International Journal of Salt Lake Research 7: 171–180, 1998. © 1998 Kluwer Academic Publishers. Printed in the Netherlands.

Divergence of Parartemia and Artemia haemoglobin genes M. COLEMAN1, M.C. GEDDES2 and C.N.A. TROTMAN1 1 Biochemistry Department, University of Otago, Dunedin, New Zealand; 2 Zoology Department, University of Adelaide, Adelaide, South Australia 5005

Abstract. Parartemia is a genus of brine shrimp endemic to Australia which is related to the more widespread and economically important Artemia sp. The expression of a multimeric haemoglobin molecule in Artemia is well documented but in Parartemia only trace levels of a possible haemoglobin have been observed. In this paper we describe the DNA sequence of a domain of a haemoglobin molecule in Parartemia. The derived amino acid sequence suggests that the possible date of divergence about 85 million years ago of the two genera predates the divergence of the C and T polymers of Artemia haemoglobin. This date would correlate with the physical and temporal isolation of Australia in the late Mesozoic. Key words: Artemia, divergence, DNA, haemoglobin, Parartemia, protein sequence

Introduction The brine shrimp Artemia is a much studied genus of entomostracan crustacean which is found in both natural and man-made saline lakes throughout the world. It responds to decreased oxygen partial pressure by haemoglobin synthesis. In Artemia the haemoglobin is a multi-domain, multi-subunit molecule which is found free in the haemolymph and is protected from excretion by its large size. The presence of this haemoglobin gives the shrimp a bright red colour. Another genus of brine shrimp, endemic to Australia, is named Parartemia. There are at least eight species of Parartemia (Geddes, 1981) and the genus appears to have existed in Australia for a much longer time than Artemia (Geddes, 1980), being found in many ephemeral salt lakes around the continent, whereas Artemia distribution is restricted to only a few man-made saltworks. The two can be distinguished by the larger size of adult Parartemia and the different shapes of the male second antennae and penes and the female labrum and egg sac (Geddes, 1981). Examination of a common species, P. zietziana, has shown that it generally exists at lower salinities than Artemia and does not show any distinctive

172 colour change when raised at higher salinities. Despite this, P. zietziana was shown by Manwell (1978) to contain a molecule of approximately the same size as the Artemia haemoglobin and having a similar absorption spectrum, although it exists in much smaller quantities (0.01% of haemolymph proteins compared with 1–2 per cent in Artemia). Thus it is believed that Parartemia does retain the gene for haemoglobin production although the protein is only produced in small amounts, and production does not increase in response to environmental stimuli such as an increase in salinity or a reduction in oxygen partial pressure. The haemoglobin molecule of Artemia is a dimer comprising two 9domain polymers (Manning et al., 1990) which associate to form a quaternary structure. Each polymer is encoded by a single gene representing 9 successive globin domains which have different sequences and are presumed to have been copied originally from a single-domain gene (Matthews et al., 1998). Two different polymers called C and T exist as the result of a complete duplication of the 9-domain gene and these different polymers associate to form either of the two homodimers, or the heterodimer. The existence of these 18 different domains provides a unique model in which the process of evolution at the molecular level can be examined, where each domain has evolved under presumably identical selection pressures (Trotman et al., 1994; Jellie et al., 1996). The two polymers of Artemia haemoglobin may have diverged approximately 60 million years ago since they differ from each other by mean values of 15.5 per cent at the DNA level and 11.7 per cent at the protein level (Matthews et al., 1998). Little is known about the date of divergence of Artemia and Parartemia and similarly, little is known about the structure and functionality of the Parartemia haemoglobin. In this paper we show that despite its low level of expression, the Parartemia haemoglobin has a domain structure comparable to that of Artemia. We show that, from sequence considerations, this molecule appears to be no less functional than that of Artemia and that the two genera may have diverged about 85 million years (Ma) ago.

Materials and methods Sample collection Parartemia were collected from ponds at the Penrice Soda Products Dry Creek Saltfields, situated on the east coast of the Gulf of St.Vincent, near Adelaide in South Australia. Although Artemia and Parartemia have been shown to coexist in a small number of these ponds at intermediate salinity, for the most part they exist as separate populations, with Parartemia occupying

173 the ponds of lower salinity (Mitchell and Geddes, 1977). Parartemia were collected from such a pond in the months of July and October 1996 when the mean salinity was 138 parts per thousand and mean oxygen content was 3.6 mg/litre, with a pH of 8.04. Adults were collected by towing horizontally a 160 mm net in areas where the brine shrimp had been concentrated by wind action. These samples were either kept alive to be used directly for DNA extraction in Adelaide or flash frozen in liquid nitrogen and stored at –70 ◦ C for transport to New Zealand for DNA, RNA and protein extraction. RNA isolation and analysis Frozen brine shrimp (3–5 g wet weight) were thawed at 37 ◦ C and total RNA was extracted using TRIzol (Gibco-BRL) according to the manufacturer’s instructions, taking the usual precautions to prevent nuclease contamination. RNA integrity was checked by running samples on a 1.1 per cent formaldehyde gel and RNA size was examined by Northern blotting using an Artemia haemoglobin cDNA probe (Southern, 1975). cDNA library construction Total RNA was used to generate cDNA using a Capfinder cDNA kit (Clontech) according to the manufacturer’s instructions. Briefly this involved using a modified oligo-dT primer, MMLV reverse transcriptase and a patented generic oligonucleotide to the eukaryotic 50 cap to produce first strand cDNA, and then using PCR primers to amplify the cDNA before addition of adapter linkers and cloning into the bacteriophage vector λgt11. Screening of cDNA library One million cDNA clones were screened with heterologous cDNA probes corresponding to both C terminal and N terminal regions of the Artemia haemoglobin gene. Permissive conditions (hybridization and washes at 58 ◦ C) were used to allow for the expected mismatching between the sequences. Sequencing of clones The inserts of the λgt11 clones were amplified by PCR using oligonucleotides designed to flanking vector sequences. The products were directly sequenced using an ABI automated sequencer. Sequence fidelity was verified by sequencing in both directions and using different PCR products as templates.

174 Data analysis and sequence compilation The GCG (Genetics Computer Group Inc) sequence analysis package version 7.3 was used to obtain amino acid sequence by translation of the cDNA sequence, and to determine DNA homologies.

Results and discussion From the initial cDNA library two positive clones were isolated, both of which corresponded to the C-terminal portion of the molecule. In this paper we present the cDNA and the derived amino acid sequence for domain-9 of the P. zietziana haemoglobin (Figure 1). The Parartemia globin sequence is named P, followed by the domain number (e.g. P9), in concert with the numbering of the Artemia T and C sequences. For the region coding for domain P9, the Parartemia cDNA aligns successfully with the equivalent regions of cDNA for both the C and the T polymers of Artemia haemoglobin. The degree of identity between Parartemia and the two Artemia cDNAs was 79.3 per cent between P9 and C9, and 77.5 per cent between P9 and T9 (Figure 1, Figure 3). At the amino acid level after translation, the Parartemia protein was 84.8 per cent identical to the Artemia globin C9 sequence and 81.3 per cent identical to T9. Adopting the molecular clock calibration commonly attributed to globins, namely about 1 per cent divergence at the amino acid level per 5 Ma (Dickerson and Geiss, 1983), the average amino acid difference between Parartemia and Artemia of 17 per cent would equate to a common ancestor about 85 Ma ago. Such calculations are inevitably very approximate, more so in the present case because it is not known whether the multi-domain, two-polymer structure of the brine shrimp globin has imposed additional constraints on the rate at which mutations have accumulated. This consideration, together with the fact that the 17 per cent difference is uncorrected for superimposed mutations (which would add about 1 per cent) points to 85 Ma being a minimum age. The divergence of Parartemia and Artemia (17 per cent protein difference) exceeds by a comfortable margin the difference between the T and C polymers of Artemia (11.7 per cent protein difference), suggesting that divergence of the two genera happened at least 25 Ma earlier than the origin of the T and C genes by duplication of a nine domain ancestral gene. Consistent with this conclusion, only one polymeric globin gene has so far been found in Parartemia. This deduction is not precise, however, since Matthews et al. (1998) have reported clear signs from an analysis of all 18 domain sequences that the molecular clock applicable to Artemia globin genes has run erratically.

175

Figure 1. cDNA sequence from the 30 end of the haemoglobin gene from Parartemia and alignment to the C and T polymers of Artemia.

176

Figure 2. Alignment of the Parartemia haemoglobin domain P9 amino acid sequence to Artemia domains T9 and C9. Sequences of whale myoglobin and Chironomus globin III, of which the crystal structures are known, are shown for alignment. The C9 sequence is identical to T9 except where shown; all other sequences are shown in full. Letters A-H identify the globin structural helices in standard nomenclature, with turns shown as AB, CD etc. Termination codons are indicated by an asterisk.

Pivotal evidence that the divergence of the two genera predated the duplication of the T and C ancestral gene in Artemia comes from an analysis of the amino acid sequences in terms of protein structure. A peculiarity confined to domain 9 alone in Artemia is the deletion of one residue between the first and fourth amino acids in the C-helix (as distinct from the C polymer), such that the consensus -histidine-proline-glutamate-tyrosine- (HPEY) becomes histidine-glycine-tyrosine- (HGY) in both T9 and C9 (Figure 2). This deletion is in the critical haem environment where the proline is practically invariant in vertebrates and invertebrates; its substitution is rare and its deletion exceedingly so (Kapp et al., 1995; Vinogradov et al., 1993). Molecular modelling has confirmed that the distance between these first and fourth amino acids can theoretically be spanned with minimal disruption by the single glycine (Trotman et al., 1991), although it is not known whether domain 9 is still functional. The existence of so rare a deletion in both the T9 and C9 domains of Artemia but not in Parartemia is convincing evidence that Artemia and Parartemia diverged before the deletion happened. Parartemia domain P9 is

177 Domains compared

Length of comparison

DNA % difference

Protein % difference

C9, P9 T9, P9 C9,T9

513 513 513

20.7 22.5 15.5

15.2 18.7 11.7

Figure 3.

itself unconventional, however, in having a highly unusual serine substituted for the anticipated C2 proline. In other respects the Parartemia domain P9 reflects the Artemia globin’s conformity with the universal globin fold whilst sharing its idiosyncrasies. In the following context the letters A to H are the conventional helices of the globin fold, followed by the amino acid position number; bracketed letters are one-letter amino acid code as in Figure 2. Sequence features that key the Parartemia sequence into the generic globin alignment centre on the mandatory F8 histidine (H), being the proximal histidine bonded to haem iron. Also notable are the widely conserved A12 tryptophan (W); the B6 glycine (G) which may assist close packing against the E helix; the E7 distal histidine on the opposite side of the haem iron; and the H8 tryptophan. Tryptophans, being large and rarely mutated, are good alignment landmarks. Features characteristic of Artemia globin domains and confirmed in Parartemia P9 include an unusual hydrophobic leucine (L) at A14 confined to domains 7, 8 and 9; an invariant phenylalanine (F) at B10; an invariant but otherwise unusual tyrosine at C4; an invariant glycine at F5; and a rare occurrence of histidine at G4 in the majority of the domains (invariance refers to the Artemia context). Domain P9 also has a carboxy-terminal extension of 17 residues beyond the end of the H helix which is identical in length and sequence to that of C9 and has only one conservative isoleucine/valine (I/V) substitution compared with T9. The linker sequence between the domain-8 H-helix and the domain-9 A helix, designated HA1-HA16 in Figure 2, is also notably identical between P9 and C9, with T9 having two differences; leucine for isoleucine is conservative, whilst isoleucine for asparagine (N) is not. In seeking to confirm that only one class of globin gene exists in Parartemia compared with the two (C and T) in Artemia, a number of different Parartemia clones have been partially sequenced. A maximum of 2 per cent difference has been found, in contrast to the 11.7 per cent difference between C and T. This level of difference is considered most likely to represent polymorphism. However, the possible existence of a second Parartemia gene or a pseudogene has not yet been ruled out.

178 In Australia Artemia exists only in a few man-made saltworks at coastal positions whilst Parartemia is found in many ephemeral salt lakes in the interior, giving rise to the idea that Artemia is a relative newcomer to the continent (Geddes, 1980). It is interesting that the proposed date of divergence (85 Ma) for Parartemia, a species endemic to Australia, and Artemia, a species with a world wide distribution, coincides with the timescale for the geological separation of Australia from New Zealand and Antarctica in the late Mesozoic (Keast, 1972; Hayes, 1973). Thus the geological evidence does not conflict with a date of 85 Ma for the separation of the two genera, which in turn provides some support for the accuracy of the figure of 5 Ma per 1 per cent amino acid divergence. Conclusions Manwell (1978) showed that the putative Parartemia haemoglobin had an electrophoretic mobility very similar to that of Artemia haemoglobin, which has nine domains. The present data show that Parartemia does express a gene for a haemoglobin molecule with a domain structure and sequence resembling that of Artemia. Taken together this suggests that Parartemia contains a haemoglobin molecule similar in size and structure to that of Artemia. In Artemia each domain has a haem-binding capacity (Moens, 1982), and the haem-binding motif is evidently conserved except possibly in domains C9 and T9 where one key proline is missing. In domain P9 of Parartemia this amino acid has not been lost, strongly indicating that its loss post-dates the divergence of the two genera. However, the highly unusual substitution in Parartemia of serine for proline invites the speculation that the possibly unique deletion of the residue altogether was no less tolerable – perhaps more tolerable – than the accommodation of an unconventional serine. In Parartemia the complement of haemoglobin present is a hundred-fold less than in Artemia but it is not known whether this is controlled at the transcriptional, post-transcriptional or translational level. It may be that the control differs at the DNA level and that a signal that up-regulates transcription of the haemoglobin gene in Artemia is not recognized in Parartemia, resulting in only a basal level of expression. Alternatively, the RNA transcript or the nascent protein may be quickly degraded. The function of the haemoglobin molecule in Parartemia is a matter of speculation. Artemia haemoglobin has well-characterized oxygen-carrying parameters (Moens et al., 1991) and the Parartemia molecule is clearly almost identical structurally. On the other hand it exists in such small amounts in Parartemia as to make its contribution as an oxygen carrier questionable (Manwell, 1978). The cDNA and derived amino acid sequences indicate

179 that it possesses a viable haem-binding motif which supports the idea that a minimal amount of the molecule is still necessary for haem transport. It was suggested by Manwell (1978) that Parartemia sacrifice the ability to make large quantities of haemoglobin because of the high energy saving this would yield, allowing them to compete successfully with Artemia in nutrientlimiting environments. Large scale production of a nine-domain polymer such as the haemoglobin of Parartemia would indeed have high energy costs for the organism. That this structure has been retained may indicate that the divergence of the two genera has been accompanied by a divergence of haemoglobin functions.

Acknowledgments A research grant from the Marsden Fund, Royal Society of New Zealand, is gratefully acknowledged. We thank Professor Douglas S. Coombs for valuable discussion about continental movements.

References Dickerson, R.E. and Geiss, I. 1983. Hemoglobin: structure, function, evolution, and pathology. Benjamin Cummings, Menlo Park, pp. 97. Geddes M.C. 1980. The brine shrimps Artemia and Parartemia in Australia. In: G. Persoone, P. Sorgeloos, O. Roels and E. Jaspers (Eds) The Brine Shrimp Artemia. Vol 3. Ecology, Culturing, Use in Aquaculture, pp. 57–65. Universa Press, Wetteren, Belgium. Geddes M.C. 1981. The brine shrimp Artemia and Parartemia. Hydrobiologia 81: 169–179. Hayes D.E. 1973. Seafloor spreading in the Tasman Sea. Nature 243: 454–458. Jellie, A.M., Tate, W.P. and Trotman, C.N.A. 1996. Evolutionary history of introns in a multidomain globin gene. Journal of Molecular Evolution 42: 641–647. Kapp, O.H., Moens, L., Vanfleteren, J., Trotman, C.N.A., Suzuki, T. and Vinogradov, S.N. 1995. Alignment of 700 globin sequences: extent of amino acid substitution and its correlation with variation in volume. Protein Science 4: 2179–2190. Keast A. 1972. Continental drift and the evolution of the biota on southern continents. In: A. Keast, F. C. Erk and B. Glass (Eds) Evolution, Mammals and Southern Continents, pp. 23–87. State Univerity of New York Press, Albany, U.S.A. Manning, A.M., Trotman, C.N.A. and Tate, W.P. 1990. Evolution of a polymeric globin in the brine shrimp Artemia. Nature 348: 653–656. Manwell, C. 1978. Haemoglobin in the Australian anostracan Parartemia zietziana: evolutionary strategies of conformity vs regulation. Comparative Biochemistry and Physiology 59A: 37–44. Matthews, C.M., Vandenberg, C.J. and Trotman, C.N.A. 1998. Variable substitution rates of the 18 domain sequences in Artemia hemoglobin. Journal of Molecular Evolution, 46: 729–733

180 Mitchell, B.D. and Geddes, M.C. 1977. Distribution of the brine shrimps Parartemia zietziana Sayce and Artemia salina (L.) along a salinity and oxygen gradient in a South Australian saltfield. Freshwater Biology 7: 461–467. Moens, L. 1982. The extracellular haemoglobins of Artemia sp.: a biochemical and ontological study. Mededelingen van de Koninklijke Academie voor Wetenschappen, Letteren en Schone Kunsten van Belgie, Klasse der Wetenschappen 44: 1–21. Moens, L., Wolf, G., Van Hauwaert, M-L., De Baere, I., Van Beeumen, J., Wodak, S. and Trotman, C.N.A. 1991. The extracellular hemoglobins of Artemia. Structure of the oxygen carrier and respiration physiology. Chapter 8. In: R.A. Browne, P. Sorgeloos abd C.N.A. Trotman (Eds) Artemia Biology, pp. 187–219. CRC Press. Southern, E.M. 1975. Detection of specific sequences among DNA fragments separated by gel electrophoresis. Journal of Molecular Biology 98: 503. Trotman, C.N.A., Manning, A.M., Moens, L. and Tate, W.P. 1991. The polymeric haemoglobin molecule of Artemia. Interpretation of translated cDNA sequence of nine domains. Journal of Biological Chemistry 266: 13789–13795. Trotman, C.N.A., Manning, A.M., Bray, J.A., Jellie, A.M., Moens, L. and Tate, W.P. (1994) Inter-domain linkage in the polymeric hemoglobin molecule of Artemia. Journal of Molecular Evolution 38: 628–636. Vinogradov, S.N., Walz, D.A., Pohajdak, B., Moens, L., Kapp, O.H., Suzuki, T. and Trotman, C.N.A. 1993. Adventitious variability? The amino acid sequences of nonvertebrate globins. Comparative Biochemistry and Physiology 106B: 1–26.