Gene, 137 (1993) 2655270 0 1993 Elsevier Science Publishers B.V. All rights reserved. 0378-I 119/93/$06.00 265

GENE07500

Isolation and nucleotide sequences of the genes encoding killer toxins from Hunsenula mrakii and H. saturnus

(Recombinant DNA; secretion; signal sequence; KEX2 protease; killer spectrum; killer yeast)

Tetsuya Kimura”, Noriyuki Kitamotoa, Ken Matsuokab, Kenzo Nakamurab, Yuzuru Iimurac and

Yukio Kite”

“Food Research Institute, Aichi Prefectural Government, Nishi-ku, Nagoya 451, Japan: bLaboratory of Biochemistry, School of Agriculture, Nagoya University, Chikusa-ku, Nagoyu 464-01, Jupan; and “National Research Institute of Brewing, Kita-ku, Tokyo 114, Japan

Received by A. Nakazawa: 30 April 1993; Revised/Accepted: 16 July/26 July 1993; Received at publishers: 9 August 1993

SUMMARY

The HMK gene, encoding a killer toxin (HMK) of Hansenula mrakii strain IF0 0895, and the HSK gene, encoding

a killer toxin (HSK) of H. saturnus strain IF0 0117, were cloned and sequenced. The HMK and HSK genes encode

precursors to killer toxins of 125 amino acids (aa) and 124 aa, respectively. Both precursors have an N-terminal signal

sequence of 37 aa which may be removed by a signal peptidase, and a propeptide which may be cleaved off by a KEX2-

like protease. There is extensive homology between the aa sequences of HMK and HSK with the exception of the

addition of one aa residue in HMK. The HMK and HSK genes were placed, separately, downstream from the yeast

GAL10 promoter and introduced into a mutant of Sacchuromyces cerevisiue that was resistant to the HMK. The

transformants were capable of killing sensitive yeasts in medium that contained galactose with killing spectra similar

to those of the donor strains of the toxins. These observations suggest that both killer toxins were synthesized and

secreted from S. cerevisiue cells and killed sensitive yeasts, perhaps by the same mechanism as that associated with the

donor strains and, moreover, that the difference in primary structure between the two toxins is responsible for the

difference in their killing spectra.

INTRODUCTION

Killer yeasts secrete into their culture medium polypep-

tides known as killer toxins that kill sensitive strains of

yeast. Many of the killer yeasts belong to the genus

Correspondence to: Dr. T. Kimura, Food Research Institute, Aichi

Prefectural Government, Nishi-ku, Nagoya 451, Japan. Tel. (81-52) 521-

9316; Fax (81-52) 532-5791.

Abbreviations: aa, amino acid(s); bp, base pair(s); H., Hansenula; HMK, killer toxin(s) of H. mrakii; HMK, gene encoding HMK;

HSK, killer toxin(s) of H. saturnus; HSK, gene encoding HSK;

kb, kilobase or 1000 bp; nt, nucleotide(s); oligo, oligodeoxyribo-

nucleotide; ORF, open reading frame; PAGE, polyacrylamide-gel elec-

trophoresis; S.. Saccharomvces; SDS. sodium dodecyl sulfate;

SSC, 0.15 M NaCl/O.OlS M Na,citrate pH 7.6; YPD, 2% glucose/2%

peptone/l% yeast extract.

Hansenulu. Yeasts in this genus, such as H. mrukii,

H. suturnus and H. beijerinckii, produce killer toxins that

act on various species of yeast, and they are immune to

other Hansen&z killer toxins. The HMK of H. mrakii

IF0 0895 (HM-1; Yamamoto et al., 1986a) and the HSK

of H. saturnus IF0 0117 (HYI; Ohta et al., 1984) have

been purified previously. Their molecular masses and aa

compositions are very similar. Furthermore, these two

killer toxins show higher thermostability and wider pH-

stability than other killer toxins. These findings suggest

that HMK and HSK are derived from the same ‘ancestral

killer toxin’. Although the aa sequence of HMK and the

molecular mechanism by which this killer toxin kills

sensitive yeasts have been described (Yamamoto et al.,

1986b), information about the killer/immune systems of

these killer strains is limited. To study the structural and

266

functional relationship between the toxins, we have

cloned the HMK and the HSK genes and compared their

structures.

EXPERIMENTAL AND DISCUSSION

(a) Construction and screening of genomic libraries of

H. mrakii and H. saturnus Synthetic oligo probes were designed by reference to

the previously reported aa sequence of HMK (Yamamoto

et al., 1986) and they were used for Southern blot hybrid-

ization and screening of a genomic library. Genomic

Southern blot hybridization with the 32P-labeled probes

yielded only one band with HindIII, EcoRI and BumHI

digests (data not shown). A phage h genomic library was

screened with these probes, and seven positive clones

were obtained with overlapping patterns of restriction

sites. All of the positive clones contained the same 3.6-

kb Hind111 fragment, which was identical in size to a

Hind111 fragment identified by the genomic Southern blot

hybridization. One of the clones, hHMK08, was further

characterized. A detailed restriction map of the 3.6-kb

Hind111 fragment in hHMK08 is shown in Fig. 1A. The

nucleotide (nt) sequence of a 2.4-kb EcoRI-Hind111 frag-

ment that included the gene encoding HMK was deter-

mined (Fig. 2a).

While HSK has already been purified and charac-

terized (Ohta et al., 1984) its aa sequence has not been

reported. The killer spectrum and other features of HSK

are very similar to those of HMK, suggesting that the

primary structure of HSK may be very similar to that of

HMK. Indeed, two probes used in the screening for

HMK gave only one band upon hybridization with

HindIII, EcoRI and BarnHI digests of genomic DNA

from H. saturnus genomic DNA. Therefore, these probes

were used to screen a phage h genomic library. Six posi-

tive clones were obtained with overlapping patterns of

restriction sites, and all of the positive clones contained

a 4.0-kb EcoRI fragment that hybridized with the probes

and was identical in size to the EcoRI fragment identified

by genomic Southern blot hybridization. One of the

clones, hHSK03, was further characterized. A detailed

restriction map of the 4.0-kb EcoRI fragment and the nt

sequence of a 2.3-kb EcoRI-BumHI fragment that in-

cludes the HSK gene are shown in Fig. 1B and 2b,

respectively.

(b) Sequence analysis of HMK and HSK

The nt sequence of the 2.4-kb EcoRI-Hind111 fragment

of hHMK08 contained an ORF (Fig. 2a; from nt 1

through 375) capable of encoding a polypeptide of 125 aa.

B

1.0 kb

Fig. 1. Restriction maps of genomic clones around the HMK and HSK

regions. Restriction map around the HMK coding region of a Hind111

fragment in hHMK08 (A) and the HSK coding region of an EcoRI

fragment in AHSK03 (B) were analyzed. Black boxes and hatched boxes

indicate the sequenced regions and the ORFs, respectively. Methods:

Phage AEMBL3 genomic libraries of H. mrukii IF0 0895 and

H. suturnus IF0 0117 were constructed from genomic DNAs that had

been partially digested with Srru3AI and were screened by hybridization

with 3’P-labeled synthetic probe I (S-GT(G/A)TTCCA(G/A)TT(C/T)-

TTG(C/T)TTCA-3’, where (G/A) represents G plus A mix and (C/T)

represents C plus T mix, a mixture of 16 different 20-mer ohgos corre-

sponding to the anticodon strand of the aa sequence from Trp’” through

ThrZh in HMK), and probe 2 (5’-ACCATCCA(G/A)TGNAC(G/A)TT-

3’, where N represents a mix of all four deoxynucleotides, a mixture of

16 different 17-mer oligos that corresponded to the anticodon strand

of the aa sequence from Asn3’ through Val-‘” in HMK), in

6 x SSCj5 x Denhardt’s solution/l % SDS/250 ug per ml of denatured

salmon sperm DNA, at 47°C and 41’C for probe 1 and probe 2.

respectively.

The deduced aa sequence contained a sequence (from

Gly3’ through Lys”‘) that was identical with the directly

obtained aa sequence of HMK (Yamamoto et al., 1986a).

In fungi, the consensus sequence around the AUG start

codon for translation of mRNA was reported to be

UCA(C/A)(A/C)AUG(G/U)C (Ballance, 1986). The se-

quence around the first ATG codon in the ORF,

CCAACATGAA, was similar to the consensus sequence

of the site for initiation of translation in fungi. In addition,

a translational termination codon, TAG (nt - 192 to

- 190), was detected in the region upstream from the

putative site for initiation of translation, ATG, in frame.

Therefore, the first ATG codon in the ORF may be used

as an initiation codon for translation. A sequence,

TATATAAA (nt - 111 to - 104), with some resemblance

to a TATA box (Breathnach and Chambon, 198 1; Chen

and Struhl, 1988) was detected in the region upstream

from the putative initiation site of the ORF. Analysis of

the region downstream from the ORF revealed the pres-

ence of a tripartite terminator-like sequence (Zaret and

267

-1566

-1560

-1440

-1320

-1200

-1060

-960

-840

-720

-600

-460

-360

-240

-120

1 1

121 41

241 81

TACCTGATTATGTGCACTGTGACCCAMCA~TAGCTGTGA~GMGCAGMC~MTACTTGTGTA~CATA~AGCTMTGTCCACT~ATGG~ACA~C~CAGCACT YLIRC KN CDPNTGS CD,VKQN,NTCVG I GA2NVRVMVTGGST

GATGGGMGCMGGGTGTGCTACMTCTGGGMGGCTCAGGATGTGT~TAGATCMCCACM~TGTTGTCC~CCMTAC~G~GCMCATCMCACAG~~CTACA~CGCTCT DGKQGCATINEGSGCVGRSTT”CCPANTCCNINTGFYIRS

361 TACAGACGTGTGGMTA~TGATTMCTATATCMCCGTMTGGC 121 YRRVEe

461

601

721

841

TATAATTTACGCTTT TGAAT TGMGGCCT GAMTTTATACCTCTTTTGATATAGT~CACTATGATTATATAGTGCGTGTGTGA~GCTAGAG~~CA

GIV\MCTTCCGTATCGTCTGCACTGACATTGCTCMGCACC~GCAGTCTCCTGATATGTCAC~MGATCMGMTAGA~T~AGATGTAGCTAGTGT~ATCTGACMT~

GGACTCCTMGGTAC~CGGCTGAGA~C~GACT~GA~MTTTTGCTCMGATTCTGTGA~CAGCC~TACCAGCGACCTATCCATGCACGMGTACTCTCTA~TT~~GC

CGACAMGCTTATCGAAAAA 662

HindIII

-1111

-1060

-960

-640

-720

-600

-460

-360

-240

-120

1 1

121 41

241 81

361 121

yRGCyEAGTAGGGAGMC TMGTCGMGATMTGGGGCTTGGGATGGATGTGTCGATTTCGAGAG ACCT CAGAGCTTTTATTAGGCATATCTAAT *

461

601

721

641

961

1061

1201

Fig. 2. The nt sequences of HMK (a) and HSK (b) and deduced aa sequences. The nt sequences were determined for both strands of an EcoRI-

Hind111 fragment of AHMKO8 (a) and a BumHI-EcoRI fragment of hHSK03 (b) by the dideoxy chain-termination method (Sanger et al., 1977). The

A residue of the first ATG codon in the ORF is numbered 1. The nt sequences corresponding to the synthetic probes are underlined and the probe

number is indicated. A TATA box-like sequence is boxed. The tripartite termination sequence is shaded. The aa sequence determined by direct

sequencing of the protein is underlined (b). The nt sequences of HMK and HSK have been submitted to the DDBJ/GenBank/EMBL Data Bank

with accession numbers D13445 and D13446, respectively.

Sherman, 1982), TAG...TACT...TTT (nt 498 to 556). The aa sequence from Asn3a to Asn62 was identical to the

These results suggested that the 2.4-kb EcoRI-Hind111 N-terminal aa sequence determined by direct sequencing

fragment contained a structural HMK gene. of purified HSK. The first ATG in this ORF may be used

The nt sequence of the 2.3-kb BamHI-EcoRI fragment as a start codon, given the similarity in terms of sequence

of hHSK03 contained an ORF (Fig. 2b; from nt 1 to HMK. The calculated M, of the mature HSK was

through 372) capable of encoding a polypeptide of 124 aa. 9543, which is larger than 8.5 kDa estimated by SDS-

268

TABLE I

Killer spectra of HMK and HSK secreted from transformants and donors

Seeded strain” Killer type Transformant or donor strainb

ScHMK’ H. mrakii (IF0 0895)

ScHSKd H. suturnus (IF0 0117)

S. cerevisiae ATCC 60782

S. cereoisiae ATCC 36900

S. cereoisiae ATCC 36899

C. glahrata ATCC 2001

H. anomala ATCC 36903

K. marxianus var. marxianus ATCC 36907

P. memhranaefaciens ATCC 36908

H. anomala ATCC 36904

H. mrakii ATCC 10743

K. marxianus var. drosophilarum ATCC 36906

KI

K2

K3

K4

K5

K6

K7

K8

K9

K10

+ + _ _ + + _ + + + + + + + + + + + + + + + + + + + + + + +

_ _ _ _

“Authentic killer-type strains classified in terms of their killer/immunity properties (Young and Yagiu, 1978); S.. Saccharomyces; C.. Candida; H.. Hansenula; K., Kluyueromyces; P., Pichia. ’ + indicates killing, -indicates the absence of killing.

‘Transformant carrying the plasmid YEp-HMK.

dTransformant carrying the plasmid YEp-HSK.

Methods: To select killer-resistant mutants. YPD plates containing HMK (killer-YPD) were prepared. YPD medium in which H. mrakii had grown

overnight was centrifuged at 19 000 x g for 10 min. Bacto yeast extract (0.1 g), bacto peptone (0.2 g) and glucose (2.0 g) were added to 100 ml of the

supernatant. The medium was filtered through a 0.22~pm filter and mixed with 25 ml of autocleaved agarose (7.5%) in a water bath at 4O’C before

pouring into plates. S. wreoisiae 851824 (MATa, leu2, ura3, trpl, pep4, cir+: obtained from Dr. Elizabeth W. Jones of Carnegie Mellon University,

Pittsburgh, PA, USA) cells, either mutagenized by ethyl methansulfonate or without its treatment were cultured on killer-YPD plates that contained

HMK. Well-grown colonies were picked up, and their resistance to the killer toxin was tested by cross-streak test (Fink and Styles, 1972). HMK from nt -72 to 862 or HSK from nt -66 to 865 were amplified by polymerase chain reaction with appropriate primers and subcloned into the

yeast multicopy plasmid vector YEp51 at the BarnHI-Hind111 site, downstream of the GALlO promoter. Transformation of yeast was carried out by

the LiCl method (Ito et al., 1983). Killer activity of transformants and donor strains was assayed by the cross-streak test.

~13~ (R. mrakii)

* *

* “l.zl u I QQGCAIIWEGSGC(T~GRSTTMCCPGDTCCNINTGFYIRS

121

I:

Fig. 3. Comparison of the aa sequences of the precursors to HMK and HSK, deduced from the nt sequences. The aa residues that are identical in

the two sequences are boxed. An open arrowhead indicates the putative site of cleavage by the signal peptidase, and the filled arrowhead indicates

the putative site of processing by a KEX2-like protease. The hyphen indicates a gap that was introduced to facilitate alignment.

PAGE (Ohta et al., 1984). However, Western blot analysis

with HSK-specific antibody revealed that HSK had the

same mobility as HMK during SDS-PAGE (data not

shown). In the 5’-upstream region of HSK, a TATA box-

like sequence, TACATAA (nt - 110 to - 103) was de-

tected. Analysis of the region 3’-downstream from the

ORF revealed that there was a tripartite termination-like

sequence TAG.. .TACT.. .TTT (nt 384 to 449). Thus, it

appeared likely that this 2.3-kb BarnHI-EcoRI fragment

encoded a structural gene for HSK.

(c) Expression of HMK and HSK in S. cerevisiae

In order to examine the functional expression of HMK

and HSK in S. cerevisiae, a DNA fragment including the

ORF for HMK, from nt -72 to 862, and a fragment

including the ORF for HSK, from nt -66 to 865, were

separately inserted into YEp51 downstream from a

GAL10 promoter in the correct orientation. A condi-

tional promoter, such as that of GALIO, is suitable for

the expression of toxic proteins in S. cerevisiae. The killer-

sensitive strain S. cerevisiae BJ1824 was transformed with

the resultant plasmids (YEp-HMK and YEp-HSK).

Unlike the Kl-type killer toxins, the preprotoxin of

HMK and that of HSK did not confer immunity because

the transformants carrying plasmid YEp-HMK or plas-

mid YEp-HSK had a suicide phenotype on the galactose-

containing medium.

For expression of these genes in S. cerevisiae, we first

isolated mutants of S. cerevisiae that were not sensitive

to HMK, since HMK killed all the available laboratory

strains of S. cerevisiae. Six mutants, which were strongly

resistant to HMK, were obtained and designated KIT1

through KIT6. All of these mutants were also resistant

to HSK, suggesting that the molecular mechanisms of

the killing by HMK and HSK are identical. Detailed

analysis of these mutants is in progress in our laboratory.

In this study, the mutant KIT6 was transformed sepa-

rately with YEp-HMK and YEp-HSK. As shown in

Table I, the strain carrying YEp-HMK was capable of

killing a sensitive strain on the galactose medium, as was

the donor strain H. mrakii IF0 0895. The strain contain-

ing only the vector plasmid was incapable of killing any

of the strains. This result suggests that the ORF on YEp-

HMK encodes the HMK-specific killing functions. The

transformant harboring the plasmid YEp-HSK was also

capable of killing sensitive strains, although it was not as

effective as the donor strain. This difference may be due

to the lower levels of secreted toxin in the culture medium

of the transformant. However, further analyses are re-

quired to clarify this issue.

(d) Comparison of deduced aa sequences between HMK

and HSK

The deduced aa sequence of HMK was aligned with

that of HSK (Fig. 3). Comparison of the deduced aa se-

quence of the HMK precursor and the aa sequence deter-

mined directly from purified HMK revealed the presence

of an N-terminal extension composed of 37-aa residues.

The precursor to HSK also had an N-terminal extension

of 37 aa. The structure of the N-terminal portion from

Phe3 to Pro” in the HMK precursor and that from Va13

to Pro’i in the HSK precursor suggests that these regions

are hydrophobic. The most favorable site for cleavage by

a signal peptidase is after Ala19. The aa sequences are

strongly conserved from Met’ to Se? between the pre-

cursors to HMK and HSK. In addition, these putative

signal sequences have KEX2-like cleavage sites after two

basic aa, Lys36-Arg 37 These features are similar to those .

of the prepro-peptides of MFcll, MFcl2 (Kurjan and

Herskowits, 1982), KILMl (Bostein et al., 1984) and

KIL97 (Stark and Boyd, 1986). The mature HMK and

HSK toxins were 86% identical in terms of their aa se-

quences. These results can be ascribed to the fact that

killer/immunity properties of H. mrakii and H. saturnus are very similar (Nomoto et al., 1984). These findings

suggest that HMK and HSK may have originated from

the same ‘ancestral killer toxin’ and that the killing mech-

anisms of these toxins are the same. A major difference

between the aa sequence of HMK and that of HSK is

that HSK lacks an aa residue that corresponds to Thrso

in HMK. In addition, the aa sequence of HSK differs

from that of HMK at 11 sites. Expression of the genes

for killer toxins after site-directed mutagenesis in S. cere- visiae should reveal the aa residues that are important

for the killing spectrum and the mechanisms responsible

for the toxity of these killer toxins.

(e) Conclusions

(I ) Genes encoding killer toxins from H. mrakii and

H. saturnus were cloned and sequenced. These genes en-

coded ORFs of 125 and 124 aa, respectively.

(2) The deduced aa sequences indicate that these killer

toxins are synthesized as larger precursors with N-

terminal prepro-peptides of 37 aa that are followed by

mature proteins of 88 and 87 aa, respectively.

(3) Expression of HMK and HSK under control of

the GAL10 promoter of S. cerevisiae in a mutant of

S. cerevisae that was resistant to the killer toxins was

associated with the same killer spectrum as that of the

original donor strains.

ACKNOWLEDGEMENTS

We are grateful to Dr. Yasuhiro Ohta, Marukin-Shoyu

Co., Ltd., for providing us with purified HSK. We also

wish to thank Dr. Tetsuro Yamamoto, Nichi-Nichi

Pharmaceutical Co., Ltd., for helpful discussions.

REFERENCES

Ballance, D.J.: Sequences important for gene expression in filamentous

fungi. Yeast 2 (1986) 50-54.

270

Bostein, K.A., Elliott, Q.. Bussey, H., Burn, V., Smith, A. and Tipper,

D.J.: Sequence of the preprotoxin dsRNA gene of type I killer yeast:

multiple processing events produce a two-component toxin. Cell 36

(1984) 741-751.

Breathnach, R. and Chambon, P.: Organization and expression of

eukaryotic split genes encoding for proteins, Annu. Rev. Biochem.

50(19X1)349-383.

Chen, W. and Struhl, K.: Saturation mutagenesis of a yeast his3 TATA

element: genetic evidence for a specific TATA-binding protein. Proc.

Natl. Acad. Sci. USA X5 (1988) 2691-2695.

Ito, H., Fukuda, Y., Murata, K. and Kimura, A.: Transformation of

intact yeast cells treated with alkali cations. J. Bacterial. 153

(1983) 163-16X.

Fink, G.R. and Styles, C.A.: Curing of a killer factor in Sacchuromycrs

cerevisiue. Proc. Natl. Acad. Sci. USA 69 (1972) 284662849.

Kurjan, J. and Herskowits, 1.: Structure of a yeast pheromone gene

(MFa): a putative a-factor precursor contains four tandem copies

of mature a-factor. Cell 30 (1982) 9333943.

Nomoto, H., Kitano, K., Shimazaki. T.. Kodama, K. and Hara, S.:

Distribution of killer yeasts in the genus Hunsenula. Agric. Biol.

Chem. 48 (1984) 8077809.

Ohta, Y., Tsukada, Y. and Sugimori, T.: Production, purification and

characterization of HYI, an anti-yeast substance, produced by

Hunsenuiu suturnus. Agric. Biol. Chem. 4X (1984) 903-908.

Sanger, F., Nicklen, S. and Coulson, A.R.: DNA sequencing with chain-

terminating inhibitors. Proc. Natl. Acad. Sci. USA 74 (1977)

546335467.

Stark, M.J.R. and Boyd, A.: The killer toxin of KIuyveromyces Luctis:

characterization of the toxin subunits and identification of the genes which encode them. EMBO J. 5 (1986) 199552002.

Yamamoto, T., Imai, M., Tachibana, K. and Mayumi, M.: Application

of monoclonal antibodies to the isolation and characterization of a

killer toxin secreted by Hansenulrr mrukii. FEBS Lett. 195 (1986a)

2533257.

Yamamoto. T., Hiratani, T., Hirata, H., Imai, M. and Yamaguchi, H.:

Killer toxin from Hunsenula mrakii selectively inhibits cell wall syn-

thesis in a sensitive yeast. FEBS Lett. 197 (1986b) 50-54.

Young, T.W. and Yagiu, M.: A comparison of the killer character in

different yeasts and its classification. Antonie van Leeuwenhoek

J. Microbial. Serol. 44 (1978) 59977.

Zaret, K.S. and Sherman, F.: DNA sequence required for efficient

transcription termination in yeast. Cell 2X (1982) 5633573.