Partial amino acid sequence of penicillinase coded by Escherichia coli ...

4 downloads 248 Views 911KB Size Report
Jun 12, 1978 - R. P. AMBLER AND G. K. ScoTr*. Department of Molecular .... A F R L DR WE PE L NE AI P ND ER D T TM PA AMA T T L R K L LTG E. 5. 6.
Proc. Natl. Acad. Sd. USA Vol. 75, No. 8, pp. 3732-3736, August 1978

Biochemistry

Partial amino acid sequence of penicillinase coded by Escherichia coli plasmid R6K (P-lactamase/protein homology) R. P. AMBLER AND G. K. ScoTr* Department of Molecular Biology, University of Edinburgh, Edinburgh, EH9 3JR, Scotland

Communicated by Walter Gilbert, June 12, 1978

ABSTRACr Direct studies of the amino acid sequence of an Escherichia coli plasmid-coded penicillinase (penicillin amido~lactamhydrolase, EC 3.5.2.6) are in complete agreement with results derived from the translation of the DNA sequence of a related plasmid, apart from a single amino acid substitution. This penicillinase from a Gram-negative bacterium shows 30-35 identity with functionally similar enzymes from Gram-positive bacteria. This paper should be read in con unction with the report of the DNA sequence of the gene [Steliffe, J. G. (1978) Proc. NadL Acad. Sci. USA 75, 3737-374LJ. Penicillinases are enzymes that inactivate the penicillin and cephalosporin antibiotics. They have a wide distribution in bacteria (1). Most of the enzymes act by hydrolyzing the ftlactam bond, and so they are often known as f3-lactamases, or more systematically as penicillin amido-fl-lactamhydrolases (EC 3.5.2.6). Several Gram-positive bacteria produce inducible extracellular penicillinases, and large amounts of pure enzyme can readily be prepared from constitutive mutants. The amino acid sequences of proteins from Staphylococcus aureus (2, 3) and Bacillus licheniformis (2, 4) and of one of the two distinct enzymes in Bacillus cereus (5) have been studied and are sufficiently similar to one other that a common evolutionary origin and mechanism of action is indicated. The normal enzymes from Gram-negative organisms are quite different in their physiology of production, because they are generally noninducible, cell-bound, and produced at relatively low levels. Although some chromosomally coded enzymes have been recognized in Escherichia coll (6), the enzymes in the enterobacteria and the pseudomonads are often coded for by resistance transfer plasmids. The enzymes vary among themselves in substrate profile (relative activity against a range of penicillins and cephalosporins), molecular weights, and other chemical properties (7, 8), but without sequence studies it was not possible to decide if the enzymes were evolutionarily related to the proteins from the Gram-positive bacteria, or even among themselves. We chose to study the enzyme from the plasmid R6K [formerly known as RTEM (9)] expressed in E. coli strain W3310, because the yields were better than for other enterobicterial penicillinases considered. A method was developed (10) for the isolation of the protein in gram quantities in apparently homogeneous form, and structural and mechanistic studies were started (11). The same enzyme system is also being used by Knox et al. (12), for the study of the three-dimensional structure of a penicillinase by x-ray crystallography. The preliminary studies (11) produced much information on peptide compositions and sequence, and suggested a possible mechanism of action for the enzyme (13). Indications were

obtained of a limited sequence similarity to the enzymes from the Gram-positive bacteria, but unfortunately the information was insufficient to allow the deduction of long lengths of sequence. Recently, as a result of the interest in the tertiary structure of the enzyme (12) and because of the prospect of the DNA of the gene being sequenced (14), the investigation was reopened, and a tentative and incomplete sequence was derived (Fig. 1).t After the completion of the determination of the DNA sequence of the gene, some further amino acid sequence experiments were performed, to check and extend the original postulate. The final extent of the amino acid sequence evidence is shown in Fig. 2. The differences between the tentative sequence shown in Fig. 1 and the sequence derived from the translation of the DNA sequences (complementary strands) of the amp gene of plasmid pBR322 (15) are summarized in the caption to the figure. All the differences except one are considered to be errors in the tentative protein sequence, caused by overinterpretation of preliminary results in our attempt to maximize the information presented in Fig. 1. The remaining difference (Lys/Gln at residue 14, Fig. 2) would seem to be genuine, because the amino acid sequence evidence for this region is good. The two investigations used different plasmids, isolated from the wild in different organisms from widely different geographical locations, but this Lys/Gln substitution is the only difference that there is any evidence for, although silent mutational differences may well exist. Amino acid sequence determination The enzyme was purified as previously described (10). The amino acid sequence was investigated by the general methods used for other proteins in this laboratory (3, 18). The protein was denatured by oxidation with performic acid and then digested with a protease, and the peptides formed were fractionated by gel filtration followed by high-voltage paper electrophoresis and chromatography. The peptides were then analyzed quantitatively for amino acid composition and purity, and peptide sequences were investigated by the dansyl/phenyl isothiocyanate method and by exo- and endopeptidase digestion. Amide groups were mostly assigned from peptide electrophoretic mobilities. Digests investigated were with Staph. aureus protease (1 !tmol), chymotrypsin (3 ,mol), trypsin (3.9 Mumol), and pepsin (9 imol). Performic acid converts tryptophan to a mixture of intensely fluorescent derivatives that cannot be quantitated in peptide Present address: Department of Biochemistry, University of Auckland, Auckland, New Zealand. t This information was sent to J. Knowles at Harvard University in March 1977 and kept by him under seal until the DNA sequence of the amp gene of plasmid pBR322 had been determined by J. G. Sutcliffe (15). At this stage (October 1977) the two sequence postulates were compared. *

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. ยง1734 solely to indicate this fact. 3732

Biochemistry:

Ambler and Scott

Proc. Natl. Acad. Sci. USA 75 (1978)

3733

1* H P E T L V K V K D A E D K L G

A R V G Y I E L D L N S G K I L E S F R P E E R F PM M S T F K V L L C G A V

2 V L SR VD AG Z ZZ LGR R HIY SQ NDL V EY SP VT E KH LT FG MT V 4 3 R E L C S A A I T M s d n t a a n L L L T T I G G P K E L L g h h v t d n m T

A F R L DR WE PE L NE AI P ND ER D T TM PA AMA T T L R K L LTG E

6

5

L L T L A S R Q Q L I D W M E A N K V A G P L L R S G L P A A F d i w

(7)

(7)

(7)

k/

(8)

/I V V I Y/S G A G E R/G I I A A L G P p b k s g r/T T G q s a t m D E R N R 9 QI A E I G A S L I K W H

FIG. 1. The amino acid sequence postulated for the penicillinasefrom E. col R6K before the DNA sequence of the amp gene of pBR322 had been determined (15). The one-letter notation used is that recommended by the IUPAC-IUB Commission on Biochemical Nomenclature (16). The sequence is arranged in lines that correspond to those in Fig. 2, in which the evidence for the final sequence postulate is summarized. Lower-case letters indicate residues whose sequence was not completely established from peptide studies. The numbers above the sequence refer to points of disagreement between the first amino acid sequence postulate and the translation of the DNA sequence, as discussed below. (1) The DNA sequence shows Gln not Lys, but the peptide evidence (Fig. 2) is strong, so a real difference between the two plasmids is suspected. (2) The DNA sequence shows -Ile-His-, not -His-Ile-. The peptide evidence was initially misread, in part because of insufficient hydrolysis time for a stable peptide bond in a dansyl NH2-terminal determination. Subsequent experiments have confirmed the DNA interpretation. (3) The presence of the third leucine residue was suspected from considerations of protease specificity. (4) The tripeptide -Thr-Ala-Phe- was mislocated because of a misreading (Thr for Leu) of a weak dansyl spot from a phenyl isothiocyanate degradation. Subsequent experiments (Fig. 2) have confirmed the DNA interpretation. (5) The Asn ascription (rather than Asp or Asx) was wrongly made because insufficient care was taken in assessing electrophoretic mobility results. Homology considerations also corrupted the interpretation. Subsequent experiments (the partial characterization of peptide T212aF1, Fig. 2) have confirmed the DNA interpretation. (6) In the original dansyl/phenyl isothiocyanate degradation experiment, a glycine and an alanine residue were transposed. The tryptophan residue was missed from its correct location, although the proximity of such a residue was realized. Subsequent experiments have partially confirmed the DNA interpretation. (7) No peptide evidence was at first found that would order these three blocks of sequence. After the DNA evidence became available, it was recognized that peptide C26c (Fig. 2) partially confirmed the order. (8) The tripeptide Gly-Ser-Arg (T32b, Fig. 2) was found in all tryptic digests in large amounts. It was confused with an expected peptide from residues 230-232 (part of peptide T15b, Fig. 2), and neglected after the fist attempt to quantitate it had failed because of an analyzer defect. (9) The DNA sequence shows -His-Trp, not -Trp-His. Carboxypeptidase A digestion of the intact protein demonstrated the occurrence of these amino acids at the COOH-terminus, but suggested the order shown. The misconception was helped by the difficulty in interpreting the dansyl NH2-terminal results on small peptides containing oxidized tryptophan residues. Subsequent experiments have confirmed the DNA interpretation.

fragments. Such peptides are amenable to phenyl isothiocyanate degradation, but purification losses are higher than for other peptides of comparable size. The NH2-terminal sequence of the protein was investigated with an automatic sequenator (18) (Beckman model 890A), using an NN-dimethylbenzylamine program (19). At the first cycle both histidine and leucine phenylthiohydantoins were identified, but after this an unambiguous sequence was obtained up to residue 26 (Fig. 2) that completely predicted the results subsequently obtained from overlapping peptides. NH2-terminal leucine had also been detected by the fluorodinitrobenzene method during the preliminary investigation (11). The electrophoretic mobilities of peptides derived from the NH2-terminal sequence (Fig. 2) indicated that the imidazole group of the histidine residue had an unusually low pK value. Extent of the amino acid sequence evidence The amino acid sequence derived by translation of the DNA sequence (15) is shown in Fig. 2, together with the evidence for the sequence from peptide studies. The amino acid composition of the purified protein is compatible with the sequence (Table 1). Published estimates of the molecular weight (8) are about 20% lower than the value of 28,900 derived from the amino acid sequence.

The deduction of the amino acid sequence (Figs. 1 and 2) was of course helped by considerations of "homology" (3), but independent evidence for the sequence and for overlaps was obtained except for the parts listed here. The whole sequence is covered by-peptides isolated from at least one of the protease digests, but there is no peptide evidence for the overlaps at positions 112/114 (where the presence of the third leucine residue was suspected but not proved), 196/198, and 245/246. The evidence that the blocks of sequence 210-215 and 216-218 are arranged in this order rather than the reverse is very weak. The amino acids in the blocks 105-111, 130-135, 183-189, 200-205, 226-232, and 241-245 have not been ordered, but the compositions of the blocks are known. The sequence of residues 200-205 seems to have been determined incorrectly in the preliminary experiments, with residues 199 and 203 transposed and the tryptophan residue (204) missed. No peptides corresponding to any part of the leader sequence (residues LJ-L23, Fig. 2) have been identified in any of the digests. The penicillinases from some of the Gram-positive bacteria have "ragged ends" (2, 5), and a similar phenomenon may be responsible for the presence of some NH2-terminal leucine in the R6K enzyme preparation. A small proportion of the molecules may start at (say) residue L19, although there is no evidence to support this hypothesis from trace residues at subsequent steps of the automatic degradation experiments.

Biochemistry: Ambler and Scott

3734

7(GIn)

Proc. Natl. Acad. Sci. USA 75 (1978) (5)

T317a(5)

T32a

T26at(12)

5

T211a(11)

7 7 ~ 1L 7 7 77 -7 7 Met-Ser-Ile-G ln-His-Phe-Arg-Val-A la-Leu>-I e-Pro-Phe-Phe-A Za-A la-Phe-Cys-Leu-Pro-Va2-Phe-A Sa-His-Pro-Glu-Thr-Leu-Va1-Lys-Val-Lys-Asp-Ala-Glu-Asp-Lys-Leu-GlyLi L2 L3 L4 L5 L6 L7 L8 L9 L10 Lll L12 L13 L14 L15 L16 L17 L18 Li9 L20 L21 L22 L23 i 9 2 4 8 5 6 7 3 10 11 12 13 14 15 16 C310b (14) -7

-7

,

sequenator

L

C27a (9)

"W

-7

7 -7 -7 r

-7 -7 -

--7

7777T

aF59 H L"= 7 (8)

T210b* (17)

L7

-

-7

7

-7

-

L7

-

q

T310a(18) 7 -7

7

-7

7

T35at (21)

-7

-

7 -7 7

T27aii* (22) 7

--7 7 7

-Ala-Arg-Va1-Gly-Tyr-I1e-Glu-Leu-Asp-Leu-Asn-Ser-Gly-Lys-I1e-Leu-Glu-Ser-Phe-Arg-Pro-Glu-Glu-Arg-Phe-Pro-Met-Met-Ser-Thr-Phe-Lys-Va1-Leu-Leu-Cys-Gly-A1a-Val17

18

i9 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55

C44 (14)(

-7 -7

--7

sequenator

C37aitC C413(1_ L

M E

- -1

L

'

6) C34a

-7

C34a**(1)

F27b* (5)

_ 7 7 7> -7

'I

E,-

C4

-

-

7

-7

C514

1H

-

F410 (14) -7 7 -7

I

"T

C34c(1) C48a(10) C39b(10) --7 -7 4 77 7 7 -C2;bvt**(4) -7 -

-7 -7 --7

F510**

-.7

-7

(13)

-C_--IL ~ C2*

C5

chymotrypsin

T112b** (2)

t

1 I7 7-7 1 A i --7 -7 --P 4 4

-2ai7 -Leu-Ser-Arg-Va 56 57 58 59

7 TllOa* (6) T26bi (5) --7 -'V --P --7 7 --7 -7 --7 -Asp-A la-Gly-Gln-Glu-Gl n- Leu-G ly-Arg-Arg- I e-H s-Tyr -Ser-Gln-Asn-Asp-Leu-Va l-Gl u-Tyr- S er -Pro-Val-Thr-G lu-Lys-H i s- Leu-Thr -Asp-G ly-Met-Thr-Val-

60

61

62

63

64

65

66

67

4 b;

-.I17

69

70

14

71

72

73

74

75

76

7

C414 (20) 7 7 7

C31b (1)

~~~C310at* (19)

ffi -7

68

-7

71p

77

78

79

80

81

82

83

84

85

87

89

88

90

91

92

93

94

C28a (13)

/

-

thermotysin

Staph. protease Hi

H4b

F3**

E:; -4 u:a7 7

a t--

86

T318* (2)

H

44'

H6* i L,,;

F511 (17)

F36* (8) 77 -7 I7

-7-7 -7

-

T16b-

77

(11)

-Arg-Glu-Leu-Cys-Ser-Ala-Ala-Ile-Thr-Met-Ser-Asp-Asn-Thr-Ala-A la-Asn-Leu-Leu-Leu-Thr-Thr-Ile-Gly-Gly-Pro-Lys-Glu-Leu-Thr-Ala-Phe-Leu-His-Asn-Met-Gly-Asp-His95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133

C313a (2)

E;-7

-7-7

-7 -7

=7

H3**

C37aiii (4)

C47c

C213t (11)

a

E;--1L;

subtilisin

S7

,

C57a (10)

a

x--,r--4L 7-7

C26bit* (3)

L

C27biv* (4) thermolysin

a7

i

-7 -7 -7

--

I

H2*

It 1

H1**

F38** (2)

FIG. 2. Evidence from peptide studies for the amino acid sequence of the penicillinase from E. coli R6K. The sequence shown is derived from the translation of the DNA sequence of the amp gene of E. coli pBR322 (15). Independent evidence from peptide studies exists for all the sequence except for the residues shown in italics. No parts of the postulated leader sequence (residues Ll-L23) were detected in peptide experiments, but there is evidence from at least peptide compositions for all of the rest of the sequence. Peptides derived by digestion with trypsin (T) and pepsin (P) are shown above the sequence, and by digestion with chymotrypsin (C) and Staph. aureus protease (F) below the sequence. Peptides formed by subdigestion are shown with an arrow from the primary peptide (H, thermolysin; S, subtilisin). Full lines indicate peptides analyzed quantitatively, with substandard analyses if marked *, and particularly bad analyses

A peptide containing a third cysteine residue [with partial

Asx-Met-Thr-Cys-Glx-Glx(Leu,Ile,Asx2,Phe,Pro,Lys)] was isolated from tryptic digests in both the preliminary and the main experiments, in the low but not negligible yield of 2%. We suppose that this might be a high-yield peptide from a

structure

contaminating protein. The sequence in Fig. 2 contains two cysteine residues in the

penicillinase

enzyme

molecule, with

a

third residue in the

leader sequence. Unlike some other enterobacterial penicillinases that contain cyst(e)ine, the R6K enzyme is not inhibited by p-chloromercuribenzoate (7, 8), which suggests that the residues may be linked by an -S-S- bridge. The missing pieces of amino acid sequence evidence could be obtained by straightforward further experimentation. If complete splitting of methioninyl peptide bonds was achieved with cyanogen bromide, residues 162-245 would form a frag-

t~ ~ ~ ~ ~ ~ ~ ~ ~ 7

Proc. Natl. Acad. Sci. USA 75 (1978)

Biochemistry: Ambler and Scott

,T313a* (2)

T312a (1)

L,;

--- 7

3735

1

7

-P

Tl7b (14)

T214 (12) p-1A7 -rTr Va2iTA__uAs I -7p-7 --7Ul7 I P7 G-7 P7 rO71 ol -..1 -.---7 7 -77 -Ma Z-Thr-Arg-Leu-Asp-Arg-Trp-Glu-Pro-G lu-Leu-Asn-Glu -Ala- Ile-Pro-Asn-Asp-G lu-Arg-ASp-Thr-Thr-Me t-Pro-Ala-Ala-Me t-Al a-Thr-Thr-Leu-Arg-Lys-Leu-Leu-Thr-Gly-G lu-

169 170 171 172

134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168

C112a*(3)

(5C412(11)

C37bi -7

-7 r-.

7

~7

-7

)

~~~~~~~~~~~~~C214a* (1)I

C312b* (2)

F514 (28).

F'AR** (9)

s4 7

L -7

P332 (12) -7 77

A

.

F515 (23)

L-7-7

--7

-)

Fl

P261 (3)

b~- -

7

,,

E-7-7 777 -7 iri

-7 --7

Staph. protease I-,

T27bi*(1) a|T24bi(5) a-Leu-Pro-A Za-G Zy-Trp-Phe- I 1 e-Al a-Asp-Lys-Ser-Gly-Leu -Leu-Thr-Leu-Ala-Ser-Arg-Gln-Gln-Leu-I~e-A sp-Trp-Me t-G Iu-A Za-Asp-Lys-val -Al a-Gly-Pro-Leu-Leu-Arg-Ser-Al T212a** (2)

173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 C35a(3) C26c(4) C29t(7) C47a*(2) i 7e * / 17 *

tpypsind

T3* 1 ~3-

* T1 .1.I

1

177. L7

I

i

T2*

,

-7

L71

7

P244b (2)

I 7

L-7 -7-

T37ai(4) 7

7

chyrnotrypsinAT15b(19)

T32b(15'1

,

7

r:0 -=4

L=.

-7

-

-7

7 -7

':2

C5

-7

7

7

-7

1

LT3917

TIVcore*(6) 7

7

-Ala-Gly-Glu-Arg-ly-Ser-Arg-Gly-Ibe- I le-Ala-Al a-Leu-G l y-Pro-As p-GZy-Ls-Pro-SeyArg-I1e-Val-Va1-Ile-Tyr-Thr-Thr-Gly-SerGZn-A 1a-Th2-8let-Asp-Glu-Arg-Asn-Arg212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242' 243 244 245 246 247 248 245

C28biii(3)

C26c(4)

7~~~~~~~E:

P-0

C210b** (4) --7

- 7

77

L-7 C316b* -7 (2) --7' '

F45* (9) 7 7

P244b (2) -7

(7) T45**-

T97ai+** (9))

K;-

7

7

7

-Gln-I le-Ala-Glu-Ile-Gly-Ala-Ser-Leu-Ile-Lys-His-Trp

251 252 253 254 255 256 257 258 259 260 261 262 263

t

C210b** (4)

,

F43 (14)

F45* (9) 7

C32** (6) 7

-7

.7

-7

7

-7

(with 0.3-0.5 mol/mol of contaminating amino acids, in most cases glycine and serine) if marked **. Broken lines indicate qualitative analyses. Under peptide lines, - indicates end groups and subsequent residues revealed by phenyl isothiocyanate degradation identified by the dansyl method, substandard if shown --. Under the sequence, - indicates residues identified as the phenylthiohydantoin derivatives by thin-layer chromatography or gas/liquid chromatography after release from the protein by automated degradation (17), substandard if shown --. Peptides marked t were examined by carboxypeptidase A digestion. Above the peptide line is the peptide name, followed by a number in parenthesis which is the percentage yield of the purified peptide. The experimental details and sequencing standards are similar to those described in refs. 3 and 18. ment that was about twice as large as the next biggest peptides, and so should be quite easy to isolate pure by gel filtration. This fragment would contain most of the parts of the sequence that are still unconfirmed. However, because of the demonstrated accuracy of the DNA sequencing method, we do not think that the effort necessary to confirm the whole sequence is scientifically justified, and we can think of no new biological information that is likely to come from such completion.t

t In view of the astonishing speed and accuracy of the new DNA se-

quencing techniques (15), it is worth trying to assess the effort necessary for conventional protein sequencing. The preliminary experiments were carried out as part of a 3-year course of study leading to a Ph.D. degree (11). The subsequent experiments, which provided most of the results shown in Fig. 2, occupied a research group (R.P.A. and two research assistants) for about l/6th of a year (based on the proportion of annual amino acid analyzer throughput used) spread over about 4 months.

3736

Biochemistry: Ambler and Scott

Proc. Natl. Acad. Sci. USA 75 (1978)

Table 1. Amino acid composition of E. coli R6K penicillinase

Sample Residue Glycine Alanine Valine Leucine Isoleucine Serine Threonine Aspartic acid Asparagine Glutamic acid Glutamine Phenylalanine

1 21.6 24.6 11.6 30.4 11.7 14.4 19.5 24.5

2 21.6 26.1 11.4 30.9 12.1 14.7

3 20.4 25.1 14.1 30.7 14.3 14.2

4 Sequence 22.3 21 25 24.6 13 11.8 30.2 31 15 12.3 15 15.7

17.4

19.1

19.0

25.5

26.0

25.1

20 16

8 20 8 5 5.3 5.0 5.4 5.4 4 4.2 4.0 4.1 4.4 Tyrosine 4 Tryptophan (2.0) 3.5 2.0* 2 Cysteine (Lost) 9 Methionine 9.1 9.0* 7.1 9.5* 12 Proline 11.6 11.8 11.2 11.9 11 Lysine 12.0 11.8 13.4 12.0 6 Histidine 4.4 5.2 5.0 5.5 18 Arginine 14.8 16.9 16.6 15.5 Total 263 Results are shown as residues/molecule. Samples were hydrolyzed at 1050 for 24 hr (samples 1 and 2) or 96 hr (samples 3 and 4). Sample 4 was hydrolyzed with 3 M mercaptoethanesulfonic acid, the others with 6 M HCI. Samples 1 and 4 were of native protein, samples 2 and 3 of protein that had been oxidized with performic acid before hydrolysis. The samples are from three completely independent preparations of protein. The results were calculated on the basis that 2 (Gly + Ala + Leu + Thr + Asp + Phe) = 126. * As cysteic acid and methionine sulfone. 28.4

28.8

28.8

27.6

Comparisons with other penicillinases The sequence of the E. colh R6K penicillinase can be aligned with sequences of enzymes from three Gram-positive bacteria (5), and very few gaps need be postulated in obtaining a good match. About 20% of the sites are occupied by identical residues in all four sequences, and these matching residues are distributed throughout the length of the sequences. There are some regions of particularly conserved sequence, while the COOH-terminal 20% of the E. coli sequence matches the Gram-positive sequences less well. A full appreciation of the importance of the regions of identity and similarity awaits the elucidation of the three-dimensional structure of one or more of the penicillinases, and further developments in knowledge of their mechanism of action. The E. coli enzyme is not much more different (30-5% identity) from the Gram-positive enzymes than the latter are among themselves (down to 40% identity). It is interesting that the pBR322 leader sequence shows no similarity whatsoever to the NH2-terminal region of the membrane-bound form of penicillinase from B. licheniformis (20), a region of unusual amino acid composition.

Plasmid R1 [originally known as R7268 (9)], the source of the amp gene in pBR322, was isolated from the wild in SalmoneUa paratyphi B (9) from a patient in London who had probably been infected while in Spain. Plasmid R6K (also known as RTEM) was isolated from E. coil in Greece (9). There is probably only a single amino acid difference in the proteins coded for by the two plasmids, and it is interesting that at this site (residue 14, Fig. 2) the R6K form resembles two of the Grampositive sequences (Staph. aureus and B. licheniformis), while the pBR322 form resembles the third (B. cereus). The nature of the amino acid substitution (Gln/Lys) suggests that tbire should be a charge difference between the R6K and the pBI322 penicillinases. A study of R factor-determined 4 lactamases by analytical isoelectric focusing (21) did not distinguish between the R6K and R1 proteins. It would be interesting to know the amino acid sequences of several of the other enterobacterial penicillinases (8). If organisms are constructed that will produce these proteins at high levels, direct amino acid sequence investigation is still a competitive strategy. It is a pleasure to thank Dr. J. R. Knowles for having instigated our useful and enjoyable information exchange with Dr. J. G. Sutcliffe. We thank Mrs. M. Daniel and Mrs. E. Taylor for technical assistance, and Dr. J. Melling (Microbiological Research Establishment, Porton, England) for encouragement and for the supply of enzyme. The work was supported by the Medical Research Council. 1. Citri, N. (1972) in The Enzymes, ed. Boyer, P. D. (Academic, New York), Vol. 4, pp 23-46. 2. Ambler, R. P. & Meadway, R. J. (1969) Nature 222, 24-26. 3. Ambler, R. P. (1975) Biochem. J. 151, 197-218. 4. Meadway, R. J. (1969) Biochem. J. 115,12. 5. Thatcher, D. R. (1975) Biochem. J. 147,313-326. 6. Lindstr6m, E. B., Boman, H. G. & Steele, B. B. (1970) J. Bacteriol. 101, 218-231. 7. Jack, G. W. & Richmond, M. H. (1970) J. Gen. Microblol. 61, 43-61. 8. Richmond, M. G. & Sykes, R. B. (1973) Adv. Microb. Physiol. 9, 31-88. 9. Datta, N. & Kontomichalou, P. (1965) Nature 208, 239-241. 10. Melling, J. & Scott, G. K. (1972) Biochem. J. 130,55-62. 11. Scott, G. K. (1972) Dissertation (Univ. of Edinburgh, Edinburgh,

Scotland). 12. Knox, J. R., Kelly, J. A., Moews, P. C. & Murthy, N. S. (1976) J. Mol. Biol. 104,865-876. 13. Scott, G. K. (1973) Biochem. Soc. Trans. 1, 159-162. 14. Maxam, A. M. & Gilbert, W. (1977) Proc. Nati. Acad. Sci. USA

74,56-564. 15. Sutcliffe, J. G. (1978) Proc. Nati. Acad. Sci. USA 75, 37373741. 16. IUPAC-IUB Commission on Biochemical Nomenclature (1969) Biochem. J. 113,14. 17. Edman, P. & Begg, G. (1967) Eur. J. B yochem. 1, 80-91. 18. Ambler, R. P. & Wynn, M. (1973) Biochem. J. 131,485-498. 19. Hermodson, M. A., Ericsson, L. H., Titani, K., Neurath, H. & Walsh, K. A. (1972) Biochemistry 11, 4493-452. 20. Yamamoto, S. & Lampen, J. 0. (1975) J. Biol. Chem. 250,

3212-3213. 21. Matthew, M. & Hedges, R. W. (1976) J. Bact. 125,713-718.