Construction of Zn2+/Cd2+ Hypersensitive Cyanobacterial Mutants Lacking a Functional Metallothionein Locus*

Eukaryotic metallothioneins (MTs) have been exten- sively studied, but the precise functions of most of these molecules are not yet fully understood. Prokaryotes are often more tractable for the analysis of gene func- tion and we report here the generation of mutants of Synechococcus PCC 7942 (strain R2-PIMS) deficient in the MT locus, smt. Viability of these mutants, des- ignated R2-PIMS(smt), reveals that prokaryotic MT performs no "vital" role (such as donation of metals to metallo-proteins) in Synechococcus. Ra-PIMS(smt) has reduced (-5-fold) tolerance to elevated Zn2+, with detectable hypersensitivity to Cd2+. Restoration of Zn2+ tolerance was used as a selectable marker to isolate recombinants derived from R2-PIMS(smt) after rein-troduction of a linear DNA fragment containing an uninterrupted smt locus. These smt-complemented cells also exhibited restored Cd" tolerance. Hypersen- sitivity to Cu"+ was not detected in R2-PIM8(smt) indicating independence of Cu2+ resistance from smt mediated metal (Zn"'/Cd2+) tolerance. A role proposed for MTs' in all of the organisms in which they have been detected is the sequestration of excess amounts of certain metal ions. The specific metals sequestered by MTs vary for the structurally distinct proteins/polypeptides occur-ring

A role proposed for MTs' in all of the organisms in which they have been detected is the sequestration of excess amounts of certain metal ions. The specific metals sequestered by MTs vary for the structurally distinct proteins/polypeptides occurring in different organisms (reviewed in Kagi (1991)). Several lines of evidence also suggest other functions for some MTs, in particular those in higher eukaryotes where the inducing, and co-ordinated, metal ions include Zn2+.
Zn2+ associated with animal MT is highly labile, a necessary attribute for an intracellular Zn2+ donor (cited in Bagi (1991)). Donation of Zn2+ to some apo-enzymes has been demonstrated in vitro (cited in Zeng et al., 1991b). In addition thionein (apo-MT) can inactivate the ZnZf requiring transcription factor Spl (human) and also acquire Zn2+ from Xenopus laevis transcription factor IIIA in vitro (Zeng et al., 1991a(Zeng et al., , 1991b. It has thus been proposed that modulation of thionein biosynthesis, or intracellular distribution, could af-* This work was supported in part by Research Grant GR3/7883 from the United Kingdom Natural Environment Research Council. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisernent" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
$ Supported by a studentship from the United Kingdom Natural Environment Research Council.
§ A Royal Society University Research Fellow. To whom correspondence should be addressed. Tel.: .
Lower eukaryotic (fungal) MTs bind copper in vivo and their synthesis is regulated by copper. The high frequency of homologous recombination in Saccharomyces cerevisiae has been exploited to construct strains deficient in the MT gene, CUPI, facilitating detailed analysis of M T function in this organism (Thiele et al., 1987). CUPl performs no essential role(s) required for cell growth, differentiation, or normal Cu2+ metabolism. CUPl mutants grow with normal doubling times in standard low Cu2+ media and are capable of mating, diplophase growth, sporulation, germination, accumulation of copper, and accumulation/activation of a copper requiring enzyme, copper-dependent superoxide dismutase (Thiele et al., 1987). However, MT-deficient S. cerevisiae are hypersensitive to elevated concentrations of Cu2+.
A number of studies have indicated that some prokaryotes contain MT-like proteins (cited in Silver and Misra (1988)) but only a Synechococcal M T sequence has been reported (Olafson et al., 1988). This MT increased in abundance following exposure of cyanobacterial cells to elevated concentrations of either Cd" or Zn2+, but not Cu2+, and the purified protein was associated with either Cd2+ or ZnZf (dependent upon the metal administered to the cells) with copper as a minor component (Olafson et al., 1980). Polymerase chain reaction fragments corresponding to the prokaryotic M T gene have been generated from Synechococcus PCC 6301 genomic DNA (Robinson et al., 1990), and we have recently cloned and structurally characterized the M T locus, designated smt, from Synechococcus PCC 7942 (Huckle et al., 1993). The locus includes the MT gene, smtA, and a divergent gene, srntB, which encodes a trans-acting repressor of transcription from the smtA operator-promoter (Huckle et al., 1993). The abundance of smtA transcripts increase in response to elevated concentrations of a range of trace metal ions (Huckle et al., 1993). Gene-fusion experiments indicate that the smtA operator-promoter is most responsive to maximum permissive concentrations of Zn2+ compared to other metal ions (Huckle et al., 1993). Following expression of SmtA as a recombinant fusion protein in Escherichia coli, the pH of half dissociation of different metal ions indicated a particularly high affinity for Zn2+ in comparison to mammalian MT (Shi et al., 1992). Amplification (Gupta et al., 1992) and specific rearrangement (Gupta et al., 1993) of the smt locus has been detected in Synechococcus PCC 6301 cells selected for growth in elevated concentrations of Cd".
We report the generation of smt-deficient mutants via insertional inactivation/partial gene deletion and phenotypic analysis of these cells.

EXPERIMENTAL PROCEDURES
Materials and Methods-R2-PIM8, a small plasmid cured derivative of Synechococcus PCC 7942 (van der Plas et al., 1990), was cultured under constant light (100 pmol of photon m-' s-' photosynthetically active radiation) at 32 "C in Allens medium (Allen, 1968) with 0.012 g liter" citric acid, the omission of NapSiO3.9HzO and supplemented with 30 pg of DL-methionine ml-I, 5 pg of streptomycin ml-', and 7.5 pg of chloramphenicol ml" as appropriate. Solidmedium used for plating contained 1.5% agar (Bacto-Agar, Difco Laboratories). Modified Allens medium contains 0.77 p~ Zn2+ and 0.32 pM Cu2+. E. coli DH5a was used as a host in all recombinant plasmid constructions, transformed cells were grown on LB medium (Sambrook et al., 1989) supplemented with 100 pg ml" carbenicillin, or 34 pg ml" chloramphenicol as necessary.
Restriction enzymes were supplied by Northumbria Biologicals Ltd., Cramlington, United Kingdom (UK)or Boehringer Mannheim Ltd., East Sussex, UK. Taq polymerase was supplied by Stratagene, Cambridge, UK, or Promega Ltd., Southampton, UK. [a-"PIdCTP (14.8 TBq mmol") and nylon (Hybond N) filters were obtained from Amersham International, Aylesbury, UK. Phosphoramidite derivatives of all nucleotide bases were obtained from Applied Biosystems, Warrington, UK. Other chemicals and antibiotics were purchased from Sigma, Dorset, UK.
DNA probes were prepared from an smtA polymerase chain reaction product (described by Robinson et al., 1990) and from restriction fragments derived from pJHNR49 (a SalI/HindIII genomic smt fragment in the vector pGEM4z, described by Huckle et al., 1993), radiolabeled with [cY-~'P]~CTP according to the procedure of Feinberg and Vogelstein (1983).
Plasmid sequencing was performed by the dideoxy-sequencing method of Sanger et al. (1977) using an Applied Biosystems 370A DNA sequencer, as described in the supplier's protocol (Model 370A DNA sequencing system; User's Manual Version 1.3A, October 1988, Interruption of the smt Locus-The scheme for inactivation of the smt locus is given in Fig. 2. R2-PIM8 was transformed with linearized plasmid pRECSU which consists of smt flanking sequences interrupted by E. coli plasmid pSU19 (a derivative of pSU2719 described by Martinez et al. (1988)) containing the chloramphenicol acetyl transferase gene (cat). The smt flanking sequences in pRECSU, obtained from pJHNR49, include a 340-bp HindIII/SacI 5' fragment and a 1073-bp PstI/SalI 3' fragment, cloned into the SmaI/SacI and PstIISalI sites, respectively, of the pSUl9 polylinker. XbaI was subsequently used to linearize pRECSU prior to transformation.
Plasmid DNA isolated by alkaline lysis was used to transform R2-PIM8 to Cm' essentially as described by van den Hondel et al. (1980). Transformants were selected on Allens agar plates containing chloramphenicol. Individual Cm' transformants were serially streak-purified at least three times to ensure segregation of homozygous mutants.
A 1775-bp SalIlHindIII smt fragment from pJHNR49 ( Fig. 2 A ) was used to complement pRECSU transformants, and transformants were selected on Allens agar plates supplemented with 20 p~ Zn2+.
DNA Isolation and Southern Analysis-Genomic DNA was isolated from liquid cultures in late logarithmic to early stationary phase using part of a protocol described previously for the isolation of nucleic acids from plant cell cultures (Robinson et al., 1988), but excluding CsCl gradients. Genomic Southern blotting was performed using 10 or 20 pg of DNA for each restriction digest, followed by standard agarose gel electrophoresis, and transfer to nylon filters. Filters were washed to a final stringency of 0.5 X SSC, 1 g liter" SDS a t 65 "C (Sambrook et al., 1989).
RNA Isolation and Northern Analysis-Liquid cultures were grown to mid-logarithmic phase before incubation for 2 h in the presence or absence of 2.5 p~ Cd". Total nucleic acids, isolated using standard techniques (Dzelkalns et al., 1988), were analyzed on 15 g liter" agarose gels, and visualized with ethidium bromide to allow approximate quantitation. Equivalent amounts of RNA from different cell extracts were denatured with formamide a t 65 "C and electrophoresed in a 15 g liter" formaldehyde agarose gel. The size-separated RNA was then transferred to nylon filters by standard techniques (Sambrook et al., 1989), hybridized to smtA probe and washed to 0.5 X SSC, 1 g liter" SDS at 65 "C.
Recovery of Integrated Plasmid from pRECSU Tramformants-Genomic DNA isolated from pRECSU transformants was digested to completion with SalI to release a -7.8-kb fragment containing integrated plasmid pSU19. Total restricted DNA was ligated a t low DNA concentrations (50.01 pg pl-') to favor circularization, and used to transform E. coli to Cm'.
or CuCI2 under standard growth conditions to determine the minimum inhibitory/maximum permissive concentrations of these metal ions. Cells were inoculated at a density of 1 X lo6 cells ml" and monitored for growth daily, for 14 days. Subsequent experiments quantified the effects on growth of selected (from the previous experiment) concentrations of metals as a function of time by measuring optical density of the cultures at 540 nm. In all cases the assay was carried out in triplicate and then repeated on a t least two further occasions (at least nine analyses).

RESULTS
Evidence of Chromosomal Localization of the smt Locus-Synechococcus PCC 7942 has two plasmids of -8.0 and 48.5 kb (Lau and Dolittle, 1979;Laudenbach et al., 1983). Southern analysis of R2-PIM8 DNA probed with smtA identified the gene on unique SalI, HindIII, and BamHI fragments (Fig.   lA), confirming its presence in the small plasmid-cured strain.
No hybridization was detected following probing of pPLANB2 (an E. coli plasmid carrying the 8.0-kb plasmid of Synechococcus PCC 7942 (Laudenbach et al., 1983) with smtA (data not shown)). pPLANBa (plasmids 1 to 7), carry seven BamHI restriction fragments of the 48.5-kb plasmid (Laudenbach et al., 1983). There was only faint hybridization to pPLANBa2 (data not shown). Hybridization to PstI digested pPLANBa2 is weak relative to an equivalent number of copies of pJHNR49 and a 213-bp PstI smtA fragment was only detected in pJHNR49 (Fig. 1B). Probing with the weakly hybridizing -2.0-kb fragment from pPLANBa2 identified the homologous sequence in R2-PIM8, in addition to weak cross-hybridization to smtA (data not shown).
Interruption of the smt Locus-The plasmid pRECSU (Fig.  2), containing smt flanking sequences interrupted by E. coli plasmid pSU19, was generated. The smt flanking sequences in pRECSU are separated by 371-bp in the smt locus which includes the smtA and smtB transcription/translation start sites and promoter sequences, which were therefore absent from pRECSU. The orientation of fragments in pRECSU were confirmed by sequencing using both forward and reverse M13 primers. R2-PIM8 was transformed to Cm' with linearized pRECSU. Stable Cm' transformants were selected on Allens agar plates containing 7.5 pg ml" chloramphenicol. After several rounds of streaking to isolate cells homozygous for an interrupted smt locus, a number of transformants were grown in liquid culture and then plated to obtain single colonies.

FIG. 2. Insertional inactivation of the smt locus. A, a 1775-
bp HindIII/SalI fragment of Synechococcus PCC 7942 genomic DNA including the 168-bp protein coding region of smtA (narrow diagonal shading) and the 366-bp protein coding region of smtB (wide diagonal shading) was obtained from a size-fractionated SalI/HindIII genomic library. A 1073-bp 3' smt flank (long arrow) and a 340-bp 5' smt flank (short arrow) were cloned into pSU19 to generate pRECSU. The orientation of the cloned smt sequences resulted in the interruption by the vector DNA (including cat) of the Synechococcus PCC 7942 smt flanking sequences. The smt flanking sequences in pRECSU are separated by 371 bp in the smt locus which includes the smtA and smtB transcription/translation start sites and promoter sequences, which are therefore absent from pRECSU. XbaI was used to linearize pRECSU prior to transformation of R2-PIM8. B, the upper diagram represents the smt locus of R2-PIM8. Transformation with linearized pRECSU should result in homologous recombination with cat directed into the smt locus (lower diagram).

-~
The structure and homozygosity of pRECSU transformants, hereafter referred to as R2-PIM8(smt), were confirmed by Southern analyses (Fig. 3). A -5.8-kb SalI smt fragment in R2-PIM8 DNA was not detected in RP-PIM8(smt) DNA probed with part of the diagnostic-deletion region. Upon prolonged exposure a faint band was visible at -7.8 kb in R2-PIM8(smt) DNA, due to weak cross-hybridization to pSU19 (data not shown).
A -0.9-kb PstI fragment containing smtB in R2-PIM8 DNA and a -3.1-kb fragment in RZ-PIM8(smt) DNA was detected upon probing with retained smtB sequences (Fig.  3B). The latter corresponds to the anticipated size of the smt locus containing pSU19 with concomitant deletion of a 371bp region. Probing with plasmid pSU19 identified -3.1-kb PstI, -2.7-kb HindIII, and -7.8-kb SalI fragments in R2-PIM8(smt) DNA as expected, confirming the site of integration of pRECSU (Fig. 3C). An anomalous additional band was thought to result from homology between the probe and other sequences within the genome.
Plasmid Recovery-A -7.8-kb plasmid (pJSTNR4.1) was generated from a SalI fragment of R2-PIM8(smt) DNA by plasmid recovery in E. coli. The restriction pattern (Fig. 4) is that expected for digestion of DNA containing the interrupted srnt locus. The smt-derived regions of pJSTNR4.1 were se- quenced using the M13 forward and reverse primer sites within pSU19 and were identical to the known genomic sequences, further confirming the correct site of integration of pRECSU.
Analysis of smtA Transcript Abundance-smtA transcripts were only detected in RNA isolated from Cd2+ exposed R2-PIM8 and not in R2-PIM8(smt) (Fig. 5), consistent with the smt mutant status of the latter. Following culture of these cells in the presence of 1.5 PM Cd2+, Southern analysis confirmed genetic stability of RP-PIM8(smt) with no reversion detected upon metal exposure (data not shown).
Phenotypic Analysis of RS-PIM8(smt)--The proportion of cultures of R2-PIM8 and RS-PIM8(smt) growing in Allens medium supplemented with a range of concentrations of CdC12, ZnC12, and CuCI2 were monitored, and minimum inhibitory/maximum permissive concentrations of these metal ions were determined for both strains. Growth of R2-PIM8 and R2-PIM8(smt) was subsequently examined as a function of time in response to the selected metal concentrations (Fig.  6, A, B, and C). R2-PIM8 survived in -5-fold higher concentrations of Zn2+ than RP-PIMS(smt). A higher tolerance to  2 ) and RZ-PIM8(smt) (lanes 3 and 4 ) incubated for 2 h under standard growth conditions in the absence (lanes I and 3 ) or presence (lanes 2 and 4 ) of 2.5 p~ Cd", was electrophoresed on a 1.5% agarose gel containing formaldehyde, transferred to a nylon filter and probed with smtA. B, visualization of the rRNA bands indicating the quantity of RNA in each lane.
Cd2+ was observed for R2-PIM8 in comparison to R2-PIM8(smt), however this was only detected after a prolonged lag of >148 h (Fig. 6B). There was no marked difference in the minimum inhibitory concentration of Cu2+, and similar growth rates were observed for both R2-PIM8 and R2-PIM8(smt) at selected Cu2+ concentrations (Fig. SC).
The structure and homozygosity of the smt-complemented cells was confirmed by Southern analysis, using the 213-bp PstI diagnostic-deletion region as a probe (Fig. 8). The restriction pattern was as observed for R2-PIM8 (Fig. lA), confirming reintegration of a functional smt locus. Fragments hybridizing to pSU19 in R2-PIM8(smt) DNA (Fig. 3C) were not detected in smt-complemented cells (data not shown).
Growth of smt-complemented cells in Allens medium supplemented with increasing levels of CdC12, ZnCl2, and CuC12 was identical to that observed for R2-PIM8 (data not shown).

DISCUSSION
Previous circumstantial evidence suggesting a role for prokaryotic MT in Zn2+-homeostasis includes; an observed high Zn2+-affinity (relative to equine MT) of a recombinant GST-SmtA fusion protein following expression in E. coli (Shi et al., 1992); MT induction by, and association with, Zn2+ in Synechococcus sp. (Olafson et al., 1980); an observed maximal induction of expression from the smt operator-promoter in response to Zn2+ (Huckle et al., 1993). Here we describe the production of R2-PIM8(smt) mutants which lack a functional smt locus and show reduced (-5-fold) tolerance to Zn2+.
smt-mediated restoration of Zn2+ tolerance can be used as a selectable marker for transformation of R2-PIM8(smt). Zn2+-hypersensitive R2-PIMS(smt) was transformed to normal Zn2+ tolerance with a linear DNA fragment containing the smt locus. All of the resulting Zn' colonies exhibited restored Cms, growth curves indicated Zn2+ and Cd2+ tolerance characteristics identical to R2-PIM8, and Southern analyses confirmed reintegration of the functional smt locus with concomitant loss of pSU19. Thus, the Zn2+-and Cd2+-hypersensitive phenotype of RB-PIMS(srnt) was solely due to loss of the smt locus. I n vitro Zn2+ transfer between transcription factors and higher eukaryotic apo-MT has implicated the latter in Zn2+ homeostasis as it relates to the regulation of gene expression (Zeng et al., 1991a(Zeng et al., , 1991b. Zn2+-requiring transcription factors are not well characterized in prokaryotes, and it has been proposed that prokaryotes may have avoided the "hidden costs" of the precise Zn2+ homeostasis required for maintaining Zn2+-binding transcription factors (Luisi, 1992). However, intracellular "Zn2+ buffering" remains a requirement of these organisms, and reduced tolerance of RP-PIM8(smt) to elevated Zn2+ reveals such a function for the SmtA protein.
There was no marked difference in Cu2+ tolerance of R2-PIM8 and RZ-PIM8(smt). Olafson (1986), observed a protracted growth lag in cells exposed to elevated Cu2+ but no coincident MT synthesis. Energy-dependent copper efflux has been proposed as an alternative mechanism of Cuz+ resistance in Synechococcus sp. (Olafson, 1986) and in another cyanobacterium, Nostoc calcicola (Verma and Singh, 1991).
In conclusion, these data show that the prokaryotic MT locus, smt, is involved in Zn2+ and Cd2+ detoxification, but performs no vital role required for growth in non-metalsupplemented media. The mechanism of smt-mediated metal detoxification is unknown. By analogy to eukaryotic MTs, SmtA may serve as an intracellular "sink" for excess metal. The possibility that SmtA may be part of a more dynamic mechanism of metal homeostasis (e.g. involving enhanced metal ion efflux) is also the subject of continuing investigation.