for sorting in the laboratory using an EPICS 753. Frozen cells were sorted over a period of about about 20 minutes immediately after thawing to avoid loss of autofluorescence, which can start to occur at about 30 minutes (Vaulot and Xiuren 1988, J. Dusenberry and R. Olson, unpublished observations). P. marinus populations were flow cytometrically defined according to their red fluorescence (660-700 nm) in the absence of orange fluorescence (540-630 nm) and by their characteristic forward light scatter (relative to 0.57 gim beads) when illuminated by blue laser light (488 nm) (Chisholm et al. 1988). Bitmap sort windows were drawn to separate P. marinus from other constituents of the photosynthetic planktonic community when P. marinus populations were unimodal, and to isolate members of flow cytometric subpopulations when multiple populations were observed. Approximately 106 P. marinus cells were collected in each sorted sample. Figure 2. Contour plots for flow cytometric distributions of P. marinus sorted from Sargasso Sea (a-c) and Gulf Stream (d-e) field samples. Sorted (sub)populations used for genetic analysis are outlined. Red Fluorescence signals of the dimmly fluorescing cells in the 40 m Sargasso Sea sample (a) (processed in the laboratory) were enhanced by increasing the laser power to 1.2 watts PMT voltage to 1500 volts. Other samples (processed at sea) were processed using laser power of 1.0 watts and a PMT voltage of 1100 volts. To minimize the possibility of contaminaton standard beads were not added to samples used for sorting; distributions (e) and (f) are from reference samples. P. marinus cells were identified by their characteristic red vs. FALS signature and absence of orange phycoerythrin fluorescence (orange fluorescence data not shown). Sargasso Sea samples: (a) 40 m, (b) 70 m, (c) 120 m. Gulf Stream samples: (d) 50 m, (e) 85 m, (f) 135 m. Red Fluorescence: in vivo fluorescence of chlorophyll excited by 488 nm blue laser light. Forward Angle Light Scatter: forward angle light scatter of 488 nm laser light, a proxy for cell size. Sargasso Sea Gulf Stream

d rm 1-i a

c

" l i ; I, 1 I? I I I !1 1' II1 1 -

-1T1IC~ 1 e

1r

1

I 0.57 gm beads

l ' l tllltl11ll1lllI11lf! !11 fII l I | ......

log Forward Angle Light Scatter log Forward Angle Light Scatter Isolation of genomic DNA. A modification of the protocol of Li et al. (1991) was used to prepare DNA from sorted cells. Cells from the flow cytometric sort were concentrated above 0.2 gm pore size Durapore membranes using Ultrafree-MC filter units (Millipore) spun at 2000 x g and washed twice with cell suspension buffer (0.5 M NaCl, 10 mM Tris [pH 8.0] 10 mM EDTA). Cells were resuspended in a final volume of 100 pl cell suspension buffer and lysed by addition of 12 tl 0.5 M DTF and 6 pl 10 M NaOH followed by a 10 minute incubation at 65 0C. After addition of 120 pl neutralization buffer (90 mM Tris [pH 8.0] plus 0.5 mol/l HCI), DNA was precipitated by addition of 1 tl 20 mg/ml glycogen (Boehringer) and 0.6 ml cold 100% ethanol. Precipitates were stored at -20 0C for several months. In the laboratory, crude DNA precipitates were collected by centrifugation at 16,000 x g for 30 min at 40 C, washed with 70% ethanol and resuspended in TE (10 mM Tris [pH 8.0] 1 mM EDTA). DNA was stored at -200C.

Construction of the petB/D clone library. Portions of the petB and D genes and their intergenic region were amplified using degenerate oligonucleotide primers containing inosine (Knoth et al. 1988). Priming sites correspond to highly conserved regions in petB and D amino acid sequences, with forward primer PPETBD314 (5'ATGATGGTIYTIATGATGAT) corresponding to nucleotide positions 274 to 293 in the Nostoc PCC7906 petB coding sequence (Kallas et al. 1988), and reverse primer PPPETBD1160R (5'-CCRTARTARTTRTGICCCAT) corresponding to nucleotide positions 64 to 83 in the Nostoc PCC7906 petD sequence (Kallas et al. 1988). 100 pl PCR reactions contained DNA isolated from approximately 105 sorted cells, 5 U of Taq DNA polymerase (Promega) and a PCR reaction cocktail containing 1 x Mg-free PCR reaction buffer (50 mM KCI, 10 mM Tris-HCl [pH 9.0 at 25 0 C], 0.1% [wt/vol] Triton X-

100), 1.5 mM MgC12, 200 pM (each) dATP, dCTP, dGTP and dTTP, 0.5 pM PPETBD314 and 1 gM PPETBDI 160R. After initial denaturation for 2 min at 940C, thermal cycle parameters were 1 min 940 C, 1 min 470C, 1 min 720 C (5 cycles), 1 min 940C, 1 min 520C, 1 min 72 0C (10 cycles), I min 940 C, 1 min 570C, 1 min 720C (25 cycles), followed by 10 min 720C. PCR products from pairs of replicate reactions were pooled and separated by electrophoresis in a 1% agarose gel. Products in the 400 to 800 basepair region of the gel were excised and purified with GeneClean (Bio 101). 1/30th to 1/1000th of each primary amplification product was reamplified for 30 cycles of 1 min 940C, 1 min 570C, 1 min 72 0C, followed by 10 min 720 C, in reaction cocktail with 5 U Taq polymerase added during the first 570 C incubation. Control reactions lacking template DNA gave no amplication products for either primary amplifications or reamplifications, as judged from agarose gels.

Pooled reamplification products from pairs of replicate reactions were excised from a 1% agarose gel, purified with GeneClean and half to all of each product was inserted into pCRII vector (TA Cloning Kit, Invitrogen) according to the manufacturer's instructions. OneShot Competent cells were transformed using half of each ligation mixture. Both white and blue colonies were picked from LB plates (0.5% tryptone, 1% yeast extract, 0.5% NaCl, 1.5% Bacto-agar) containing 40 gtg/ml X-gal and 50 gg/ml of either ampicillin or kanamycin. Plasmid DNA was isolated by InstaPrep (5prime- 3prime) or standard alkaline lysis (Sambrook et al. 1989). Plasmids were digested with restriction endonuclease EcoRI and the digests analyzed by agarose gel electrophoresis. Clones containing 400 to 800 basepair inserts were chosen for sequence analysis.

DNA sequencing and database entry. Double-stranded plasmids containing PCR fragment inserts in random orientation were sequenced using Sequenase version 2.0 (USB) and primer PTASP6 (5'GATCCACTAGTAACGGCCG), complimentary to vector sequence flanking the pCRII cloning site. DNA sequences were entered into a sequence alignment editor (Olsen 1990), translated into amino acid sequences and aligned with a petBID sequence database according to conserved regions in the amino acid sequence (Widger and Cramer 1991, Greer and Golden 1992). In order to obtain sequences across the intergenic region for all clones recovered, clones giving sequence at the 5' (PPETBD314) end of the amplified petBID fragment, and which did not match sequences already in the database, were sequenced from the opposite side of the pCRII cloning site using PTAT7 (5'GAGCGGCCGCCAGTGTGA), which was closer to the intergenic region. This procedure yielded some sequences spanning the entire length of the cloned fragment, while others covered only portions of the 3' end of petB, the intergenic region and the 5' end of petD. Sequences, which were determined in a single run and in one direction only (Adams et al 1991, Adams et al. 1992, Kahn et al. 1992), ranged from 71 to 562 basepairs in length. Clones with sequences which did not align with petB or D were eliminated from the analysis.

Sequence comparisons. Fractional sequence mismatch values were computed using the computer program of Olsen (1988, 1990). The phylogenetic tree was inferred using the Phylip version 3.4 neighbor-joining program (Felsenstein 1991) and Kimura genetic distance calculations (2:1 transition:transversion ratio) (Kimura 1980). To minimize effects of sequence determination and PCR errors, alignment gaps were omitted from sequence comparisons.

Population genetics calculations. Values for E(Sn) were calculated using COMPAH90 (E.D. Gallagher, University of Massachusetts/Boston). RESULTS

Sequence diversity in the combined dataset. The combined dataset for all eight sorted samples contained 191 clones with inserts homologous to petB and D. Among the 191 clone sequences, 71 alleles were identified as deriving from genetically distinct individuals in the sampled populations (Table 1).

The 71 alleles were identified according to a two step process, employing both qualitative and quantitative criteria to distinguish true genetic differences from sequence mismatches arising from PCR and sequence determination errors. In the first, qualitative step, sequence comparisons were used to identify 59 sets of clones among which intergenic regions varied in length from 19 to 91 basepairs, and in sequence to such a degree that homologous nucleotides could not be identified for use in sequence alignment. Differences among these 59 sets were deemed qualitatively different from PCR and sequence determination errors (which were nonetheless undoubtedly present). Most of the 59 sets of clones having unalignable intergenic regions contained single or identical clones, or clones exhibiting minimal differences attributable to error, but three sets contained clones having sufficient mismatch (up to 19.1%) to suggest that they had been amplified from genetically distinct, though related individuals. In the second, quantitative step, a criterion of 3.5% mismatch was applied as an upper bound for sequence differences considered indistinguishable from PCR plus sequence determination errors in order to subdivide the three sets of clones and add an additional Table 1: Frequency of genetic alleles among cloned amplification products from flow cytometrically sorted samples. Allele / Sample G50 G85Br G85D G135Br G135D S40 S70 5120 TOTAL P. marinus 1 Allelel 12 14 2 Allele2 3 Allele3 4 Allele4 5 Allele5 16 5 6 Allele6 1 7 Allele7 8 AlleleS 9 Allele9 10 Allelell 11 Allelel2 12 Allelel3 13 Allelel4 14 AllelelS 15 Allelel6 16 Allelel7 17 Allele19 18 Allele21 1 19 Allele22 1 3 20 Allele23 1 21 Allele25 1 22 Allele26 3 23 Allele27 2 24 Allele28 1 25 Allele29 1 26 Allele30 1 27 Allele31 1 2 28 Allele32 1 29 Allele33 1 30 Allele35 1 31 Allele36 1 32 Allele37 1 33 Allele38 1 34 Allele39 1 35 Allele41l 1 36 Allele42 1 37 Allele43 1 38 Allele44 1 39 Allele45 1 40 Allele46 1 41 Allele47 1 42 Allele48 1 2 43 Allele50 1 44 Allele51 1 45 Allele52 1 1 46 Allele54 1 1 47 Allele56 1 1 48 Allele57 1 1 49 Allele58 1 1 50 Allele59 1 1 51 Allele60O 1 2 52 Allele61l 1 1 53 Allele62 1 1 54 Allele63 1 1 55 Allele64 1 1 56 Allele65 1 2 57 Allele66 1 1 58 Allele67 1 1 59 Allele68 1 1 60 Allele69 1 61 Allele71 4 62 Allele72 1 63 Allele73 1 64 Allele74 1 65 Allele75 2 66 Allele76 1 67 Allele77 1 68 Allele78 1 P. marinus Total 23 23 24 23 23 19 28 24 187 No. different Alleles 6 14 6 6 16 8 15 21 67 E(S19) 5.3 12.1 5.2 5.4 13.7 8.0 10.5 16.8 10.9 Chloroplast-like 69 Allele20 70 Allele24 71 Allele40 TOTAL 23 23 25 24 23 21 28 24 191

Abbreviations: G50, 50m; G85Br and G85D, Bright and Dim clouds of 85m double population; G135Br and G135D, Bright and Dim clouds of 135m double population; all from the Gulf Stream depth profile. S40, 40m; S70, 70m; S120, 120m; all from the Sargasso Sea depth profile.

E(S19), estimated number of different alleles for a constant sample size of 19 clones.

83 12 alleles to the dataset2. The subdivision of these three sets of clones according to the quantitative criterion resulted in the total of 71 alleles (Table 2).

Pairwise comparisons among alleles identified by subdivision of sets of clones having similar intergenic regions show a mean fraction of sequence mismatch at third codon positions (fractional mismatch at third codon positions/3 x fractional mismatch at all codon positions) of 0.82, significantly different from the expectations of random PCR and sequence determination errors (this thesis Appendix II).

A prototype sequence (the longest and least ambiguous) was chosen to represent each allele in sequence comparisons and aligned with other sequences in the database according to conserved regions in amino acid translations (Figure 3, this thesis Appendix IV). Translations of prototype sequences were highly homologous to petB and D sequences already in the database, with all translations containing phylogenetically conserved amino acid residues, including histidines at positions 186 and 201 in the Synechococcus PCC7002 sequence, hypothesized to participate in heme binding (Carter et al. 1981, Widger and Cramer 1991, Hope 1993), when sequence was available for these positions. The most striking differences among alleles were in the intergenic region, which differered in length and sequence among alleles. In coding regions pairwise nucleotide comparisons among the 71 alleles showed that they differed by an average of 24.1% ±4.9% (Figure 4a), with most differences occurring at third codon positions which differed by 51.4% ±9.5% (Figure 4b). The more highly conserved first two codon postions differed by only 10.8% ±6.8% (Figure 4c). Since the overwhelming

2The 3.5% upper limit criterion for sequence mismatch among clones assigned to a single Allele was arrived at by an optimization protocol which considers the fraction of sequence mismatch at third (silent) codon positions for sequence comparisons along a scale of total sequence mismatch. Within the bounds of theoretical limits, the criterion divides the set of pairwise sequence comparisons within the three sets of alleles having similar intergenic regions into the two subsets having the greatest possible difference in their mean third codon position to all codon position mismatch ratios (this thesis Appendix II). Table 2a. Fractional sequence mismatch at all sequence positions, including intergenic regions (x1000, below the diagonal) and fraction of sequence mismatch located at third codon positions (fractional mismatch at third positions/3 x fractional mismatch at all codon positions, for coding regions only) (x100, above the diagonal) for pairwise comparisons of prototype sequences for Alleles having intergenic regions alignable with Allele 1.

1 2 3 4 5 6 7 8 9 10

1. Allel - 082 089 094 085 100 090 090 081 080 2. Alle2 126 - 092 083 079 080 078 080 078 073 3. Alle3 158 124 - 091 085 086 084 079 089 080 4. Alle69 057 133 160 - 084 100 083 085 083 078 5. Alle71 080 167 173 095 - 100 079 080 071 072 6. Alle72 059 148 134 051 057 - 077 080 071 059 7. Alle73 073 156 169 070 077 076 - 085 060 078 8. Alle74 148 166 190 153 181 154 143 - 071 084 9. Alle75 081 157 155 077 072 048 086 155 - 057 10. Alle76 119 154 191 129 180 147 142 142 164 -

Table 2b. Fractional sequence mismatch at all sequence positions, including intergenic regions (x1000, below the diagonal) and fraction of sequence mismatch located at third codon positions (fractional mismatch at third positions/3 x fractional mismatch at all codon positions, for coding regions only) (x100, above the diagonal) for pairwise comparisons of prototype sequences for Alleles having intergenic regions alignable with Allele 4.

1 2 3 4

1. Alle4 - 102 101 079 2. Alle41 132 - 102 091 3. Alle77 042 138 - 079 4. Alle78 045 160 060 -

Table 2c. Fractional sequence mismatch at all sequence positions, including intergenic regions (x1000, below the diagonal) and fraction of sequence mismatch located at third codon positions (fractional mismatch at third positions/3 x fractional mismatch at all codon positions, for coding regions only) (x100, above the diagonal) for pairwise comparisons of prototype sequences for Alleles having intergenic regions alignable with Allele 17.

1 2

1. ALLE17 - 071 2. ALLE31 075 - Figure 3. Intergenic region sequences for 71 petB/D alleles recovered from the Sargasso Sea and Gulf Stream field samples (this work, Genbank accession numbers ), cultured P. marinus strains Med4, SS120, FP5, MIT9107 (this thesis Chapter Two) and MIT9313 (this work, Genbank accession number), Synechococcus strains WH8103 (this thesis Chapter II) and PCC7002 (Brand et al. 1992), cyanobacteria Prochlorothrix hollandica (Greer and Golden 1991) and Nostoc PCC7906 (Kallas et al. 1988) and chloroplasts from Chlorellaprotothecoides (Reimann and Kueck 1989), Marchantia polymorpha (Ohyama et al. 1986) and Zea maize (Rock et al. 1987). A potion of the sequence alignment is shown, with the final six codons of petB at left and the initial eight codons of petD at right. Intergenic region sequences are arbitrarily represented as contiguous with petB, with alignment gaps inserted between the end of most intergenic region sequences and the beginning of petD. Intergenic region sequences for P. hollandica,Nostoc PCC7906, C. protothecoides,M. polymorpha Z maize are truncated. Designations on nucleotide lines are clone names. Allele designations are indicated on the translation lines below. i r-) "-4 -4 Eý Eýf - E E- 1E- E-4iE-4 i iE- i z E-4i E4 i i0iZ 4) CC 5. 1CDMC I 0i 64 C, IM. eel b44 I I14 8 0

E- 14 UE- H E4 E- 1-4UE4 >U -4 U -1 -4> X H o

E4 -1E4 U I V

0, CA t ~ I 4

-4 2 .*~2~ *2~ X~2CE-4l

L)(d$ iL) 44 to E-4I i iU E4. L 044 4 4 I ,4I4- (9 V I I

II 4, 4 4~ U 1

0 ERI I $4(d 04 1 4 1 1 1 1 U 1 4 464 4-f 4 i E- 4 U

4 U 4 C 4 - 4 E IU4 C

(i 4-4 4 4 i 4-)

4 41 Ty U UU U U r.4 U 0 L 944 Cl) U U I I 44.40 - f< U U 4 4. ~I~ 4 34 I4 I4 0 U c 4 4 f--4 P. . 4 E- U I. I

OEV 4P L~4 -

ti E-i 0 -H0( Uo 06 E 0 4 run U .U: .U: U.UWI.)R U: 4E- .U: · ) u U UE 00 oasiUE- EE E. E- P 4 0 U P U1 5U * U'4.U:4 010.-(a -- E--I, I I ER 1 E41 R, - p 10 L- CR E-4 I 1 Cl 1 E- l--IE - n 4L 4t4 - I44 14 f4 II I 1 0.0 E-4E44- 1 E4-4E ,-4 u a4 i m ua . a.ua.ci L -1 1 ' 1 6u0 uIuu M ..u4I. U~.~t.~ l . 1 l

0.- 4E-u4vl u I U I U IU -4 4 JJ14 ~U . V1 .1F U 4U4 1u 1U4U4U1U4U4V 0 4 04 4UUUU4U4U:.:U r3 'U

00 ~ý E- E414~ E-E4~ ~)C)CI

0 %0 In 00 0% C4 r, 0t n cq cn 0% 0 r- OD m' 0 .4 .-e4 1 0 .4 .4 4 .- 41N 1 (" tq NO40 LA LA 0ý LA 1-4 0.-q to 0 0 M 1 4 .- 4 eq f %D ý4 ý4ý 4c -4- qLn iD1 - 04G4 o000, -4 eq eq ml Ln w D O 41-o0O 400% 0 "% C4 4 m4t'. viIA0 a - 0 0 m .- 4-4-4 .-4 .q 14 .4 14 -4 t C30 % C04 e C'l) W444 eq ('I ('C4 cq 0 0 0 0 w0IV W 0 )40W 04 0W 0W4) W4)W40) 4 4)M40)140L4)W0 0 0000000M 000000Cowo )0 0 r 0-0 -4 0 .-0-14ý M M 40r-40 M-4 0 4M 4-r 0M-4 0-4-44a M - 0 40M-4CO-i 40M-4 i) '-4 Q) 4 ('4-4 LA r-i LA r4 V) ,4 LiA4LA -i LA r-4 W toM4 La PC4LA .4 M~ 4 Lo .4 00 ,4 GO O O4 .4 OD .4 OA CD 4 D.-4 D4 4- 00 4 CO 00-4- CO4 OD4 -4-4ao4 ml-4 -4 PC-4 -4 14 1-4( - 4

87 E4tE4i E-4 i E-4i P E-4 E-4 1EI41- ' E1 E a 4,-- E-4 i . ,E 4 'E- '0 'E,

a ý 4lt

a- _-,EIOII - - ) E-40 - i ) iE-

UI jI cI I I I L) IL) I) s EI) IjI . I I I UI) I C.) I0 ) I

*I b4,I g I4 I S4 x I b I ý b I444 :i w w I~ w I b, x I I M, I I f Id If .4 E, I , I .4 .I IEý blI E114b44~

.11 1 (4 1I9 It I E 'II E4 g a I 0h

I1 I r l 11 '10

I~I

-4 -

EEa) E4 v L)

EI 4 u

lcll~~ i 10000 IIIUIII0LOL 0

Pp n c p E-44~C C U CItC 2 2

L)54- r-) I- I4,

0 L E41 L) t 4 e -4 .I e I N0 ' 0 0 001 00 .- -410~-4N0 e 01I~01U) -4 00 e 0 -4~- r~ e ~~) ) 0 ) ~I100 NC1010 e .4 ~eqeq i1 0 () ~0 ~0E4 eq c~1~ e~l jI) jI) I) I) ) U) U') U) ')E) 1 ~E 0 ~ 0 0 0 0 0 O oo o o o o o U).-,4 Un.- -IU)- UU -IU' .IU) 4 ).. 0 a- .I -I0 I -40 - 0 a- 0 .l .I0 -I0 - 0 a- q -4eq.I q - e ,I q~ieqaI q)- a-1 -I1 -Ij) a4 M - - aIa- -IaI I a a- -I a- - -4 N aI -IN a- a4 N - N aI -IN a- -Iai - .I.4 aI - ai.- -I.I 4.- - ~ ~ E)E-~aI

E88 I EI-

0-1010-41 01 E- IPI - - -iE4iE4iH iE4iH tU tU iH iH eP iE4iPApE40 WI~~~~~~ 10 0 0 0 0,0 0 0 0 0 0 0 ~ ~ 0 0 0 0 0

L)I E I tjI C)I r)I )I L)I L)I luI r)I LI I IL Ilu Ir) IU UI UI )I cI EjI UI tjI uI U I . I U I II I I I LII 1 I I II I I I9 I I I ED a E9

I 1 IdM 2 W Id 1g 1' I X I I I I W,

4 I4 d hd b4 X Iý I I I V I E4 3 .4 a4

I. I I F EIE

IA c WM IIfI I 0 I CA IA I I IA I0 IF I' W A

(iI

II 4

tI I I L) I I U

I I,

V~~ E-I4 '

P 4 I )

E44

U 0U~

4 t E4

1 E4' 4IEIE E4 Ii E' a 4

E44 1, E--Iý

P.0 0 0 0 0 0 0~ 0 0 riOo0~~'4

.5 89 U I a '4

041N

to

8= mo • r )41 .4

0 -4 8 -•-, i

el o)

E4 0 ..0

E-4 0- >

N N~

E-ii-4 to ( C'J

13 0 . Figure 4. Frequency distributions for nucleotide mismatch in all pairwise comparisons of allele prototype sequences. The number of nucleotides in each comparison varies according to the lengths of the different sequences. a) all codon positions, mean 24.1%, std. dev. 4.9%. b) third codon positions, mean 51.4%, std. dev.

9.5%. c) first two codon positions, mean 10.8%, std dev. 6.8%. o

0 S

0 & 0

92 0 o

0* I/1

CV

0d 0 o0 0 0 Z0 0Z 0n 0 trn

suospJIdimaoa uanbaS as!miied jo aqtunN QU)

0 -

0 0 0 I majority of sequence differences among alleles are located at evolutionarily unconstrained sites, most mismatches among alleles are likely to be due to evolutionary processes. Thus, even though some prototype sequences undoubtedly contain errors, the alleles named in this study serve the purpose of identifying genotypes present in the natural populations.

Phylogenetic affinities of the cloned sequences. Neighbor-joining analysis including twenty of the longest prototype sequences revealed that most alleles belong to the P. marinuslmarineA Synechococcus cluster, within which P. marinus and marine A Synechococcus lineages are not well resolved (Figure 5) (this thesis Chapter Two). The presumption that alleles branching with Synechococcus WH8103 in the tree actually derive from correctly sorted P. marinus cells, and not from contaminating Synechococcus, is reinforced by the observation that Type 6, an allele which branches close to Synechococcus WH8103, is highly similar to petB/D sequences from cultured P. marinus MIT9313 (99.7% sequence identity over 260 bp) and P. marinus MIT9303 (98.3% identity over 272 bp) (Figure 5b). The possibility remains, however, that some sequences recovered in this study do, in fact, come from mis-sorted Synechococcus. To reflect this possibility, alleles which fall into the P. marinus/marineA Synechococcus lineage should be considered as presumptive P. marinus alleles.

Three alleles, numbers 20, 24 and 40, do not cluster with P. marinus/marineA Synechococcus, but instead form a phylogenetic group branching below the divergence of chlorophyte and land plant chloroplasts (Figure 5). The position of this cluster is reminiscent of the phylogenetic branch point for rhodophyte, pheophyte and some green algal chloroplasts in 5S rRNA, 16S rRNA and psbA phylogenies (Van den Eynde et al. 1988, Maid et al. 1989, Douglas and Turner 1991, Markowicz and Loiseaux-de Goer 1991, Oyaizu et al. 1993). Thus, these alleles may have originated in plastids of Figure 5. Phylogenetic relationships among 20 petB/D alleles cloned from flow cytometrically sorted populations in the Sargasso Sea and the Gulf Stream, cultured strains of P. marinus and other members of the oxygenic phototroph radiation. (a) tree inferred by neighbor joining using 163 nucleotides at the first two codon positions of petB and petD. Alleles 20, 24 and 40 form a cluster allied with green plant and algal chloroplasts. The remaining clones cluster with P. marinus and marine A Synechococcus WH8103. (b) tree inferred from 77 nucleotides at first two codon positions in petB and D to include short sequences from P. marinus cultured strains MIT9303 and MIT9313. Allele 6 is 99.7% identical to P. marinus MIT9313 and 98.3% identical to P. marinus MIT9303 for all sequence positions. -C

Cl

Oso 0 .tU

ta

LmII =J1

97 c d

OCO i-7

6

U

0 6i U

COO

0 -

W4

Iv cryptophytes or coccolithophorids, both of which were present at the Gulf Stream and Sargasso Sea stations (R. Olsen, pers. comm.), or perhaps in plastids of the newly discovered marine chlorophyte Ostreococcus tauri (Courties et al. 1994), which may have been liberated during sample concentration. Liberated choroplasts lacking phycoerythrin would match the flow cytometric definition of P. marinus used for cell sorting and could therefore have been included in the sorted samples without machine error. The three chloroplast-like alleles were omitted from the population genetic analysis, leaving 68 P. marinus alleles in the dataset..

Phylogenetic affiliations of alleles not included in the neighbor-joining tree (because of short sequences) were assessed by comparison of mismatch frequencies between these sequences and selected sequences used to infer the tree (Table 3). All alleles not included in the tree were more similar to at least one member of the P. marinus/marine A Synechococcus cluster than to sequences outside the cluster and so were included in the population genetic analysis.

Allele counts for P. marinus in sorted samples. Clones recovered from each of the eight sorted samples contained between 6 and 21 P. marinuspetB/D alleles (Table 1). Correcting for unequal sample size by the method of Hurlbert (1971), the estimated number of alleles present in a sample size of 19 clones for each sorted sample varied from 5.16 to 16.83 (Table 1).

To assess the completeness of sampling for genetic diversity, rarefaction curves were generated for each sorted sample and for lumped samples containing both members of flow cytometric double populations, constituents of entire water columns or the entire dataset (Figure 6). For a sample which contains a good representation of the genetic diversity in its parent population, a plot of E(Sn) (the estimated number of genetic w mcc m Ln oN .- m .- m m m m -v r- m c w H cc w m c ) z L m " -, :N :

D Co Co l k (N C) ko 0 0 0 ) 0m 0 H N ON cc N L 0' C>O% C(N T T %D l - Oc C• n) Ho L c 0k H HC m N O N C) N- 0 H c H L :) H H4 N -4 H H r H H N H 0N(NO CC H H P) (N H H (N 04 (N H H H H H (N H (N

CY) N- N- k cc m ON H4 (7% LA N -V' m m N- H4 N %D cc (D ) CN k-c H C) cc CD 'T~ c mcq (N M (0 0c 0 rN N 00 m N 000 q m0 mc 0cc0 0 N CD H N C) HHH H H H H H H H H H C) H(N (NH H H H H H H H H H H H H H (N

H r H N N A 4 N0 m cc c c N Hm c> 0 0L m 0N Nr m N0 c, 40 o N c 0c 0L0 LAL0 N u 0 0N 0 N N N 0LU H H H H HrC)H H H H H HO C)H H (N H H-H H H H H H H H HO C H H H H

O o 0 0 rn0M0••00 "N H- n (N L) N- H cc " LAm m LA m (N••••0 0 LA m 0 cc H cc ) LA ccu mY) cc Nr-b0 o r cc ma cm LA co LA CN C) m cc C) H4 C) cc cc H- (N H- w cc cc cc LA H q cc H (N H cc C) LA 00 N- C) T m~ cc H H HO C)H H HOC)OC)4 H H HCo C) C) CH H H HC)H H HOC H4 C) C)Cc ) H H H

" 0 0 cc LA m cc0 LA H N N" c N C A Hk c H c 0 H N 0m m "N (N C)> H H - m OccCNr- - " opW H c LA rH A N H N LA v c> c) C) (O) N- c H HH H H H H C H H H C)H H H H H H H H HH H HOC HH H H H C

0c N c0 c0 m L 0N 0 0 c 0 0N N 0cc 0 0 0 0 0c m H Nr-fc4 LA cc Pm m LA - " m cc w N cc H N- cc cc m c) cc (N (D q m c) cc q a C) 4 1 C co c) H1 c) c) c) H c) C) c) C) c) (N C) H C) C) H C)H H C)OO C) C) C) H )H H

kD cc m Nr- LAm c No C) C ) m m) c) -- cc cc (,4 cc ýT oc N- N- cc " c ý- c -v cc C) LA IT

W 0N C) r- O c0 c0 c0 cc 0 c0( 0 0c c -( C) m 0 0 C) C) Z) C CH r-ACC)1 -4 C) C) C) C) H H C)

cc p cc qv LA LA cc k cc 4 LA H- H- N- (NN c c Ho m-n rN r- c c c o LA cc ccm ( cq ch H- C) LA cc 0 rN 0c m cc ) H1 %D 00 N- Nv cc H Hco H LA c) r-) c c c r-c)Nm Nr C) n C) C) C) H C) H H C) C) C) H ) H H H (N CD C) H H H C) C) 4H C H C) H NC)H H

LA m~ (y m cc LA LA cco cc LA cc cc N- (N cc cc ý cc (N N q(3) ccc LA cc cc cc N H m- %cc) LA N- 0c (N H % H- c0) r-N U)LA 0) O c cc ccO cc c C) y C) H- cc (N N cc0 N- CD 00 cO C) C) C) H4 C) H C> C) C) C) H C) C) H C) H H C) C) H H H H C) H c) C) c) H C) C)

N LA ) I Ch cc ) H cc LA (N (N " cc Ln cc LA ccm c ,) c> ccw cc ) oc -. C) " nm (N "N cc r) LA LA cc oc ) o N cc (N cc N- (N LA cc cc cc H1 (N cc cc r cc cc LA 0) cc) H- C) C) C) C) H C) C) H CD C) H H Co C> H C)HC>) C) H HC)HC) C) H C)O C)H C)O O H H

(N N Hi -N H m L (" ( N " w T m c C) m c c v c C) c N T mT L0 m m LA "N LA M C> M cc cc " Cc " (N LA LA cc cc cc cc H C) N- LA n cc o Ln m) H- c) cc C) C) C) H C) C) H C) C) C) H C) C) H C) H4 H C) C) C) H H C) C H C) C) C) C) H H C)

N N LA N " -v cm T C)ON> Hi cc cc cc oocc c q cc k N N- c N- r L Lko cO LAmc N H mc H H- Hq m mc H HOHH c LAOHOH cc cc cc cc0HO ýZ IV cc HOHHOHHm m LA HOOCD C)O rr HOHOHo N- cc cc OHT HAH LA -N c H ) LAH cN LA N cc cc C) H -A P) LA 1 N c -1 HL -4 c

H v D ) H H H (N (N (N (N o) c>m P)P m P) ) rm P) mm LA

100 (n In In co r- r- r- -v m r(N D C)ý -A C " (D a)m) ma c) r- o to ) In n cN H In ,o - o r- a) a) In rO (N HW (NO H0 (NH 0 H m• H mN NH H 0 H H H H• HQ • H H Hm• (N

m zr '0 c N- N- c) ý m~ In In Hi c) ýz -z (N w0 a c> 00 0 M 0 0 0 0 i0WN 0 0 n H NH H H 1-4H H M N H H H H H H H H H 4N

In In ui C) c> N- In In c(4 ,T 'ý) c) N- -ý N- H- a) C) H ay) c> %D) - o) In m m' o) N- r- ,o n In -- 4 (N m' Hi

m m o 0 0 0 < 0 0 0 • "• HIn (NH H-0 H H%D (N H mHH (N r-)NH HH4- % H r H In H m H - HIn c,4 H H HN

C M ) C)0 C)k0 0 ýD (N 'C m C) H N ( ( H- C) a) rI)H (N a)N I-al Nl C) ON -:I (N (N CD C) N- Cn H H C) H H H C) H C) C) H C) H H H H H C) H NU)

SC N H N a ) H N H ( N 0 M -T H i (N C) N In C) C) N M' M H (N C) OD N H H H H H H H H- H H H H H H H C) H tlj "N-- m CC) NO m C) m In ko ' ' m ' In -Hv (N Hr- ) In ')m a) N H (N4 'IT N C) N ' a) In H H C') a) CD H C) C) H H 0 ( C) H C ) H H H H C C) C N 0

a)

N O HIt) C) C) C) a) N (N c) T (N ý H ( C)OON T m C) a)m' ) a) 00 wt O. a) (N C) C) aNL)In O)a) IC) N wt (N (NH H H H H H H (N (N (N H H H H H H N q )

C v -zC'm) c4 C) rn cc) -v c,4 H a) a) In H N I) a)e Sa0) C) ) C a C) H - O N" 0 0 0 C) H 0 N H- C) H- C) H- C) H- H- C) C) C) C) C) C) H- H- C) C) Hl 04 O4

U4J0

H H C) H ( C) C) (N H H H H C C)O H H C) C) H

a3) c) m' C) C) a) HA I) C) I) c) N- --; I) In H- a)% (N In Hn C)M a)C) 0( H CH C)C) H C>(N H4C)N (H 4 0C) 0C) 0C) 0C) 0C) 0nC) C

4-4r-.4

kC oC) rn C) N (N ) %C) m' a) (N kC m - - c) H r N In q) a) a) (N %C) a) N N C) C) 0(a) N H -C' r-HN -: C) C) C) C) C C) C) C) C C) C) C) H 0 •) o H,4 H H H H H- O~0 H

qz' N- (N C) C) 0) N- w m' a) (N It) N C) -T C) (N N In m N I . a) m C)oT CD " m C) C) m m I n N H- C) C) C) H C H N H C) -H H C) C) C) C) C) C) H

9 0 (N maN C N N- mO "e ' ( m -)z N I kC)m r-N n It CO N- a) C) -I HA N- m' N a) C) a)N HA N H a) O) CF) N- H (N A H H (N H H- H

c•.a() o (N - IC) N- CO a) C) .- (NC4 M N- a)0 a) H (N n kt) N- 0) A oe

Q4 ý

101 Figure 6. Rarefaction curves for clones recovered from a) individual sorted samples, b) lumped samples containing both members of flow cytometric double populations or constituents of entire water columns and c) a lumped sample containing the entire dataset. The upward trend of all curves indicates that the number of alleles in all populations are incompletely sampled. E(Sn) (the estimated number of alleles for n, the number of clones in a random subsample (Hurlbert 1971).

102 IE

e4 kECE

Co

vE

C Eh C= _

(us)3 (us)•

CI

Cn

co

C 6 omI W4," An 2k. ECW z C ir C %n oC (us)3 (us)g (us)3

t,

E= N 0 0 0 E LE U, o % U- 0 0.00 0 0 U,. 0

r- V In V M f4 - C V- N u qaCs), I 1-V In V IMr4 C (US)3 (uS)a

103 '0

0~ cE z

(us)O (US)a

0 Ts

0 0u e

0 0 0 o In 0 (us)a

104 I I I I

S

C/)

f 4 - 0

0 if

rt)

- 0 0 0 0 D t •,•• • (US)t',

105 variants for subsample size n) versus n approaches horizontal as n approaches the sample size (Hurlbert 1971) 3. Curves for each of the eight sorted samples, and for lumped samples show upward trends of E(Sn) with n, and no indication of curve flattening. It is therefore apparent that many more genetic variants were present in the Sargasso Sea and Gulf Stream populations than were recovered in this study.

Distribution of alleles among sorted samples. Of the 68 P. marinus alleles in the dataset, only 16 were found more than once, with 12 appearing in more than one sample. Since rarefaction curves indicate that populations in the different sorted samples were not exhaustively sampled, the absence of shared alleles is not considered evidence of dissimilarity between samples. However, the shared presence of alleles is positive evidence of similarity. Scoring alleles according to their presence in the different samples reveals no consistent pattern of shared genotypes, but instead suggests that all of the sorted populations drew membership from a common gene pool (Table 4). Of particular interest is the finding that eleven alleles were shared between the Gulf Stream and Sargasso Sea depth profiles and that two alleles were shared by members of the double population from 85 m in the Gulf Stream, and one by the pair recovered from 135 m.

3 This assumes a large parent population with a relatively small number of genotypes. It is possible to imagine a parent population in which every individual has a different genotype, in which case rarefaction curves would never flatten out, even for samples approaching the size of the parent population. However, even for this imagined situation an upward trend in the rarefaction curve for a finite sample would indicate that unsampled variants were present in the parent population.

106 0 8 E ** * * *

o

* *

0o

o 'oso

w * * * *

00 *

* * * ** * -4o0 EI

czo C

* * *

4-0 - co p04

4)

107 DISCUSSION

Field populations of P. marinus are genetically heterogeneous. We find that field populations of P. marinus are genetically heterogeneous, according to the distribution of "presumptive" P. marinus petB/D alleles cloned out of flow cytometrically sorted field samples. The P. marinus alleles must be considered only presumptive because they fall into a lineage containing both P. marinus and marine A Synechococcus cultured isolates, which fail to form distinct phylogenetic clusters (Figure 5, this thesis Chapter Two). Thus the P. marinus identity of these alleles, presumed because of the flow cytometric purification, cannot be confirmed by phylogenetic analysis, and it is possible that some may have originated in marine A Synechococcus mistakenly included in the P. marinus sorted samples.

Between 6 and 21 alleles were present in the set of clones recovered from each of eight flow cytometrically sorted samples, with the presence of large numbers of additional alleles implied by rarefaction analysis (Figure 6). The set of alleles recovered within a water column is phylogenetically diverse within the P. marinus/marineA

Synechococcus cluster, with both Sargasso Sea and Gulf Stream water columns containing alleles which form phylogenetic clusters with distantly related cultures isolated from the Pacific and the Sargasso Sea (compare Table 1 and Figure 5). The structure of P. marinus populations implied by these results is consistent with the detailed findings of Selander and colleagues for the population genetics of Escherichiacoli, in which most of the genetic diversity of the species is found within local populations, and in which cell lineages are globally distributed (Selander et al. 1987).

Oceanographic implications of genetic diversity within P. marinus populations are twofold. First, the physiologic activities of P. marinus in the water column are not

108 simply described by the properties of one or a small number of cultures isolated from the same geographic area, but by the combined activities of a phylogenetically diverse set of cells, some of which are similar to isolates from distant parts of the world. Like SS 120 and Med4, these genetic variants may exhibit physiological differences which would give each a selective advantage under different environmental conditions. The second oceanographic implication is that adaptive changes in P. marinus populations over the annual cycle and with depth are as likely to be due to shifts in their genetic composition, due to differences in growth rates of indigenous genetic variants under varying environmental conditions, as to intracellular adaptive responses (Wood 1988, Wood and Leatham 1992).

P. marinus populations recovered from the Gulf Stream and the Sargasso Sea water columns share eleven out of the 12 alleles which were found in more than one sorted sample (Table 4), despite the large scale hydrographic differences which distinguish these two water bodies. The large number of these shared alleles suggests, again, a similarity to the population structures of E. coli, in which populations in distant locations share alleles.

Flow cytometrically defined subpopulations from two "double populations" recovered from the Gulf Stream exhibited overlapping sets of petB/D alleles in our analysis, with the pair of subpopulations at 85 m sharing two alleles and the pair at 135 m sharing one. This result implies that these flow cytometric subpopulations are not genetically different, but may owe their distinct properties to processes such as synchronized cell division (Vaulot et al 1994) or mixing of cells acclimated to conditions in different water masses. Since the subpopulations of the two double populations were incompletely resolved, however, it is also possible that the shared alleles could result from the presence in either sort window of cells in the "tail" of the other subpopulation's

109 distribution. Also, it should be noted that multiple populations observed at DCM's in the Atlantic are transient phenomena (unpublished observations of R. Olsen), whereas multiple populations at Pacific station ALOHA are a stable feature of a permanent DCM (Cambell and Vaulot 1993). Multiple populations at ALOHA may therefore have a different origin from those examined in our study, and may indeed contain genetically different subpopulations.

Genotype frequencies among PCR clones derived from natural populations using degenerate, inosine-containing primers reflect both gene frequencies in the sample population and PCR amplification bias for or against specific alleles. It is therefore speculative to draw inferences from gene frequencies in datasets such as the one at hand. With this caveat in mind, it nonetheless bears mentioning that Allele 1 and Allele 1-like sequences predominate in clones recovered from the shallow mixed layer at both sampling sites (Figure 7), while at DCM's Allele 1-like sequences are present but do not predominate. Since HPLC and flow cytometric studies find that surface waters under stratified conditions contain low chl b2/a2 ratios (Goericke and Repeta 1993, Cambell and Vaulot 1993) similar to those found in P. marinus Med4 (Goericke and Repeta 1993), and since Allele 1 belongs to the phylogenetic cluster containing MIT9107 and Med4, these results may be correlated. The results of this study are therefore consistent with the hypothesis that populations at different depths in stratified watercolumns draw their membership from a single gene pool, but that gene frequencies vary at different depths, resulting in populations exhibiting different chl b/a2 pigment ratios.

110 Figure 7. Frequency of Allele 1-like alleles in sorted samples. The frequency of alleles having intergenic regions similar to Allele 1 (Alleles 1, 2, 3, 69, 71, 72, 73, 74, 75 and 76) is scored for samples at different depths in the two water columns, lumping Bright and Dim flow cytometric subpopulations at two depths in the Gulf Stream profile. Although this inference must be considered speculative, the high frequency of recovery of Allele 1-like sequences in surface water samples is consistent with a predominance of these alleles in P. marinus populations in the surface mixed layer.

111 4 U) 4 0) (U r) 0 r E 0 c

0 0

C)

CO LO 0 C c- 0

0 0 0 0 T- T o- 0) °i OO 0 60 0) -C C_) 4-0)

O 0i-

C- CT U,- 0- 0)

L_ Li. 0 O o U

LO o If) '- 0 0 o o a 0LiO 0 LI.10 -C a- 0 (wU) L44da] 0 -C 0

112 REFERENCES Brand, S.N., Tan, X. and Widger, W.R. (1992). Cloning and sequencing of the petBD operon from the cyanobacterium Synechococcus sp. PCC7002. Plant Mol. Biol. 20:481-491. Britschgi, T.B. and Giovannoni, S.J. (1991) Phylogenetic analysis of a natural marine bacterioplankton population by rRNA gene cloning and sequencing. Appl. Environ. Microbiol. 57:1707-1713. Campbell, L., and Vaulot, D. (1993). Photosynthetic picoplankton community structure in the subtropical North Pacific Ocean near Hawii (station ALOHA). Deep-Sea Res. 40:2043-2060. Carter, K.R., Tsai, A. and Palmer, G. (1981). FEBS Lett. 132:243-246. Chisholm, S.W., R.J. Olson, E.R. Zettler, R. Goericke, J. Waterbury, and N. Welschmeyer. (1988). A novel free-living prochlorophyte abundant in the oceanic euphotic zone. Nature, 334(6180):340-343. Chisholm, S.W., Frankel, S.L., Goericke, R., Olson, R.J., Palenik, B., Waterbury, J.B., West-Johnsrud, L. and Zettler, E.R. (1992). Prochlorococcusmarinus nov. gen nov. sp.: a marine prokaryote containing divinyl chlorophyll a and b. Arch. Microbiol. 157:297-300. DeLong, E.F., Franks, D.G. and Alldredge, A.L. (1993) Phylogenetic diversity of aggregate-attatched vs. free-living marine bacterial assemblages. Limnol. Oceanogr. 38:924-934. Geiskes, W.W.C., G.W. Kraay, A. Nontji, D. Setiapermana, and Sutomo. 1988. Monsoonal alternation of a mixed and a layered structure in the phytoplankton of the euphotic zone of the Banda Sea (Indonesia): A mathematical analysis of algal pigment fingerprints. Neth. J. Sea Res. 22:(2):123-137. Giovannoni, S.J., Britschgi, T.B., Moyer, C.L., and K.G. Field. (1990). Genetic diversity in Sargasso Sea bacterioplankton. Nature 345:60-62. Goericke, R. and Repeta, D.J. (1993). Chromatographic analysis of divinyl-chlorophylls a and b in samples from the subtropical north Atlantic Ocean. Mar. Ecol. Prog. Ser. Goericke, R. and Welschmeyer, N.A. (1993). The marine prochlorophyte Prochlorococcuscontributes significantly to phytoplankton biomass and primary production in the Sargasso Sea. Deep-Sea Res. 40:2283-2294. Greer, K.L. and Golden, S.S. (1992). Conserved relationship between psbH and petBD genes: presence of a shared upstream element in Prochlorothrixhollandica. Plant Mol. Biol. 19:355-365. Hope, A.B. (1993). The chloroplast cytochrome bfcomplex: a critical focus on function. Biochim. Biophys. Acta 1143:1-22. Hurlbert, S.M (1971). The non-concept of species diversity: a critique and alternative parameters. Ecology 52:577-586.

113 Kallas, T., Spiller, S. and Malkin, R.C. (1988). Characterization of two operons encoding the cytochrome b6-f complex of the cyanobacterium Nostoc PCC7906 and highly conserved sequences but different gene organization than in chloroplasts. J. Biol. Chem. 263:14334-14342. Kimura, M. (1980). A simple method for estimating evolutionary rate of base substitutions through comparative studies of nucleotide sequences. J. Mol. Evol. 16:111-120. Knoth, K., Roberds, S., Poteet, C. and Tamkun, M. (1988). Highly degenerate, inosine- containing primers specifically amplify rare cDNA using the polymerase chain reaction. Nuleeic Acids Res. 16:10371. Lewin, R.A. (1981). Prochlorophytes. In The Prokaryotes. Vol 1. (eds. M.P. Starr, H. Stolp, H.G. Truper, A. Balows, and H. G. Schlegel). 256-266. Springer, Berlin. Li, H., Cui, X. and Arnheim, A. (1991). Analysis of DNA sequence variation in single cells. Methods 2:49-59. Li, W.K.W., Dickie, P.M., Irwin, B.D., Wood, A.M. (1992). Biomass of bacteria, cyanobacteria, prochlorophytes and photosynthetic eukaryotes in the Sargasso Sea. Deep-Sea Res. 39:501-519. Li, W.K.W., and M. Wood. 1988. Vertical distribution of North Atlantic ultraphytoplankton: analysis by flow cytometry and epifluorescence microscopy. Deep Sea Res. 35:1615-1638. Moore, L.R., Goericke, R., and Chisholm, S.W. (1994). The comparative physiology of Synechococcus and Prochlorococcus: influence of light and temperature on growth, pigments, flourescence and absorptive properties. Mar. Ecol. Prog. Ser., min press. Morel, A., Ahn, Y.-H., Partensky, F., Vaulot, D. and Claustre, H. (1993). Prochlorococcusand Synechococcus: a comparative study of their optical properties in relation to their size and pigmentation. J. Mar. Res. 51:617-649. Ohyama, K., Fukuzawa, H., Kohchi, T., Shirai, H., Sano, T., Sano, S., Umesono, K., Shiki, Y., Takeuchi, M., Chang, Z., Aota, S., Inokuchi, H. and Ozeki, H. (1986). Chloroplast gene organization deduced from complete sequence of liverwort Marchantiapolymorpha chloroplast DNA. Nature 322:572-574. Olson, R.J., Chisholm, S.W., Zettler, E.R., Altabet, M.A., and Dusenberry, J.A. (1990) Spatial and temporal distributions of prochlorophyte picoplankton in the North Atlantic Ocean. Deep Sea Res. 37, 1033-1051. Partensky, R., Hoepffner, N., Li, W.K.W., Ulloa, O. and Vaulot, D. (1993). Photoacclimation of Prochlorococcussp. (Prochlorophyta) strains isolated from the north Atlantic and the Mediterranean Sea. Plant Physiol. 101:285-296. Reimann, A. and Kueck, U. (1989). Nucleotide sequence of the plastid genes for apocytochrome b6 (petB) and subunit IV of the cytochrome b6f complex (petD) from the green alga Chlorellaprotothecoides: lack of introns. Plant Mol. Biol. 13:255-256.

114 Rock, C.D., Barkan, A. and Taylor, W.C. (1987). The maize plastid psbB-psbF-petB- petD RNAs encode alternative products. Curr. Genet 12:69-77. Sambrook, J., Fritsch, E.F. and Maniatis, T. (1989). Molecular Cloning: a laboratory manual. Cold Spring Harbor Laboratory Press, New York. Selander, R.K., Caugant, D.A. and Whittam, T.S. (1987). Genetic structure and variation in natural populations of Escherichiacoli. In Escherichia coli and Salmonella typhimurium: cellular and molecular biology (F.C. Neidhart, ed.). Am. Soc. Microbiol. pp. 1625-1648. Swift, H. and Palenik, B. (1992). Prochlorophyte evolution and the origin of chloroplasts: morphological and molecular evidence. In Origins of Plastids (R.A. Lewin, ed.) Chapman and Hall, pp. 123-139. Vaulot, D., Marie, D., Olson, R.J. and Chisholm, S.W. (1994). Prochlorococcusis highly synchronized to the diel cycle in the equatorial Pacific and divides up to once a day. submitted. Vaulot, D., and Partensky, F. (1992). Cell cycle distributions of prochlorophytes in the northwestern Mediterranean Sea. Deep-Sea Res. 39:727-742. Vaulot, D., Partensky, F., Neveux, J., Mantoura, R.F.C. and C. Llewellyn. 1990. Wintertime presence of prochlorophytes in surface waters of the North-Western Mediterranean Sea. Limnol. Oceanogr. 35:1156-1164. Vermaas, W.F.J., and Ikeuchi, M. (1991). Photosystem II. In The Photosynthetic Apparatus: molecular biology and operation (L. Bogorad and I.K. Vasil, eds.). Academic Press pp. 26-111. Widger, W.R. and Cramer, W.A. (1991). The b6/f complex. in The Photosynthetic Apparatus: molecular biology and operation (L. Bogorad and I.K. Vasil, eds.). Academic Press pp. 149-224. Wood, A.M. (1988). Molecular biology, single cell analysis and quantitative genetics: new evolutionary genetic approaches in phytoplankton ecology. In Immunochemical Approaches to Coastal,Extuarine and Oceanographic Questions (C.M. Yentsch, F.C. Mague and P.K. Horan, eds.) Springer-Verlag pp 41-73. Wood, A.M. and Leatham, T. (1992). The species concept in phytoplankton ecology. J. Phycol. 28:723-729.

115 Chapter 4 EPILOGUE: RESPONSES IN THE LITERATURE TO THE PUBLICATION OF CHAPTER ONE: MULTIPLE ORIGINS OF PROCHLOROPHYTES WITHIN THE CYANOBACTERIAL RADIATION

116 Since its publication, "Multiple Origins of Prochlorophytes within the Cyanobacterial Radiation" (Urbach et al. 1992) has generated numerous citations and has contributed to the formulation of new theories about the evolution of pigments in cyanobacteria and chloroplasts (Bryant 1992, Bullerjahn and Post 1992). This Epilogue is a brief review of published responses to this paper. The conclusions of Chapter One relevant to the discussion are (1) according to 16S ribosomal RNA (rRNA) phylogenetic analysis the three known prochlorophytes and chloroplasts each fall into separate lineages dispersed among the cyanobacteria, and (2) this polyphyletic distribution suggests chl b photosynthesis was "invented" several times during the evolution of cyanobacteria and chloroplasts (Urbach et al. 1992).

Most authors citing this work have no argument with the first of these conclusions (e.g. Bryant 1992, Bullerjahn and Post 1992, Cavalier-Smith 1992, Palenik and Haselkorn 1992, Swift and Palenik 1992, Cretiennot-Dinet et al. 1993) which echoes an inference drawn by Turner et al. (1989) and by Morden and Golden (1989a, b) from analyses of the evolution of Prochlorothrixhollandica, and which agrees with the analyses of Palenik and Haselkorn (1992) and Lockhart et al. (1992c). However, several workers have pointed out that the independent invention of chlorophyll b in each of several lineages is not the only evolutionary scenario to fit the data, and that the distribution of chl b among cyanobacterial and chloroplast lineages could also result from lateral gene transfer (Palenik and Haselkorn 1992) or by descent of the entire cyanobacterial lineage from an ancestor containing both chlorophyll b and phycobilisomes, with subsequent loss of phycobilisomes in some lineages (prochlorophytes and green chloroplasts) and chl b in others (most cyanobacteria) (Bryant 1992, Cavalier-Smith 1992). I wholly concur with these observations, and with the suggestion that information on prochlorophyte chlorophyll biosynthesis and chlorophyll a/b binding proteins will help to clarify this issue. To this end, Bullerjahn

117 and Post (1992) cite findings that the major chl a/b binding proteins of Prochloronsp. and P. hollandica are of similar molecular weight and are immunologically cross reactive. They combine these observations with sequence data indicating that P. hollandicaantenna proteins are different from chloroplast light harvesting complexes, and form the hypothesis that photosynthetic antennae in P. hollandicaand Prochloronsp. share a common origin separate from that of green chloroplasts. Future work in this direction, including analysis of the P. marinus photosynthetic system, will surely reveal a fascinating evolutionary story.

One research group, including P. Lockhart, A.W.D Larkum, D. Penny and coworkers does, however, take issue with the phylogenetic branching pattern inferred from 16S rRNA sequences in Chapter One (Lockhart and Penny 1992, Lockhart et al. 1992b, Larkum 1992, Lockhart et al. 1993). Their skepticism comes as an extension of their work showing that branching patterns inferred from protein-encoding genes are confounded by nucleotide substitution biases (detected as differences in G+C composition at third codon positions, see Chapter Two). Lockhart et al. (1992b) argue that phylogenetic trees inferred from cyanelle, prochlorophyte, cyanobacterial and chloroplast protein-encoding sequences are invalid because patterns of similarity in third codon position G+C content are congruent with the inferred phylogenetic branching patterns. They argue, in addition, that 16S rRNA trees, even those inferred from sequences without detectable G+C bias, are also invalid, based on the following premise: Although the effects of substitutional bias upon inference are best demonstrated by analysis of protein-coding nucleotide sequences, bias (a genomic effect in eubacterial lineages) must also affect variable sites within rRNA genes and may also have an effect on amino acid substitution (Jukes and Bhushan 1986; but see Prager and Wilson 1988). Inference from such datasets (particularly where long edges link taxa: e.g., Giovannoni et al. 1988; Morden and Golden 1989; Evrard et al. 1990; Kishino et al. 1990) may therefore also be misled. (Lockhart et al. 1992a)

118 Larkum (1992) proposes that, since phylogenetic trees linking the Cyanophoraparadoxa cyanelle to chloroplasts and showing prochlorophytes dispersed among the cyanobacteria are invalid, it is reasonable to propose that prochlorophytes are monophyletic and and share an ancestry with green chloroplasts.

Lockhart and coworkers make a valid point in saying that phylogenetic trees, inferred by methods which assume constant substitution biases across taxa, may be in error when calculated from data having varying third codon position G+C content. Their extension to 16S rRNA sequences quoted above is a subtle argument, but also may hold true. Resolution of these phylogenetic problems must await the development of methods demonstrated to be insensitive to nucleotide substitution biases. However, the failure of P. marinus strains to group with chloroplasts having similar third codon position G+C content in trees inferred with psbB and petB/D sequences (Urbach and Chisholm in preparation, this thesis Chapter Two) argues against an evolutionary link between P. marinus and chloroplasts.

I'll close with a quote from Bullerjahn and Post (1992) which is appropos: If one reviews much of the literature on the prochlorophytes, it becomes clear that taxonomic classification and the inferring of relationships between organisms can be a risky enterprise. Some microbiologists have been eager to show a direct relationship between prochlorophytes and the chloroplast, whereas others were determined to place them among the cyanobacteria. The underlying problem in such exercises lies both in the danger of overemphasizing one aspect while belittling others and in the import of externally imposed bias into the judgement of a certain property.

119 REFERENCES Bryant, D.A. (1992). Puzzles of chloroplast ancestry. Curr. Biol. 5:240-242. Bullerjahn, G.S. and Post, A.F. (1992). The prochlorophytes: are they more than just chlorophyll a/b-containing cyanobacteria? Crit. Rev. Microbiol. Cavalier-Smith, T. (1992). The number of symbiotic origins of organelles. Biosystems 28:91-106. Cretiennot-Dinet, M.-J., Sournia, A., Ricard, M. and Billard, C. (1993). A classification of the marine phytoplankton of the world from class to genus. Phycologia 32:159-179. Everard, J.-L., Kuntz, M. and Weil, J.-H. (1990). The nucleotide sequence of five ribosomal protein genes from the cyanelles of Cyanophoraparadoxa: implications concerning the phylogenetic relationship between cyanelles and chloroplasts. J. Mol. Evol. 30:16-25. Giovannoni, S.J., Turner, S., Olsen, G.J., Barnes, S., Lane, D.J. and Pace, N.R. (1988). Evolutionary relationships among cyanobacteria and green chloroplasts. J. Bacteriol. 170:3584-3592. Jukes, T.H. and Bhushan, V. (1986). Silent nucleotide substitutions and G+C content of some mitochondrial and bacterial genes. J. Mol. Evol. 24:39-44. Kishino, H., Takashi, T. and Hasegawa, M. (1990). Maximum likelihood inference of protein phylogeny and the origin of chloroplasts. J. Mol. Evol. 31:151-160. Larkum, A.W.D. (1992). Evolution of chlorophylls, light harvesting systems and photoreaction centres. In Research in Photosynthesis, Vol. III (N. Murata, ed.). Kluwer Academic, pp. 475-482. Lockhart, P.J., Penny, D. (1992). The problem of GC content, evolutionary trees and the origins of chl-a/b photosynthetic organelles: are the prochlorophytes a eubacterial model for higher plant photosynthesis? in Research in Photosynthesis, VoL III (N. Murata, ed.). Kluwer Academic, pp. 499-505. Lockhart, P.J., Howe, C.J., Bryant, D.A., Beanland, T.J. and Larkum, A.W.D. (1992a). Substitutional bias confounds inference of cyanelle origins from sequence data. J. Mol. Evol 34:153-162. Lockhart, P.J., Penny, D., Hendy, M.D., Howe, C.J., Beanland, T.J. and Larkum, A.W.D. (1992b). Controversy on chloroplast origins. FEBS Lett. 301:127-131. Lockhart, P.J., Beanland, Howe, C.J. and Larkum, A.W.D. (1992c). Sequence of Prochloron didemni atpBE and the inference of chloroplast origins. Proc. Natl. Acad. Sci. USA 89:2742-2746. Lockhart, P.J., Penny, D., Hendy, M.D., and Larkum, A.D. (1993). Is Prochlorothrix hollandica the best choice as a prokaryotic model for higher plant chl a/b photosynthesis? Photosynthesis Res. 37:61-68.

120 Morden, C.W. and S.S. Golden. 1989a. psbA genes indicate common ancestry of prochlorophytes and chloroplasts. Nature 337:382-385. Morden, C.W. and S.S. Golden. 1989b. Corrigendum: psbA genes indicate common ancestry of prochlorophytes and chloroplasts. Nature 339:400. Palenik, B. and Haselkorn, R. (1992). Multiple evolutionary origins of prochlorophytes, the chlorophyll b-containing prokaryotes. Nature 355:265-267. Prager, E.M. and Wilson, A.C. (1988). Ancient origin of lactalbumin from lysozyme: analysis of DNA and amino acid sequences. J. Mol. Evol. 27:326-335. Swift, H., and Palenik, B. (1992) Prochlorophyte evolution and the origin of chloroplasts: morphological and molecular evidence. In Origins ofPlastids (R.A. Lewin, ed.) Chapman and Hall pp. 123-138. Turner, S., Burger.Wiersma, T., Giovannoni, S.J., Mur, L.R., and N.R. Pace. (1989). The relationship of a prochlorophyte Prochlorothrixhollandica to green chloroplasts. Nature 337:380-382. Urbach, E., Robertson, D. and Chisholm, S.W. (1992). Multiple evolutionary origins of prochlorophytes within the cyanobacterial radiation. Nature 355:267-269.

121 Appendix I

PRELIMINARY ANALYSIS OF GENETIC DIVERSITY IN GULF STREAM POPULATIONS OF PROCHLOROCOCCUSMARINUS BY DIRECT SEQUENCING OF PCR PRODICTS AMPLIFIED FROM FLOW-CYTOMETRICALLY SORTED CELLS

122 Chapter Three details an investigation into the genetic diversity of natural populations of Prochlorococcusmarinus from the the Gulf Stream and the Sargasso Sea in which P. marinus were sorted from field samples by flow cytometry and the diversity of petB/D sequences in the sorted samples assessed by cloning and sequencing individual molecules from PCR amplification products. This Appendix describes a preliminary investigation which preceded the cloning experiment in which PCR products sorted from Gulf Stream field samples were directly sequenced. Autoradiograms of the resulting sequencing gels revealed comigrating bands at numerous nucleotide positions, especially in the intergenic region between petB and petD, indicating that multiple petB/D sequences were likely to be present in the amplification products.

METHODS

Sample collection, flow cytometric sorting and DNA isolation. The DNA preparations used for this preliminary experiment were the same as those used for the analysis of Gulf Stream populations in Chapter Three, and originated in samples collected from 50 m, 85 m and 135 m at 370 30.68'N, 680 13.69'W during July of 1993. Samples collected from 85 m and 135 m each contained a pair of flow cytometrically defined "double populations," subpopulations of which were sorted into separate samples, resulting in a total of five sorted samples (Figure 2 in Chapter Three).

PCR amplification. Sequences at the petB/D locus were amplified as described for the analysis of petB/D sequences from cultured P. marinus and Synechococcus, using biotinylated forward primer PPETBD314 and unbiotinylated reverse primer PPETBD1160R (this thesis Chapter Two). PPETBD1160R also served as a primer for sequencing. Conditions for PCR amplification were less stringent than those used for the cloning experiment (this thesis Chapter Three) and produced sufficient amounts of PCR

123 products to serve directly as sequencing templates. Templates were prepared using streptavidin-coated magnetic beads (Dynal) and sequenced directly with Sequenase (United States Biochemical), both according to their manufacturers' instructions.

RESULTS

Autoradiograms from direct sequencing of petB/D amplified from sorted samples revealed multiple comigrating bands which differed in numbers and intensities among the different samples (Figure 1). Most dramatic were the results of direct sequencing of petB/D from the 85 m Bright and 135 m Dim samples in which multiple comigrating bands were frequent in the intergenic region above the petB/D start codon and were also visibile at some third codon positions in petD (Figure 1, lanes b and e). Multiple comigrating bands were less prominent, but visible, in sequencing reactions from the 50 m amplification (lane a) and could not be detected in the 85 m Dim and 135 Bright (lanes c and d, compare to reactions for a P. marinus culture in lane e).

"Major" sequences could be read above the relatively low or absent background of comigrating bands for the 50 m, 85 m Dim and 135 m Bright samples. All of these major sequences were similar to Allele 1, the most frequently recovered allele in the cloning experiment (49 out of 191 clones in the combined dataset for Gulf Stream and Sargasso Sea sorted samples, 22 out of 139 clones for Gulf Stream samples alone, Table 1 in Chapter 3) (Figure 1, lanes a, c and d).

DISCUSSION

The presence of multiple comigrating bands in sequencing autoradiograms can result either from sequence heterogeneity in the template DNA or from suboptimal

124 reaction conditions which allow premature termination of nascent DNA chains. The latter cause would not, however, be expected to produce a disproportionate number of comigrating bands at third codon position sites and intergenic regions, a result which is dramatically evident in autoradiogram lanes from the 85 m Bright and 135 m Dim sorted samples (Figure 1, lanes b and e) and, less dramatically, in lanes from the 50 m sorted sample (lane a). The results of this analysis were therefore consistent with the presence of multiple petB/D alleles in at least some of the flow cytometrically sorted field samples, and prompted the cloning and sequencing investigation in Chapter Three.

The investigation in Chapter Three revealed that, in fact, all of the PCR products from the Gulf Stream and similar products from the Sargasso Sea contained multiple petB/D sequences. Consistent with the results of direct sequencing, the 85 m Bright and 135 m Dim samples contained the largest variety of petB/D alleles: when corrected to a constant sample size of 19 clones these samples contained 12.1 and 13.7 alleles, respectively. The 50 m, 85 m Dim and 135 m Bright sorted samples yielded 5.3, 5.2 and 5.4 alleles for the same constant sample size (Table 1 in Chapter Three).

There was, however, one type of comparison in which the results of the direct sequencing and cloning experiments did not consistently agree. The major sequence and the most frequently recovered allele were not consistently identical for each of the sorted samples: while the 50 m sample yielded a majority of Allele 1 clones (17 out of 23 clones) consistent with its major sequence, the sets of clones recovered from the 85 m Dim and 135 m Bright samples were mostly Allele 5 (17 out of 24 and 16 out of 23 clones, respectively), with only a minority of clones belonging to Allele 1 (three out of 24 and two out of 23 clones, respectively) (Table 1 in Chapter Three). This discrepency may be attributable to two potential causes: 1) Allele 1 may have been preferentially amplified under the less stringent conditions used to produce PCR products for direct

125 sequencing as compared to the cloning experiment, or 2) Allele 1 may have been preferentially chosen as template during direct sequencing reactions. Discrepencies in the identities of major sequences and most frequently recovered alleles for the different sorted samples reinforce the necessity of mentioning the caveat in Chapter Three: genotype frequencies among PCR clones (and major sequences of PCR products) derived from natural populations using degenerate, inosine-containing primers reflect both gene frequencies in the sample population and PCR amplification bias (as well as sequencing reaction bias) for or against specific alleles. It is therefore speculative to draw inferences from gene frequencies (or major sequences) in datasets such as the one at hand.

126 Figure 1. Autoradiograms from direct sequencing of petB/D PCR products amplified from flow cytometrically sorted P. marinus from natural populations in the Gulf Stream (a-e) and from unsorted, cultured P. marinus MIT9313, isolated from the 135 m Gulf Stream water sample (L. Moore pers. comm.) (f). The arrow marks the initial petD ATG codon, above which is intergenic region sequence. Dashes mark the positions of third codon positions in petD. a) 50 m, b) 85 m Bright, c) 85 m Dim, d) 135 m Bright, e) 135 m Dim sorted samples (Chapter Three). f) unsorted, cultured MIT9313.

127 L~YP

mAJ REFERENCE

Hurlbert, S.M. (1971). The non-concept of species diversity: a critique and alternative parameters. Ecology 52:577-586.

129 Appendix II

JUSTIFICATION FOR THE 53.5% SEQUENCE DIFFERENCE

CRITERION FOR SEQUENCES ASSIGNED TO A SINGLE ALLELE

130 Introduction. In Chapter Three, genetic diversity in natural populations of Prochlorococcusmarinus was investigated by comparing DNA sequences of cloned PCR products amplified from flow cytometrically sorted field samples. The genetic locus examined included the 3' end of the petB gene, an intergenic "nonsense" region and the 5' end of the petD gene (the amplified locus, including both gene fragments and the intergenic region, being referred to as "petB/D"). Clone sequences (which were inferred from single gel readings, and not confirmed by sequencing complimentary strands) were categorized into Alleles, or sets of clones within which sequence differences were indistinguishable from PCR and sequence determination (PCR+SD) error, corrresponding to genotypes present in the field samples. For ease of analysis, each Allele was characterized by a single clone designated as its "prototype." Among the 191 clones analyzed there was a great deal of sequence diversity, with 59 sets of clones exhibiting intergenic region sequences that were so different in length and sequence that it was impossible to identify homologous nucleotides for use in sequence alignment. These sequence differences were easily identified, despite the undoubted presence of PCR+SD error, as ones which originated in genetically distinct individuals in the sorted populations. Three of these sets of clones, however, contained sequences exhibiting sufficient mismatch to suggest that they were amplified from individuals containing different, though related sequences. These sets were subdivided into Alleles according to a •3.5% total sequence mismatch with prototype criterion for clones included within an Allele. The total number of Alleles arrived at by this process, including the sets of clones with unalignable intergenic sequences, was 71.

This Appendix presents arguments in support of the •3.5% total sequence mismatch criterion used to group clones into Alleles in Chapter Three. Initially, bounds are set on possible values for the criterion using a theoretical calculation based on published values for Taq polymerase error rates in PCR and an ad hoc estimate of the range of sequence determination error rates. Next, evidence is presented from the distribution of closely related 131 clones among field samples suggesting that some sequences within these bounds do differ as a result of natural variation, and not simply due to PCR+SD error. Finally, a practical method is described for optimizing the choice of criterion using the fraction of sequence mismatch occuring at third codon positions, and the method is applied to the dataset.

Theoretical confidence limits for the frequency of sequence mismatches between cloned PCR products attributable to PCR+SD error. During the course of PCR amplification, replication errors give rise to PCR products containing one or more mismatches with the original template sequence. The frequency of errors within individual PCR product molecules can be seen to follow a Poisson distribution, with variance equal to the mean, despite the fact that the frequency of errors in whole PCR reactions conforms to a Luria-Delbruck distribution (Eckert and Kunkel 199 la,b), and thus has a very high variance'.

In the final, cloned PCR product the mean error frequency (f) is equal to the probability that polymerase error has caused a nucleotide to become mismatched with its progenitor in the original template (Eckert and Kunkel 1991a,b),

1 The Luria-Delbruck distribution was first proposed to describe the widely varying frequency of mutant cells in replicate bacterial cultures grown under non-selective conditions in the classic fluctuation test (Luria and Delbruck 1943), but the distribution is equally applicable to the frequency of errors in replicate PCR reactions (Eckert and Kunkel 1991a,b). In a Luria-Delbruck process the rare occurrence of a mutation (PCR error) during an early generation (cycle) causes the mutant (error) frequency in a small number of cultures (reactions) to be substantially greater than the mean, giving rise to a frequency distribution having a high variance. This is because the final number of mutants (PCR products) resulting from an early mutation (error) is amplified through the process by a factor of 2x, where X equals the number of generations (cycles) between the mutation (error) and the analysis. Thus, in a Luria-Delbruck process the number of the round in which a error occurs influences the final error frequency attributable to that error. In the case of the distribution of errors within a single PCR product molecule, however, the situation is different. In this case, each nucleotide is descended from its original template nucleotide through a number of rounds of PCR replication, each with a probability of error. Every nucleotide in the chain is subject to this process independently, and the final error frequency attributable to any error is equal to 1/L, where L is the number of nucleotides in the chain, and is independent of the number of the round in which the error occurred. If the probability of error is assumed to be equal for every nucleotide position, errors within the chain are seen to be rare, independent events with equal probabilities of occurrence. Thus the frequency of these errors conforms to a Poisson distribution. 132 nm f=_ (1) 2 where n is the number of PCR cycles and m the probability of error per nucleotide per cycle2. Since these errors are distributed as a Poisson distribution, the variance for the frequency of errors in individual PCR product molecules is equal to the mean, and confidence limits can be calculated for the frequency of PCR errors in individual product molecules using known values for the polymerase error rate. The expected frequency of differences between two cloned PCR products that can be attributed to PCR error is equal to the expected frequency of errors in a cloned molecule twice the length (1)of the region compared between the two products (neglecting coincidence of errors at homologous nucleotide sites and assuming the two cloned molecules are not amplification products of the same template molecule, which would decrease the expected frequency of their sequence differences). Accordingly, the 95% confidence upper limit (L95) for the frequency of mismatch between two cloned sequences attributable to PCR error is

L95 = f + to.1[21-1 'f (2) where t0.1121-1] is the critical value for student's t distribution for a one-tailed test with a = 0.05 and degrees of freedom equal to 21 - 1.

A term may be added to expression (1), the mean frequency of polymerase errors in cloned PCR products, in order to account for sequence determination errors. This gives nm f =_-+g (3) 2 where g is the rate of sequence determination errors in errors nucleotide- 1. If sequence determination errors are assumed to follow a Poisson distribution, this expression for f may be used to calculate 95% confidence intervals for PCR+SD error using equation (2).

2The probability that a molecule has inherited a DNA strand containing an error introduced during the previous round is m/2, hence the 2 in the denominator. This factor applies to the product of the final PCR round as well, although in this case it reflects the probability that the progeny of a mutant strand in a cloned heteroduplex will predominate in the final plasmid preparation. This formulation neglects the effects of superimposed errors. 133 Sequence determination errors are estimated ad hoc at between 0.0 and 5.0 x 10-3 errors nucleotide-1.

Application of the theoretical confidence limits to analysis of the data in Chapter Three. Mean, standard deviation and upper 95% confidence limits for the frequency of sequence differences expected in cloned PCR products compared over 71 basepairs (the length of the shortest sequence in the dataset) were calculated using the lowest and highest reported Taq polymerase error rates (Eckert and Kunkel 199 la) (corrected to reflect nucleotide substitutions only3), over the estimated range of sequence determination errors (Table 1). These upper confidence limits provide theoretical bounds for the value of the criterion to be used to declare sequences different or indistinguishable. If the lowest combined error rate were to apply, sequences in Chapter Three having alignable intergenic regions and differing by up to 3.1% of nucleotide positions would be considered indistinguishable from PCR products amplified from identical template sequences, at p = 0.05. However, applying the highest reported error rate raises this criterion value to 18.5%. Thus, the 95% upper confidence confidence limit for the amount of sequence difference between cloned PCR products attributable to error is shown to be sensitive to estimates of the polymerase and sequence determination error rates, the true values of which are unknown for this experiment (Table 1).

3 Values for the range of Taq polymerase error rates inPCR are available from the literature (Eckert and Kunkel 1991) but require correction in order to be applicable to data analysis for Chapter 3. In Chapter 3, apparent insertions and deletions are not counted as sequence mismatches (i.e., gap weights were set to zero) on the assumption that these mismatches are likely to have been caused by PCR+SD error. Accordingly, Taq polymerase error rates used to calculate theoretical upper 95% confidence limits for PCR+SD error in Chapter 3 have been corrected to omit insertion and deletion errors using the original data, where this is aplicable. Specifically, the highest reported Taq polymerase PCR error rate has been corrected from 2 x 10 to 1.6 x 10- 4 errors nucleotide- 1 cycle-1, using the original data (Dunning et al. 1988), while the lowest reported error rate is unchanged because insertions and deletions were not observed in the original experiments (Goodenow et al. 1989).

134 0 a) 1o4 A *1 4-

0d

oi 8 .00 a 0i0 0

coou

0

CUD4.1

0

4- %n25 O) .c -000.

f~o0

00 a.) LAua.). .40. - a) t: 44 ao. c7 U) 0 .. 000 4- t

0- 0 W 20.Cl)E- (.1)

- 4 78 UC ) Q,, 00 -L4-

po cn 04 * Cd 0-~t 0 ~Cn.

cdu 0 4El

Ev LV I...# C.)

135 Applying the calculated extreme bounds for the criterion to the task of lumping indistinguishable sequences into Alleles using the set of clones recovered from the Gulf Stream and Sargasso Sea yields a total of 74 Alleles in the combined dataset for the lowest error estimate and 59 Alleles for the highest one4. While the impulse to err on the side of conservatism would dictate the choice of •18.5% mismatch as criterion for inclusion within an allele, independent evidence suggests that this criterion would lump together a number of cloned sequences which do, in fact, derive their differences from genetic heterogeneity in P. marinus populations. A lower criterion, while still above the 95% upper confidence limit calculated from the lowest reported Taq polymerase error rate, can be estimated by considering the pattern of distribution of closely related sequences among field samples and the ratio of their mismatches at third codon positions to mismatches at all codon positions in order to distinguish sets of related sequences giving useful information about genetic diversity from sets in which members differ mostly due to PCR+SD error.

An upper limit for the sequence mismatch criterion suggested by the distribution of closely related sequences among field samples. In the dataset of Chapter Three four 100.0% identical sequences belonging to Allele 71 were recovered from four different field samples: the 85m Dim and 135m Dim Gulf Stream samples and the 40m and 70m Sargasso Sea samples, and therefore were cloned out of separate PCR reactions. Allele 71 is highly similar to Allele 1 (8.1% mismatch), which includes numerous identical clones distributed among several field samples. If the 18.5% mismatch criterion were to be applied, Allele 1 and Allele 71 (as well as several others) would be lumped together as members of a single Allele. While it is not unlikely that a single Allele 71 clone could have been produced from an original Allele 1 template, it is highly unlikely that the identical set of errors would

4 Application of the upper bound criterion calculated using the highest combined PCR+SD error rate causes all sequences having alignable intergenic regions to be lumped with each other. There are 59 sets of intergenic regions which cannot be aligned with each other, therefore application of this higherst possible criterion value results ina dataset containing 59 Types. 136 occur in four separate PCR reactions and four separate sequence determinations. Therefore, at least some sequences in the dataset that differ by 8.1% represent true genetic variants and not the results of PCR+SD error.

Optimization of the criterion for distinguishing sequence mismatch attributable to PCR+SD error from mismatch due to genetic diversity by considering the fraction of sequence mismatch at third codon positions: a practical method. Evolutionary change at third codon positions is generally faster than at the first two positions because degeneracy in the genetic code allows substitution at the third position of most codons without changing the encoded amino acid. PCR+SD errors, on the other hand, should be random with respect to codon postion. A high fraction of sequence mismatch at third codon positions (fraction of mismatch at third codon positions = fractional mismatch at third codon positions/3 x fractional mismatch at all codon positions) is therefore evidence that a pair of cloned sequences differ mostly due to true evolutionary divergence under selection for a conserved amino acid sequence, rather than due to PCR+SD error. Conversely, a low ratio is evidence of the activity of random processes like PCR error and inaccurate sequence determination.

In an ideal dataset, a pair of cloned sequences which differ due to PCR+SD error alone should have a fraction of sequence mismatch at third codon positions equal to approximately 0.33. In an actual dataset like the one in Chapter Three, however, there is likely to be a significant amount of scatter in values for this ratio at the low end of the total sequence mismatch scale due to sampling error associated with the small number of mismatches in highly similar cloned sequences of finite length. Additionally, in a real dataset some sequences differing by small increments of total mismatch may, in fact, differ due to evolutionary processes, but these individual comparisons cannot be distinguished from PCR+SD error because sampling error obscures differences in their fractions of mismatch at third codon positions. The effect of the presence of these small evolutionary differences will

137 be to increase the average fraction of mismatch at third codon positions above the 0.33 value for comparisons at the low end of the total sequence mismatch scale. In the face of this complexity, the task of setting a total mismatch criterion below which sequence differences are considered indistinguishable from PCR+SD error (for experiments in which the error rate has not been experimentally determined) becomes a process of examining the distribution of fractional mismatch at third codon positions along the scale of total mismatch and identifying a point of discontinuity at which the sphere of influence of random errors gives way to one of evolutionary divergence. At this point there will be a maximum difference between the means for fractional sequence differences at third codon positions for sequence comparisons above and below the criterion. This transition point should fall between (or perhaps slightly below) the theoretical upper 95% confidence limits for mismatch attributable to PCR+SD error, determined according to the range of probable polymerase and sequence determination error rates.

Optimization of the criterion for sequences in Chapter Three. For comparisons of clones having alignable intergenic regions in the dataset of Chapter Three, a plot of the fractional sequence mismatch at third codon positions (in coding regions only) versus total mismatch (in coding regions plus intergenic regions) exhibits a marked biphasic appearance (Figure 1). Consistent with the expectations of random error due to PCR+SD superimposed over a smaller amount of difference due to evolutionary processes, comparisons at the low end of the total mismatch scale have a lower average fraction of mismatches at third codon positions than do comparisons at the high end, but this average is greater than 0.33. A criterion chosen to distinguish domains of total sequence mismatch in which the dominant cause of sequence mismatch is PCR+SD error or true genetic diversity should divide the dataset into two subsets having the greatest possible difference in their average fraction of sequence mismatch at third codon positions. A graph plotting this difference for

138 Figure 1. Fraction of sequence mismatch at third codon positions (in coding regions only), plotted against total mismatch (including coding and intergenic regions) for comparisons of clones with alignable intergenic regions in the dataset of Chapter Three. Dotted vertical lines are theoretical upper 95% confidence limits for percent sequence mismatch (L9 5 x 100) attributable to PCR+SD error for the highest and lowest combined Taq polymerase and sequence determination error rates (Table 1). Datapoints below 1% total mismatch are omitted, and some symbols represent superimposed data points. Solid curve: fourth order polynomial fitted to all datapoints. Dashed curves: upper and lower 95% confidence intervals for the solid curve.

139 12.

1.0

0.8

0.6

0.4

0.2

0.0

-I0 0 10 20 Total mismatch (%)

140 hypothetical criteria placed along the total mismatch frequency scale at intervals of 0.1% mismatch shows a local maximum at 3.5% total mismatch, between the theoretical bounds imposed by published values for Taq polymerase error rates combined with estimates for sequence determination error rates (Figure 2). The value 3.5% total mismatch was therefore chosen as the best possible criterion to distinguish sequences differing due to PCR+SD error from those differing mostly due to evolutionary divergence, for the dataset of Chapter Three. This value is consistent with a Taq polymerase error rate near the low end of the range of reported values5.

Reexamination of the graph plotting the fraction of sequence mismatch at third codon positions versus total mismatch frequency (Figure 1) using the 3.5% criterion to divide the data into two subsets and analyzing each subset independently for the mean and its 95% confidence intervals, illustrates the contrast between sequence differences attributed to PCR+SD error at the low end of the total sequence mismatch scale, and differences due to natural variation at the high end (Figure 3). The mean fraction of sequence mismatch at third codon positions for comparisons to the left of the criterion equals 0.48; for comparisons to the right of the criterion the mean is 0.81. 95% confidence intervals for these means do not overlap, indicating that the criterion chosen by this method separates the sequence comparisons in the dataset of Chapter Three into groups which are significantly different in their fraction of sequence difference at third codon positions, and therefore in their ratio of random to evolutionary sequence differences.

5 It should be pointed out that this analysis would be improved by an additional sequencing effort, providing full-length sequences confirmed by sequencing in both directions for all clones having alignable intergenic regions. It is predicted that the number of sequence comparisons yielding ones and zeros for third codon position/all codon position mismatch ratios would decrease with increased sequence lengths and accuracy, reducing scatter at low total mismatch values in Figures Al-1 and AI-3 and perhaps altering Mean R - mean L values at low total mismatch values in Figure AI-2 as well. An additional sequencing effort before publication of these results is being considered. 141 Figure 2. Optimization of the criterion identifying the upper bound for total sequence mismatch considered indistinguishable from PCR+SD error. Datapoints are differences between mean values for the fraction of sequence mismatch at third codon positions for datapoints in Figure 1, grouped to the right and left of hypothetical criteria placed at 0.1% intervals along the total mismatch axis. Dotted vertical lines are theoretical upper 95% confidence limits for percent sequence mismatch (L95 x 100) attributable to PCR+SD error for the highest and lowest combined Taq polymerase and sequence determination error rates (Table 1), and therefore represent theoretical bounds for the criterion. The curve is a second order polynomial fitted to points between 2.0% and 5.9% total mismatch to aid in identifying the local maximum, which is marked with a solid vertical line at 3.5% total mismatch. The criterion is therefore set at the value of total mismatch dividing the sequence comparisons into the two groups having the most different mean fraction of sequence mismatch at third codon positions possible within the bounds of the theoretical limits.

142 0.4

S0.3 Z"

9 0.2

0.1

0.0 0 10 20 Total mismatch (%)

143 Figure 3. Data from Figure 1 divided into two groups by the criterion (solid vertical line) drawn at 3.5% total mismatch. Data to the left of the criterion are comparisons between closely related sequences which, on average, exhibit low fractions of sequence mismatch at third codon positions emblematic of sequences which differ mostly due to PCR+SD error. Data to the right of the criterion are comparisons between sequences differing by greater than 3.5% total mismatch which, on average, exhibit higher fractions of sequence mismatch at third codon positions, characteristic of sequences differing due to evolutionary processes. Error bars enclose 95% confidence intervals for the mean of all datapoints to the left or right of the criterion.

144 1.2

1.0

0.8

0.6

0.4

0.2

0.0

-0.2 0 10 20 Total mismatch (%)

145 Conclusion. This Appendix presents both theoretical and practical methods for discriminating between true natural variation and PCR+SD error in the analysis of cloned sequences amplified from populations of mixed genetic composition. To my knowledge, both are currently absent from the literature. The practical method is applied to the data of Chapter Three for sorting clone sequences having alignable intergenic regions into Alleles corresponding to genotypes present in natural populations. Even without this analysis, the dataset of Chapter Three shows abundant genetic diversity, with 59 sets of petB/D clones having unalignable intergenic regions; differences among these sets must be attributed to natural variation and not to PCR+SD error. Application of the practical method to the analysis of clones having alignable intergenic regions adds 12 additional Alleles to the Chapter Three data used for population genetics analysis. The Alleles produced by subdivision of sets of clones having alignable intergenic regions exhibit an average fraction of sequence mismatch at third codon positions of 0.82 when compared to related Alleles, significantly different from the expectations of random error. Therefore, even if their prototype sequences contain errors, these Alleles serve to identify genotypes present in the natural populations and are legitimately included in the population genetics analysis.

146 REFERENCES Dunning, A.M., Talmud, P. and Humphries, S.E. (1988). Errors in the polymerase chain reaction. Nucleic Acids Res. 16:10393. Eckert, K.A. and Kunkel, T.A. (1991a). DNA Polymerase Fidelity and the Polymerase Chain Reaction. PCR Methods Applic. 1:17-24. Eckert, K.A. and Kunkel, T.A. (1991b). The fidelity of DNA polymerases used in the PCR. In Polymerase chain reaction: a practicalapproach (eds. M.J. McPherson, P. Quirke and G.R. Taylor). IRL Press, Oxford, pp.. Ennis, P.D., Zemmour, J., Salter, R.D. and Parham, P. (1990) Rapid cloning of HLA-A,B cDNA by using the polymerase chain reaction: frequency and nature of errors produced in amplification. Proc. Natl. Acad. Sci. USA 87:2833-2837. Goodenow, M., Huet, T., Saurin, W., Kwok, S., Sninsky, J. and Wain-Hobson, S. (1989). HIV-1 isolates are rapidly evovling quasispecies: evidence for viral mixtures and preferred nucleotide substitutions. J. Acquir. Imm. Defic. Syn. 2:344-352. Keohavong, P. and Thilly, W.G. (1989). Fidelity of DNA polymerases in DNA amplification. Proc. Natl. Acad. Sci. USA 86:9253-9257. Luria, S.E. and Delbruck, M (1943). Mutations of bacteria from virus sensitivity to virus resistance. Genetics 28:491-511. Saiki, R.K., Gelfand, D.J., Stoffel, S., Scharf, S.J., Higuchi, R., Horn, G.T., Mullis, K.B. and Erlich, H.A. (1988). Primer-directed enzymatic amplification of DNA with a thermostable DNA polymerase. Science 239:487-491. Sarkar, S. (1991). Haldane's solution of the Luria-Delbruck Distribution. Genetics 127:257- 261. Tindall, K.R., and Kunkel, T.A. (1988). Fidelity of DNA synthesis by the Thermus aquaticus DNA Polymerase. Biochemistry 27:6008-6013.

147 Appendix III

COMPARISON OF NUCLEOTIDE DIFFERENCES AMONG CLONES ASSIGNED TO ALLELE 1 TO KNOWN PATTERNS OF NUCLEOTIDE

SUBSTITUTION ERROR BY TAQ POLYMERASE

FOR DATA OF CHAPTER THREE

148 This Appendix provides an analysis of the pattern of sequence mismatch among clones assigned to one "Allele," or set of clones within which sequence differences are considered indistinguishable from PCR and sequence determination (PCR+SD) error according to the 3.5% total sequence mismatch criterion (Appendix I), in order to investigate the hypothesis that their differences arose through Taq polymerase error during PCR. Results suggest that different sources of sequence variation, possibly sequence determination error or natural genetic variation, contribute to sequence mismatch within Types, in addition to Taq polymerase error.

Sequence mismatches among clones assigned to Allele 1, the allele containing the greatest number of clones in the dataset of Chapter Three, were compared to nucleotide substitutions characteristic of Taq polymerase error. Allele 1 was spawned from the subdivision of a larger set of clones having alignable intergenic regions (Appendix I) and contains a total of 44 clones, 11 of which are identical to the prototype sequence. 11 additional clones, 10 derived from the 70m Sargasso Sea sorted sample and one from the 50m Gulf Stream sample, contained the identical G->A substitution as their only mismatch with the prototype. This nucleotide substitution was present in the remaining S70 clone sequenced across the site in question, as well, and may have originated in a natural variant of Allele 1 P. marinus present at 70m in the Sargasso Seal. If this nucleotide substitution is not attributed to PCR+SD error, then at least 22 out of 44 (50%) of the clone sequences classified as Allele 1 are error-free. Among other mismatches, T- >C transitions were the most frequently encountered substitutions (14 out of 44 total substitutions), with C->T almost equally abundant (12 out of 44), followed by A->G (5 out of 44) and G->A (5 out of 44)

1Although it is possible that a significant fraction of pet B/D sequences cloned out of a pair of pooled PCR reactions, each beginning with about 105 sorted cells could contain the same PCR error, especially if this error occurred at a hotspot for PCR misincorporation (Keohavong and Thilly 1989), it seems unlikely that all of them would have the same error.

149 (Table 1). Detailed studies examining Taq polymerase errors have proven that most Taq polymerase errors are A->G or T->C transitions (Dunning et al. 1988, Saiki et al. 1988 Tindall and Kunkel 1988, Keohavong and Thilly 1989, Ennis et al 1990, Eckert and Kunkel 1991). Therefore, the type of sequence mismatches observed between clones assigned to Allele 1 and the Allele 1 prototype do not conform to expectations of variation due solely to Taq polymerase error. It therefore seems likely that natural genetic variation and/or sequence determination error contribute significantly to sequence variation within Alleles.

150 Table 1. Nucleotide mismatches between clones assigned to Allelel and the Allele 1 prototype sequence, excluding G->A at position 1243.

Nucleotide Identities: Number of Number Prototype->clone Occurrences of Sites T->C T->A T->G C->T C->A

G->A G->T A->G A->T

151 REFERENCES Dunning, A.M., Talmud, P. and Humphries, S.E. (1988). Errors in the polymerase chain reaction. Nucleic Acids Res. 16:10393. Eckert, K.A. and Kunkel, T.A. (1991). The fidelity ofDNA polymerases used in the PCR. In Polymerase chain reaction: a practicalapproach (eds. M.J. McPherson, P. Quirke and G.R. Taylor). IRL Press, Oxford, pp.. Ennis, P.D., Zemmour, J., Salter, R.D. and Parham, P. (1990) Rapid cloning of HLA-A,B cDNA by using the polymerase chain reaction: frequency and nature of errors produced in amplification. Proc. Natl. Acad. Sci. USA 87:2833-2837. Keohavong, P. and Thilly, W.G. (1989). Fidelity of DNA polymerases in DNA amplification. Proc. Natl. Acad. Sci. USA 86:9253-9257. Luria, S.E. and Delbruck, M (1943). Mutations of bacteria from virus sensitivity to virus resistance. Genetics 28:491-511. Saiki, R.K., Gelfand, D.J., Stoffel, S., Scharf, S.J., Higuchi, R., Horn, G.T., Mullis, K.B. and Erlich, H.A. (1988). Primer-directed enzymatic amplification of DNA with a thermostable DNA polymerase. Science 239:487-491. Tindall, K.R., and Kunkel, T.A. (1988). Fidelity of DNA synthesis by the Thermus aquaticus DNA Polymerase. Biochemistry 27:6008-6013.

152 Appendix IV

ALLIGNED SEQUENCE DATA

153 w El qi) • oa 0 CN 1 0 Cmu0moo m Om ,-U O4 1-4 m C)DO mm N>1 0 ý4

-H• cco - n :D = =) =)O - o04t .- ,•m J: 0000000 uuuu ci)t-lU) *H0-,-0i cy)-cCO (9(90(90(9(

UU) :DC n

I I 0c00 I I I I I I l I I I 00 I I 0 I 00 co UUUuuUu00000•0 c) 0 m 3 010000000000000 ,1-.Wo r;loc4 4 i -,4. kD 0>mci• w- CU H 4 ,-U)- W E-1*1- 0,-SZ KC4 F:Cg4 4C 4 l<44 4 0 O: 0000000< I ; z c f< f< 000I t < c0(90 OI Ohct< - UUUUUUU=UUUUOUOD00000000000000~- Ou 000000000000000- U -. tj*-H 04 0'-4,-. 4-o1-4 m

< < oc < (9l000(9(ouuli 00l I l 0 01< l I $I UUUUUUUUUUUUGUUui • ou• UUUUUUU0000000 000000>OOOOOOOOUUUUUUUUUUUUDDO• I IIIIUIou UUUUUUUUUUUUUOUU 0000000 000000000000 0 UUOUUUUU 0000000 uuouuuo-uuuououuuD00000000000000

(1) ."0-0 mn Mc cUNS-U U 04 00000000~0000i #£,E0 .-o. .00 (9000000000000(5- Q-)a4o 00 0000 la)0 Uo0 0:ý- O 000000 000000000- -E 00U000 00000U' 14 Ctý 4 00 0•0 -H 0C U)ciJcciý4 OD ' 0 O0 00 0000 00000000000000 0- o-~-ci'4 00.:0Q4 C0• 00000 00000-4 ci ) ýcN 4-co)4cc) cur- UUOUUUUUUUU UUUUUUUUU•UUUUUUOUUO UUU 5 o U000000009 00000000000000000 ý41.-44-) -- 0 MN -4i) -Qci) 'ANC' c>r-> 00 0000 4-40w0 00 0000

0 0 . N- - 00 000 ' Z0 mVam 4)4 o u u uu 4-3) 04 04 00 0000 0 0000ou00000000000 0 0Ocm ~2 rM 1.4 -H r. o u ou u 1-) (accý4- UU UUU

> 4444E 0000 u : ai)n -Ha 04 <4Ot u 00 (99(((000

0 r,0 1) E'Q *110U)- ) U) -H - 0 ci) M U~4-4- 0 o

E M r-I" - c>1 04 N0 0mmo>C m Oc0-40 • C Cmm mo uO 000000000000000 .4oT n % - 0 co w f IXo (9 (9(9( 9(9(9(9(9(9( 0 *0Ea :2 E- N 1)4 ' (a0 4 - 0wHa 00000000u 0 .- 0)a)H4 n XX O C . 4 U (9 X9(949 9. 9(9M94(9aý -4>, 0 OcC mX U EA in N:R: (n .4U0 <') Xm a4 -4I-4-4-4-4-40-4 CD)(9c(9mj-99%9-c9omc D( C DC ) DC 9 DC)C 004T ) o0) 4 00 0- --Ný qr4 ml---- r-iCN)-40 U~ca0 aA I --- 4-4-4 -4 -4 -4-4-4 -4 -4 -4-4-4 -1 o-,-iMCo OiJ cU' S r-4 CN r4I E N- C, - 0 S40 m 0o uOOm S0 C CMI ON0>-4UO A-) -4 (l0 C) C M Mr- >1 4

W . 4U 0O~r Mf Sa4ý4~~ ~rE-mo 10E-Hm = r 44u 44 W =ý :3:M MM 4 XC1O C4'0E- 3O -4W - m04m 0• M~a •ZHmbMoo4m AUNM> 0 UUUUUUUUUUu UUU UUUUUuuu4 88888888 80 UU U U U U U U U U V U U U U r-4 U UUUU U U U U UU U U U U r4 UUUUUUUUUUUUUUUH- 00000000000000UUUUUUUUUUUUUUU00r4 u. I- 0 U UU BBCUU U U U U U4 U 4 U4 UUfBC 4 U4 r4r 00000000000 U -I UUUUUDDUUUUU u -1 u ý4

uuuuuuuuuuu UUUUUUUUUUU 8888• 8888• 08 00000800 u UUUUUUUUUUUUUUU-4 C,--'u 1-4 BBuH

U r~-4 0U-4OU'

UUUUUUUUUUUUUUUUUUUOUUU U u UIUUU 0 I 8UUUUUUUUUUU

UUUUUUUUUU UUUUUUU•uuu OH OUUUUUUUUUU UH uuuuuuuuuuuUUUUUUUUUUU 880080088008 0UN UUUUUUDDUUDUUD=H ý vouuoovuuooovuuv- U8U88U8888U UUUU<<<<<<

('- 4:-.

U4.3 LI U'-.

o " m HIo C r-i ('4 r-I 4. o0-4 U 0 0 0) U oL)fu 0.0 .1111.1 0(' A4 HOE X XO4 w4 L) -. u C14'ko.Q -4 O'- w rc m.n 0 U O.0 U tn wH Oi m = r-i r'4 mI r Ln %o0r- w cm 0> 4 "' r'1 q Ln w. S" m v L %or- wo 0 " m - L vAr-4 T-. ,-4 v-4 v-4 r-4 00000000000000004 - r-4 ,-4 r r --I -IA -fr-4 ý-4 v-44 44ý§" 4 r4 -4 0 CO4('41 ...... C). C>4(D4 C>4(40C ...... C>1(4 C'4 (' C>4 ('C>> 0' aO4-0 -4 a) EN m Fh4O 0D C my U O O-4uo o . o7 0 r-4 U 0 .- 4 moo mmr->1 0 ,-4 ýo -Q -4 r-4 J0 MKcllýu a POczý: I40 -a a cc, E x M:< u 000000000000- UUUUUUUUUUUUUUUM UUUUUUUUUUUUUUU•- U0

0UUUUUUUUUUU•U0 88 8888Fc PC Pi 0:4 88888oz 4 UUUUUUUUUUUUUUU-4 888888888888~88i000UUUUUUUUUUUUUUU• 0000000:4000 UUUUUUUUUUUUUUU- •UUUU•UUUUU• 000000000000000 UUUUUUUUUUUUUUU,-l UUUUUUUUUUUUUU UUUUUUUUUUUUUU

UUUUUUUUUUUUUUU •U UU C)U C C)U U = D UU U0UUUUUU~UUUU~~UK-i UCDU uUUUu-4

UUUUUUUUUUUUUUUU 00000000000UUUUUUUUUUUUUUU 00 000 0000 'A UUUUUUUUUUUUUUU,-UUUUUUUUUUUU U-

UUUUUUUUUUUUUUU• OOOOQUOQ©U0004U0000C000ut00 0

UUUUUUUUUUU888888888888Uuuxc F<

000D0SU UC)Uu 0U U 000UUU)U 000-UUC-4

JUUUUUUUU UUUUUUl 0:4000000000000000 ououuouoCuDUUUU~UUU MUUUUUuUUUUU c0uomUU'-4 -4

U n n n M n n -4 0 0 u uu 00U 0 >4zuUC> 0 0 0 u 0 00- o II 10I 00001UU 1UU UI n 00000000000 UUUUUUUUUUUUUDUU- 00000000000 000DUuUU 0000000000'UU UUUUUUUU-4 0UUUUU4U4UUU0,-iDýDO ý: O 3= =000000004 ):)= u u n -4 UUUUUUUUUUUUU~UU4DODOOD DO•ODUDD

000000 00000 0 OUUUUUUUUUUUUUU- 0 000000000 UUUUUUUUUUUUUUU-4oououuouuuuouuuu-

cci UUUUUUUUUUUDDDD 00000l0000 ii cci W)e C:)oa C) E r- M I" AZ .- I0000 (N -1 c: 0 OmCl ONCD - 0 N 0D-4 U 0 ID .0 r-4 moo mm r- >,l- 4 0 m > U)J- mee"w=xFCf< r,"HH 0) a) H 14L wOHa4 U4

-44--4ý4 -4 -4-ý4 -4-4-ý 4-ý4-4-4-4 -4 0000000000000000 0000000000000000 0--fAo 0 w r-Ivww r- w-- w lr w LAL L L L L L L L L L L L L L L cJO 'V z ~ m ~ ~ o-uc a

in 4E in a4 U -L Cq nCD C CD(r 0 zr0)D Cm 1 D r Uio lA H

~000000000000040H 000U000000000000-

I It I I I I I I I I I I I I 1

80000880808000000- 0 00 00 0 0 0 0 00H o< 4 4 OtOtl t c ~ :cx :D:: :))nD :: MDH UC)uu -)u uu 4 FC 0: DC440: i l<0: 40 f

:3 C Cd C) UD U C) D CD UD CD UD C =D CD

0 u CD CD CD CD CD CD CD CD CD z u' ' u

00 C C CDD8 (-) 0000000( 00 000H: CDD D-I OOOOOO:D OOOD OOOOOOOOOOOOOOOHUU U D D M :D :D000 o<

D00C-)U U u u u 0:4 u U-4 u u uu u u U U UU u uUCD- mD:00 000

000000000 00U000H 0000000000 00- < 04CCD

0 .)C 0 0 0 0 0 0 U () 0 0 0

0000000000000<0

u uU uu uu u u -i C. 0 0 0 0 0 0 0 0 0 0 U 0c < < < < DDCDCDD

m0 0) 0)>~ EN 40) - CO(nLn 01%0-A 0 ::: 0 U~-) 3-) H- m .- i wx .o r- -i %.o,Q -iH 0 'o'(3- in 0C0 0c 0)~~I o 04 ý4'0 E- in : ý4 E-4ulC:3 v.0) CJ)4E (TI H H ~ ~n~ ~ E ~ (0 r-H( " m I-l in I'

HomoHH) C) C) C) in CD N C) C> C0 H (D HC) C) H inH C.HE H HHHH H H H H H HH H H H HHHH HH HH H H H HHH 0000000000000000 0000000000000000 r-40 r-4 -40 0'0 r- r 0'0 r o i a J-i 4)a ) C" r= -4 . N E V e-n C,4N-4 C C0 mn MoC> 4 UO 0 0 o omm o 0C-4CUOO J f enooC>D MM > A mCC m m r-enr->- O UNDaQU-W.04 OH7 X XO Q U-4a U

UU4U 4 4 oc4 4 4 ý4 mUm4 ()UUUUUUUULUU()U -4 CooJUUUUUUUUUU U4-

UUUUIUU .)U LUCUUUUIr4UUUUULUUUUUUULU

guuuuuuuuuovuuuu-4uu

ooooo oo oo o 4 ý:8888on :D 0000000,-AZ:)-4 rC,

~~~~~~ULU L) UIL U U )U )riU U U a 4 aI 0 0 I I 0 I I 0 I-4 I 0 1

U U Z:)U U L) U LU

~~UUU-4 UUUUUU)UUUUUU UU UUUULUUUU UUOUUUUUUUUUUUUUUUUU-i

C u uuuuu u ::u = u U -i t 00000000000 000004 UUUUUUUUUUUUUUU~ 000000l0~ ~ - UUUUUU ~~C")UUUUUUUu U 000000 UU 000,00-4 - CO000o 00

UUUUUUUUUUUUUUU0- DU U U UU U O O UUUUUUUUUUUUUUUZ-4UUUUUUUUUUUU=UMi UUUUUUUUUUUUUUUiUUUUUUUUUUUUUU.-

00 0 00zr-Iu 0 U:4440 MI :D Z'3 :3D n :>U U UUUUUUUUUUUUUUU.-40UUU00UUUUUUUUUUUUi W- UUUUUUUC UUUUUUUIq =n=nDnUD M: :n=4U

UUUUUUUUouuUuUUU-4 UUUUUUUUUUUUUUU

4000000): 000000o o ' -4 4 <4444 UUU UU UU UU rU-i UU U U UU U U moom22222n=6 2 22 r-I 8888u88u888888880uoouoooouuuouuuu,4 888888888888 2 424 0 V u uUU )UUUUU u u uu L -4 ,,,,,,,,,,,,,- I 0IoUUUUUUUUUUUUuu UUUUUUUUUUUUUUUr-4 u

UUUUUUUUUUUUUU-i888888888886 UUUUUUUUUUUUUUUi88888888 o Z0) r a en u u -1 V-' m N e mn m 4C'D.-n

Z-)=)= )D )Z ..-UU0 4 UU000000004 U 0 0000000000000000

UUUUUUUUUUUUUUU0-4f U U U UUUUUUUUUUUUUUU4 00000004 4 44 4 44i4 0000-4 4J4* 41 r4 :D cnmC << m r r- r-Hvr4>1 ý40 4Jr= rý mnr- C)C) 0 00 MM ,->1ý -

mm24 z3mo MM m04D4 v a m o m > 4) TI a b h M U O W e a a K M M~ r-C" ~MVLO N - LU o -4N n L t r D h D ic,4m cn% Wr MM ýe0>4UC rEEý234 rIT4 r- r- 1-4ýý 14 -4 r-4

v-c o r- oo-o-4o -4o ro o- v4w4o o vo 4 , ric4o o- ov -oo o-i-4rf o o4r4o C> 40 C>Ce C e 0 ýCl(em 4 mý >C>C m Cm D >C C > )C>C C ) ) Cmý( ) >C [a0OO DO D0 DC 0C 0 DO 0O DO %4 mONO Y mO N0 %0 Y ) o 0 CN,40 : 4- N-o o0mmm C ) moMMr UOO1ý 0 o 0mm 0 .: m oo mmr U E-A moo N ) 0 Onvc- mao-iw OtIEmZE-co 0o F: L

UUUUUUUUUUUUUUUM

UUUUUUUUUUUUUUUiOBOOO000000O000m 0)wHxDOODOOOODODOOODO 04<< C:C4 4UUUUUUUUUUUUUUUM OZ4<

UUUUUUUUUUUUUUU. UUUUUUUUUUUUUUU-4 UUUUUUUUUUUUUU 000000000000000 00000000000000- 000000000000000 uouououuuououua ouuouououuuouou UUUUUUUUuoUuu-i UU UUC C U U UUC)U U U UU UU 0OO 0O'H00000000

< I IcL :C5 I I4 r< I I I I I 1

,:Cxc F:4F 4 04 04 xt 4 FC - 0000000000000 00

UOOUOUOUOUOUUUUU 000000000000000UUUU DU 888888888888888~ nU nU U UU 000cUUUUUUUUUUUUUUUrH Oct<0000-0r-i

OOOOOOOOOOOOUUUUUUUUUU)UUUUUUUUUU UUUUUUUUUUUUUUUMDDODODODOODDDDH COOOOOOOOOOOOOCH 000000000000000

UU UU UUU UUU UU U U

DODOODDOODODODOOUUUUUUUUUUUUU 9 0: c <

OCCX 0000KC oo< n444 44 04444 4

U 4 4 0044 0 00000 UU UU U UUUU U UUU- 4~~~~2 Oc 0: -c

0)0 01) E N- 0 ommo1 ao'o -lUO 41 o 0mm 0 moo mmN4 z 4 .mt moo 0 mm (In0a0 0000-40 O am ma.-wN 0)0r) O) oo IU )Coo1o

o. ý rooE-oooooE- ooor4P4P o o 4oc %-H t3 )Ha n xý ~ C-I C>) mD(D L3 D N C) C) C0 C- C') m> D CD C) a) -H-- 000000000000000 CI)o 0 U.) 0 N4 C m E r- m cq -4 l 0 Dmm L oI)O-mm 0 I o0-1 0 4J -1 moo mm r- >,I 4 - moo. mmr- ,4

O 2ý a4w a4wHH Wr a Me M M OlA Xa

00 0000 0 000000000 u uuu 0 000000400 O 00 0000

U 0uouuouu 000000000000000 0 08=0 8888

0 U UU000000000 U Uu U0UU

0 U00000000 U 88 8888

U00000000000000U00 000000 u0 U 0 0004<00 0cýP

: 0 00 0000 000000000000000-4 U U UUUU

0000000000 ::)444 :: 44ý:00000 0 1 uU0 00U 0000U C U U

=0:40: n:8#Uc9U :=n4< 0. xD UUt

U D OD0000000 0 M =)UUUDUUU U U Uu uC)UU II 000000000000I ~ l l i i 4 0 U 0000000U u0 UUU0 U

000000000000000uu-I00000000000DU 000000000 000000 UUUUUUUUUUUUUUUl uuu HM 'II MI 1110 1Il 0 0000000000 0000 11 CI 0 0000000 000000000000000I- < I<<< 0auuuooi 00000000 UUUUUUUUUUUUUU0000000440000000 U0 00400u0Uu0u0u

UO U UUU4 UUU404 0 l UUUUUUUU 0000000000U0000U-400000000000004004 1000000000 0000000 C

00 0 0 0 0 0 0 0 ouu000000uuuI100000000 - #:CZ#C , I C C t -44-cfýO4 Zi II'00000000 :)= n :)n = 0 =

000000000000000UUUU00 Unun oUU0 Z.4 0 UUUUUUUUU'0000000

000000000000000- 0 000000000 SI 004 00z 000000000000000000<- Ft FC 00 o 0 mo<

ID-- ID n0 00 =

uu u u Ou ID>. D 0 omHmEr4uu a'U uoO UO M Ub ~ 30 uuuuUOm0 0 Uuo UO0m M Usd 4iE 4 N- 4 mwm< oocPI mmc->-.-<.. 4c'4 04 Aoo: O N mam~ mmN>1uD H (DD 1 ) )C E~- (N-4~.N~~. ~ rcE'rm\ m14 mwa C m4 u-3 1-44-D4) M -4-14-4-4 -4mC)C r-)-4M4--4-4-4 -4H4-4-4-4--4-4-(-4-4-440--rAk -44r-I4 0-4-4-4-4-4--4-4w -4-i4-4-4-4r-4-4-4 -W4 ý4OO0 0 0Mr 0 ,1ý 0 0 Q 0M )4r 00000 - -0000-0 00ýýu a0000 OVIO (1) ca o,-i u 0 0U)0lz in3mn 0. N (a0 .1.1 a4 W EMlEU ý4 CU00 LV rAýE 01CC lFCUC aO

C0 000 0 0000o0 U DOU

000 000 000 0U 0U 0U UUU u

U 000 0UUU U()C

ouU OZ00 000 U 88 U U C):EU:00 U U 000 00 u D000 M =) Soo o,C, 00 0 0 0 UU 0 0 0 UUuO 00, UU 0 00 UUUUUU

U00E<<0

0 0 UUU U

8 f

UU

UU UU UUU UU UU00 UOU 000 00 000 DU:Dl EUEDDE 00 000

U ) 4JJU) o 4 U)c 004

U)-- S0 0mm o ,-I UO0 Co omin ONC-0 0 W .- moo m m N >, 14 0 ) moo mmr >0 Ov*o me.-m -4 .0 Q r-4 ý4UO E-4 11 3: E-4oo E2 0 -, DOHamu< ilýEUUu M00E 9uam(x zzc3 l ý4 rO mýu) 3: E- N r- cý o4 w u m o in mn 0)0U a m 0000000000000000 eO inini ) n in in inmn in ý rIn in" i in I'Di r-I r-iO r-U -A Cm CHA C C CDCDCCDCDC>Cmr-i r4I C r-i -,O 0000000000000000 C) > . w0 ") IlE "U)IrI lr11 I V r-rITImIV q c T- oi - o o o- rA 4r- rI - ri -4r4o- oý 00 -9 W C14NCNC C0 mmo --tot H

V , 4.H- 00- r 00i-i.L i- (C)f-4 F-4 Un an co(1)H wC) 0 4J x mxHHAEL. EL0 U Um

Co 00 000CC r-1m..HXX8-- -0 4J -44 N0 1 rOoE-4 1-4,-44 F4~Lfl n .- 4 rH m >,) >,xýc U U ta) ra-H -H x

0r- E-'C Pt it,It aýaL =xU ()IU-LID~ICID'CUI U U r-O I: (NI U - FUI,(-

U ) IUU U E-U U U OCUr- U 'U 018,9,9,9,E- 14E D4 1E4E E4 P L3

U 04iU U ,U4404 It U iE-i I C>i-U U -I- E U

U 'U IU U Iu I

0 - o E--'Cu t, 'F-' ,-4 UH'-I

U iUU 'U IiEi U i i EIU > Q0 0E- 10 E- sE UIUIUIUIUIU u I U I i ý4- U U U U iU I *0 U< 00080oupyUUUU 1 0 1 1 1 1 U Q E- 4 ON UIUIUiUIUI a) U H U aU 1 a0144C 'U) a xU a 'U< )a

u U I~ U IF--'I U U toU U I 0202020203

U H IHIU I-I F-' U EU IF- itE-' i u 'HE-.I CoD Utooa) o"4a UL-- UEn UU13 a4 UUUM aUa

CýM oUlUIUIUIUIZ U i:U 4.)0 r U Q) 1 E- 13.E--4 '- E14F4 F-'F- IF-I E-' 00 0 0 0 0 0 00 0 0 000000000000000000 DI~C)l CD P U- EU F. IUE-. I F-' U > UE U E-' r- E-.Ell (At -iE i E" Ci E- i1 C) , ý)CV)-4UIUUluUIU4 Uri x U U-A U- C)0 C)4)41 0-I- 0IONCYr - 3)a rNA -f L ý1 r r Q1 N - rc r5EýEL -)rI - O >0 > ý ýýU M M -I- QOOC) Hý ,In( XXU0 4JL) -)U r r0M .0 ~ X U ) I ui I() oI C)i 'Xý 24I E : E n1 ý 4 W' C:1 ' r, -rLnkr D h(D- "() num r 0 0 C r- EU-4 1U 4 ! 1uiU -4 -1U 4Li 1 U- -4 U 14 r- --irirIrU I Ar UAI4ý14 ý, ,!- I U ýI, EI)a)U H.0 iE-E- E C)C)C)(C C ) ) )C CU(DC UC D U))CDC)C rýr r4 -,!ý! - - -I -Iq l -U I- UAri U I ý U U IU I'U iE rug* 4-1 a LOU) u N< c I tN E- - Em-4 a)-IE4 0 t *H a)I'o) i )a) a)a~~i~ )~ U UJJ U L0 C) a) a a) -UU H c : ; F N 0N 'u Lf ,>Xco) u-- (-n mn.~.~cwU Ena) -HAI 004 U-to oo

I> *H0'(iSN ~ --a) u 4 oC a 40 r-iN.0X N N -4NI z- Ln in r-(14 UU CMC) a) a C u 4ELI)W C)

E O ) -"v-P(-4 " T M r M M0 (D

L9 0 O1C0I *HC) tOo 0OH ELU L) 00 00 -4-,4 x .- I r-4 N CICO00ý Q C) I ..) i -4 - I 10%t - r-IN w w -wOq ciai IN N -4rWMW I 4f- N 4OQ N N VO V E- E-' m LA ýi-4~-w w >, >4.0. U-u oU m id- -4 LAnLA -4 H1 H a4 a4MCM = taU U 4 -) -UCNCV U ZN ZtutM14 1 L) U LJ En EnaC)

U I uIU I iE-4

UI IttBtGI 'U I C E-"C

Ut IUIUIUI E44 ý'U ML' >4 U BUF E- - 104

>04 I4 to to U10 a10 I-)C. u I EI 11 1111 II 12 4- WL 0:1 W 04I .t U i EII I

II IUIII III 1 0111IIIItC.I ) E- >

I~ I I E-i i E4-,, >UIIL I I 010U I

IE-,, I ,,. L IE U a4 I E- U E-UI U I 11I U u l U I I EI I tI il11ll I E-,UUI I cIN 14 IN C4N C%4r4 IN4IN I4N 4 IN I4" C14 C14I4 N N N Ut tQ b a t B>>BP>H u•14 0 1 1 U I It, UE-' IIUII E-i E-' IIIE-4 0I I t~ I .llI l

P 040 '0'

'E0 tI >4 0' UE-4UiU-U11 IIOIII U- UtU Ut U tI tU tU b lE- IE- I ul 5 1 I 11 It, I

I I I I I CUC,

D I I U EUU10 E4U HI UIUIUIUIOI

E- E- H >E-4>E-4

U- 'U U.1E 3

00000000000000

U UE-4S00 ~ E-

O0 IOU)~C CDNII 00NN M X0tu00Ia N N ININ000 N N 4) >4** -44~-4000L O -W 0 ýNN d1 0Co mm -4 X X tu r-I -A U -) -4 I C) C>C o 4- 4J -H -,1 0% CN N- ) ) IRV"IV O WO 1t 14 N- N- .0 . N N UG0 >1, Iol ON 0% C14C414 :>,..C: U U to (a -,H -, IVtC4l H H A a En En = 0VU -4 1LA U N CD141 a U U4J.J4-UU 0 z aU 0-4• En n 14a4a4a4fr4 m .- t~~~ DN~ mNm~Lat 0 -i N m v LA wD Ný wD m0 v-4 v-4 r-( 4 v-4 r- -4 .-t v-4 v-( C'4 r-4 -4-4 -4-ý4-4-r4-4 -4 -4-N En-40 V-4 r-4 W-4r-4 V-4 r-4 V-4 r-4 -4 9-4 -4 -4 -4 r-4 W-4V-4 -4 -4 r-4 r- -4-4rl-4 -4-4i--4-4-4 -4 -l4-T- 4-T-14-4-4V-4V-4 -4 000)a D00C>000000000000000 (Da D C CA (n m m ml m mnm m en m m m m m m m m mým mn OH 1J)U) (J)C) C :CNNC oN (DCo m -1--H Hec ocul ý -H.HX X -I,-1moo 11 -HEl) -H E) IJ X VVHHCI)I C1) "\ (N .- 1 AJ 4-) -4 .H HH-I 000C40 rO 1 >1 a: HqE- "I Ul i Omm 00 >1 -140-HCYC >1 ýa U U (a R -H- )a-H- a4 a4 MEjx U 0u J Lna0) (1 a4r4 n U U JJ U U r- 1- r0 M

U -, U UW UE U E- < < I I

UHI HIt) U H U. i U I C) L E -• a : I I a I 0 i i i E U I U I U I U 4UU U0U 14U10 4 H I 01010000000 10 10 I U0 IUFUIIUlUiIPIBI 'H IHI H UUI~iIal~il p E-I u .. I:i Iit i Ii

ZI Z I • 4I F , iZ ,

P IbE- E-• Ip 0000000000010 1010101 H -HU I U CH0 E-01000101 E- E-. Pa P 1 Oc0- 'H 0 0 U) UE-~ I al 04H OU'a l CE-) Hu 101010101I - I - I ý I 010 01011 S - 0uIH IHuau' a' 01011UI E10 I H>PHH>H>H> 01a:10I018 U H- H >'•1010 H'a 101 H 10 'H H101 0000000000 00808T8T8(U E-J i a 101010 'a: U>HZU>U>U> - i it)U(. U U HUIHulIPiIa:0IUIPi IUIi 4101819191UIUIHIOIE, F-4o a: 010101014 <: a41IUDUIOd - <: I

Hi~li~ilit S1010 i HIUIUIa

p I (-) U U) U OctU I 0 I 010 IU urbiVOUiPlullHIHIHIHIHI

U0108 H U 'IU101'Hi 0' F9 t9 (E1

UIUIUIHIHI HU 1010.10 1 0 4HP0H 1U0 U 0 1 1 U0 10 1

U a U a4 U a4Ua: r,)a

U128121 F: UH1UIUrIU I U I I U U I 010101010004000000 121 F4 1: 1 F)4 11 1:It 010101010I 4 1010 'H a: P IU lU I I U I 410E 'Cu W, 4 W ,

010101 U I 101U IU Ia UIU:IUIHIHIH>H><>H>H> SH 10101H4

CEU0o000000000 1 0 E14 l Eo.u(9 Eu,10 1 l< < Uc:ý B II1 uE E- uu- C

WUEn NN" 4 T4-0 c = )mm-1 -HXX7.CUU - r-4 00%m ml) EJl)Ello-H VHmm-H-Hw.- l- aO "4U000E MX U U 4-3 EnM)L

CD C> CCm> D NýC>1 CD C) .) C) CD CmC N)C) OD 0>(

000000000000000000004 00000000000000000000 mm m m m0mmmmmmmmmmm 0 -H 0o INCl CNI N N L9 W 00 m-H -H X CD0000 IN C) C )4-) ~J-4-)-0J 0) CoN N-r Co Co o" 00 u OMfU ' W I-I4m--Y AAnu-) II W4.r

UIUlUIUIUI 0 )-, w HEn -Tll 'ITININI (14CN 00000000000 H-U I U U I E I OcE- <19 010101f<

S010101 ) 01010100302070?03 H I 0 10 10 '0 0 I 8101010l1l 10~~~E- ~LnL)ri- ~rO E, Do ý ,c U (o ro -i- x HctE- If 0: I4 . I UEU IU I I 0101010101

.0 Hi0 l Hui l 101E- u u U- i U E, F- - E,-iI Ol•l•l•l•lI2 E-4L U01 luH I>,,,F l(-)IU , I O0A IUIUI U• lU U'OI• 01t,1010101 0110101101

01EU>U>U>U>0> H I, H 00101010 I - 'Oi 00H 1l010 HOI Uý E E E4 34E141 0c,80000 HIOIbUIUIi. H 0I 10 0 I0 I P-• E-• P P- P

000000000000000 F -IH-,, OllIOi0lUI I I I 01 - I UIU~Ibl0 HI

H I 010100101I ,1 i H I I H:I H H,:: H:• Hi•: I H>H>H>HiH>,

01010141 1 u P u0000 101014E IE4P 4 iI. 8 4 E >I E,_ E.. > Eý > E'- >

U I • i •UI I

010101101H>U Hcý> H-> H > 0101 H- I H p 01010101I-. i . EiE'HI I< up 01C1I< u9 I iF I 010101014Hi E, 0 10E- E, I UIUIUIUIUIu o P E ' 0110101101 010101010 I 0 U I00U ~ r HI 010 H I H

Blblblblbl010 '0 I It, 0 'o'o'o UI 0000000000 01110181 0101010101UH 00 0H- UI~1~IUIbI 00000000 OUIUIIUIHI 0101010101010101 H H0 H~IUUiI>I HHiHHIHxt101I

E-X X~11I

c- uU < )wU$:I OHc UNC)I c UU CA

0 0 0 0 0 0 0 SI 0 0 0 0 1 0 0 0 1C0 0 0 01 ( >:1-)1 - z z z1 E4 0 > 0I01 HC101 wL9 1 1 l 14 14 1 EýIU I 0 0 10 C>0 Cl(1) -1 f101 10101 01% " i i w 4 - -- 4 N 1 H 101010101 '-4~~~~rOEýL ~ ~nr4- ~ rdo0. ~P Nm 1,c ~~>N E-4oC E-'10 >ooa - ENC Iu 4) ~ AU)x~ Hx U4);H1 ,-l .H .8 1(,1) Hl 41- ý.1,a:(3r

Ur-IrI-A A -1r4- -10A q

r-4U T- - L- lT1rIýA A U-- Ulr N N C.~U CoC CC)CC)CCC>0C I CDC>CC)(C>< I lu U 'UC. Co 00 koko ko%1 o Dýok 10%110k %0I'D I'kD ko IN I Lf) ,-i - 'U'GHHm i 0ta4 U)Um U 4-O

00o0 000o0o0 000000oo N, Ný N> ND ND N, N N N N N ND ND N N N> N ND ND ND 0L9 .-10 0-,'• toil C'N N N -. -4 x x It C)4.-00 000M0J0 0 -4 - X X. 9Ma q 4 -4O r-4 .-l N0000NH -1.MM4 $4 t- 00 S b 0 'Mom 'ar&&LA rU O E E~- L 4 -4 w w > >Ih.c 4 i~ U m V 4 -j - §O 'V U 4Ji-i-iU U 9 'UM'U 0 123Hi4 A X V U Li 4 ý- 4a U 9.nn 3 E En w 4 ##M~mmz5Uwwm 13 A a m 4=

, 8 Li Lii1 8 I" ,I~I Li i Li'Li i • LL'4E-'tLi 44-r41 Li 1 Li 1 1ii iE-~i ~ iE- , Li wUUcl440 p 40x 0m U I IU I E-I IU I EII II I I I I 'Li L Li I I I41>1>4 >I I >000 I 1 U>1 >1 4 1 i -I I UtII I I i

Li Li3I Er- E-' E-' L

i ifU &E- U 1 ) I U I t IU " I I Iu Iu I I I IDID 11 0 1 1 01 E4I E - I4 U ' i I I IE I

ru~c uZ

u I u Iiu I U Li E-4

1 1 E- U)I ul)u Li 1 1-'4l '1'I >E-'~ EU 'EU- >E- >E- > E- > (i~ lU N E-1iE- I I

IE- iU IC i' IN• Li L i E-' Ei i E,IUE ~~ ~zuE ~ UC~ E ( f

i L i LLI UI u I L) I uI ~Li~Li~L1 8 1 E.a 8Li 518 4u~LE-'Li U w-L14 4-LL x-EI0 E-010Li4 41 I I E- E-'IU II ~ b A4 u 34 13 u 04 1 U io116r L) 13'l. U IL4ua III U4 'I I u IuI U I u I U IUL I u IL0 I :Au IL i i Ii I IUIUIUi F, 1 U4 1 FU 1 4 1 E-1Uu 4 1 Icx i co f iI) LILi Li u Of4~Li I LI I 'LI l4t Oc I 4 1I LI U IL I tiLI'LIQILi ElI- i IE u I I ' E- I I U U I L) E4 I E4I E I E- I L E- Ii U i ) E - 0 01I U 1t UL i iI --l u P LiU E4I IciI iI4IciI 0I0i EIUEn0 i nL)I E ['U ni UEl U iv ll 0E 0E_G= i I E4-1>I.E4 E1 : A4 E LiE- Li Li IE-14M: 1ý F :4:, E- E- a: E-IF• KF 0>1>E->1 I E- I F: 41 >, , U I U UI I I E- I ,, 0I u ý,Z u ý,f<4 0 ý,,9iiu 41 L)1 f 0 10 ,E,,0 ,r A A: I'I N'1,I r : ()c101 ~ a4 Ua 404 U4 U 04Ua4 U1LIU04 U14 I f I UI U I Q I U I L)I I UI >W- 101010ý0:iolI I u -4I E-4 U I 14 0 1 l I U ( iII E IUiF E Li I Ii LiU i IU SE-4 -4 iU i E-U E '- '- 1 E'<-I' 10 1 'U< 'I E4 'E I N -oi 1414tUIUI1IUOOOOOOOOOHUH0WUE4 1UIU1E-U4UPUE U)i E-i 1 Li1 1 E-4 u i I 10E- 1 01 91 9 : E41 P BUf< U I LUI UI B 14 1 4 4 4< I 4-IU 4 I1 IU I

mmN -H-to) to C11NC'4W- 4rC C-N CoW c'JN14rINI 01-i H04 W 0 0 00O 4J4J 13 -.4 -- 4 M~ M t- t- $3-.4 U''U' ( "4N r -.4 Co Co I' 14 (- tN Q £ 4l)4J dova m IW LA v-4 w-4 OD OD >I >1.9 Li Li U 10 9 '0' E-N- N L -4 MOW >1>X.t ULi. Li U -,A -1 0 0014 CU )AM UU4J4 U iLi r r W0ý-4 H a4 11 M M X U01-U 4J 4J U U 9 'U mm M ELALl. w mOOc 0: m- m 4e0 L '0aN Oc o~ 0 M-,4 v-4 ('4m' qw LA 'o t- wO 0m 0> v-4 N' m~ v LA w0 r- w m 0 ,-I r- .ir-I V-4ri ý -.I 4.- r4 -.4 "' -.4 .- i T-4 T-i ,-4 V-4 T-4 .-4 4 r-I rq

-4 r-4 -4 v-4 r-4 ý4 v-4 v-4 r-4 v-.4 -- v-4 r-4 r-4 .-4 r-4 v-1 -f T-4 v-4- Co 40cO C> DODO C>oo CO CODCODO C>o CO COo CODOODOODOCOO00OOD0 CO o CO COoo CO CO COo CD COo Co CD C> ~0%0% 0% al ON0m ON ON ON 01 ON ON ON 0h 0) Ch Om0% 0 OM-- UW ON , C> C> m m00 w0000t ti.J .-ai v-I 0000-4(> a 4-H -W .tv V .-4 Colo co o Al rtIt E-'E' u) Ln 4-r-i Co oD5,5,>.z .q 0 0 md md -.-i -i H4ELA 12 4) H HO 04 a (nEn x L)UJJ1 'o r-10 Z Id It

0 U I IU IU IUE-4I P I 141414EI-IM 10101051 E- E-4181 101 -1 E4 E- MU1 S1414 411 441P4> r-I EAU 101 010 UI 0010101 O0010 01 00 Di 409 10 01 00 U10 1E4 w 1 E409 001010101010101011II~~ I) I IU U I I fU I 1E04I E- E-I1 1 010iF IE- E- 1 E- lE-4 1 1 01 010I 1010141 1 IU 0101010101010'0" 0 0 1 0 U) 0 C 0 ' 101E-4 U IU I U 0 1 U1 101101

101 0 0 0 i0 Ix EIOI IE 1IU0 E101 I U 101 10 E-E-401 101 U I UIEU U 0I I U I 101 010I110101010lI E- I UE-. 1> > >1010101 O-E-> U0 IE-4 IE- I 101E4001010101010O E-4 E-4 H l4~ 101 0i ý0IW0' ' Oh'>-I E'U U Oi U U U U IU I I 01UU010 IU'101 I UI 44g 4E-E I IEE-' z z z 0 w 04Q 0 W 0 1 10A 010 0 101 1 Z,4I L4 I U4 UI 1I t-i i1 U I W I 0 IE-' I10 1 0 I l0I0I1l1l00I14tb40 U CI 1L0 101014114 1 U IU IU 0I8I8l8I0I I8'8I 0101 iE- 1U0 E- E4 0 E-40 E-I E-4

S IiIB 1010 101 IU 1 lt I I l~l~l tU 10 U 10 > 40HU >11011 E > 0101010101a101

E-4 101IIU I 101 I U rlI u U U IU U I~ SU 101101 0 10101 0l101C l1 101 8 p U 6i8 E-86 484 1 II U 101 E-4 01 E-' E- O E-' -E' ol I ol a ,1 1 E-- 1 ý1 1 SU

U 14ILr 0011 1 1 14 01010111010100 o i00, Im C~ E-0 UE 0 U E-4UE- 0 HP 4 U 6 1 01 ý ý : 1 1 8 -41

10 11~ a4 I a S10101 ~ 10 10101ID 4 14 1 IU I UI 101 10100. 101 HUlHL)l E-4 101 W, 10 aU 1 pU 0 1010 1i

101

H P' WE- W E-4 [IE-1 -- E E-'I0110101010101 U -U IE- > - >E-4 >- 0101010101001010101010- E- > -4 I F-' I U- I u- I 0 1010 101 C II-I1 I I I r N I IN U10 U1010 U U UPP101IF' I O IO~ 1 FW14 I 1 0.1U 101 E E2 Z 20.02 M M BL) U DU I U I r I I I I E-14 I I E- UI '- I 0 0 129 0 0U

-1300 E-L, ý I4 ý IL4E,, u 0E 1 I I-) I DBI U I E, 1 E- I E4 I I ' I > > > >U 4 > 24 4 0 101

I 1r-14 1 4JId r- a t) 4 14q

T1rI 40C Da41 "U-4 M 1 ýr 4 1i UV 4)4 400C -r-i-v-I x x q -s' 0 - .4 N N 10 -vi IN FO E•H LA LA ,-4,oi wO w 0 -. H• LA LA v-IC>SCmO ) C War a4a4 En u u41 J-l0U UZ t t Z MlE94 0 3: a4 4a4 C4Or44 m E14El. En Ef E 0 4 4 4 m m 040 - Id m- wN r- wd mA 4DNC0 W-4 (In -W L%0md - wOO 0) T-i IN m~ Id LA k r-0N CO aI 0 i IN m -wi LA w. N- CO m 0 12O> rh 0 r4 r4Iw-4 v-4 v- v-4 v- i v-4 vi I4 r-4-4 ý-4 r- r-ir-4-4 T-iv-i v-I r-i ,4 v-i vi v-i r-iv-i INý v-I v-ir-4 v- v4ir4i v-i vi T-4 vi v-I v-i v-Iv-I v-4 v-i r-4v-4 0000000000000000000vI r-4 00000000000000000000 v-I vi v-4 vi v q-1v-i v4 vi v-Iv-4 v-i v -Iv-i r-4vi r-4v-4 v-i v4I v v-4v-i v-Iv-I v-iT-4 v- v-4 v-iv-i r-iýv- v-I r- r-4v-I viv- vi T-4 v-i V4 vi r4 vi -4 vi vi v-I vi v-ivi r-4vi r4 r4 -4 r-4 n uaw-H o o -H -H X .LJ00-< 4.J-,A m m o 0 - 0 0 a nU) ý 0- 01OD00 >1£ F-. i MN (.H a: oUUMJ x x (TIU1mmcoo H E.

C9( 1919( IH UH H 1( IP 10 1E I U)i1 > U)H .1 1 Uu N U NQz m C- I0101 UI 101 rl I U I C- I ui I (9(IUIUIUI(9l(9I( I(9I(9 1( OIOIE-I8IIOilIBKIOiK U E191E i E P EtE- i i BE- i E-4(9 (9 P(9 (9 E0

I HIHIE- I 10 14 I

U4 Uipi101 0I0 0 1 1 0000000000UlglbibrBI 0 Ip I p pI UBUPUBUBUB(9191(9191(91 E- I(9 E-I ElHi 01010101010 10101 0I I1 (9 I I - 1,E D I H149IH 1(9101 I- I U0404040404 0000000000u u I U U H> H> E- H H .. IQ Q.1 0IUI•IUIBI ~ C1 (9 0'010101101 H ' 'E- iE- E- I

HIHIHi(9I IHIHI -IHIHI I I IP HU U1 UIUIUIUIBI

(E(E- >1UH(H0 9(H(E.9E. H 1(91( H 1(1U 4 ~I{ I 04 1 14 1 14 1 I 00100010'0 O H I~ ( >I(9I( P> H>QQQ0UPl I( > I 9101011 jI1 i1 1 1 Eý(91(1 -1(914Eý iOi (9HU9H9H(9lUIH 1(90 1( 'HI UIUIUIUiPi (91(91C40 wE-0 '(91ol 0 1 14 18 1 1 (1 (91 1 (18 1 :01 KS HP>#E-4 (D000000000000000 > U 0 U MU 4 U xE U '10 10 1 10 01911011(91(91(91(9110O UO: 0; (11)F 4 l4 11"k p CA~~~o( 8 m 8313 1 E- i E- iF-4 E-i (910

'OE-. 0101 1010 4141 10101 1010 OlGlOlolol 00000000 0000 0 0 H-1 1 1( U I 4 14 14~~~1041 1 1 1 1 1 O4 14 1 I 100Esý Oi I U IU I lHI I 00 00000000000000AR 4 I U1 E-1 E-,I> IU U U U ES Ol4l0l0l4l Iý : ~ -U EU U- U EB 91 b 8 u U U

Eý 1, I E .'F-I Iý U I1 Ii10U I i E-4 8 E,> E->ulus

1010 (9 (91t 1 uI UI I 101 10 I U I U I u 4 KS1(91(9 f, PE, f3 p 14Ep 2 U 19 14 UPBU 00~1 -)m '< 010 mo10 0U KSI K O0o oi'<- 0 o: lol w C u U2KUS0 E < KE- 114 10

4

0 0 00 (N ( (,4 N N *H-H X .HH l o0 mH--XX •'0.-H.H C) 4J 4J r-4 0C1N N 0000000000000000C) C)01 (NN.1. U) U)- xI-i- -N;T 1-4N'C) N0 000 (N 'cH-1 -. Ci) CD C-01)C >.1A->.1 -H (9 (9rr r-l() H - in in ,-I U0 E)-4F4 U>1 U>1. C14 <.l 13. 4L.En UD LiU)U to 049I0-1:~ 0H() a ) "DN 0.1 01 I 0,H (N l r-U) . r- N N101 tulU-,.- 0 cooooooooo.o-oooooo ,H (N m I'mm. N-- Ct)010 NH(N m I'D r-N COz01m. rH -4 .HI .41H 1-1-4 -A ,-H rH 1H (N ZH XH LHLH.H14 MH MH . MH MH 4 aH a4 $.H

r-4 r-I .- I r-H .Hr-I .H .H .-4 .H -4 ,HI r- rAH v-A.HI r- -4 v-A. 00000000000000000000 oe0(00 .,-H (N (N (Nq(N (N (N (N (N (N (N (N N ( (N ( (N (N (N (N (N mý ml mv flm m 1 m m M m m mn m m m (ý ml m ml m 12O1.U)>.a44-1 .- .- 1-4 -. 4 . l -A. ,-1 -4 r- -.-4 . .--4 4 -1 -4 r-,- A .4 H -4 .- 1 .H .H.Hr-H -4 .H .Hr-4 .-4 .H r-H ,-1 .HA.H -1 .HHrA r-4 -1- 00 co mm X~w o 0 N N m -EH(AEll_1 x x 'cl, t-T. .-4 '1 ,..,N 00 CooC> 1 Hm 0 0 900 -4-- 00N C 0C) 4- J -4 -1 0'O\NC N El) E)w (01%Cl C14CN ,- A En) El)ý4 N4 r- N- .0 N N .-4E-4 -V --: mON 000300H"-4 -14 w] w) Ii 4O.NNGJWý4NrNr Q N N rO ro AE A -i r-i Zw>>U U MI Ea w H -H '0 '0 E-i E-4LA W r-i r coOD i>, >i.C £.( U U ra M - *H " Pl)E n4IQ0-CUWW U U.J UU C liEl

E U U - 0 E-. i F- 0 E- U u u 0 < U F- U U - U -< - E- F- 0 U UEUEUEUE 0 0 0 0

SU [3 Uu uUC E-~ E-~ f U E-' 0 U EU - E-4 ' E U E& EU 41O l 1010:1 1 011 O E- E40 0 0000 c 09 0t9F F- F- 03- 08 3F-UF-F- < [3 o I4 Q 1:gcl<12 U eC U0 u EU C u u [. UU

0 ~E 0 F 00 UE S( F 0 E-

U F 0 U U E - 0 0 : Q< 9 -'0 0 uF- U 0 0 00 F-

U 0 - U UE < 9U U 9

,-4 E I0 EU'

0UE- EU1010UE- U E-. >E- I1 - 0 U S E,08 C 0 C <0 E• [9 L9 E U , 0 UEu u UE EEE -4© UE U U) U9UID :FI U iE- O• 8 L9 i 9 I[9 E-L1 L4 -E 2 (D L4E-O 4 E-I - EU EU EU u u9 , xC U U EU '- 0 U 0 EU 0

U EU IU I CU I

D m m-4 H s-4 H - -T -I' H -4 > C C)C. -) 4-a)1 0)r-r 0)a U EUU EU E U E UE rO-cE- N-4 OLnm E N vq ýioo- 1o >U C >,i.aur)mn UEUF 44 U - -14 ý EU0H-N N ElUF-U,,U,,U,,UF- C a4 4C1 XlfMM3 ~ ~CWWa U 4E-Ui ELa4aI~ U i o E U E-u 0 U 0 1•0 • r-4~8 n% -m m >n-T i" m-,n%,0 1 10 -mci>8 C)$ý9O 0U< 0:4WUoOf O r-4Pr- lr U ---UE-- HEF- - ~ Ar4bE- u<<C~ DC)CC>C 10)CCDDC.CC)C 0 8 8 OEUEFIOEOEG 'IV0 -. 1 r- I~ Uh.'rI0Z.ý 0VqU U)UEL0 ; 0 0T- OEOjE{MO 01010l p010U IU14 EU- lIEU4 E E14 P I [1 r-riv- 4 - -I rf - i - E--4 I ý I E-l iri 0 ýl 0zE-iel - E-E3 U-IUOUOEOE 01 cy8 c1 c ~U EU EU EUEU EU000 00EU EU EU0914 E IF I- gF-- F- U1) U1)UI1U U- F- F-' -' F ~01EUE10E 0 018 EUuE U- EU4 EU E-4E

rlEU EU I F- I 0000000000 U- U (-)iE-E f< OE>EOEOE0> 00000000

004 EHX-4i l--4LL -

0000 *HHEO'N- EDEl) '0' F-LALA~-4,-E ( E - E-4 ~UE-> U Eli-H- IC)~ I ~ 0U U~~ ~ U JJ UU 9 1-)iEl

ED>' N N aX H -HEn U- Do-J rsr E-4 E- r4 '0't--A CE) CE] El) l) H-4Q U Us- OuE/)-4 U)U04

r-$ ( m ýr LA 'XE N- M (h 0D -4 "- M ,T LA %,EN- M) M 0D

-44444---- -4 -A -I r-4-4- -4 -I -I -A (A 0 0 -H LA LA LA LA LA LA LA LA LA LA) LA LA) LA LA LA LA LA LA LA LA 04 4- r-4 -4 -4 -I 4--4 -4-4 -4 -4 -4 -4-I-4 -4 -4 -i -4 -4r-4 A-iA" jO o 0 00 m m X 00 ooH 0000.-H mmN H00oo 90 C0 -1:1 -IT 0) "N C' I-JO NN "e a wo mo o 04 U n - H w) > > .-H .-H 44 4M M 304aUU m>>-•.00- rz 0 OH I-i0U) w0 H-N -HAN .,-Iq I-i( " 0"UA CA 0 H .--H (Ou OOHHU)~mI-JI-UU 0 -H> - > (uA' O4o 4-4 4-4 (A U)

0C-i o* UI. IU I U I 0c0 C>U) UIUIUIHIt)' 0 10 u l 1 I u Hý 'UI U 'H-I I E-' i I UIUIUIHIUIH 'E-. 'H E " 0

'U EE- E_- E- E-H 1 14 14 1 1 UUIUICDIIHlIH I U..H'IH'IH- U -' I rHU-, I UHUHUHU1U1 ' I I C I E' I -- I E E 0 U) En ) En . E- 4HHu U ' U 'E I P- u u _ 1

E-I i E-i E-i E-I i -i I ,F- ,E:. I H(:- E.I (,: 4.I[.) F rH4 PF © U-.

0 -1U --- o U)%SO ) ,ii uI E- E-E-• 'E E-9 EE-+ -'P i o X CD 'U It EUE Eti I I t-4 ::I . I *H ' E-4 'H E-' It) I -H 0.01,...-H UHUHUIUIUH >0 H P-. E-, E-,-. E-,-- u-< b-iFCiE " i 'U (.- i i UU2U2UEU(U( IUE I -U p ~c I P i-i t -00 H A- •. E-+ i -+ -(4 0 UIFG, E-+ I I I U m ItA-i 0H~ H UIU 14 l4'U UP DI i i i MA4 -4 E- E 134 1 0U u'U u Hý i F0' ' UUIU 10 'C' 0 ' I~iU0 U I , 4-i F

PIUIU) CDuU)::1: 'CD'UI u p 'H U CDUE- BHE- PIH 'H P > H E-4'HH I CDUEýUJ i'U i UI'U• UI 'HJ E'F U H•-E.,I F I

'H 'U 'H 'F UH'H'H'H'H' ' U ' '•H 'H' Q E- ' EE-'H H>UHUHUHUHU'U'HIHIH' BlUIUII~I0 1 0 1O10 UO FCC1 U 'U '-U E-HU -E ON' UHUHUHUHUH E-,U'UIUIU'UI F... F- P.->--- P--->- -- H'H'CDOu UHHHHHHHHHH H H- 'H 'H UIUIU UI8' EH E- 'H E- 0(1A E-U'H'UIHIH'U-H'HUsIHU'U 14 F-1 F04E F 14EE- 0 144U I1 u 1u u CIUU4UE 1 < cYU H 44W FUF- E-i -iE, E - N U 1 1 1 1 1 E, F11 P > -1> > E > C100i 44I ý 1-1) 0) : : E U<- '11 'U H =U ,1 UHUHW44W HuxtH>HPHPH>>> : uH:.C)C/U'H' E- > ~ C U 1 -)14U aU r-)U2U12 0 91 U OO(1U t-.I u 'I IC I I u : U : ; .0 8H ig P8rU AI CD ' 0 u F E- E-1 -H OrdC),r, u u u ý - - q'i- U 'u H 'H EH

0 04 4-10., 0

000So 41E v 1..r:A

'(a Cl 0 'C 4i oB 0 m NoN m_ CNI r--t- (.,4C-N O 0 E4- O ID0>J 00 Xx C> ooý4 C ý mmXX-iDC- -q-C> ooN NIý - -H ,-E 0000- ooH c HH 000) ý4 ý '0 H m f S1H H C A-i AIN SOMNN OHHQ A En a4 r Ua OOHHQ.Q.U)mI-I-UUU 0 -H- X R : z 0 00cso -O0a U0 H C) I Ar-4 N- Ci 0 - N 1 C "0 0 H 4J > I-ULo -H 0 .. HD 00)D >0 N0 C ) H0 C) 0Z 0 C) N C C00 CI0 00000000000000000000 0 \D O ,- HHH H H H H H H H H H H H

-HAU a) 4 C)- (N N 0ý4 0 4 0NC00 00 m-XX 00Cý C) mmXX00UU~ , OOxxcýý A- CC0 0HC- -H C)C -4-4 0000HC D )0000 0I) q) .;r -;I. C O""r-4C4 -. 0I 0 N "- -I r -J WWNN UU > -H.,-H '0' E-oFH n in r-4 r-i mo ýCý U Uu EnjU) > > -H- -,-i Xiw MlHHZHH)UCH " O( m A UU H-H-Iw x2C fu Q,) ")0i (n EnU li J u)UU 0 0 -H -H 1a 0) 4- 4- M -: r 4 0 124 a

4: F4 i4 U 14 E(5I E<5 I4 F i ti <1<14:4414:1 F > E4> meHHQF> F> m >mouU > (_ O-4> F>HHH> E4> U U IU 'H i 4:4:1414:14:1 H

14 :14 0141:14: 11

HU 'HHU HiU Hc ICiH

4:14:141U iU iU U' UE-4Ui(iUi4: ý0 E-4 C40 E 11 1EU I(4 00F w 01 1 1) (U) S1 0 1I 0I < 1: 0:' 14 K4 I 0KC I 1 fI:l I

H iH 'H H 'Hi H, U 'H PU PUC'PU P U P- P UH I U H 'Ur1I F U Fb4:

UUiUIUIU'U'U'U'UIU'UI U 4:'( '4 l4:iH HIH E-4 (DI

HP ý'H PLý14 4 H 'HPEý 14 U 'U4I914u9E 4'H uE 14PiE-I 0H14P4 aH 141 la o<01 I 1 ~ I 8~ IU U B9 0U

U U U- H C: PHUUUs 'HHI u 'H I H' p(5ý l( I FUUuI H F '(5'H UiU: 11410 IUiUI L3 0 H 'E- H 'H H (5) ()i - (5)i U U U U 'U'U H IU E-4 P(5'(P'(5E-4 u u (u(F( 1P 5P FI IFIUI ~ H Hi0ig:( C E-4 000 HI20 181 IH' '44 U4:U4:U4:U4:U (5' ( '(5i5l(5 UIUIHIHIUIUIUIUiUIUIFi i~ E, 4 EU'P U UHUi '(5'ro'H'CC l E-4 H 'H 'U i'U U H 'H 'HP ' E 1 EE>lU > ELL.P1HIHIU'H'UIUIUIUIHI 14Ff4 -E14 P 14E 4 HI

UI:C :f :IUI H>)UHUHUHUH 'UI H (5io UFUIU'UIH'UHIHI 5 '(5'H'U'H PUIUUUI(54: (5

(0)i 14 1 '4: I(< 4: 4

u u(5 U U4 H 'H P 000000000000000000F 0 CEN x0 0 UIUI-ibmUIm 40-0 0 O0-H-H000000W Ui mU 0 b ib i l~ u lu l 0 0B C)X4 x$:4C) c Fil n xC)c)C4:Kt:4:4 ,! r-I -1 (D C) ) ) -- (D C: 0 ý ý4 (1 ror>PF-V 0)ON >nP>r- r4 ri A wr--Aý4 >4 uu - r AJAJWw> >W .,N - a " m = lj-) uu 00 --I a r 4-44-0 4 M M 3::'12W ' 4 a Z 12 1-IX 1 O- -wm n -z n;o c>-i"r-4i-dl r- r4) :rLnkDt-oD4 4 - - ý a - Cc

ý4rI ý l UIr4rI r r ýU iU IU UIiý 4ý Aý I4rfý C) C>P(D CDC DC D< DC C DC >C m r) r l ) r) n l)rl)(1 m m m i)m l)m E--4-4E-44E P 4 P - -4- 4C 0 0 0 000000000000000 -4E44444----- H -4 P-i P H E-4 H

'--Il- 0 0 e J 00 C> ~ ~ mmxxoo()r ~ C)~ a)C TH 0000 D -H -H 0 UZ - rO rO F FH m 4 0 ) )NN '0VHH)U)-w 4Cn wUU W > > -H -H W) 4a) F4H Q 4 En (A X X 4J U U 0 -H -H 0) ' XEEXm X444EM R Zra 4a4aA -i "O m~ -,z n ) N- mOCO 0D - " in qu) %o N- CO CO%0 -4 -4 q-4-4-4 -4 -4 -q T-4 rACO

-4 -4 i-1 -4-4-I -1-4 q-i 4-4 -1 -4 -4 -4 -I -i -i -i -4 0 - CD.o CD. C> COo CO CO o CO CO o CO COoo CO CO COo CO COo CO COo Co COo CO 04O A ~ 1-1OO4 N C- 00 mmXX0oUU ~ 00INC) u00 u 00 0000 J0 10 0 e r-H -4 00 0 0 1-. I4 ' -4 F-A C> C>CW> >,-I -41 0C) 0 0 >4 $4 (1) - ) -IT NP m 0) 'O 10 F- F-rn 00r 4) 4t)>> 11 H1 -4 "- ". m. m) x) x -L J1 u 0 0 -H- -H4Im 11 UUO 0 0

E-F - 4 >- p> E-4 >p p >

0-4 tO0101010 tO tO1

I I I I , I IF-tF- F I'E- tot SF--Et tF -I E-t Eýo iFtotutu1t u)t_ U14E110 a, I u) a4 t uut01 aU 1010 uu a, urF 04tO t u I Uto1toI

010>- E F>1 F 10to O~ to - tu Fý> - u .LIOF t, ,- t- tF1-1 u~~~~ ~ ()IL) ~ I u I ) ' t (j u ,cc~ -)Iu ~ -)~ I E iI~ E4 o tF- IF-Ft to tF- tO I F-IF-I

tdi O ~ 101010 1 I tiib FE- E-iU UIF-, -UF- F IIF I U) E-4I u - II u I U101- 10 IU I 40 I C)14U t~tI 1 0 1 0 1 l< ct FC u z c UU U.FE- E-U U iFE-4U to U U,t Ii UE 1F-- i >1 E-i I E- E-4 iI OI--U ~ E t 0 F-UM 0wF-IF-wto 0 tc P B 1 Oc o s u s 1 1 0101 tutu 101 0II1 ol

I L) ; j 1 r) I~U F- IF UtF- Q14IF ctQu 14 L IF uF- 14 IF1 411 t FIF-t0 f< C F- to F-U F-E, 10 t 0 U 1 E IF-IF-I < F tI I c~o I tU t-IUIOLI FIF- F -0 10101w0U~E to tutIF-r)P wuIF- - toý101 101w01C401WCIF- tO IýUEU u u0 u 0 IuI0 ) I u uI > >'F>F->t >F>> F> -'>F- I E g E 4 F 41 , 1 :I Er, ,<ý E tlýx4 I < I 14: I I I pI u0 0 u 0 0 8 0 > > >4 > > > > >4> >

E t 00000000000000000 itýi E- F- IF- I I 0 101 1 1 F- E-4F- KC1 4 BUO O ooc4 I 4,4FC4 L,4-F4-I<--~ ,L, ' EL,' 01UP -F-IP)IF>-F>a4 EO, E- E- - i E4 u i i u -) i U iE- E- I u8I U I u lIp Ix 8 EI0 ,U0:F- 0 U(90O UItU(9 UU0 (9 UU)I U (9t4t:t UI U UE0 I 0 FC 14 1 14 1 Q O4 J1 r) 4 Fý U ( U E E4> > >>E> > 818U 10 0 4 E40 E F, > E,> UmUGUGU EUU Eý> Eý E 1 : ý : 1 #0 1 4 Pc1 : p o41 :1 < I r-) I0< I u L)0 I tutu1E IE tOU IF IE IFtO IF-< IF- I E-H8 E-> >H E- H -HE4 4 H 4141 18 10 01 01 01 19

F-i UiFI Ui ~ 4U I IF-ý I E OF-tF-0tOItot0IF-IF-l01E- i0 F- ~ 0

0A000> E, >E, > E 000000HOJH41E-'> 0U OUttE

NN ('-C' 00 r14 (N OO E D OO mmXXooUU ) C C)m m .x .1x Co00 o - ' Coooo =-H - 00000 0 C') 0 0 --:r ID' m0)0 " " - -4'-tL NN-. Li WJJ))N N 000.-H -H 00 Oi w 1) N N ,'El,o LA r-i 0 w)in) > > -H "N 4H -1 14 -4 N N '0'0 F-4 F- LA '-4 C) co c0. LA) LA) 1-4 ,-4t2X0000 w) > > -H1 -H- 0.0E 0."- a-0 M x) X) LJ 4-1 0 0- 0 0 H- ItI) U)9Zmm Ja11. 0 -H-H-- 4 MM 4-4 - WDLNH 3:4"1114ZZ O 44 4-4 U) = 4-1 4-3 'I "N C) qv' LA %DtN- m m)010 (n"H Kv LA %.o r- wX 010 0 -A ,-Ir- r -I. -A ,-4 A -1 -4,-4 C" N 10 r-4 -i r ID LA '-1) N(0 0 0 QDU)A-) H ýi4 .-4 ýl-I-H r-I.- i .-ý-4 I r4-I .H .-4 ýq r-I v-1 ,-I .- I .-i r-4 .H OH 00000000000000000000 00000000000000000000 COO w I' D'w D'ID I' D'ID I' D'ID I' D' D'IDw ID LAL L L L L L L L L L L L L L L L L L L 0 0 40co00 r- m cmm N CV o JCo 00 X cooc 0:t•.•..-,0 0 >4 W -W-O0% N)" v- - J4J 0.) N N - -4 C C 4W1C 0 0 ) )NN rO'0E- E-4 LnUn V- ,C,0C==U U CO CO > > -4 -4 'ti E-4 E-' L L qr -4 w wD ._ .4 U U L1 Lq > > .-_ .,q " aHEn En 1A-)U U 0 0 -4- In S0H H a a w Mx = 41 4 UU 0 0 -j-H4 (a fa

S10 U 1 01 01 BI

I I I

U IL I4U0 UI11 2 1I U 14rU I I Irl u !MMI~~II I I 1 1E4g0I E-4 IL 0- FCI I0 L-4 0 1 11 1 U I E 4 oq I oc I U I E- I ILI I1 U010 0 E 4 I

~~~~ 'U U~ L- U It UI~i~i-~i ~ E- I~ I I IU I I uIt I I- EI It O I U UI I II I , 4I E-I L) 1 I U IU 1I I E-UE--•- i E-lu 1 0 I Ii EU• 1: 1 - 1 E-tI E- i E- E E-0 5E- Fi E 4

ua4 u # 04 a 04 u a4 u a4 u ua ua

IL) 10 i I I I-I I, U II 4 14IU0 1tU 1810 'U U 14 U II U0in 10 t0 E- l 0 1t E-, 1U totI IU1IUIE-,0 1 114uB ;u IP IB Iu 10 I -I 8 E8t8H8E-48H0., E-4 4 41 10 4 1 4 U1 14 0I

UIl I U I u I uI u I IBu ýx owl U.-4101 b10 b 1 U 0 410 1 f< udI 6u

E- I 0 -10 .10 10' 1 0 I u IU I 0 0 2Ito I0 E- 103U CL 11II I 1U IC. I U I U I U I U I

L)I I I I U I1u L I1 I I U ~ ~ ~ 4 ~14 c 1UU 4Ut 1Ut

ru EI U 1 l l U0U I I I I IL I I I I 1 1 I OC I I IU1 1 1 1l b 104 1

U 3 E-U10 0 U0' E-H0 U E-U E-

E- E-I 14L) E- 1i oi I

NN C'4 OO4 4 o o mm X o o C- - C'4" 0 x xoo 1 C1 0H 14-4 0000CC C -HI -,4 00C -0):t W 4J 2.0e00 ~1-14C>~ Q2l~ ~ m~ UU 0 w844 l- 1.4or-r- 5-.4 vv If aiý m C r 4N t J 4J- 0N N 10 '0 E-4E-' mn n r -iHD ww >)N 0 00)0 H H C4 a~() En M) X 4J 4J ) 0000 -11 *'1 In In w WHH aiommoo )-J)-UU W0) M XM4 M444uMuM 0)I-4 -OtNi4C n" l -O NC LM- %D4r- W CDO0M. -4 N' m' qz in %D- w mD 4o~ 1-1V-1 7-1 4- H -4 1-4 T-4 H -i (' r-4 -4 r-4 -4 T-4 -4 -4 -4 -4 -4 "-

-4 T-4 -4 -4 -4 -4 -4-4 -4 r-4 V-4 V-4r-4 -4 -4 V-4 -4 -4 r-4 -4 r-4 -4 -4 r-4-4 -4 -4 -I -I -4 r-4 -4 r-4-4 -4 -4 v-4 -I v-4 14 " 000000000)C>C)C>ClCD000000C>C))00000o C 000000000000000 C Cl00 0 0-,.-0 w w- wO %D w. '0 w. %D0w. %D%0 %-0%0 %D0%P0 %D %D0w. %D %D o oN - r-N roN N tN N N N N N N N r- 0N'-00 0N'00 00 mmXX00UU ~ 00 XX00oUU~ -1 0C>oooC> -- H 0000 r-I1-4 C)C>C>(D-AH-H0C0C)0 0 ý () () NN JMMJWWGNN '00 E-4E-~ Ll um -i -i CD cc) .. c: U U w] Ln > > -Hi-H ''0E-ý E- L Ln-fl A- oo oo c.0.0 U U w) u) > > -H--H 1)WH H "~ am m) x) x 41 -1 UU 0 0 -H--H )u r 00)H bXo 4444M4 amm ~m~ x) m£3x) 4-ýaýa ~41 u uU:1U 0 0 -H -H-> 0) 0)j M GHHc M 4

'U

E-

'U I C iE-E-' U IU '_9<

'ci 1U I E4 'U

E- 14EU 'UFCC

E- 'lC t7ý 1-o<

U 10 U 14

'C,:8 pcCU 10 'c U_9 C) 'c 'F-- C) 'UabU(9 'I'U0 'U 'U'U 'I'U'U 'U 'U 'UCU 'U 'U 'U rE-' I 'i 'c 'C, 'ci'U 'I 'U 'U 'ci Eý '(9 'c 'U 'U 'U 1-' 8 'U U0 'U

lU

IU 'U 'U 'a: 'ci 'a: 'F-- 'F-

0 0 NC'- (cq 0 0 )>. 00 mm~Xo0Ouu C~oC) n r c)(D u u>ýý U41 -4 (Co. 0 000C-H0C 0 0 ý4 4 as0000H-00000-4-~00 0)-)al-VH ) ("O--qr4 N414C41J - ý4r a)J a) N N JNJJ 0)JGNN NNH-1N-1O Ln OCCU()E 9'0-o 4-4 :r 04 a'UU0U4 >-a-4 E0-' F-' Lu Ll .- 4 .- 1 M CD .0.0 U U U) W] > > -,A -,A H a)H a) U3n x~ x. J u)uU 0 0 -H- -H- M) 0)

M~ w']mLfl ý) r- WDC 0)C H C'-i M ' UlC) C'- OD (3) 0 -H 4r- -4H -4 -4 -4-4 r- 4 r-I,H .-4 -4 .H .HI -4 H.H, VH. '

0-40-40-4T-A4--4-40 r-4ý4r--44-4 0-4r-4-4 -44--44--44--44--4-4 -44- -4 r--4- CDCODO000 DC DCD0 DCD0000 o DC 00 O00ooDC 0 0 DCD0 oo0 CD 0D 00 CDoCoC 00OOD00 CD C) ol C) 0- 0 a-. CDa oCo C>g C-.% C ) (C> C3 C).) C>nM C-. C> C> C>(7) 0-H-O coooooco••••ooo ooo INNNI~ 0 0 00 mmXXoo0uU> V Kv Ch ON LALA CN CSIcoooCd.ooU -4-4 1-i1,4r- r- 4J~~LIOi) IDW 1W N N 4 1C4 4141414 0 aD aD En (AI x.0 =) 4-1 4 U U 0 0 .-l- -14 (a Im IDa ) P0H-E4 4 H14 C4o~1Ln " -itI U)v-M w4 X w X 4J 4~J4~.U U 0m) 0w *-l-> -H > (0-4 -H4MI

I , I IC, IS, 11

I I i lE i IIE- EI-ll I U U lE- 1 1U U :

I I lII I It I Ii I I I I I I I U a4U 4U a4U 14U 11 I I l l I I I I I I I I l I I III I I I I I I I= I I I I II I l I I iI It i l l I I I I II I I I UIC U I iU I I I I I I I I I I I I I I I i I I

Ua U U IU

I11111 I tilt It it lUtI t ilIII II 1111tilitti tltIR11t ~4Il liii tillt U U IU lillt 1i llil Ut liltI 111:11: il I I I II I I I I I I I I I Il ii it i l tl i III I I III I 1 1 1 1 1 1 I i IE I I I I It I I I I I lIII l I U I I l I I I I I I E-4 tE >4~lE'I- 11i 111111liE i IttI I llIII U tU U I i al lit till Il ii iIll I I t I I I i t I I I I I I I I I II E-' >i4 -I P' ýl l I I I l I I i I I I . t t t tII t t t I I C- U1 IE ICE-' i E-i till illIll l~l i I I I I li t I ii Iiii -l III lil ti 1111111111 ltilt 1111

I I I I I I II f U t I I I I I I Ill till, I IUI isi 1111 E-4

I I f I I I I i I U I I I I I I I II 11III1 tili UtiI ill It UE-U4 'cU i I I I I I II I II I U t I I I I t I I I UI 1 U , UI I- I I I 11111llill t III Ill it I I t I I I l lI I I I I It I I C blii I C~'zIPU' i I I I I I I I I I I t I t I I I I lit titI I I II I t tI tI I t I II l I I I I I I I I I I I I I I 111111ill IUI 11111lil l i I I I t I I I I U I t I I I I I I I1111111IlUI It tIllI I )14 L I U Ia U4 QUI1 E1 E-41i Ei 1EiE- -4 EU- i'C E E- OE-- C-'E-4 IIII11l111I'c11 IIIIII 'c~lc~l'~l~lcE-4 1 1C .1E E- .IUIC U I0'c IU I I I I i II I I I I Ii I i l l i t C U E t IC-' IC-< I I I I I Il l I I I I t I I I I I I It MUMMME E-M I I iI III t l I I I II I t i it I I I i l l l lI I lI I . I I lII I tII I t I I I I I I I c i I I I I I I I l I I I I I t I I I I I I I I I I I I I I l III i t t U t t t I t I I -1C1C11 U11aIUI IUI I i t I I I I I I I ~ I I I I I I I t tI ll It I II tt I t lit I IE-'IU E-I I I #CCU

1 Itlililt Ull l1Ii~t it E-R pE-18 1 11) 4 VI V U11o LI)V LO E- EI I 1 19 10 1 I I14 UA UM I II l I U- UE-E UE-U U 0 E-4 E U E4 - II I U I I I I I I I I .'C 'c 1 I 10 IC I 'c EnI c'U MtC ' EAC 5 111111l itt~ll t t ill ill U U U E4U U2)I U U, Ua4 IU 1 U I3

i i ti I i I cC i t I I I Ill i i i l lI lll II I I U I I I I I III it tIIIl itt c-Ic-i %> 0000-c4cI 000t 0%I I Istt lilit I u u E--EX ')I .' 0'r.IL0 4 u044 I I t I I It I I I I I4 I I I I t I I l I 1 ~ 10)1 E- 1 18 8 lIt till ill I ill 1111 l

11I111II1ItUl IIIIIII It lit Ililt 11111lItti -4--4- 44-44M r4-4-4--4-44-114 -I - r4 M-I -I r-f r-M I1iý0 4ýIý 1 Ir--4-4-4 1 4 r- r- 11I 11II III I l IIII 1 1 r- -IIrfý 41 I-qý IU 4- IU 4V qrI 4- r-I r I ~ T-4r- - -U - r4 4 4 4 4 -fý4 U4-4-1-4 ýA 1111111 Iii 1111111 II 1111111111 ill 111111 1111111111 I 1111111 I I I I I I I I I .~ I I I I I I I I I I I I I I I I I I I I I I I I I

U E-EU ' LPL)XUUPE') - U)E- > Ec->c- 4j r1I41-4 ID I)H 0.0 ) U) 60 1 F- I U< U1 0 *H1c- I 1 -V10c' - r N I-j414 00h 00 44 co 4a LooOO.z ee uE U AEli u 0 Ocn4 0x c-nU c ' Enu' LA'0 N Ol

4A)rIl 00.c4.40014-0I -44--4 -4-4--4--4 Q)-4)-H-4"-4-4n-U-4X4X4-4-44U-4)-0-0 -4-4-([4- 0000 0000 0D 0) 0) 0) 00D 0 000 04 m -w,-Ln a r-aDchcov- c m* n,0r-a OD40

C 0*HD ýf -ir-4ý v-4 ý4 v- -4 -l900000000000000r-I r --I -1TcI -4 i T-IT-4I ý r-4

10 C) cooooooooooooooooooo C 0CoMo 4o Co C C O C>ooooC)C) ) D ) " CN NW ý4 N N (1 (1 X X 00 c 00 mm X Xo00 00C0 - -400 29285 E 2 2 t 0n •0 00 -1:rI' " a N -4,--4 14-4 r- r-J.)4- (3)2 N) ,--4 -4 M -1 r-4(N -w 1-4 N Nq -ý. q. m a% " 4( n - - U U- 10 10 E E-~H n u) r -i r-4 Co co c0 .0 u u w Lo > > -q .- I 11) 21 H- H- 0444 (n2 U) : 4 a 0 0 1)1)4-4 4-4 LOQ~U(n2 a J 04a Z Z0l . H1 1 12

U 'Fý -. H F-. i U a'U a4U a4U aU U E- F, F o E- i

I ý 1 010(

,00

U<0U

U 'U IU I I

U a4U a4U 4 U a4 U a)

UE-4 U1 1 g

U04010 U 'UU 'F-CU F-U I

0 p010101 101 H 401 U0 iI U IiU E-UE - iI UIUiUIUIH>(IE- U H a' U a E-'U a-' U 0I U I'I iUI U - l I

000U~o000004 I 0 101 I 80oUiUi~I~i•0 0 u

E- i < i P : ý : ý i 0 U xt I - 'HiýFP 01(DIlE- i E- (E- ,I E- I E- ' U i <~ i<

U -u E-4U E-4U-

SE ooE.i888~~ , U 'UU iiU U E-U ui Fi •oooo.. .o E- F-i -iE) iE- E4 - U U U Eýi F- i U U HE-- 'H 'F-

010 I10

I 0 1U

U U U I 'F-. I u C Pi i E- i E-

U E U -'I. ->E I P i PE FiI Po E •01010101 U UE-o-•• ' U U U E- UEI U i 'U E- i E, i U I'U U I'U I'U I UH)OýUHaU F-4U Ha U U H iH 'HPIPI u I u1 ~ U I I l 'U El rj C- I H I EH - I I u U 41J (d F- F24 E14'E-E . L4 3

E-- (N( E E00 .N(N ý4 -ý- N) N3 NC4ý Mx xoC 00 0: -H -H- 00 C -4- 44 000C( : a414a, 4NaN Nul> > -H -H .H Ll( II~1~' N C 0 0-4 N C1 t 0 C 30 U)0-,- 0 0 -H1-- 112 112 -4~ ~C -1 ~ - -C4'C-A -4 CD4C) 0 04 -4 -4 -4)( 12).J (N (1 r-Nn COwM~0m C> (Nr~" ,o -0N CO 0)> .-T-4-)-m- "-"-4-4 -4- ý4- -4 -4-4-141- a)a)4-4-4 N 0~ -H r-4-14-4 - -4r-4-4 -4-4-4 (N 12)4.1 H H1 1"1 1" (12 m (1 4- 412(1 u1 (1 (1 1.- (12-4 112( 0 0-4-4-4000--04-44-4-4-4-4-4-4-4-4-4-4-4 -44444------4- ý4- -4r-4 -4 -4 -4 rA-N (N (N (N (N (N (N (N (N (N (N (N (N (N (N (N (N (N (N (N (N T-4 -4-r-4-4 -4 -4 -4-4l -4-ý4v-4-ý4-4A-4-4 -4 -4 -4 -4 r- r- 00C m m 00 mmXXoo 0ý>>-ý. 0 C04CN 4 - X X 0000 -H4- 00 0 ý- ) a) -24ai4-1 tr IV "n "In En x ) (A > > -,I- 4-4 4-4 U) Un :ý 0 .-H1H rU

U I U I U I H HF F E- I E- HF FEu

9 U U 0 H~

< : 0:4 uH Uu U Si E- H U U i E- u F

E-, E-4 I4 F- I34

U U - 80U I U iU0 H

U IU uuIUiU I 0000000000000O0000 LL14 C-) UH'UH H'H' 'H Bi

uH )U4 u u a4 U a4 UF a4

UlUIUIUI uU UH u 4 uH H uiH u;F- E-' 4 u U 004U u I U U 1U 'U IF U H 'U HF $ < < I UHUIUIHI~l I U I~ 'HI I El ~~hE-li1 1 U H 'U iH, -. F Ei F E- • , a, EUp UU4Uc U IU U FU u 12u coI u u >U Fl', HUEI1

o i U 'H

H H0 ~: D ,:4 ~ 1212 : UUJI000000) I:F U F C2:x4 I E-4>I U1 E-U FU IE-4 U I E- i E-

UU U >U

101 U- I F I F .HH '4 UHUHUHU r-4 0~ 0 C)~ ~ CD uu:?Un 'HC'c

N N • 00 L -Iri0 m 'r,[- oUU ML >>., U)>'4 a) ) HmH 0"4 C4U) ro o -4-4m )U A )i '0H'0 n i0 t-i W al C) 0-. Un UT n "D r- w -H-m C'1H in %-D) N O0 0: H- ý inn %.-0 w aN c) Hq H H H- H- H- H- H4 Hi H n40N 4 ~0-4 rH C'] m - ulr- 4-0A N U)i i 0N

E4 0 00000000000000000000 - 4 -4 -1 -4 -H r -4 -4 - -4 -H -H-1 -H _ --IH -j " -,4 a)'C.' Co' >a a)*)C' C Co'C' C)'a) ( ) C)' a)ca)' a)'aD HVHw HV HV Hz HV HV H HV H H H- H HT H Hr H HT HqH v) ui n in n inm m inm in in in in in in in inm nmm r-A~~ A ~r4-i A - i I H H- H- H- H- Hj H- H4_ H Hj rO00 S(a to~S~ t CD e 0 r-I V-4 %4D 0o r-4 T-4 r-I c14 C14 C*4 m ~ k lA 0) C0 0 C 4 -4 -4 C4M m D q q i i - - r oc Le ON 4 nD r- 01, -4 -4-4 -I -4 I-4%OC40MC.)C, $414

0 0 C>r- C>ý4C - Or4M I M -I Mr4M-I Mr4M - M18I ý1 M - - -IMf4Mr4M - - -4Nr4 -1

to) 1 EnUE4U1 4 W U E4E E-4I)U) E-4UUE E40U U)UK 40 *14JI 0Iý 0 o •u u u o oouiE ra IVI [ý1 f.I lu E- ýw8E- D En ý41 4 41 E- E-4 E, E-1 E- e- ~mu a~ omo u

4 r I E II I I I 0> E4> > > , >, E-,> E->

• ii r= fir. I o< I E- E-4 EH E 4.-4ý4 4) 0E.0- I >-0E 1 >1 SD-'EE->E- >E->E>4 - IM 0E 1 'E4 E- 4 >E

44W IU4 IU ra14 1 04 I 1 4L 1 I1 1 I1 ' 1 1 1 0 1 a 01 I I 1- ) O 1 41100 41 01 41041 U1 00 1 1 00 oo..•O z oE 0p0 L): Oz 0G--0E.

U 1 1 10 0101 01 011 01I 0101 ý4El-, 0-0-0-0-0-0- >- 0E4 0 0E-4r-- X ~ E-4 0 )0 1 I 1 10 0 0 0 1 G I II 01 O4 0 118 S141F 814

I I 1 1 I I f1 El o' E'I 0I-4 I 0 1 0 01 1 1 IBI < ::'

EV00 I 010101 00 *. I10 10101I 0 'I1 0 01011 1 H I I I 0 >14 4 E-UI4U> ,44 94 -4- >1 I U1-4I E--4

U) > > 0 EE-> E 41 0E > E 4J 101S0 < U) 0.-I 1IH 14 r_ wOowowow LA' 010w 4 (NCIO 94 LA'0C- e01 0ý 0LA0 Dr sw010 a1.0moem-ICC4I B0 Pmem0 0ý 00

4.~) t-U ý4E 0 L)04 I01ua 4 uE- iw u u4 04 u am L) a4 U 4 U 4m m um 1 m 0 )a 0 U Ia L)wowm Io L)so Uo L) U II)Iu em w m m iIE-4IL emE0- 0E t0 .s: 1 >1 -i 0 1 U)-4 0UI0 I) 1 IOO L( O1 CHId0 I 0 IIbI 0 I-I L10 U) e 0 8

1-0-I0U0 U-- 0I 0 I0 B 0 0 U0 l

00U3. 000000000000000000000000000000000000000000 S4I) oooooooooooooooooooooooooooooocoooooooo

U4-10 01

E-4> >E m a2 M 0 - s Co a e\4(N

141 r-44L -4 141t-41 U)-4 Lfl -41 U -414U'-4 LA ,-A 11 r-4 141 -4 Lj1 -4 LII -4 LII -44Ln 0-4 C0> *- C0 ,-q C-4 C0-4 r C)0-4 r - 0D4 0DýACý4

SE- iUE1- i• I 1

' CDI U,CDICDI0: I CD, E,- iIE- •E- I ¢EF I E i F iI p i

SE-4 I 000000000000000000 000000 0 O u~~~~lu usuu(z u UI 4 z1 l 0O0 U

CDEC• 'CD CD' CD'C '

Ui• u i OtI I ,•< C' •,LI, Z C DI EOI I E1< lIuI, : E--XEE' UE- UEIUIUEIUE UE- UElUE E- 4 1 : U' E- u x u >110u >x u >1 E-: EýIU I,P E-4 E- > E4 -i EI u o > Cu, UIUIE

El UEi U ED I E E D E

u I E, 0 1

E-EEl -UIUI UEI E-' E->E> E-' EUii u a4 v4 rý f-4 0 UI 4UD E-E- a CD UCD

Wa~~a~ata a a E- E- -E- Ký~~~aaIm

E U- -IUE)I'I E -'IU -I

'0 IC I E-i C i U i

E- HE- Eý E-2~

m 01 0 01

Oj. 00 r-i m-00 ,o0 (CN-A veC ommmamm~f o0 ,:i ,N-It orn %De m )O ~ o% o ~~~l~l o ,-L N- L-Ion 0 U, Ii - II) I) aI) ,-i (N a) (N - (N)C (N)C (N)C) C(N a (ND (N)C 4- (N a) (Nl -43(:) 0- -40 (N -4 w~ ml a) Lfl Q) C' ) a) ( r- -1 0 - .H r-4 .- i 04 r-I- -qr r-A r-4 C4 r-I "- r-I "- r-i ,-I " rA U)C -I m CLfl En ,4 n -4 Nu)- 0 4 -4 -4 CN-4 Nrq( -

N N I _N i Eý i Eý I N-' NI I E- E- i N F-' I N i N- E-. E, Ni E-uE-' Eý E-• _N> Ei 01<112t El;14 UoU

N i~ I < u< E-•1 I r_) I4 r_) I U > E- I • N• I u • E I >1 N N >E p 1 0 I

E-EI- ~U E-uIuI E 1 > > E I E- C I a:'a<: E-E 0 O E- I E-4 1 U' u2x 1 u QýuND NH 0 0E- 8<1<1 ii• 0101 : o N 'u E-> UE-4>11 -i11E UAuN i' I 01 oi0101 • N.1I 'F-11 i U E-uE-( u- u-

0101

uD p rD E UuN E- E EN>N> > E ~ ~ I EG-' - Eu uNE> UNE 0101U - IE,> - I 0101ý rI CLi E- I E-4'U0uN I.4E- 0 I E- I N- D E--. IL < :I I U 0I~ E-' E-i i P 0 Q P Pg N u U I U I I: E_, N>N> E- 0P0 E E- > E NN ii NUE- E-i ulu' 00000I•i E-' i E-'E F40 u0 81< P00 0 10

U - 4Dr4(V Ir - ugui N iN i r--U r -urN --

m 4 n i o N- N

u Fl.

o m w to (D "•m Co '.D~ '-)N LI) Ln L '.o o o0 ko I'D o N CoI' ko k0 4 r 0C04(N,--i- (N r- (q ,- r ((A "m- r - j U) CD0i (D0 4 0 r-4 1-4 -4 -4 -4 En4 4 .4 Oct En m En 4 El) 4 -4 En4 4 4.-m --4 -n -4 -n -4 EA -4 En 4 En- 94 .4$C~ i ft4 0I 0< En Oc En U >1 U)UUaUa:S:))claL :a:laUa m 0p ,-q "N C') -o Ln '0 N- Co 07% 0D r-i "N C) - Ln %D0N- Co m 0) ,-i (N m Ln %.o N- Co m' c -4 (N -T3 LI) '.o N-m Co 1-4 (N mn -ti Co m)0o )0m'. 4m0a )0a% 0'Na. 6 6< 0o Dc 0 -c 0 o)0 A i -i4 .- 4i- .-- .-i 1 4.- 1 "-4((N (N (N (N (N (N (N (N N m m' ' ('1 m( -4 --- i 44 -Il l A - 4 -4 -4-44---4 - 4-rl-4-v4 -4-4-4l-4-4-4l -4 -4-l 4-4-4 -4 -4-4 -4 -i -4 oooooooooooooooooooooooooooooo~o~ooo~o~ooooooooo000000000 Co 00000000CDC0DC000oC000 000000000DC)ClC >Cl DC DC-CoCoClCO0 ~lC DC )(DC )C >0000C a-400 J-) ,-4 ,- .-i i- Nr (m Cl) m mI (N " ,--- ,--- 0 0 M 0Ln0C)Wm r- 00"-4--4m (mX00X kN r- - rmmm oooo.-- 0 o00 w0 0 -0i-c c0-4 c -4 V V " " LI) LIC) F- F-' F-4E- cn a U U [ ". U) >> 040U) r-4LA,_I F:4mt MoV-I Lfl,-oi _4d 1)M En a 41 H XXMMD4Li4 M M4XX H "-H =4Xý:ýa U U 4 J a4aZ4J 0 S -H 1-

U I E-4 F iF-4i Fi E Uii E-4:>E E-. E-,E-'i I E-i IF-' H I umummumumumuUlUIUlbiblBIUI101010 U• 4•U 141 m C, iC, i ' P 'C,< I Ip I f IC, 0E-H>P 0 > - > -'O>-Po>-PE> U IU (IUIU IU I U I EýC, > C,E- > 4E4> E I Eý>

01010>~~HH >' > iHI4:ICp>Fp >E- >

IC, C,UHiUHiU IC, I HO I F-' I

> E- E-4: U404: 04 > 1 4:14:1 10 1 S II I H I 4 I

U I C,i C,i OE-'E-p O P E-i u < uu• u uou0 1I P E->~~ U 4:UHI U 14: IF- II-I U S E 0E I 1 0 I I gaIHIH4 gy I F- zUIF4 u

000.U<.00<.Cl, 1C ICx4 I1 H01010101 IlU FEE4

1U41 UIUIUIUIUIUIUI 0 a4 1004U 4 a U 1 0 1 I0 0 ER I C, I i IC, IC, IC, I U UIU uU U u U u I

U1 U I I u 4 01 U UE, a4U I4U la PC a4UaE- C4Ua • lO l 00i iD I

4E1L 14 IF-P14

4:'UIE F-' P , PCICIHI 114:14:14: IH C ,I , CI 01 CI0l () I4 u U

U 4 U I U E-4OFPiE-io-

HU1U EFU i uu E- UC 4 lU E I- I U:I4p F-il4 UIHIOIHIHIHIHI 4:0'F- ý 10 ,4 F . I C C, ,C > C, C,- C cS o 101oo-o I I 0104P 10 U'H 'P U4:P P E-4 0 u E-u U EI3HIP>EIHIHIHI I>

CIC IC UC I CI, C

8S010 :8 U

CE-iCI P H PH- : F- I

N-oo m ~Lnc>%o0.-0)(c Dmm -4m mr-4 rn l00C(N (1m C LANNN-4NN cimmmoooo

41) la) 41) w r1'mNO m m M.- l-Hr- r- •4-) 0> -i 0> -i 0 --i 0) i-4 V V N Ln u) U u Hý Hý H H, co' u Ln r-4LA ,-4 -iq' - LA r-4 4L) 4) M/ u) a4ia~4 ITu -u -- u U4) mA %o N- m m 0) .- i N m -4, LA %.oN- mo) cX.0 H "- m ,v LA LAo N- mX m 0D r-, " m .r LA Lo r- m meIm m m -rq - -4 4-v -4' 4-e LALA LA LALA) LA u-ALA LA LA La) LA LA Lo LA LA LAoLA LA r-4me -4-4-4-VI mmo 4-4l -4 -4-4-4-4-4l-4oewm -4 oomem -4 -A4-1 -4 -4A-4 -4 -4ooooo -4mo~ -4 -4-4 -14-4 -4A-4 -4 -I 00 0000 00 0000 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 01 S00000000000000000000000000000000o 0-O U) 0*-- N- OD m 0 .-4 %o C) , r-I r- -4 (N N C) in)I' (N< 0) .D C> C) (N .- in .l (N m %D r-A - (N -i (N I (N C i ý1 M M "C -1 irn %D 1 '0(3 (v)-1 r c ) o 0 H.-41-4 ý4~ - in .- 4' -4 N Qo - ) o o ý4 W) (1 ) C) N( a)~41-1 (N a Q) ( a 4 .-d Q) 4 4) m ) ( M -1 iM .-4- 4 CNC)( rcI o -Co (Ncqr ma)1i r~-i ,-i in) -1 in r-i in r-1 in -1 in ýi in -HinL ýin4 -i r-41 -4 ,~--t,in" ,-..- ,Rn rq-iL ri n4i - u : 1 4 : iu) -iL A r q )

I E-' I E-' I ' E-04: IF< OlE-44 c l re i : i U 1 -H U E-4E U ri E- E-. 4:F

H> ui 0 E0H iE-, E-I0i H>-'HHHH 0 I a4U a4 U a4 U> a oi 01 EH 00l<.I r •HF- Oi U 0:I4U a:4U a~4U 0a4 Oi 0000-I E-• ,I u4 I U 4: : 04IU: r4, 0ý 4:I CHi a U I U EiU i Eý

U a4 :1U . 04 4U u 8 8 u 01>1i E:-:01 ' I H r...4:04:04E- S U a.4'U.. a'. U 13 U a4

U 13 bH U I U E.,-I> >PE- ul ', I u I 00 Oi 01 04:

E-H>81 Oi E-HI> 04 E0-c P• I P.. I r. 0i~ E-iF-'' E-, i E-• i .. i 8018o4 : F-' uH> : Oi E- > Hi Ei rH0 E-' E-> P Eý > E-'0U)

Ol~lul~lol H> r,...• -I E,.p'lI 0I-' - 'OE-uot i U i OUIO) U E-IO

1>0-4 > E-i F-E-'iouu I F-''H I u'' E. I Iu- I F->E-' u0000K HI H4 E1E- ' 04:014 041U13 E-' i U,. i E-.' i 0i1 oý4I F-iý UI 8? 0 F-' I . -0X-' C- Ia U a4 CL4 E-' i - I 0 4E-'F-' I E- E-t E-I F- '4: '4 IOE"> HI > > Eý>P>01 1 F-'OHO-'OE'> > E-' E- i E-, i • 04U04 U 4 a a C P1) U)E OE-'OE-4 10) UP0101 4u<:'HI UE1 UEý U P• K>• : . >: 0000 P > > E- > P > 0 UE- I

0) (N N ) in) (N ml C) N- a0 m C) r-I .-I kD C) .-4 .-4 r-A (N ( N '0 C) CD (N -4 - (N m '0 1-I.-4 -4(N.-A n '0 in) in in) 1-4 U-4 0) r-q Am,-4I in) %D N- O 0) -1 i .-H -4N, i--4C ra)4)%.D (N 0 NN C4 ý4 4NiNý aI) (I1) ý4W ý 4W ý 4 W ý 4()ý ý4Q)W4Q)W4Q)134)ý4W (I)a) C)OU)OQa) m () Mla oT -i C) r-40U)r4w C)-i I r-I-4 L)U) U) Lr-4U) U) r-4 U)-i U).-iU) r-4 M) r- -i M) r-i M) r-i CO -i Q 1:3Zr-ic4 (N -A in .- i un -i n U) OC in in -4 E'- 0:n F40-4in in o4 in ,4 in -C4in g4 in - in in .--I in ,-4 in ý iLf rinI r-4 in .--4 in -1 .l -i C')M- r4) .-- mn 4 O 4: O o:4:c x 1 4o a) U) - in.- i vin" enin in -; in in'0 in N- inw ino) inc) in.- i in"N inm inr inin in'0o inr- inmc) in inC) in),-4 in(N inC'T in inLn in-,DrN in in a) inCD inC)-i in in(N inM' in- in in%.0 inN- inM7%q) in inC) in-A -4 (Nr) in in ,In LIn'0n u -••• T-1 .-4 1-4 r-4 .- I •o.4 •o.• o• oooo ~ • • • .-4 •o .-- 4 'A . (N (N9 Noooooo(N (N (N N4 (N4 (N (Nooooooooooooooooo C') m' C') m mC') C)m m m IV) IV) V' qz~'l 'I

C) C) C) C) C) CD C) C) C) CD C) C> C> C) CD Co CD C)- C) C) C) C)l (D C) C) CDIC) CD-CD CD C) C) C) C> C) C> C) CD C) C) C) C) C) C) C) CD C> C) C) C) C> C> C> C> C. C) CD C) C) C) C) C> C> CD C) C) C) C) C) C) C) C) C) CD) C) C) C) C) C) C) C) C) C) C) C) C) C) C) C) C) C) C) 0-4 C>~~~ r-i -4 N- mD m r- m m ~~ m m m m T-4m LA -1 Nc W- C)M)-r- N C N (N (N m( mm mCe m m m m k0m r-()m 11 C mN Lmi- LAOm : L k X LT,ACLA C>"IT'A "4,C>r- LrC. LAI ua r-i L r-4 LAr ,- Ln -i LA r-i LA r- LA r-i LA r-i LA- 4 LA r-i LA u- r-4 c) .-zr r- o) U) a) (1) U)f U) U) C) r-4 C' ,-i C' 1- C' r r -44 r-i AC 0,A m -4 i-n KC-4 .lF:-4 o< r- 4 4- r- < r-4 04 r-i 0:4 r- om Eni Lo Oct N ~N-1- N - U) ~i) (A

HIi E' uCDD I 0 E,> 0 0 ct 0 H0 CDE>>>

0 a 1 E H, H >

FCCH>I FCC ,u I u I E- < I p 1 0 1

. i UEi I

u'p u uo uc

U UI U Hi

H- p PH UIH PHf< u uH Hý> Hl>

E>>H' U' I E-'(' > E- E 1 0 1 0 10 0 1 UI Hi i-c' HBiB CDI Hi110 1

0H'i t1t9L U0.Hit 1 BI H -4 H91 r•IE -i I1 , I ., I I H i I U' CCH' CDDCCDDCU IH1 C1D H> <>4E H E- S i 0 I CE, Hi Hi~E-ii x X4Ol 01 HI HIU10 1I U I U U I Hi HiIl ilU

U' i 'I uE : H 0019 ' u I1 I

UC) .lUEn. um CA. E- E- H''--E~ E- IYB

U 'HIC) UE- UU. 'H 'H I I < H '>H'H HE, 'H i CDCDCCDCDDCDC IH P IH -4HIE- iUH

EOct u4l BYzY P: E4 d-4iP UHUHUHUH CD C C< DC CD 'CD 'C CD CDCU' I CDCDCDCDCDCDCDCE-4 P C D CD- CD'D EC'CD

o vo a 0 C' LA .- I "T N- .- I (N4 in N- -i ,-I (N r-A N- w o m N •00 00 ODO• 0 C) ON -4 ON N (7.( ONLA , 4L .- N- C) OCD (N CD1-1 rin (N ,i mv LA tT c> LA () LA MC N- (n MC %,0C' (N (N (N (N in mn m m in m m mm,-I t-n mn-4m ' u -- I U)Q ) Q) Q )aU)QC)(Q 4) Q 4)UCI4)UQ a U a) a) ) ) LA r- LA U) AC) CLA-LAI LA r-4 LA r-i LA r-i LA r-i LA r-i LA) ,-i LA r-4 LA ri- LA r-i LA r- LA r-i LA ,- c' -i C' -i C'o U) -1UD CD UI(- ) U)q mn.-4 c m ri-n-m r-4 m- .- I mn -~i-4 -nivM i-nm i-n -4 inm -4 m r-4 qý .--I .:I -I q~ C'r--4 , C'-4 C r--4-I C' -- C' ,- -4C -4 C' r r-lC' ,-4r- <-44- 4_o' - rI4• 4

Ný w) m) C' T-i4 enmN T LA Lo N- m) mC c' ~-i4 mN .:n LAvm r-ND C) C'ipC -4(N n qr LA LAo N, Dc) (=)C' -i "N e) ý LA LAo N- m ) C'-ic (N

U) -H- CD CD CD C' CD C' C> C' C' C' rlo LA .-i LA (N u) r') ui -,:rLA Ln Ln ýo Ln r- t~o CX (NLA C') -U w N- iA LA LA) LA LA LA) LA ýC) k.0 ci 0 LA C)o kC' kC) ' .0 ý. 'D I'D -10 i1 i-i o ) NLn N- w N- N- r- ci) C> ) ciCo ai) C) (2) C) (1) C) (1) C) ai) C> ai) C) ai) CD (1) C) II) C> ) c o) a)c) C) (1) C) ai) CD (1) CPOD~ 11) r C) ,-4 (N .-4 (N .- 4(N -i 04 r-4 (N .- I (N r-4 (N r-4 (N A C-4( Nr- Enwr-4ofZ r-En F:r- NnF4(A- m- smC-- A04WFCE =-1ý lý o 4rqr4rirA-n

U < i cci:, Hi' U 'UE- S i E4'4- I E I i PH4:' 010 0' UJ IE-'H I

0' 8 8 0 00000101 8:Oi S> 0101 PH u i uoU i 01 0 E-u 4 H'

x 00002 8 00E i U'ui H-E-c PH> PU >E->i U i 0 4 U o~u E,< I' < I U I U 04 U ccc 4:' a. E-H0i E-1 04 H> u1u1 UI~i 0000a4ia4 EU I 0101 U101 u1uIcc I H-i> P >

U'IH-BI i HE- 0b>HEý E- UH-> H' 0i:04i 0'H> H1H1 E- U 4:'0'1 E1- 0000 0H 1 0xt 1I 8 ; 0101 UUU Ee>E> H>H> l UI UHOl H I 0' 0 0 H> Ei00 H101•141 0000 ci I H I c0c 1 0 1 0101 4:>Ei TI U0i01 I U U'40 4U' 0 : 0C1 0 0 00 U i U 0101O4 U 4 0101H1H> 0000U IU I 0101 00Z ai C-)E-4 a4 UH U'UM u0I01P0 0iH' 0101 U IE- I Hi ccc'cc I cci 0101'o

cci: - I c-c cc Y l< : 'ccIi cci: ' I UHEH UHUE-c H1 0'

cJJ 0 C)0 Ln '.0 0 0 -- ml 0) m m (N --I , (N ( M W M r-.N ,- 41 0rMi - c) LA ,-iLA N L"mm 4) 4.-J NI--L Ln LA L LA LA LA- %D 'D '.D %O '.--o00%oo k OD1w-c'0.--" - L 'LnONvNNLi .o U -,1 , ý4%, -kor Nr nr Dr -r o Ciw CDwCo) C), i)C w D -)D a) iDw cwiCDiQ) C)• a) CD-) mai)C,. i ) a- a,- 1O IO 0 4-1 ) ) ) (N c--I (N4 -I LA)c-I LA cI C) c-I C) cI C) c-i C)I c-I C) c-A N - c--I c-4 c-1 c-I c-I c-I c--I c-i r-4 c-I c-I r-I -- 4 c-i c-i c--I -I c-4 c-i r-i r-i ci r-i m~ c-i mc r-i LA c-INr-I Nqr-I LA c- LA c-i w En ]4:U4U4U4:):):U)4( < ns E:

C> C) C) C) C) C) Co C) C) c-4Ic-I ýl c-1 r-I c-4 c-I ci c-I cI (N (N (N (N (N (N (N (N .N (N crn m' m' m' C) m m C' M c-I c-1 c-I c-1 -l c-i c -ý c14 c --I c -Ic- r-c 4 -c 4-I r--I r-c-I A -I r-4 c--I c-I c-I c-I c-I r--i cI cI c-I r-4 r--i c-I c-4 -I c-i c-i c-I c-i c-I C) C) C) C>00 C) C) C)0 C) C) C) C)00000000000000000000000 C) C) C) C) C) C) C)l C) C) C) C) C) C) C) C) C) C) C) C) C) C)l C) 0 LA LA L LA LA LA LA LA LA LA LA LA La) LA LA LA LA LA LA LA LA LA LA LA LA LA LA LA LA LA LA LA LA LA LA LA LA LA LA LA LA) LA LA LA LA LA a40.O rnr) m l) cq-4 -4 0- 1 1 0 ,-4 r-- r-r-m-) D>o)11O 0 0 0 0, -i 0D r-4 a o0 ( (N In LA u U Fo E-4F' F o Co u u c C uiLo- -> E0 -4 A -4 2) X) En U) 4. CL.Ua4 -L4:E-4 H P4 U4al C Z0 0 U -I

1 o: 1 4 1 E--i i

UI I u I r-) I 1014- 1 4. 10 18< 1 4 0 1

UIUIUE i FFF I 41'010101 0 o U muoI UIC I u

-, A- o u F F F E-AH UIUIUIU UFuummu I i C I'U I E-4U > IUUIEI E-4> UE-4 >IF-_ .- F F 00u F F U ' F U i rU UC4 UFF-Hu iHE- U EC, FU0>U > I,i~i.: UE4 '.H , U ..>>,

U00800080 EU EU C)M U F u

01010101010101U~u•

101 UIUIUIUIUIUIUI UISUIUI U IUI U-I uIF-.i U U E I

0 a G iEH'I F I FHI F I F I F-. I F- iI 0 001

U u I

U a4 ) U04a4 Uf 01010101010101 U IF- U U E- H-4U 9 -- U-. H -H E. E-iýý I 10I I 0 0I >1i• I UF-'iF-' U IE-.•F U IUII OI E-.i oIE-.UE-, I U F U - I F -.'U i-'I IE' F-U F iF-- #FIC'F-' u UI : 0 IO I < U)U)U U ) uU U() U I] I UA U 'U U F-'U IF- F - 'F--E i uiF', U i UlIUI~MUA U UM U'un UI 0101010101010101

U'I I UiIF10i1 FI E, I

U IF-',, OFF-' F-U-U ' - UF' IF u g I p 4 8EP UI U I(. U F U E-' U F-, u F- UgE-L

UI o:V 8C? U I4 1

r 0010101m (N (N r-cC -1-40 0mm 00 0XXU --4 -4 001-4 1- U --4 U1) U))4"TMM ) 1 -1 -N1 r-4Nr- 4 4 4 0 0U w 0D -40c> -4i' NLA-LA U U E-E-PPO UU raEx -- >> qlA,--4 U u) u) 134 Ua(0HH HH X-X UUJ ) 0O-) EAK0 4E X :I~ m 2 m.0 m42 E :zxý::E Z ,a A .a A uU)- U)-4 -40 m010 -4 (N 01 'UWLA I'D N- w1 oo1 or4 (No 01mo o.q LA o %. - w01O) -4 0> r4 0> r-4 0D H4 mi r4 IMr- m~ r-4 IM H4 m H- mD H- IM v-i mD r-i ca r- mD-i r C r-i mD q4Q i-Q) v-4N A LA r-4LA v-i Ln r-q ,q' v-i Ln r-I LA r-4 qc r-i LA H- LA r-i LA r-I LA r-i LA r-4 LA r-I LA r-i LA v-4 LA r--i LA v-i LA r-4 LA v-I LA r-i LA r-i LA r-i v--i m v-I m r-4 in r--i

1,E-4 ,o 4ýEl . E4E 3 1 E- , ,

1. 1 1 ,10 , I,, ,,•, 0 - 01010101 0 I0i01I1010 S 0 I- ) EIL.O E 4 1 E1 - m1 11. 1.E-4 I I 4,[- 0U tUI 1 UE-001E- 01 1 10101010 0 4J~Ev101 , I 01010 0I01 1018 010101 E E-4i E-4 i E t4 I I IIU41410 21814I IUUIUI 814 I 1U 0101@ IUI I S101I I I 0' I I U I I 0 1I U I U I I OXU I,UI I• , , I 1 0 1 1 ,1 1 1 1 L)I It41 I f< I 1 4 I1 4~ U y N : 4 .4: vI : :

I L I u I u 1 I0 I I0 I I

UIU,IUU U PC i U 4Ui U I E'-, I 1E1

U I IU U I IU1 IU IU

, 0• 1 U,•-0- I-u I 01 E-4aE E , ,1 ~~~s 4 ,4 t t liu t

U ii, U -, "-u u u m E-I Eu1LIm 4u Fa u E4 &1,I u ' U IU2I¶ UI I I I IU I tU I) U E-4i:U U UUI 4. 1 U IU1 IE4 1iUE-4i 1)41 10 1 0 1UIE- U I U 1 1 U 1 E4 UI 0 )0 (AEn0 CA2 A0 t 16)91 1.6 E - IE 1 9 E-@ 91 9 9 E- E-i u I u LI I I I - 0.) U -4u - #ccI L) I - I E<-- UI 4) 10I I I I I I I UI IU I I u P I Eu ,I I u U0,1 I Iu -I u E-1i4E 4u4I I ) Ex4I u 14I Eg E-4i0 EL 4EU E1 di V9i81014101 1 1 1 4 41 >111 U - 4 g eol 4I UJU t a, ý ,I 4 0 1 1 1u I UI 1U I

1 1 1 F 1I W1 14 o~~M11 I~tlI0 L)1I I a Ez4 E-0' ý 0 0' 8 0

0 kD Eh OD 3 CN r- m LA 4 1m 01 t OD M 0 r-4 V-I %D C 0 -I v-I v-4 N N N M kD LA LA LA S D 0 N T-1 - NS m %D i r v q r-iLm K r-i%o -i0 D0 0 (D4N4 m LA uJ-L NT40 N .imT4q LA ko t- 0 C 'h v-I-4 v-ý r-4 r-4 T-4 T-I r- -i v-IDN CSO1 NI 4 1-4 C4 C4 f 4)- En CIDPC OD 4 00IO DO D4O D4O 04O D4O D40 D 4 UoIDCDI 4- cDIc -4

0ij -H 0oooooooooooooooooooooooooooooooooooooo0ooooo° 134-i4S 0ooooooooooooooooooooooooo00 0000O00000000000000 OW C•o mW.n • •• ' N 0 4-4 • N • N ~H ••00O• (N o -o No 0 0) kL al r- OD 00 MD ON OD 0) Ch r-4 0) (N C a) U) -1,D rIn r-lN-i OD wD m' N 0) -i r-A M (N M Wi -In 0C, I mD) Mn r- M OD I 0) CN (N (N m m m Mmea () m - T. Cl ) ) a) (: ( ) Q 4)f)w C) 1z) a) rQ04) C () f) a ) w C04) 01) 0a 10) 010a0 ) 0) 0 10) 0) in r4 wn ,-1 n r-4 wn -4 In r-i In r-I U) -4 I) r-4 Lfl -i In -4 In -4 In -i 0) -i 0> -40c r-4 C, ,-I 0> r-4 0: -4 c0 -4 0) -4 0D ~-i c0 -ic -4 M r- M r-4 M r-I M ,-i M -4 Cv r-q M' r-A M r-I (1 r4 fn r- m r-A m r-4 .4. r-4 01 r--4 -4N - r-lN - -4N -4N - N- -I N- -4 N- -4 N- -A

E- C.-4Hý D. E, U.- E- E- F' H) E-H i H i

E,LýE I Ezi E-F F>H

U IUU U IP I E-4, SF ' 1 4:I1 - I,,:10

>10 P,. -I E- -I E-4':P.-i E-4H- i U L) a C- X4 U r-i 0 1 0I•- I E, E-,HH'H 3 i'0'

r,> E- E- > E-4 E- > u) I u. 4 E-'i SE E u E•oc E U4 E- E- P -, . p I p. I FF-' I P- -Ii CLi Ui 4 3

L)E- U E- 0404 F EýejUz EIII UX ) E- E- 00OH--I E-4EF 0 I I C E- E-i i ,,:: 1 - E-. - U E- i u

0I0IF F E-1l U U I IFE- t- E- >1F-l Ecr > >i E0 > U:E0, -14:U 4:1Eý E HoIo E- 0-10Ei i E- UM UXE-oH O- 4: E-' I C-)E-4H. I UE-' I 0F OXOF-0uou) OU Ol•E- E'

E-.iI U (.)E, 4: Ei HE- HIH xt:

En

0 10 C) I1 E-Hi H

0 0~

01

,. w.

H 'Hi E-H> H

0> ch N- -4 ( m) N '-4 CD mD ON w m' m' C> 1-1 "N m m C' n -" C- --- (n•0•'KTIM LIn0 CO InM ýI'C'N m C- mD CnlO INC (N m m m m m) 01-- 4 -01T-r 0 0' -IT- 'IT 01 In 0'-- n ,n-IT01 % -V %--ILfl -44:O C ) -0iC) -AC) 0 ) 0l D 0-- ) 0- ) )C0 014C ) r -q mmammmo r-4 m -4m.- -4 m r-4 m F-A Un -1 man M i 01' -101T Nqr-4 Nl -4 N-r N 1- N -I N-r- r-N l -4N a ~ ar-aI~~0: r-l4:I U)fc4 $ccCfm 4: (A : uI4 1m4:m) 4:c mi4PU) m <:

~0. NCD C' 0)C-4 C> C) C1 I CD C'l CN C'0o Cn CD In CD NDC,'C C.4 0 (DNCC') 1D In In N> CDC' C0 C. C) C') 01' InCn N>CCCD C>0ý (C- 0 4-1- 0C> C0C> In CoIn In In In In In In In CnID Co Cn C>I> C> Co)CD C> > D N N DN N N ND C0 CD C) CDCD CD C0 C> C>CD C0 C0 C' 0D 000 0D 00%0 0' 0o 0D 0D 0' 0o 0D 0o 00 ID%00 0' 0o 0o 0o 0l 0. 0 k00 k00D k0 k0 k0 k0 ý0 10 k0 0o%0 0D0 0Dý0 10 0. a4J -1 m m *-U m m4 - N -4 1 ammmmmmr-wmeme Nmcmeommo'w N Ln So ko o oo oo rm r C) 0) a) ()CaCa ) a)QJowow wm1 (1) N C, 2a) (1)0 (N -4 (N -4 -(N r-4(N r- (N-4 - (N A (N -14(N-4 "- -1 mI r- (- D)-4-q 0) -A 0> r-4C0 - 0> r- C>]q C' -4 .- r-4 -4 C4 En4-4 M- ,4 (A 4 In OX- n pc4En -C4 n f;4C) -4 -1 x LI -:4NA :4 En :4 U) -il ) -<

F 4 I E l -I ý E4 E4 E 4 E E-4i E4 - i1- 'H' '- E- ( EHH > - >H U010 i

U U (U IU I u E- I o4 P I I

u E- E H

U Ua U 04U -U U a(U a4U 04U 0 U Ui i U U U UI U I U U) U i U i U ' u

(

E- . SU41 (U3 -I 004

E,' L14 L4 f3. 44a Ll 4 4 iU U UtU •U E U E U o E- .,2: (U FCC U UQ.AU U I ol LI UUIU IUu UI u uiU i U U E-U E- IU I U u- U U IIU i Cl X4 4 H I H4 E-E-. 'E-' EE-4 U u I I U ( HUUH• UEHUHUEHU UEH

HI U HHI Ei- U HUIUI U)0 U u u (- u u o u

u0 I u U 'I U1 1' U U E-4U C. ' H , -.E-- EH- Li U' H' u I I I E OH H E-4Ipp2 -CU4 .UEC U E-2 U EP U4 U E E2Ul U PHUi

I Ui U E- P i U

U H UUI HE-. 'UIUHiEH UH

4:1 I I1 iH H 0 ~00000000

0 1 0 01 C U) CXD CD0 00 c000 <0i I 4:' 4: 4' CD0N C 0o0 U E Ea U I U' U' •JE-• E - 0 i U .I U 1 H:

0 H > I- > E> 0i U0

1-4 OD 0> .: LA %D 0)% 0) -4 (v ( " o m N N- 0) 0Y .r-m mr) mo > )o A m U) 0LA U) U) LAn U) LA U) U) %DU %D (2)0a )0 (2 )0C (2)0 ()2)0 (2D (2))0 (2)0 (2D)0 (2)0C U)L)UC ) >a)C) 4) m a)a)4) a)a)NUN N aN) a)N a 0ý -4 (N r-4 (N -4 (N -4 (N (-4N,-4 (N -i (N r-4 (N r- (N -4 (N (N r-I (Nr- (N .-I (Nr- LI) - LA).4 0 4 0 r-4 0D r-I 0D .- 0 r-4 Nl -4 En -4 (A -4 U) -4 En4,4 En4 4 En -4 EA -4 En - (A4 < En w-4 r-I .- I r-4 '-4 -4 rA ~-1M' r- -4 M" r-4- tAN f Nr-4 r-4 LA)r- Lfl -4 Eno0 U nC 004 m -.4 Ln kD - m) 0m 0o -I (N m .* LA %DN- w) 0m 0 -4 (N4 m -,T m ) N. 0)m y.-4) (NACm - LTU) ND 0) 0M 0m r- CD ) z LI ) N, 00 Ua-1 ooooOO•O•ooooooooooooooooooooooOoooooooooo 00000000000000000000000000000000000000 C)OOOO0o0ooooooooooooooooooooooo0oOOOOOOOOOOOOOO U>Cp C) CD C) CD C C) C)) Co C) U)C> (D C) UD C)U) U) U)CD UD)C> C> C) C) C CD) UD UDC) UD(D UD UD UDC) C) C) C) CD C CD C) C) 040 4 OJ CD

-4 -4 U U mmrl)m (N 'OO-4 -400 m0 0 -t-OOc OO XXUU U) U) m t--C OD C>( c, ~mmmocoo-.H.H0OOq -I m m co CDu X oýU u w gUUUa)•• t lU) a).T~] C h M M M r~O~-4-i,-r--r--ýU U4J4 J 0 0W W) N N

o 0-4- '0 U 0 E E-P U -) ] -H "Z r-i tn -4 (1) 4) Un](Ci] Q C).4L M (U " " H Ho X X U U_ j 4-J 0n 0.0.0En C -H r-H> (U _ M

UuA• U) ,d U R •••,tIU U A~tibttiDeU U UIU IU U) IU 0IU UUU

U U I E- U UU U 14 ,E- U -t E-4 >E- > p >E > > >l> p > p> E- E- E

m u Xu 4u u ou

Ut CUUIUIUIUIUIUU IUIUIUo

Ut UI--' Ut I 'U IU E- UI I UE- I iLU

u ou ,oum o o oe om me eeom ou E-~~ o u o u IC Iu IL Iu u IulE-t

UUU-U , UUE-]UUp< , p =0 E-UUIUP,F , UUE- , ]UP ,IXE, ) U U U C om)u u uCD mCD um~ u >',E E -u, u:>u u oum Oct

r) 000000c o0om0E Ch0Ug 0000CU(A 0(u! L)m0 cmU( o ) (Ao0om 0 < xiC4<:=' IC Oci x#c <9 'E : 'ý < xI Oc xX I

u :lul u I lull10lul U i Eý

<

• •• • • •oooo Y .• 0 0 •

U lbI E U I U U 1 & -tE-U U

N oOH Umcmoo XX iu U U U) U rN- UJ mmmmoooo HEO- 0 U E- C U U) E U-H 1]) 4 PIC Fl' : OC4 I $rI-Ott I X4J I00w Ut N N

*00- 00<0u00000<0z00C0000 u0000000ooooo o 0 T U <0000

Oj- ::

C0 m C ( r- 0 L' ( r- OD C- (N-4 -4 kD 0 C) ,--4 - C'J CN mN I'D Lfl UL

4) 4) L) a)U 1 4q • AQ - )ý a) ý 4W ý • )ý 1Jý4 4()ý )W ) ()U4W a) Co)0wm q) Q q O 0l - C r4 D .- I C r-Im -q m 4 -4 H mI H- m M o - 1 " r LH r- Ln r-HLfl t~ r-qLl r-lLfl ,-I qT r-4Lfl ý-! I1.-I ul -4 in ,-4Ln r- un r- n r-4 in r-j n r-Hin r- L rin in .- Lflr- Lfl -4 in ,-l4 r-4m ,-1 - v 0o 0 a 0 A

I I I I I I I j I I I E-II IE- I E- I I U I I I I - I 01 I I I I I I I IE- OI E I O II I E I O I I I I PI I P C I

P~ ~ E-U E- 10 P H i iEiU i U0 C)o~~P I E-

I~ I0"~ ~ 01 P 1 0 I0

IUI oo

U U I xE P I I EI

ooi o io iE o41I '-'EUoo ~I ~ 10~ I : I~,F ~ P I I - ' -E-4 E-,EI~ E i ' - E i u - o~

000 0 IFU 1 1 11001100 0100001010 0 I E-4 ~ ~ i i IF41- F- IE IE IEP IF I -IF FI -I 0-I

U E-H4 UHFHFH'HEHi4 HIEI H-HPE-

1010101 U X 1010101U10101P1u10101P10 01 )101 0g 01 : ) '(< 1 r-1 E- : 0 1 IP I U) U)U U U UIU)U ElU-)UIU 0-H IJ-u - uE U U Uu U U P I t I iF4 ( E- PH U 0p -~ p~ PHP P (N IN (NPH (NE- P NPE-4 I UIuH~ (~ 14 o PoPN r-J IN

r 0 01 C 0D In o t 10 -4 CN m t '-4 .-4 (4 -4 Nr

4) 010 Q 1) Q 4)1 0 01 01U4 401I 0 4)14 Q-4 Im Na) -4 01 (' L a) 1 ( dI 1 0 1 4 0 C C) 0

('4 (' ('4 r SU lb t r4tU I i s 4 4 1 t o s LA r- LA I LA .-i LA.- A LA r-4 LA .-4 LA LA r-40 LA r-4lU LA r-iLA r414 -0 A 0U4 0ý -0 C) CI0 C) 0 0 A c0 r4 0 i i M I4 M -IM4 -I (nrI M r4 fn -4 Mbr-4 M r- ý ic -Iq14M - t1 - V40 r- I r-rIPr- IPr- r- I r- I r- r rI B I 14 1f I 4 0 I 4 r- PCv- s E4 i

UI I c I I 0 0101 I E4 I i i IE-4 - E I ) I

I uc u L)

E4I0IUIE) ' u E-4'cE 0 I E-I I E.E IEB 1 i ~-4E-4 -EP- 4u E - E- 4 U E-HN, I0 4: : 1 1 I E- 0I UIO: I- IUE-I4 E-~ I UE4 EIU 1E-4EIUEoE-- I oi I UI 10I I~I P I1~~ I I4 IPZ IEP I 10I I 0I ~ UlE4IU ~ L I ~ 0 I UIU UI I1JI I UIUI ~-'IU: UII I: U I UUIU uU I- I- U E-UIUI I UU :BPiiIU IU p0 1 r0U 1 I 'I 'r I UIO B I UIUIUIUIUIUI I II1 U E- 4 FII 100 EI- ' I I IE144I IU "' IE 14 Iu 1E 1'4 10 II 10I 14 I I E-,I E 4 I 14I 1 1 1 I 14 0 0 (40 1 L) r u U'I 1 1i IPL) 1 u4 9P Bu lE-0I 4 E- E-l B 0 1 0- LA - t' N pci1 UE- E-404r- Iii N I E-4E4 E-E-4 I NE- i ruU i9II 1 EUE- 014 L4 14 1u4 E-4 i IU - I IU IB E4 IIIP 1 u I J 1E :E:I I E0-U o I I~~E-4 I ~IU E-I1 0 0 01 0 U 0 0 0 0 10 00 E u u U IE I I

I P P I E-4 E- E4 'UU"U'AI) IA1LALALLALAALACCOCOCC 1COOCOCCONNN UNNNNCDCCDCDCCDCDDCDC01I010 I 00000 0 00000 0 I 0 0 0 -4EHHU 0 0 E- 000000000000- E004HE-4 L E-'-4-4u- l UIPU I UI4 UI UI UIUIOIU:0 Pg U I Ir u) I U E-4 U ~I4 1U1IE- 1 U iIE- E-, E-4IU IUI I IP l I U I t- Us.N N NN N E 0NN4 1 0 E- 2 I E- IPL I E- U 0 L 1 L PE-NNNNNNNNNNNNNNNNNNN I 0UB I )I 1 O4 9 4I u01 UBI 10 I E-U 1 I 0 IUE10 1 0 I Il I d I 1 P I L) u I u I iEI LI IN L) I I 4I 1 8 01 I L)1 H I1 uB B I 1 0r) I08 1B u I I 4 I 0 1 P I 1'U 23I '1P9 0'I E-P 0 F -Ig, I' L)a a 4 U0 a U a410 U04 U 134IL 04L) a4 L)aU 04U IU a4 U U a4U aU aUU UaIua ua E-14 04 1 1 IL E4 i U4 i i1i P E-U U P 1 044 1 u 1 10 - 44I 14IL I II1 141 Pic L)14u U1 I I I L aU1 n EnU En En1u UU M01) u AI IU) n EnuU U)L)EnU InU)E) E u0U) 1) En U EnL) En L 1 E-,I 10I E, E-, P El0 1 E-4 E1 0 E10 E1 0 ..0 ..II0 10 1 H -0,81 H E4H E-4HI E- H 1 IH0 0H 0I 14 r I I-H -41,6 E40 000000IE4 E- E- E- I E- UE-4H I u L)IU I Iu I u I r1i IU I u I U I u I IL)IU I L)rI PI 1~ I I I~ I ~ I ~IN101 I I ~ 14 4 101 I140CI 1040 I 14 04 1 1041 I I 1 IU U U U 1 1 E-40IU E- E- T EI-T 0 1 U U U 1u L) I I z0II U IB U u Iu 1u I I U U U IM I II I IP I I E4II UB I E- EýHIP UI 10 U IU 0IOc I <0 10 <0UIP < 1 4 ,BI

0JE- 5 E- I0 U IU IU " EU U UIJ E-1 'U 'U E-4 IU IUE- U I U IU I t, E I Ef E-413 11 44 E-4 E-4 E- 4E- 1 P14 E-44 E- D. E4E3 E- 14E- D 1 4- E 0E-4 I I0 I IUP E- I E- I E-II E- E- E- L) I U 1 10 U rI r-)1 I U I C I1 1 LU U 1u E-UUiI E

ICPCoo1 ooo 4o1 UE 0oo41o4 1oooo ooooo oooo Ol1 4- O r-I cD 0> v LA %D C>0 -4 Nn-Vr CID Ch m~ ý- m i-LA LA LA L LA LA LA Ln %D k.0 ýo 10 .DO %D %D k wk N~N'LAn- or 4m r- r k r r 0 •• •• W•• 0) 0 0) 0 40 ,-4 C'N ,-IN - C') ,4 C%3.-4 C') ,-4 C13.-4 C14`4 c') 4r-i .*-4 64),-14 C' -4 c,4.- i N' A C'-4c4 c'4 .-i LA .-4 LA .-4 0) A40 -4 0) -i40 -i c -4 N ý.-I r -4r -44 T-4 r -4 T-4 r -4 rA-4 -4r- r-44 r-4.-I ,-4ý r-4-4 .-4 ,4ýl-4 ,4r4-4,-4 -4 r--4 r-4 r-4r-4 r-4 r-I-4 M -I M l r-4-4 r4LN .4i N 4 LA r-4LA -

uIE' E- lii lEu-I S E-I I4 I I I E I I I I I I I I I I I I I I I 0 14I IP I 10E-4I I I I I I iE- IU I IEI P I 1IU IBI0I1E- 0:4I IP 511111111111IIII I I I I I I I IE-4 I II 10 II t I I I I I I I I I I I I I I E-4 '.. 1 . 1 0 10 0 I I 41 I U I Ic IU iE I I I I I I I I I I I IE 14 ~M E I V I I I 14E- EU I EU':I I I III 1 I I I I I I I E-4 9 14E-4 E- E-4 I E-4I I I

LI I E--iu U~IE-I -I EI uIM 4 IL! ll IB 10 1 B I 4 I IB 1 I,P I I 1 Ii 1E-E- i E-LE I E-UI U I p0IP :b10 'LU ' E-I- IE-' ''U Ie IE UE- I E l Ep-1 ' IE-'I EP IIEU I E- 1 0 I E-4, I IP I I 4 4 1 Ep I P 0 1J 1 IUJE 14j 1 uII E-4 0 I10 10 1 P E-4 E- E, P E- 1 1 14 UIB 1 0 I I 4 14 14 I 14B -1IB104~~f -1I3 E- -I E- .I I pF.3 I 9 10~~~~ I 1E-II I 0 4I 14 I 0 I E- 0 E-4 l U :0E- Z P 0 p0I P E- E- viP E- L) E-4 E- 0U U 04 U 04 U a4 U14U U 04UaU04 IU a4U 04U a4 U IU 04U M 'E4 1'06 IE4I E4§O I ' 8 aE0 4 U uI U I I41u I I u I u I I U I u I U uI t3)I 4 4 E- i 110 E41 14E- 1 -I14 -4 p o 14 14 14 uI 10 1 uI I0 u 10I UBI Il< P I P I I P 1 R 01 I 1 0 14 En1u En EUU1 Enu AuI IU nuE nU EU cn u AuU) Iu E U0r4 s U ',auauau,uauaua,I I1 u I 104I u4I I141 • UEg4 4 1 14 14E 14'E- 1ZI '<4 uUluCI•uUII :1 uU: IUu 1 guu::I Uut I I1 I u I u I U I u I U I U I u I U I u I Iu I U I

0L) I uI 1,u Iu, 1 u II, I • ,1u U L) IU1 u u I01 U0 1 U u 1 u0 U L) S I u I u I I I1u E- I E-4 I I 1 H4f-4I E-: E 4 :10< 0 : 'E< 'U I: I 0 ~ I E- I 1 P E--I B13 u Iu I U IU IUp I 0 0 1uE P 0 1 1 U E-i U

WE4 E-4li. E- E, E- I I E-4 I : U I fUI U I u U I I L) 10 I L) 4•J, I uIUu IU I U I L) L) u L I U I U bL) L

.11UP .1 JlLU) 911 . IEJ.1 .1 .13 P U I0 LU U I 10 1 0 I t EUI-

E-4 6 4'E< 14 1 4 1 1 4E"- 1 4J

En 0 >000 C)o00 C 0 00) 00 00 00 00 00 0O 0 00 0)0 0 0 0 0 0 0 0 C'r-4N IN ' .- 4 C4) ,- N' ,4 C4)r-4 C') r '-4 r-C 0 ') -4 C4') C') .4 C4) e4 C4) rA C') -4 N' r4 "' r- LA r4 LA r-40o -40 .)I 0> -40C> 40a 4 to -4444En E4 44--444g-44---44--444fcw4E E )4E f,-P4n4t --nOtE 4-E4-4 n4--4 -4 #c LA04 En -4 N) ,40 LA 4 La g

0)- m~ qv LA% r-0NOD m~ 0 .- 4 m' -cr~ LA %oN- OD ai 0 r-4 c') mf v LA %oN- OD Ch 0 .-4 C') M' V LA %D N- OD cm 0) i 4C') -o~ LA %.DN- OD 0 r-4 -4 -4 -I4 -4 -4-4-4-1 -I 4-ý4-4A"I4-1-4-4-4 -4-4-4-4r-4r-4-4 -I -4 -4v-4v-1 -4-ý4-4-r4-r4-- "4 r-4-1 oOOOOOOOOOOOOOOOOOOOOOooooOOoOooooooooOO0oooo0 0~4** cooooooooooooooooooOoooooooooooooocooooooooo000 O N N> C> N>N N N N N N N N N N N N N N N N N N N N N N N N N N N N N N N N N N N N N N N O-i '-1 '-l4 1- 4 - oo--mmooXXUU•mmm C 0 HNOO0 0 •• • •mamoooo.•- 0 o • •• a) TTr - a% a --i- r- ý4) J 00 0 ) ) N N o r-4> -i ' 0'0 o4 (N LI) m u I E- H E-4 H o' c u U -. 0= En) E-4 r-4 > > -1-4H ') ,-4 Lfl '-4 0) l) U)E U)n(n a1 a4 i4 to I-4 4-!I-I x U U 41 4-1 0 0 .c r_ -H - 40 rot

' ' ' : •1E E,10 EU 4 H H 10 : EýI U

UIUIHI a I uIUUIUH o-- • uI Ul, lIU 'O UIUi 1'10',E U) 4o U,U i,0 P c I • U I I Up E-4I I U I I I I I U0: 10 O< 1 91 IB I U 10 0:1 o 0 o u 1 0I 0 I - u U U p U E') 'P 1 E')4x I 1:I P 10I' 0 E-) 0 - E,1I r U0 I 1P E,14 1 0 4 1 4 I

U 000 00000< 00 00000000U 0 0 E0 rE- 10E- I F:4 I -f 1U - I 4 I~I~~ 0Ii c ' UII H '~

IU IU I I I E- I IUu IC I I IU I 1 I 1 I U0I U0 I Fe,0IBIP 4 10 1 I E-1410140, 1 10 El 1)E' 10lE- H I1 S E--I4 I E 1 E-4 E.1 ~ 1 pUB114 z- 0~~' I 10~~~~IU 1 I E-4I 1 10: 10 : U L74 10 00:4 I: U0:1bIB 101 uP I I0 1 r-) IEI:B p I 0BpII 9 10 1 p-IB I Z I 1 U I Fl.IB u I Oct IB U I IB oc

E<-. H: IE E-' I IiE E-' H 0 UH 'E- 0 1 U CU) 401 U 0IH H I :H -< I UHU4 IH- E- II0 U I E-I pUH I Hp K0I I E- 4 1I4 P4 E-I E-4 1 0 14 14 4 1

Uo-U I0 Or- 'U Ii0 IUI0 1I E-IIBUU10 I p 10I 0p U Up< - 10U 0U PU U 'U oPI

UI I U I I I I F IU Ip IU IB H-E-I1 2

UI1 iHI H0 1i I IUIUIUU U IUiHbIHIlU

U )UU PU :IE- IU u 1IU ;U IU IU IU I p E-I P4 E-4E- ~I ~ E4- Io4 ~ 14E 1IE4IE- 0~ pE IE1~ 0 014~ P - I PI P I

U U0 1 10 1 u 0 1 0 1U IO HP IP 0 IU4 Io 1 0)1 U 124U14 U 4 U 4 U 4 U 4 U 4 U 4 U 4 U 4 UI 4 U 4 U 4 U 4U a S U UB U0 UB I b U rU iu L) 1U 1U (-) UB u 0iJ I - - ZZUU Ud IoUm amIm E- m . w- m m U IU Iua IEm m u I a zH I u I KC1 4 1 u KCHH H1 UI U Iu IU I L) I u I U I E-4L2-Hr)P l. L0 4 EU)W Ll0P l. 4 IX PW FI' 44P I 4 PW - E0 PL0 PCUO E-1 0:0I, p- ,EI E0 P. P0I P I 00F I I E O u .CX04X<=F4 X4

I I I I I I

I I I I I I I I

I I I I I I

I I I I I I

I I I I I I

I I

I I I I I I I I 'U 'f 'U

'U II

IU li

'ci: 'U ILI

'U

C0 a 0 U) OD O)C4 rN m LA (N (n ON r- O(0 0 ,-0 -1 C> 0 -1 r-4 (N (N m ý LA LA Ln ia) >, INCc m k 4 (N e 14 -4 %.0 IawDC> (N ) 'IVm LA UH4 -4 oý rIMcN I: uLA 'o N- OD m~ ,-i -4-I .-4 .-4 1-4 r-4 1-1 'A I'D "Ch CN W N ý4CN-4 (N Ii) 4)4)(1 a)ý 1 4a 4a - )WQ 4a 40 )Wa WV4Q W(1 cW() cia 1) CI~)() ma)OQa)mI) 0 .4 0 0o or-4 o40D - M r-4 -4 C) or-4 ) ,-IoM ,-IM) r- a -IM -4 M) -4M) r- M -1 M -I M r-4C) r-AC) r-4( r-4m -4 L)A LU- ,o' r-iLA r-4LA r-jwz -4LA r-4LA ,-iLA r-iLA -4 LA r-jLA r--4LA r-iLA) ,- Li r-iLA r-4 LA I) W9o u) ,- L-4 LA -4 LA r-i LA r-4 q -i mw r-I M~ r-A mw -4

r-i (,4 m~ ý LA D Ch 0 Ul) -'-4 ,or )c0 -4 (N mn -- LA , r- co mf00) ~-i(N m z LA '0Nr-0 c)( 0D 4 (N)m T LA '0o N, (I o) 0) ,-i c c) -T Ln LA -- 4r-IA4 -I - -4 v -4 -4Nr4 C "-4 " 04(N N C MM C- C-(v r- mm Mr q r-j--V Z14-

ODD00 C 00 0O0 D0 OD00000ODO00 00 D 0 D000000000DO D DO D0000 0000 D00 D00 00 00 00000 a 0)0 )0 ) )0 )0 ) )0 )0 ) )0 )0 0 )0 0)0)I'a m)a (mo00) o (o0m(m 0)m0)IX ((00( 0 ((00 0 0 - Do aN) C C0 Ln 'A4 'K0 Nt- 'A CN m N -4 -4 CN .- M) %D M r- OD OD OD a) a) c> a) r- a) c,4 a) m)c) Ln ~-i.o -i r- .- i c) a) m) N o -i .- i mn (N r LA -ý3 c) La mnLo ONr- ay)ook Cz) C% CN (N N m M m mc Y) M " m -r k 0) C: =1wQw11wlI lwQwC w LA mA ~ ~ ~ ~ ~ ~ ~ ~c)9= ) Q) c)a 01) a) (1) 1) 01) r-4Au -i -4 mLA.4LA r-4 mA -in-4 LA-4LA-4 mA r- LA -4 LA ,- ulL-4 -40 c) -4 0> -4 c> ,-i C> -1 C> -4 (Z) -4 Cý) ,-4 c> r-4 C) ,-q C) rn) -4 -i Mr'-4 ýi r' m r-.4 M r-4 mn r-4 M r-4 M .- m .-q m~ r- m V-1 m r-A -.4 -f I4 r-4 qz ,-A N- r-4 N- r N -4 rN i r-N- -4NH r-4N- r-N

I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I-, I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I-, I I I I I I I I I U I I I I E-4 I I I I I I I I I I I I I I I I I I I I I I I I I i EI I I I I I I I,.I I I I I I I Eý I I I I I I I I I I,.I I I I I I I I I I I I I E-I I I I E I I I I I I I I I I I I I U E I I I IEI I I I I I I I I I I I I I I I I I I I I I I I I E I I I I I I I I I E I I I I I I I I I I I I I I I I I I I I I I I I I U EI I I I I I I I I I I I I I I I I I I I I I I I I I I I 'U 'U I I 14 I I It-) I I 'U I I I I I I 'F-' IH I I I I 'H I I I I I I F-' I I I I I I 'U I I I I I I I I 'U I I I I I I I I I I 'U 'F-' I I I O 'F-' I I I I 'U 'U I I 'U 'H I I H 'U 'U 'U 'U F-' 'U U 'U 'H 'U 'U H 'U ltD 'F-' F-' 'H 'F-' 'F-' E- It-) 'F-' 'U 'H '$4: 'U ',ci: 'H 'U 'U iH 'F-' iH 'U 'H 'F-' 'H ltD

'F-' 'U it, 'F-' I'F-' 'F-' (U 'U 'U IU 'U 'U 'F-' 'H 'U 'F-' ltD 'F-' H 'H 'U It, 'H 'F-' 'U

C) LA ,-I ' N (N m N- rI -1 (NN ) %.0 m) m) N- a) ) a C> m) .- 1 a) N3 m) m' a) LIA ,-1 %0 r-q N- .-- ao a m) " CD~ i .- m (N~m LA -v c, LA m %.oo) N- m' a) LD c) LN (N (N (N m m m m m m (l m (24-IW 0Q) 4) a) (1) (1) ci) 0) (1) U-H LA r-4 LA r-4 LA r- LA r-4 LA -4 LA r- m ,-i LA -4 LA LA -4 r-i LA ,-i LA r-4 c) -4 D r-4 co -4 C) r-40 C -4

C00 D 0 0C)00 ODC) 0 00 C)00 0 C)OD C)0 Co CDC) C)C> 00oCoCC Co) C>C) C C)C C) C)Co) C) C>CDC)0( C)C) ( C) C)(DD CD C) CD C) C) N- a) a) c) ý-4 ( m ý~ uA koD N- D a) a)) p)D r-i4 D (Nm 0 O O LA O aON- OD a ) ODM OD O 0 aC a ) 00 0 O a) a) 0 00 0 0 0 a•a a a a• a • •a • LA ••aoa~ooaoa) C r-4 2o" mw ýV LA %D N- OD a) C) .- C4(N ~ LtA I'D N OD Ch C. r-I "- -:r t LALA LALA LA LA LA LA)LA LA Ln %DLA ko %D kD %D %L0 oaGo LO LDonwn • o • oHLDHLD LOHLDLA inHoLA N' N- N- N- N- N- N- N- N- N- aOD O) a)00 aOD 00 O) O) O m mO

000DC C 000> Co C> 00 CCo 000 D C) C) C>C)C C> C) C> ) C) C. C) C> C) C) C>) C) C> ( C) C) CDC) C C ) ) OU C)C>) C ) CD C)0(D( -'-- 0 Oj..-I 0'a T-4 O 0 q Ln %D CA 0 V-1 m -W N OD ai m r-I m

r-Ln LA Ln LA LA U LA LA %D %D%0%D %DO %D %DD0 w0 ko - o rI'0r LAnr % r r 00WO00W00 000(300oW00o0 o00o0 0 W0 WoWCDWMQ W ( 03 0) ( 0) 40 r-I "~ ,-I N' .4 N~ r-I 04 uI N~ r-4 C14r-I C14u-I CN u-I r4~ uI 04J u-1 N' r-4 N~ u C4~ ,4 "~ u-4 "N u-I LA -I wA u-I 0> u-I 0> u-I 0> u-I 0 r-4 0 u-4 (En -44 rlt-4 u-Ip r-Inn4C u-I r-I u-I u-In u-I4 E r-4 Pu-I u-n -I Hn

4J CD 0D L '0 O% 0 -4 r-N 0 O m (n

r-LA LA LA LA LA LA LA '0 %0 '10 %D0 %D '0 %D%'Dý 0'0-~ D t C0 m r- %D r- Nr- 0 -,0 4- )C)Wo)om wo woC) Cowo wma) mmwmowa C w Cý4 m ) 0) 0) 4) a) 4-J momeoooooooooooooooNN•Nr-4 C14u-I o0 N r-4 N u-i% ýI C-4 r-4 0 r-I N'u-i ('4 -i ('lui F4r-i u -I •c%4.I( -I 64u-4•Noooamm c,4u-i 64lI LAr-i LAu-Imamooooom 0>r-I 0) u-I0 q 0ou-I 0 -I to-0 Ný u -Iu- r-4 r-4u -4 u-I u- -I u-I T-4 u-I uI u-I u-4 uI u-I uI r-I r-I u-4 u-4 r-I r-I u-I u-I u-4 r-I u-I u-I u-I u-I Mn r-4 Mn -I LAu-I N- r-4 r- -i LA r-ILA r-i

mn v LA %o N- wD m~ 0 u-i m m -w LA '0 N- co (7% 0 u-i c4 m qv LA %oNr- OD 0h u>-I N' mn -w LA %D N- OD Ou 0> u-I ('4 mn q LA %0 N- OD 0 J- m~ ON (A O'h (AOu 0 000000000)CýC>ClC>C a- u-I u-I u-I r-4 u-I v-4 u-4 u-I .- I ('4 ('4 0'(4 C%]('4 N 4 ('4 C4'4 M' Mn Mn Mn enM en en M a4u- v-1 u-I u-I *-I u-I u-i u-I u-I u-I u-I uI uI v-i .I uI r-i r-I -I u-4 u-I uI u-I .-I -4 u-I u-I .- I -1 u-4 -4 u-i u-I .- 1 u-I 1- uI u-I A- u-I O000000000000000000000000000000000000000000000 on O 0 O 0oOOOO 000DOOOD00DOOOOOOD00OOO 000000000000000 CD D D D D D D D D D D D D D D D D D D D D D D C C C C C C C C C C C C C C C C C C C C C C C r-A r-i4 ý4 mammrwr-Ar o c- " LnLn u. u EU) - ýE E D0 , Lqn 4 r-4 > >: cz r-i U, r-4 a) la) En En a4 a4 (0 '1 H0 H X U U 4-)~ 0 0 C00-H-

I I I I I I I I I I I I I I I I E-I

I I I I I I I I I I 10 I I I I I I I I 1E I I I I I U

I I I I I I I I I I I I I I I I

'U10 lCDSU I I I I 0

Eý 10 E-4I 1'U I a I E,

I I U' 12U I I 'U 'U

I I I

ru E-4 u I I OrIU I I I u I EAI I E-I I I I u I 'U I '4:11< U C)I 12E, u 'U E-4 PI 'U 'U 'u E-U 12 'C 'U 'U4E- 'F- 'U

8

u '4:

I

r oo-I O • -ooo •4 0 • • • • 0000-

00000 0000 0000 00 I- o: A4) C> 4)-V-CT- V rro c'-4U,0) in 0 (n o -q r-r- I- -44L4- 00a)a)- 0- ,-T r-4U -4 0) Q) En U) a4. Ci. UII0 (0 F-.H HE-4 E-. H- F- H E-ýH E-. H X0' Oa= UC UU_U 4-4c .0.0 4- 0En 0.0-HH W)- - > > o) 0> r-i N . m q-w ,o r- w mD c>0-4 "' m zr. ,o r- WN CD -4 In"H ,, Ln ,o r- m 0 4-) 0)0a0 4 CD~~~~~~~ Co U>C) U,C>C U, U,>U U, U,) U, )C U, U,)C )CD CDU) CD U) C, U) C

OU C) C) C> CD CD CD (D (D C) CD Co CD C) C> C) CD C) C> C) CD C. C) %10 (n OD 0) U- m M C 0 0 o-C-4 L 0 l 1 IN IN IN m m ko 0) 0l N ý-4 r- IN1 o T1 r4"T4cn l l )r Dr- -r40) mA CDaLCD-a ýl" I " m -I LA m 4m-4 m 14--4z LA %0 N 0 0) 14~-4 1--l- ,-I r-i r-4 ,-I r-i r-lAI'D CAIN Cq 4 " ý4 q INC a) a)) () ) W a) )a am a aa oD a) m a)ma) C> ffA 0> .4 0o r-I 0> r-lcmQ riC-I .a r-aco C)m ý4 'I ml -4 EI r-Ica r-af 1 co ,-4CO r- COA r- -4) -4 "N1 -4 A LA ,-H LA ~-4 qP .-j LA -i Ln -4j-z ,-iu .-i LA r-ian -i Ln r-ju) - LnrA an r-a LA r-aan r- LA r-LAu -A LA - Ln r-4an .- Ln r-I,-a r-q m~ r-4a) r- m --A

i io a Ua Ma a em o

P UU E- U P 3' U Uaaai~~ E-

4 1 4 1 UUFIUI I 1 S, 111c 1 1 I 1C

Q~UI U I U I U I U U U U I ~a U U P U U U P u

I~iUIUaUaU IUIUIUIUIUUIIUIU UaUa SaIU0: 1 0 0 Ua11

.1l l l CUUa IUIUIUIU uauf4 a C 1~ lu i u llUlua I a a a I lucuI F I I :CXluaýI IU Iluau: IO IU'U I C);

fl .I, -1 F 'p 'Ia.1 U Il I 1 I I l l PIi u P i u i l i u Ii i i I i

I U 4UC14U al Ua4Ul l a4UaUaU0f f 1l,l Ul I P a fII4Ua U l ai U i I4uaE f IIuU l a4ua) a4ul 0 aa4 Ua4 a U* a 00U 40:4 r. a4 u a4 uP a4I PU Ua a4 a4u 0 a4 00 00 00 00 0 0 OZC00 Up Up: p L) : , E- : P 1 u#r!) 1 UP U o4 1f- itu a41UP U la4 uU 1r,i U au Eu l a4uE()li a4UHuiI Uua4| |l 4 U1r| i()| a4 Ua al 00a4i E)a4U i Ua aI U I Ua I U I U U aU U U Caa0 U U Ua aaaaaU j U U I U a a a a a a a a a a a a a a a a a I I I I a a ~~~~ aaaaaaaaaaaaaaa ~ ~ ~ ~ aON ~ 1 Ia P a ~ ~ E- .11a a a a a a a a a a a a a a a al a- P Pal P -I P -I P P P P P P P a~~a a a a U U a~~~~~~~a a a ~~~: I- U a I- U 1- 9 R U P u P aLP a4 ap a Papl E- aP- ap-P u< P P > P P > x a< a a ac a~ a a a : 1 1 a En a (n a w a x aCx a x U m U m atx U m U m a m f< x a~ x 0a x U m a~ U m U ma a ma U CA P a E- P U I a a a E- : E4 a a F:4a 1 a L4 i Eaa 1 a P <:a I 1 P g aI 1U 6 a a i P a #a4aau iP aoz 4aa P a a a u a ua E-a ua U P P P E-- 1P 4 mU wU mU Um Um U 0:U )0( ,UF4uUmItwU 4UMU( a E-i E au auua--44- E-4 auaE-P P a UE-a P E-- i iu Pu iE~P iE-4 aE-P UP a-iP E-iP E-

a a, a2 a a 2a a a a4 2a a a a4 a a a a a 4 a (a W a a) 4a a) -a a a 1a a a a a a aD a a) al 1a Qa a) ca a a a ai ap aq a) a) a- a> a4 aQ a- aQra aI a--aQ a- aQ a- a a- a ai a a- a a- a a4a a4 a a a a a ra a a a a a a a a a ra aý aq a- an a- an a- aq a- ana- an a- a) a- ao a- an ai an a4 aI a- a) a- an a- an a- an a- a a a a a a a ýa a a a a1 a r a A a) a) a a a a L a: a a4 a at a a: aD aI a a: aI a a0 a aD aC a0 ac aD ac a a aD a4 aDa aD a: a a a ac a a1 a a a r a a a a a a a a a a a a a a a a a a a

_4~~ ~l r q ~ ý i rIr - i rI - - - 4VA ri - - A 'A U-4 ' L - LA4 - IN- q rqN- - LA- ~LA -- - -A - IN -4 LA W 4r- -4rI 4 ý4r4N 4ý r-4m 4 0- v-4 i A r.-a 1A 04- Ir4r 4-4 Ar4-IN IN- r-aIN r-- -4 L4ýA L-A -1LAr L- -A

0 Aj

-a04

ac 0> ON CN o as 0> U) T-A 1w o o o- C'Vo mr T-4 r-I C3 m r-4 No w wD o Nr-w m m 0m> ~-i "~ m~ m, m~ m, .-i% -i t- -i wO wO a "~ 0 -4 ~-4m c4 v. m m, . 0;aU % r - m~ COw, w > ev4 c13 c4 Ce m m m m m m m m r- m ri w ýw w"w wvvv-w Lwa Lmn-vn-v %o %Dn 0 0 40 Q a 0 0 0 o 04 o 0 00)Q) 40 o 0 fn ~-4M -I U, r- U, -4 M, r-4 U, r-4 U, r-4 M -4 U, r-4U, r-I U, -4 U, .-I 0w -4 0w r-40q r-q 0- .-1 0- ,--40 -40ý -l0 - r-40-40r .-I 0 r-4 T-1 T~ -4 T~.-1 r-4 r-I -4 , -1 .- I -4 T-44 r-44 , -4 4 L. ./)4 -n4 En 4 En 4- En 4 N) g- En EN -4 N ) 4n

S I I U I I I I U I I I I I U I

P101 14 q E 10 U 01 10 4 i

S0 101 101 10 1 10 0 10E 10 10

IB I10I 1E10 '0iE'E'IE-14I1 10 0 L IU IU IU IU IU IU IU IE- E I 4 I ) IU 1 IE-1 r_ IE-IU '0I E-4 10I r_) I I 0L) I I 1 IL)I CI0 I1

tI I I tI I I I i i I i I IU I u I U I u i I I I I UC I UI I I Et I I: C II I I I I I I UI I t Ii: I i : E I I I U II U I U I u I 1 1 1 II 1 1 F : U E E U I I E- I U E- I I

I I I I I I I I I I IIII I I: I I I I I i I Ie I I ItI I ItI I IdI I I I I I I I I I UI I I I IUI I I I I II I I 1I III I 1 I 1 I 1 I IU I I I UI I II I I U H H HULUUA Ab U UUO UAUU ULUU1 U UU LUAUA• O O•

S I I I I I I I I- I E4, 0 E0 I I I I~IE)I 241#00 8 1 1 I I1 I I 141 I I I 1I I 1 rEI EE lZ4 I'Z 4io E-II0 'I 40E-) I E- I0 'q I I

f UIomoo~ooooouooooooooooooo~u~mnuo~mnuo••oso I I 10 1 I I I I I I I I U I I I IUIUIU IU U IU II IUI onoIU oo I U0 u0 I I u I u I II I u uI I u I I III I I U IUI IU IIUI I

I I II I I I I I I I I U I~ IIU I I IUI I II IU II I I I U I U I Ot4 I I I I u I I I I I I I I I u I L)I u I u I u I u I u I U I I I I I I I II III III I I II I I I U II1I II I I IU 1 4 I I I S10 P U E- U E- I > I I I U E-4IL %b1b Il)114E- Il uP xýI B I ICI EI14 U 1 4I E14 Up IxI

I I I 1 1 1 1 1 I II tII l l II I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I L) I IL IU IE I 6 ) I I I I I I I I E-4

0, CN 0> U, T-4 IR N- r-4 C1-4 m r N- -4 V-4 C14 r-I

14 a )Q 4 )a 4 )Q 4 )Q 4 Q 0 0)a 0) 40l 0)) 4) 40) 0L) 0) 0)() 4 U, .-I U, r-i U, -i U, r-4 U, r-4 U, .-i U, r-4U, r-4U, r-4U, r-iU, r-iU, r-i 0>ricýr40) r-I 0> 0) r-4 0 r4 0 I 0. 40C)4 040 co ,4 .4c>ý a)~ iv -I q-4 v-4 r1 -I rA -4r -Irl-4T r1 -4 ,-4 V1 T-4 .- rI-41 m .4 -n4 in .4 tN 4 N .4 N g- En .4 tN gI N) 4 N -4 En 4

-I- D l tm L % r D l l 4 m ll l r O 4l i I I I I rI I I II I I I W - DI I II "

0 4J IIlI O C U•o ao o o m o o o mo •4ry) "Tm tom r-"Tm v 0)- wDL m -m n Ume ' nV0e m %'e mmr-m' mm M m MMMqe0)MCo4r r- LA LA LA) LA LA LA LA 1 . ý0 %0L- k ' I'D Lr-Lko 1-4 ko %Lor- a) LLr-L.o r--r- - S(1 ) (1) 0 (1CD0)0 C), . C- 4 CN- 4-CN -4 CN-4- C.4 r-1r-4 r-4 N- -4 ,-N1-1 .-4 ,--4N 'A- I 1-1N 1- - 1 1 - .1 C1 4L -4 4L C) A -ac-c--C) Cc--lC LA- , - -l

SE I H IE -Eý E EI E I

E-I u I U I U I I I I ' I I U I E-4I E I EH E-IE- ,uu , U ,u o u u

U aE- :U E aE'E- U Um U . oPI ux w4wZx uo o u K Wuo04u 'E-. u f

k F-

a~ U I U 'H C-Uý u auE u H Ua- I aua u E, UE- u 'UE-- E-aE'H aF- -E-aE- P IE a

U P I E- E-ll lE- E-I Ui E- E-I II E-I E I I P P P EI

E- I•u E••• E• E-, I,Ul E E- -4 E- P4 'H Poa E .I -49a--a-9a(9a(IE(Iao-

a U a E- a a E- a a a U J I Pa a ýa ua a a ýa a E-a1 U a a a E-a aý a a IE I Iý i

o U~~I ~ ~ o: o

S aD II D waD waD a) C>a C a) Ca C) a) aD a3 a) a a) a1 a D a)D a)C aa)o a)Q a) a) a) a a) a ) C -1 04 r- " "a- aN" a4 aN a--a -aa- " -4a4 " aq r- "aq -4 N " -4 " -4 aq a- a4 4C - ano- a-1C aý a" a- a- a- al "I a- a- 'Aa- '4 al a- a- a- -1 a4 -1-a- a- -A a4 a- a- 4 ma1r4rI a ) I q - ri -I -I n rAa a -4

ao%D %D %0 ao%D %D aD aD %a wao o wao %a vao ko aD w aD aD aDI' aD aDI' aD waD I' al ko wao 11 %DaD

ý IAýl I-i vAI vI l r- ý4 r4rII- rI r- a- a4 a- a-'1 a4 -Aa. a4a- a- a4a- -4 a- -a4ý a- a- a- al alrIr 0 I 4.3 a a a a a a a a a a a a a a a a a a a a a a a a a a a

Q ,--A --I ý4 ý

M -ODC>oD-A--mmoo XX U U w wa 11 - r- r-O 1 ) -4-- r- 4-4JJ 0 0 a) a) c)O~-i t y-1t0,0Lf n m U U &4 E-4E- E-cE-E- Om- Uuu : UEL-A--1 > tr r-4 un 1- (1) (1)M) CA a4 a-4 1) 1 H H H H H H U U LJ 4J 0 0 .C C: -H H1 MW r0 :4X M L104 L404 a 4 4a al ZHHHZO,, UUU .O-I

~Z 0100I0 I 101IC MER X, OP X, ER X,

I pm UE- I I I

1I0101 908

U U i XH U~~I04 UUOct U OtU cH U < H I:

E- E .I IH H HI-4H pH EIH 1001 u<1 u u~u 0u4;1 E UHUHU-1-- 101011 -1 E-1 I I I I li I E4 E-' HPHIU UU E-4 p 1 S0 0I 1010

UIUUIUUIUUIUUIUUIUIUIUIUIU U U #C U H 'HF- U 'H UU :4 0 0 I I F0 0 U0110 1 if Ii I

IIu U I UI I : I : I I I. I I :)I I U I IU I Ut Q: 91 U QII : : ~ U UUI II UI U I ~ I 4U OctUU U I I I IU'1 I l I lIi I

U I U I I I i E- 1 4EUt/ EU ~r,- ýr, 1I U EUE-I iI I E IiP

I I I I I I I I I I I I 1 I I - I I I I

UI I U I U I U I U I U I UI I I U I U I I I U I I 4I II I I I'I I IUI E U IE U EI I I C I E Ui I E-4I U I I C I U2 I U I I I I0I P E IPI I I UII 10 I i i i I Q I Q I Q g ý g Q Q - Q IZ < n g Q I E : $04I 1 IS 1 U I fI :- : IZD I, I I,

I I II1 I I I II I I I1 ,11U1I I I I0 I I U I I I I I I I III0 I0 I I I IU I I II I I II I I I I I I I I I 1 I0 I 0I I 10 I I I I I I 0 I I I I I I E-II I I I U I I 0 I

I I I I I I < I I I04I I 0 I I III I 0 I I I I11 I0 I I FI I I I I < I I I II

S I I SI I I 1 I U I I I I I I -Ha4HHHHHHHHHHHHHHHHHHHHHHHHHH IaiI I HHHHHHHHHHHH t I O IDI I I1 CI0 HHHHHHHHHHHHHHHH I 1 I e) m coI > U x U L ) a) 3 I I I I ~ l I I I II-ljI

I- r- I- I- ID mI> mI> mI, mI m M C) -' I 0 0 I4 I- I I I-A I I) ) 1- "C I) I) r- r- M M MI4 M I -4 -A r- r- I 4I 0 a) a)

I4 rI IArI rI r- I- rI -4 rI q l VA l IA l VA l Il r- ri I-U- rI rI I4 I- I- rI I 40-,C H HUPUBPUP HPUH H 1>U-4>P 10m(D.a W D 0% CDo> %0 CDDCCCCDk0%D o o I'D>DC k I'Dl >> W1 I'D 0 AJ4- - l oA -v 4tA r-uiA -n i-I 4- 0

I0 Li